Observations of high concentrations of I2 and IO in coastal air supporting iodine-oxide driven coastal new particle formation

Authors


Abstract

[1] Theoretical studies have predicted that concentrations of gaseous I2 and IO of the order of 80–100 ppt and 40–50 ppt, respectively, are required in coastal air to account for photochemically-driven coastal new-particle formation events to occur. However, measurements reported to date (i.e., ∼20 ppt I2, ≤ 10 ppt IO) have not supported the required model predictions. Here, we present measurements of high concentrations of I2 and IO in N.E. Atlantic marine air on the west coast of Ireland. The maximum mixing ratios of daytime I2 and IO over the seaweed beds during low tide were 302 ppt and 35 ppt, respectively. The I2 distribution was rather inhomogeneous, even at the inter-tidal zone, but closely related to the macroalgae biomass abundance. New particle formation bursts were frequently observed during daytime hours with the concentrations up to 4.5 × 105 particles cm−3 during low-tide conditions, and the concentrations of ultra-fine particles were positively correlated with the IO concentrations. Considering the constraints set out in theoretical studies for new particle formation via condensation of condensable iodine oxide vapours, the results reported here clearly demonstrate that the molecular iodine and iodine monoxide concentrations in coastal air are sufficient to meet the theoretical precursor concentrations required to drive intensive coastal new-particle formation from higher order condensable iodine oxides.

1. Introduction

[2] The impact of iodine chemistry in the lower troposphere is presently a subject of considerable interest. Numerous studies [e.g., von Glasow and Crutzen, 2007] have concentrated on the potential of reactive iodine to affect the oxidising capacity of the marine boundary layer (MBL) in a variety of ways: catalytic destruction of O3, altering the partitioning of NOx (NO2/NO) and HOx (HO2/OH), and activating chlorine and bromine from sea-salt aerosol. Iodine may also play an important role in O3 and Hg depletion episodes in the polar boundary layer. Recent field measurements and modeling studies indicate that these atmospheric processes could be enhanced with elevated tropospheric iodine level [Saiz-Lopez et al., 2008; Read et al., 2008].

[3] Since molecular iodine (I2) was suggested to be the dominant source of coastal reactive iodine in the MBL during the 2002 NAMBLEX field campaign at Mace Head (Ireland) [Saiz-Lopez and Plane, 2004], it has become well-established by laboratory studies that the emission of this molecular iodine from low tidal macroalgal exposure is indeed a most important process responsible for the observed tropospheric iodine level [McFiggans et al., 2004; Küpper et al., 2008]. Significant levels of molecular iodine have been observed at Mace Head and nearby sites (∼100 parts per trillion (ppt)) [Saiz-Lopez and Plane, 2004; Saiz-Lopez et al., 2006; Peters et al., 2005], and at La Jolla, California (∼8 ppt) [Finley and Saltzman, 2008]. Although large-scale I2 observations have not been realized so far due to the lack of analytical techniques to accurately measure this species, the measurements of iodine monoxide (IO) radicals, an unambiguous product of iodine atom reaction with ozone, have been carried out at various coastal sites and the polar regions by ground-based approaches and satellite [e.g., Peters et al., 2005; Whalley et al., 2007; Schönhardt et al., 2008].

[4] The formation of ultra-fine particles in the coastal atmosphere has been observed frequently [O’Dowd et al., 1998, 2002; O’Dowd and Hoffmann, 2005]. Positive correlations between seaweed mass, I2 emissions and the resulting particle concentrations have been elucidated from chamber experiments [Sellegri et al., 2005]. Also the role of iodine oxides in driving coastal ultra-fine particle formation has been widely investigated [e.g., Hoffmann et al., 2001; McFiggans et al., 2004]. This formation process is thought to involve the production of IO via the reaction of O3 with iodine atoms that are generated by photolysis of I2 and the self-reaction of IO to yield higher oxides (I2Oy, where y = 2–5), which finally nucleate and form new particles. Therefore, the presence of IO in the atmosphere points to the possibility of new particle formation. However, model calculations [Burkholder et al., 2004] demonstrate that IO concentrations reported in recent field measurements using long-path differential optical absorption spectroscopy (LP-DOAS) are not sufficient to account for the significant aerosol production observed in the coastal MBL, mostly because the data derived from DOAS measurements are calculated assuming a constant concentration along the light path of several kilometers which consequently can not reveal the true concentration level in the inter-tidal zone of a fraction of a kilometer. The same is true for I2 measurements which were mostly carried out by LP-DOAS, leading to potential underestimation of the concentration levels of I2 in the inter-tidal zone. Although broadband cavity ring-down spectrometer (BBCRDS) can provide “single-point” in situ measurements of I2 [Saiz-Lopez et al., 2006], the limits of detection of these spectrometric methods are in general relatively high, which makes the accurate quantification of daytime I2 still difficult and therefore obscures our understanding on the atmospheric processes of iodine, since models have predicted that even very small amount of I2 under daytime conditions could strongly affect the marine atmosphere [Peters et al., 2005].

