Speciation of 127I and 129I in atmospheric aerosols at Risø, Denmark: insight into sources of iodine isotopes and their species transformations

. Speciation analysis of iodine in aerosols is a very useful approach for understanding geochemical cycling of iodine in the atmosphere. In this study, overall iodine species, including water-soluble iodine species (iodide, iodate and water-soluble organic iodine), NaOH-soluble iodine, and insoluble iodine have been determined for 129 I and 127 I in the aerosols collected at Risø, Denmark, during March and May 2011 (shortly after the Fukushima nuclear accident) and in December 2014. The measured concentrations of total iodine are in the range of 1.04–2.48 ng m − 3 for 127 I and (11.3– 97.0) × 10 5 atoms m − 3 for 129 I, corresponding to 129 I / 127 I atomic ratios of (17.8–86.8) × 10 − 8 . The contribution of Fukushima-derived 129 I (peak value of 6.3 × 10 4 atoms m − 3 ) is

contain a significant fraction of iodine, and play an important role for iodine transportation from ocean to land. In particular, speciation analysis of iodine in aerosols is one of critical issues to understand the reactions and atmospheric processes of iodine, and develop a comprehensive understanding of biogeochemical cycles of iodine in nature.
Most of the work on aerosol iodine speciation has focused on water-soluble iodine (WSI) (Baker, 2004(Baker, , 2005Gilfedder et al., 2008). Early modelling studies suggested that aerosol iodate should be the only stable iodine specie as a by-product of iodine oxides production, with negligible iodide due to transformation into gaseous iodine (McFiggans et al., 2000;Vogt et al., 1999). However, this prediction went against several previous observations that showed significant iodide and soluble organic iodine, accounting for 10-100 % of WSI in aerosols (Baker, 2004(Baker, , 2005Gabler and Heumann, 1993). An improved model was then developed, primarily by increasing the proportion of iodide in the aerosols (Pechtl et al., 2007). Few studies have addressed water-insoluble iodine in aerosols (Gilfedder et al., 2010;Tsukada et al., 1987), which might not be easily converted to gaseous iodine before being deposited on the Earth's surface. Previous observational data suggest that water-insoluble iodine is abundant in aerosol particles, representing 17-53 % of the total iodine (Gilfedder et al., 2010). However, there is no comprehensive investigation of aerosol iodine species that encompasses aqueoussoluble and insoluble iodine species. This constitutes a significant gap in our understanding of the interaction among atmospheric iodine species.
Due to the difficulty of source identification of natural 127 I, 129 I (T 1/2 = 15.7 Ma) has proven to be an especially useful geochemical tracer both in the marine environment and in the atmosphere (He et al., 2013;Hou et al., 2007;Jabbar et al., 2012;Michel et al., 2012). A large amount of 129 I has been released to the environment by nuclear reprocessing plants (NRPs). These dominate all other anthropogenic sources, with about 6000 kg released as of 2009, from Sellafield (United Kingdom) and La Hague (France) Raisbeck et al., 1995). Anthropogenic 129 I is currently distributed unevenly around the world in water, atmosphere, and terrestrial systems (Hou et al., 2009b). Above-ground nuclear explosions also produced 129 I, injecting it into the troposphere and stratosphere. The proportion that entered to the stratosphere became mixed globally before re-entering the troposphere. From the troposphere, about one half of the radioactive particles is deposited onto the earth's surface within 20 days (Tölgyessy, 1993). Gaseous releases from reprocessing plants and accidents are primary sources of 129 I to the local environment (Hou et al., 2009a;Xu et al., 2013). Reemission of iodine as the forms of inorganic and organic iodine (e.g. I 2 , CH 3 I, C 2 H 5 I) from sea surface plays a significant role in the atmospheric budget of 129 I (Englund et al., 2010b). Gaseous iodine compounds photodissociate rapidly in the atmosphere to generate iodine atoms. Iodine atoms are oxidized by atmospheric oxidants (e.g. O 3 , NO x ) to iodine Risø, Denmark for aerosol collection. The two nuclear reprocessing plants (radioactivity labels), Sellafield (United Kingdom) and La Hague (France) have released a large quantity of 129 I to the marine and atmospheric environment. This anthropogenic 129 I discharged to marine systems has been transported to the North Sea, Baltic Sea, Norwegian Sea, and beyond, by ocean currents (red lines). Yellow and blue arrows show the westerly wind and easterly wind, respectively.
oxides, which then incorporate with particles either by nucleation or by reaction and/or trap mechanisms (Saiz-Lopez et al., 2012). As with stable iodine ( 127 I), 129 I also disperses in the atmosphere in both gaseous and particulate forms. Particulate 129 I has been used successfully in monitoring studies, and as an atmospheric tracer (Englund et al., 2010b;Jabbar et al., 2012;Santos et al., 2005;Tsukada et al., 1987). While, speciation analysis of 129 I in aerosols is extremely scarce with the only one being our previous study .
Here, we present the results of speciation analysis of stable 127 I and radioactive 129 I in aerosols collected in Denmark, in order to understand how the iodine is carried in the atmosphere and to investigate transformation mechanisms between iodine species in aerosols, as well as gas-aerosol exchange processes involving iodine. To our knowledge, this is the first paper to address the conversion among overall iodine species in aerosols.

