Atmos. Chem. Phys. Discuss., 6, 7177–7205, 2006 Atmospheric www.atmos-chem-phys-discuss.net/6/7177/2006/ Chemistry © Author(s) 2006. This work is licensed and Physics under a Creative Commons License. Discussions Processes controlling the concentration of hydroperoxides at Jungfraujoch Observatory, Switzerland S. J. Walker1, M. J. Evans1, A. V. Jackson1, M. Steinbacher2, C. Zellweger2, and J. B. McQuaid1 1Institute for Atmospheric Science, School of the Earth and Environment, University of Leeds, Leeds, United Kingdom 2Swiss Federal Institute for Materials Science and Technology (Empa), Laboratory for Air Pollution/Environmental Technology, 8600 Dubendorf,¨ Switzerland Received: 6 June 2006 – Accepted: 11 June 2006 – Published: 28 July 2006 Correspondence to: A. V. Jackson ([email protected]) 7177 Abstract An automated, ground-based instrument was used to measure gas-phase hydroper- oxides at the Jungfraujoch High Altitude Research Station as part of the Free Tro- pospheric EXperiment (FREETEX) during February/March 2003. A nebulising reflux 5 concentrator sampled ambient air twice hourly, prior to on-site analysis by HPLC speci- ation, coupled with post-column peroxidase derivatisation and fluorescence detection. Hydrogen peroxide (H2O2) concentrations reached up to 1420 pptv over the 13-day period with a mean of 206±261 pptv (± one standard deviation). Methyl hydroperox- ide (CH3OOH) reached up to 921 pptv with a mean of 76±96 pptv. No other organic 10 hydroperoxides were detected. The lack of an explicit diurnal cycle suggests that hy- droperoxide concentrations are chiefly influenced by transport processes rather than local photochemistry at this mountainous site. We find elevated concentrations of H2O2 in air masses originating from the south-west indicative of higher concentrations of HOx due to more active photochemistry. Air which has been recently polluted exhibits low 15 H2O2 concentration due to a combination of suppression of HO2 by NOx and depo- sition. We also conclude that despite being at a high alpine site, the vast majority of the air observed was extensively influence by the boundary layer during our campaign (diagnosed from high CO concentrations and the high NOx to NOy ratio) resulting in deposition of H2O2 to the surface and hence reduced H2O2 concentrations. The con- 20 centrations of H2O2 sampled here are consistent with previous box modelling studies of hydroperoxides which invoked a depositional sink. 1 Introduction Hydroperoxides play an important role in gas-phase free radical chemistry of the at- mosphere and in the aqueous-phase chemistry of acid precipitation. Hydroperoxides 25 such as hydrogen peroxide (H2O2) and methyl hydroperoxide (CH3OOH) are produced through the self-reaction of peroxy radicals (hydroperoxy, HO2 and organic peroxy, 7178 RO2) (e.g. Lee et al., 2000). These peroxy radicals are intimately linked to ozone pro- duction and loss, leading to hydroperoxides being a key diagnostic of the chemical state of the atmosphere. The chemistry of the Ox, HOx and ROx is summarized in Fig. 1. H2O2 is also a major oxidant of SO2 in clouds (Penkett et al., 1979) so along- 5 side O3 and OH, its measurement is needed to assess the oxidative capacity of the atmosphere. Previous analyses of H2O2 and CH3OOH observations have identified key physical and chemical processes controlling their concentration (e.g. Heikes et al., 1987). As shown in Fig. 1, the production of H2O2 is controlled by the abundance of HO2, which 10 shows sensitivity to HOx (OH and HO2) production, loss and recycling. Formation of peroxy radicals is predominantly through the photo-oxidation of carbon monoxide (CO) and volatile organic compounds (VOC), by the OH radical (detailed in Lightfoot et al., 1992). A second significant source is from formaldehyde (CH2O), which undergoes photolysis and reaction with OH to produce HO2. 15 Air-masses rich in NOx tend to show lower H2O2 concentrations as peroxy radicals oxidise NO to NO2 rather than self-reacting to form H2O2 (Tremmel et al., 1993; Penkett et al., 1995) and total radical loadings are suppressed by HNO3 production through the reaction of OH with NO2 radicals (Poppe et al., 1993). According to Lee et al. (2000), substantial suppression of hydroperoxide production occurs at NO concentrations ex- 20 ceeding 100 pptv. In contrast, it is calculated that NO concentrations below 3 to 20 pptv are needed for hydroperoxide production to dominate (Reeves and Penkett, 2003; Crutzen and Zimmermann, 1991; Finlayson-Pitts and Pitts, 1986). Such low concen- trations of NO can only be found in very remote regions of the troposphere lacking NOx sources. 