[5] While most of the above studies have been located at Mace Head, it should be noted that Mace Head possesses perhaps the lowest algal biomass in the region as most of the shore is characterized by bare rocks rather than notable kelp beds. This study extends the studies at Mace Head to simultaneous measurements in a nearby algal biomass hotspot, Mweenish Bay. During a campaign in August and September, 2007, elevated concentration levels of I2 and IO were observed, particularly over the exposed seaweed beds. This result provides substantial support to the strong iodine-containing new particle burst observed in the coastal MBL.

2. Field Measurements

[6] The field measurements of I2 and IO were made at the Mace Head Atmospheric Research Station (53.25° N, 9.80° W) and Mweenish Bay (53.32° N, 9.73° W), in N.E. Atlantic. The later is located about 7 km southeast of the Mace Head research station.

[7] Molecular iodine was measured by a diffusion denuder system in combination with a gas chromatography–mass spectrometry (GC–MS) method, which provides “single-point” in situ concentrations of I2 at the sampling site. Details of this recently developed denuder/GC–MS method are given by Huang and Hoffmann [2009] and therefore it is only briefly described here. Ambient I2 samples were taken by brown denuder tubes (6 mm i.d. × 50 cm length) which were uniformly coated with 11.25 mg α-cyclodextrin (α-CD) and trace 129I, at a flow rate of 500 mL min−1 for 60–180 min. The interference iodine species such as ICl and HOI were isolated by coupling a 1,3,5-trimethoxybenzene-coated denuder upstream of the α-CD/129Icoated denuder (data not shown here). After sampling, the open ends of the denuders were again sealed with PP end-caps and kept under refrigeration until subsequent laboratory measurements. In the laboratory, the samples were eluted with five 2.0 mL-portions of ultrapure water into a 25-mL flask in which amounts of 500 μL of phosphate buffer (pH 6.4), 100 μL of 2,4,6-tribromoaniline (2.5 mg L−1, internal standard), 400 μL of sodium 2-iodosobenzoate, and 300 μL of N,N-dimethylaniline were added. Subsequently, the solution was shaken at room temperature for about 120 min, leading to a complete derivatization of I2 into 4-iodo-N,N-dimethylaniline. Finally, the solution was extracted with 100 μL of cyclohexane. 1.0 μL of the extraction solution was injected into a GC–MS system (Trace GC/PolarisQ, Thermo Finnigan, Italy). A Rtx-5MS fused-silica capillary column (Restek Co., Bad Homburg, Germany) was used for chromatographic separation, and the MS was run in the selected ion monitoring (SIM) mode to enhance the sensitivity of measurements. The detection limit of the method is below 0.1 ppt for 90 L sample volume, and the collection efficiency is greater than 98%.

[8] IO was measured by two active LP-DOAS systems [Platt and Stutz, 2008]. Details are given by Seitz et al. [2009], thus only a brief description will be given here. The almost parallel light beam of a high pressure Xe-arc lamp (type: XBO 500, Osram), was sent through the open atmosphere to an array of quartz prism retro-reflectors (63 mm diameter each) and then reflected back to the receiving optics of the telescope. At Mace Head the light-path (6.8 km, one-way, 13.6 km total length) crossed Roundstone Bay to Roundstone, about 5–10 m above sea level at high tide, where the reflector consisting of 76 quartz prisms was located. At Mweenish Bay a light-path (2.0 km, one-way, 39 quartz prisms) crossing Mweenish Bay to Finish Island was established. This light-path crossed the sea at about 5–10 m above sea level at high tide. The reflected light was analyzed by a spectrometer (Acton Spectra Pro 300, f = 4.1 equipped with a 1900 gr mm−1 grating and Acton Spectra Pro 500, f = 6.9, 600 gr mm−1 grating for the measurements at Mace Head and Mweenish Bay, respectively). In both cases the detector used was a 1024 pixel photodiode array (type: S3904-1024, Hamamatsu). Spectra of the atmospheric light beam, scattered sunlight, and the Xe-arc lamp, were recorded. For each wavelength range an atmospheric measurement spectrum with a time resolution of 30 sec and signal integration of about 15 min (a maximum of 30 co-added spectra) was taken. IO was measured in the 430 ± 40 nm wavelength regions. For the analysis of the LP-DOAS data the software DOASIS was used to simultaneously fit the different references to the atmospheric spectrum using a non-linear least-squares method (for details see Platt and Stutz [2008]). The spectra were analyzed for IO absorption structures in the wavelength range between 416 and 439 nm including references of NO2 and H2O in the fitting procedure. The averaged IO detection limits are 0.6 ppt at Mace Head and 2.1 ppt at Mweenish Bay,.