Aerosol sampling
Aerosol samples were collected on 0.45 µm polypropylene filters (Type G-3, PTI, Germany), attached to an inhouse aerosol collector at Risø campus, Technical University of Denmark, Denmark (55 • 41.77 N, 12 • 05.39 E) (Fig. 1) Figure 2. Schematic diagram of the analytical procedure for the determination of 127 I and 129 I species in aerosols (adapted from . TI for total iodine, WSI for water-soluble iodine, NSI for NaOH-soluble iodine and RII for residual-insoluble iodine. during the Fukushima nuclear accident, the changing frequency increased to every 3-4 days. Seven aerosol samples were collected from 31 March to 2 May 2011 (local time, the same below), shortly after the Fukushima accident on 11 March 2011. One sample was collected from 8 to 15 December 2014. The samples were put into plastic bags and stored in the dark before analysis. The sample information is listed in Table 1.

Speciation analysis of 127 I and 129 I in aerosols
Separation of iodine species from aerosol. The aerosol samples were analyzed for species of 127 I and 129 I using a newly developed method . In brief, iodine was extracted sequentially using deionized water and sodium hydroxide solutions for water-soluble and NaOH-soluble iodine (WSI and NSI), respectively. Total iodine (TI) and residual insoluble iodine (RII) were separated by alkaline ashing from the original air filters and the residual filters after NaOH solution leaching, respectively. Iodide and iodate in the water leachate were separated using anion exchange chromatography. Figure 2 shows a diagram of the separation procedure for speciation analysis of iodine isotopes. Determination of 127 I by ICP-MS and 129 I by AMS. 127 I in all fractions were diluted by factors from 1 to 20 using ammonium depending on the iodine concentration in each fraction and the matrix of the leachates. Cesium was added as internal standard to a final concentration of 2 ng g −1 to normalize measurement efficiency of iodine in the ICP-MS measurement. Standard addition method was employed for the measurement of 127 I in ash leachates of total iodine and the insoluble fraction. An iodine standard solution (NaI, CPI International, California, USA) was spiked into the reserved aliquots and diluted with 1 % NH 3 q H 2 O to a final concentration of 2 ng g −1 . 127 I in the diluted solution was measured by ICP-MS (Thermo Fisher, X Series II) using an Xt cone under normal mode, as described elsewhere . 129 I was measured using a 5 MV accelerator mass spectrometry (NEC, Wisconsin, USA) at Scottish University Environmental Research Center, UK (Xu et al., 2013). The prepared AgI precipitates were mixed with silver powder (Sigma-Aldrich Co., USA) with an AgI : Ag mass ratio of 1 : 2 and pressed into copper holders. A terminal voltage of 3.5 MV and I 5+ ions were chosen for 129 I detection. The measured 129 I / 127 I ratios were corrected against a standard with a 129 I / 127 I ratio of 1.138 × 10 −10 , prepared from a 129 I standard solution (NIST 4949C) and an 127 I carrier. The analytical precision of standards and samples are within 5 %. 129 I / 127 I ratios in the procedural blanks were measured to be about 5 × 10 −13 , 1-3 orders of magnitude lower than those measured for the samples. All results were corrected for procedural blanks.