25 Lower concentrations of H2O2 are observed in air masses exposed to the ground or clouds due to dry and wet deposition (e.g. Kleinman, 1986; Heikes et al., 1987; Chandler et al., 1988; Gallagher et al., 1991). Rapid loss of gaseous H2O2 into clouds 4 −1 is due to a high Henry’s law coefficient of H2O2(HH2O2 is 7.73×10 M atm at 298 K, Sander et al., 2003) and rapid oxidation of S(IV) species to S(VI) within the aqueous 7179 phase. Other important H2O2 and CH3OOH sinks are their reaction with OH radicals and photolysis at ultraviolet wavelengths generating OH and in the case of CH3OOH, OH and CH3O. The aim of this paper is to investigate the impact of these chemical and physical pro- 5 cesses on the concentration of H2O2 and CH3OOH at a remote mountainous site and to understand the usefulness of H2O2 observations as a diagnostic of the atmospheric compositional system. 2 Experimental 2.1 Site description 10 The observations are from the FREE Tropospheric EXperiment (FREETEX) that took place in February to March 2003 at the Sphinx Observatory, Jungfraujoch High Alti- tude Research Station (7.98◦ E, 46.55◦ N) (subsequently referred to as JFJ) situated at 3580 m above mean sea level (a.m.s.l.). There have been three previous FREE- TEX campaigns in 1996, 1998 and 2001 at this site (Zanis et al., 1999, 2000a, b, 15 2003; Carpenter et al., 2000). The Sphinx Observatory resides in a saddle between the two alpine summits, Jungfrau (4155 m a.m.s.l.) to the south-west and Monch¨ (4099 m a.m.s.l.) to the north-east, in the Swiss Alps. This leads to mainly north- westerly and southerly air masses reaching the station. The site characteristics are described in Zellweger et al. (2003). The observatory is usually thought to be located in 20 the lower free troposphere (FT) in winter and early spring but can experience planetary boundary layer (PBL) air (Carpenter et al., 2000), especially in the summer when con- vection is enhanced (Baltensperger et al., 1997; Lugauer et al., 1998, Zellweger et al., 2003). The high altitude site is thought to experience few local emission sources, and therefore can offer ideal conditions for long-term sampling of free tropospheric air. The 25 JFJ is incorporated in the Swiss National Air Pollution Monitoring Network (NABEL), which is maintained by EMPA Dubendorf¨ on behalf of the Swiss Federal Office for the 7180 Environment (FOEN). Among other species, NO, NO2,O3 and CO are measured rou- tinely and are available with a time resolution of 30 min. All meteorological parameters are measured by the Federal Office of Meteorology and Climatology (MeteoSwiss). Due to the importance of monitoring long-term trends of gaseous and aerosol param- 5 eters in the free troposphere, the JFJ station has been incorporated into the Global Atmosphere Watch (GAW) program of the World Meteorological Organization (WMO) as a Global GAW Station. 2.2 Method for hydroperoxide sample collection and analysis Sample collection took place at the Sphinx laboratory between the 27 February and 10 the 12 March 2003. Figure 2 shows a schematic diagram of the analytical system. Samples were obtained from the gas-phase using an amber-glass nebulising reflux concentrator (Fig. 2) based on the design by Cofer et al. (1985). Atmospheric air was drawn by a vacuum pump from the NABEL inlet (previously described in Zellweger et 1 al., 2000), through ∼5 m /4” OD Teflon (perfluoroalkoxy) tubing into the nebuliser, at a −1 15 flow rate of 2.7 l min . The residence time in the inlet tubing was less than 4 sec. A sampling time of 25 min provided sufficient concentration of the solution for analysis. Collection efficiencies of 75±3% and 40±2% have been previously determined experi- mentally for H2O2 and CH3OOH respectively and were assumed in this study (Jackson and Hewitt, 1996; Morgan, 2004). 20 Once collected, aqueous samples were injected through a 0.2 µm filter onto the HPLC where hydroperoxide speciation was achieved using a C-18 reversed-phase column. Helium-degassed mobile phase was supplied at a flow rate of 0.6 ml min−1. Post-column, the hydroperoxide components were derivatised in a reaction coil in the presence of horseradish peroxidase to form a stable dimer, based on the technique de- 25 veloped by Lazrus et al. (1985, 1986). The derivatisation reagent was raised to pH 5.8 by the addition of potassium hydroxide (1 M) and buffered by potassium hydrogen ph- thalate (0.5 M). A peristaltic pump delivered the derivatisation reagent at constant rate of 0.25 ml min−1. The pH was then raised above 10 to produce the fluorescent anionic 7181 dimer form via a Nafion membrane submerged in 100 ml NH4OH (30%). Detection was achieved using a fluorescence spectrophotometer with excitation and emission wavelengths of 305 and 410 nm respectively, via a xenon short arc lamp. Calibration of the system was carried out daily. Further details of this method can be found in Mor- 5 gan and Jackson (2002). Based on three and ten times the standard deviation of the base line noise, the instrument limit of detection and quantification at 95% confidence (Taylor, 1987) during the campaign was found to be 21 and 68 pptv respectively.