[9] At Mweenish Bay, particle measurements were carried out by a nano scanning mobility particle sizer (nano-SMPS), covering sizes from 3 nm to 20 nm, and a standard SMPS, covering sizes from 10 nm to 100 nm. Both the nano-SMPS and SMPS were standard Thermo Systems Inc. (TSI) systems [Wang and Flagan, 1990], with the nano-SMPS using the TSI 3025a condensation particle counter (CPC) as a detector and the SMPS using a TSI 3010 CPC as a detector. The instruments were located about 150 m away from the low tide region and sampling was conducted down a 3 m long, equation image″ stainless steel inlet. Size resolved concentrations were corrected for diffusional losses, which were calculated based on tube diameter and residence time in sample tube. Equations were taken from Seinfeld and Pandis [2006].

3. Results and Discussion

[10] “Single-point” in situ measurements of littoral I2 were made by a coupled denuder system [Huang and Hoffmann, 2009]. Compared to our previous starch-coated denuder system [Sellegri et al., 2005; Saiz-Lopez et al., 2006] this newly developed method provides better collection efficiency by the use of more effective coating materials and better accuracy by minimizing the interference iodine species. However, it should still be noted that the sampling time of denuder used in this work is relatively long (ranged from 60 to 180 min) which does not allow measurements at the time scales at which the photolysis of I2 occurs. Therefore, the peak values of I2 might actually be larger than those reported here.

[11] At Mweenish Bay, the daytime in situ measurements taken exactly over the central zone of seaweed beds during low tide show very high mixing ratios of I2, ranging from 110.3 to 301.8 ppt with an average of 186.3 ppt. Since the inlet of the denuder was set up very close to the seaweed beds (∼5 cm) during sampling it is reasonable to treat these data as local source strength. When the measurements were carried out downwind of the seaweed beds with a distance of about 100–150 m, the daytime concentration decreased to about 36.7 ppt (ranging from 14.9 to 87.2 ppt, also partly shown in Figure 1a). This significant decrease of I2 mixing ratio observed downwind of (i.e., further away from) the seaweed beds could be attributed to the rapid photolysis of I2 that has a photolytic lifetime of 10 s [Saiz-Lopez et al., 2006], which is further supported by the large particle plumes observed during the advection crossing 100–150 m, and to the concentration dilution effect during the air mass transport from seaweed beds to the sampling site. Note that during these measurements the wind (from sea direction) passed over the seaweed beds with a speed of 10.5–15 m s−1 corresponding to a transport time of about 7–14 seconds. However, the averaged night-time I2 mixing ratio downwind of the seaweed beds (∼124 ppt) is comparable to the levels found directly above the seaweed beds (source strength) although the values drop to around 30 ppt in several episodes. This result indicates that I2 mixing ratio will reach a steady state in the coastal air at night, which could be responsible for the pulse of IO that typically occurred after dawn [Saiz-Lopez and Plane, 2004].

Figure 1.

Comparison measurements of (a) littoral I2 and (b) spatial IO between Mweenish Bay and Mace Head. Note that I2 samples were simultaneously taken in the intertidal zone about 100–150 m away from the algae beds during daytime and night-time low tide in both locations. The time (UTC) present above the column indicates the sampling period of each sample and the error bars refer to triplicate measurements of the same sample by GC–MS. Also note that only the IO values above the individual detection limits are presented here. The averaged detection limits are 0.6 ppt at Mace Head and 2.1 ppt at Mweenish Bay, respectively. Elevated IO levels were observed during low tide with intense solar irradiation. Also the IO mixing ratios were positively correlated to the concentrations of I2.

[12] I2 measurements were also simultaneously carried out at Mace Head. The selected sampling site was about 100–150 m away from the seaweed beds. The averaged daytime I2 mixing ratio (∼23 ppt) is comparable to the reported average concentration (∼26 ppt) by BBCRDS at the same site [Saiz-Lopez et al., 2006]. However, the mixing ratios during both daytime and night-time are significantly lower at Mace Head than at Mweenish Bay (see Figure 1a). This difference could be attributed to the higher biomass density at Mweenish Bay than at Mace Head and the difference in algal species composition at these two sites, i.e., dominated by Laminaria digitata, Laminaria hyperborean, Himanthalia elongata, Palmaria palmata and Alaria esculenta at Mace Head and Ascophyllum nodosum and Fucus vesiculosus at Mweenish bay [Irish Seaweed Centre, 2001], since chamber experiments showed that the emissions of I2 are related to the seaweed species and biomass [Sellegri et al., 2005].

