Investigation of potential interferences in the detection of atmospheric ROx radicals by laser-induced fluorescence under dark conditions

摘要

Direct detection of highly reactive, atmospheric hydroxyl radicals(OH) is widely accomplished by laser-induced fluorescence(LIF) instruments. The technique is also suitable for the indirectmeasurement of HO and RO peroxy radicals bychemical conversion to OH. It requires sampling of ambientair into a low-pressure cell, where OH fluorescence isdetected after excitation by 308 nm laser radiation.Although the residence time of air inside the fluorescence cell istypically only on the order of milliseconds, there is potentialthat additional OH is internally produced, which wouldartificially increase the measured OH concentration. Here,we present experimental studies investigating potentialinterferences in the detection of OH and peroxy radicalsfor the LIF instruments of Forschungszentrum Jülich fornighttime conditions. For laboratory experiments, the inlet of theinstrument was over flowed by excess synthetic air containing oneor more reactants. In order to distinguish between OHproduced by reactions upstream of the inlet and artificial signalsproduced inside the instrument, a chemical titration for OHwas applied. Additional experiments were performed in thesimulation chamber SAPHIR where simultaneous measurements by anopen-path differential optical absorption spectrometer (DOAS)served as reference for OH to quantify potential artifactsin the LIF instrument. Experiments included the investigation ofpotential interferences related to the nitrate radical(NO, NO), related to the ozonolysis of alkenes(ethene, propene, 1-butene, 2,3-dimethyl-2-butene,-pinene, limonene, isoprene), and the laserphotolysis of acetone. Experiments studying the laser photolysisof acetone yield OH signals in the fluorescence cell, whichare equivalent to 0.05 × 10 cm OH fora mixing ratio of 5 ppbv acetone. Under most atmosphericconditions, this interference is negligible. No significantinterferences were found for atmospheric concentrations ofreactants during ozonolysis experiments. Only for propene,-pinene, limonene, and isoprene at reactantconcentrations, which are orders of magnitude higher than in theatmosphere, could artificial OH be detected. The value ofthe interference depends on the turnover rate of the ozonolysisreaction. For example, an apparent OH concentration ofapproximately 1 × 10 cm is observed when5.8 ppbv limonene reacts with 600 ppbv ozone.Experiments with the nitrate radical NO reveal a smallinterference signal in the OH, HO, and ROdetection. Dependencies on experimental parameters point toartificial OH formation by surface reactions at the chamberwalls or in molecular clusters in the gas expansion. The signalscales with the presence of NO giving equivalent radicalconcentrations of 1.1 × 10 cm OH,1 × 10 cm HO, and 1.7 × 10 cm RO per 10 pptv NO.