A smog chamber comparison of a microfluidic derivatisation measurement of gas-phase glyoxal and methylglyoxal with other analytical techniques

Abstract. A microfluidic lab-on-a-chip derivatisation technique has been developed to measure part per billion (ppbV) mixing ratios of gaseous glyoxal (GLY) and methylglyoxal (MGLY), and the method is compared with other techniques in a smog chamber experiment. The method uses o-(2, 3, 4, 5, 6-pentafluorobenzyl) hydroxylamine (PFBHA) as a derivatisation reagent and a microfabricated planar glass micro-reactor comprising an inlet, gas and fluid splitting and combining channels, mixing junctions, and a heated capillary reaction microchannel. The enhanced phase contact area-to-volume ratio and the high heat transfer rate in the micro-reactor resulted in a fast and highly efficient derivatisation reaction, generating an effluent stream ready for direct introduction to a gas chromatograph-mass spectrometer (GC-MS). A linear response for GLY was observed over a calibration range 0.7 to 400 ppbV, and for MGLY of 1.2 to 300 ppbV, when derivatised under optimal reaction conditions. The analytical performance shows good accuracy (6.6% for GLY and 7.5% for MGLY), suitable precision (<12.0%) with method detection limits (MDLs) of 75 pptV for GLY and 185 pptV for MGLY, with a time resolution of 30 min. These MDLs are below or close to typical concentrations of these compounds observed in ambient air. The feasibility of the technique was assessed by applying the methodology to quantify α-dicarbonyls formed during the photo-oxidation of isoprene in the EUPHORE chamber. Good correlations were found between microfluidic measurements and Fourier Transform InfraRed spectroscopy (FTIR) with a correlation coefficient (r[superscript: 2]) of 0.84, Broadband Cavity Enhanced Absorption Spectroscopy (BBCEAS) (r[superscript: 2] = 0.75), solid phase micro extraction (SPME) (r[superscript: 2] = 0.89), and a photochemical chamber box modelling calculation (r[superscript: 2] = 0.79) for GLY measurements. For MGLY measurements, the microfluidic technique showed good agreement with BBCEAS (r[superscript: 2] = 0.87), SPME (r[superscript: 2] = 0.76), and the modeling simulation (r[superscript: 2] = 0.83), FTIR (r[superscript: 2] = 0.72) but displayed a discrepancy with Proton-Transfer Reaction Time-of-Flight Mass Spectrometry (PTR-ToF-MS) with r[superscript: 2] value of 0.39.