[13] Comparison measurements of IO were made between Mweenish Bay and Mace Head by LP-DOAS. At Mace Head, the mixing ratios of IO observed are consistent with previous measurements taken during the 1998 PARFORCE [Carpenter et al., 2001] and 2002 NAMBLEX field campaigns [Saiz-Lopez and Plane, 2004]. At Mweenish Bay, IO was measured for the first time, and the mixing ratios observed were strikingly higher when compared to that at Mace Head. As shown in Figure 1b, the maximum IO mixing ratios occur at low tide during the period of high solar irradiation, and the daily IO maximum decreases gradually with the increase of the minimum in tidal height (correlation coefficient r2 = 0.68). This diurnal profile is well consistent with the characteristics of I2 emission at the local shoreline. When elevated I2 concentrations occurred on August 29 and 30, higher IO mixing ratios were observed correspondingly (see Figures 1a and 1b). Obviously, enhanced algal exposure leads to elevated I2 emissions and consequently higher concentrations of IO. This result is further supported by additional measurements at Mweenish Bay where IO mixing ratios of up to 35.0 ± 7.7 ppt were observed by a very short DOAS light-path (500 m, one-way) which covered exclusively the exposed seaweed beds at low tide [Seitz et al., 2009]. It is noted that this maximum IO concentration is very close to the littoral level (27.6 ± 3.2 ppt) at Roscoff (Brittany, France) measured by the “sing-point” in situ technique of cavity ring-down spectroscopy [Wada et al., 2007] and laser induced fluorescence spectroscopy [Whalley et al., 2007]. It may therefore be inferred that the actual IO concentration could be 5–10 times higher than that reported from LP-DOAS measurements of several kilometers extending over open sea in the coastal atmosphere at Mweenish Bay and nearby locations.

[14] The elevated concentration levels of I2 and IO are clearly correlated to the observation of strong new-particle bursts in 14 out of 23 days of measurements (17th August–8th September) at Mweenish Bay. Figure 2a illustrates the particle size distribution spectrum measured on 30th August. The ultra-fine particle bursts take place during daytime low tide and last for about 4–6 h, which is closely related to the diurnal variation of the seaweed exposure period. The concentrations of ultra-fine particles (D < 7 nm) can be as high as 4.5 × 105 particles cm−3 during the lowest water (spring tide) and are positively correlated to the measured IO column density, as shown in Figure 2b. Note, in the correlation presented, the IO time scale is shifted by ∼15 minutes to allow advection time. Although the shift may be excessive for a transit of 150 m, it was the minimum time scale that could be shifted due to instrument temporal resolution. The correlation coefficient r2 = 0.66, suggesting a significant dependency of particle concentration less than 7 nm on IO vapour concentrations.

Figure 2.

(a) The particle size distribution of 30th September observed at Mweenish Bay, and (b) the correlation between ultra-fine particle concentrations and IO column density during the period of particle bursts (spring low tide). Note that the IO time scale is shifted by ∼15 minutes to compensate for the advection time.

[15] In summary, the present work shows that elevated concentration of I2 can occur over the seaweed beds during low tide under atmospheric conditions. The source strength is closely related to the seaweed species and their biomass. Shortly beyond the seaweed beds I2 concentrations decrease significantly during daytime, indicating rather inhomogeneous I2 distributions even in the inter-tidal zone. Elevated IO mixing ratios are also observed over the seaweed beds, which drives the enhanced new particle bursts at the concentration level of ∼105 particles cm−3. These results clearly demonstrate that the concentration levels of I2 and IO in coastal air are sufficient to meet the theoretical precursor concentrations required to drive intensive coastal new-particle formation from higher order condensable iodine oxides. From the correlations established in the current work in combination with previous observations of new particle bursts at the west coastline of Ireland [Sellegri et al., 2005; O’Dowd et al., 2007] and at Brittany (France) [Whitehead et al., 2009] as well as with the laboratory studies [O’Dowd et al., 2002; McFiggans et al., 2004; Küpper et al., 2008], it may be inferred that tidally driven new particle formation events could be widespread over the algae-rich coastline.

Acknowledgments

[16] This study was supported by the European Union under the FP6 project (contract 018332): Marine Aerosol Production (MAP), the European Integrated project on Aerosol Cloud Climate and Air Quality Interactions (EUCAARI) (contract 36833), and the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) within the graduate program 826 “Trace Analysis of Elemental Species: Development of Methods and Applications”. The authors are grateful to Harald Berresheim (NUI, Galway) and the Martin Ryan Institute, especially Richard Fitzgerald, for the support during the field measurements.

Ancillary