Results
The concentrations of total iodine in aerosols from Risø, Denmark, ranged at 1.04-2.48 ng m −3 (average of 1.8 ± 0.5 ng m −3 ) for 127 I and 11.3-72.9 × 10 5 atoms m −3 (average of (44 ± 19) × 10 5 atoms m −3 ) for 129 I, during March-May 2011 (Table 2). Total 127 I concentration of 2.36 ng m −3 during 8-15 December 2014 fell within the range in 2011, while 129 I concentration of 97.0 × 10 5 atoms m −3 was about two times higher than the average value in 2011. Much lower values of 129 I and 127 I concentrations were observed during 18 April-2 May 2011 compared to those before 18 April (Fig. 3), but 129 I levels did not show a synchronous variation with 127 I concentrations.
The results of 127 I concentrations in the studied aerosols are compatible with those from an inland city, Regensburg, Germany and from some Pacific islands (Gabler and   1993, and references therein). 129 I / 127 I atomic ratios in the investigated aerosols were (17.8-86.8) × 10 −8 , which are consistent with those found at Foehr island, north of Germany in April 2002  and in southern Sweden during 1983-2008 (Englund et al., 2010b;. However, the measured 129 I concentrations and 129 I / 127 I ratios are 1-2 orders of magnitude higher than those observed in Vienna, Austria during 2001-2002, and in the high altitude eastern Alps (2700 m) during 2001 (Jabbar et al., 2011(Jabbar et al., , 2012. For 127 I, the WSI fraction of aerosols collected in 2011 occurred predominantly as iodide, ranging from 0.12 to 0.33 ng m −3 . Minor iodate of 0.02-0.03 ng m −3 was detected only in two samples (AE11-2 during 4-7 April and AE11-7 during 26 April-2 May) ( Table 1 and Fig. 4). The highest measured 127 I − concentration (0.74 ng m −3 ) was observed in aerosol collected in 2014, and there was no measurable water-soluble organic iodine. Watersoluble 129 I shows a similar species pattern as water-soluble 127 I except that iodate-129 was below the detection limit in all the analyzed samples. The concentrations of 129 I − varied in (3.3-5.9) × 10 5 atoms m −3 with an average value of 4.4 × 10 5 atoms m −3 in the aerosols from 2011. This value is about 8 times lower than that in the sample collected in 2014 (30.1 × 10 5 atoms m −3 ). High NSI concentrations were observed in all of the aerosol samples, with an average of 0.64 ± 0.21 ng m −3 for 127 I and (13.6 ± 10.1) × 10 5 atoms m −3 for 129 I. RII is the dominant species in aerosol samples for both iodine isotopes, with concentration range of 0.34-1.66 ng m −3 for 127 I and (4.3-39.9) × 10 5 atoms m −3 for 129 I.
There is no available standard reference material for iodine isotopes in aerosols which can be used for validation of the analytical method. While, the sums of all of iodine species in the analyzed samples are comparable with the measured total iodine in the corresponding samples for both 127 I and 129 I in a range of 85-110 %, indicating the reliability of our analytical results (Fig. 5).  For the samples collected before 18 April 2011, the proportion of 129 I and 127 I species follows an order of RII > NSI > iodide. For the samples collected later in April 2011 and in the winter of 2014, the three iodine fractions for both 129 I and 127 I account for almost equivalent proportion of total iodine (Fig. 5). Compared to the former sampling period (31 March-18 April 2011), the most apparent feature of the latter sampling period is a dramatic diminution of iodine in the RII fraction, by 38.7 % for 127 I and 26.8 % for 129 I. This was compensated by a corresponding increase in the iodide fraction of the aerosol particles. Iodate, as the least abundant iodine species (< 3 %), was only observed in two aerosol samples for 127 I, and no 129 IO − 3 was detectable in all aerosol samples investigated in this work. 129 I / 127 I atomic ratios of (15.6-102.0) × 10 −8 were determined for the different iodine species in the aerosols. The variations of 129 I / 127 I ratio in iodide, NSI, and RII are similar to those in total 129 I / 127 I, with high values observed prior to 18 April 2011 and December 2014 (Table 2 and Fig. 3).

Sources of 127 I and 129 I in aerosols
Variation of 127 I and 129 I concentrations in aerosols against meteorological parameters (i.e. wind direction, wind speed, and temperature) during the sampling period shows that wind direction has a principle influence (Fig. 3). Back trajectory model analysis shows that 127 I and 129 I concentrations in the aerosols were directly controlled by the sources and pathways of air masses (Figs. 6, S1 and S2 in the Supplement). The relatively high 127 I and 129 I concentrations were observed in the aerosols collected in early April 2011, when air masses were mainly transported from the Atlantic Ocean across the North Sea by prevailing westerly winds. While, relatively low concentrations of iodine isotopes were observed in the aerosols collected in later April, when air masses were dominated by prevailing easterly winds and passed over the European continent and the Baltic Sea (Figs. 6 and S1). Marine emission of volatile iodine species (e.g. inorganic I 2 and HOI, organic CH 3 I, CH 2 I 2 ) is a major source of iodine in the atmosphere (Prados-Román et al.,   Ernst et al., 2003;Persson et al., 2007;Jabbar et al., 2012). Aqueous 129 I was carried by ocean currents and transported to the North Sea, Kattegat, and Baltic Sea, and continues to the Arctic along the Norwegian coast (Alfimov et al., 2004a;Buraglio et al., 1999;Hou et al., 2007;Raisbeck et al., 1995;Yi et al., 2012). Remarkably elevated 129 I concentrations of up to 10 10 -10 11 atoms L −1 have been found in the North Sea (with 129 I / 127 I atomic ratios of 10 −7 -10 −6 ), 10 9 -10 10 atoms L −1 in Norwegian coastal waters and the Kattegat (with 129 I / 127 I atomic ratios of 10 −8 -10 −7 ), and 10 8 -10 9 atoms L −1 in the Baltic Sea Alfimov et al., 2004b;He et al., 2014;Hou et al., 2007;Michel et al., 2012;Yi et al., 2011). Besides the directly atmospheric releases of 129 I from European NRPs at La Hague and Sellafield, the relatively high 129 I concentrations in the aerosols influenced by westerly wind might also be attributed to the secondary emission of marine discharged 129 I in the European seas (the North Sea, Irish Sea, Norwegian coastal water). The back trajectories analysis (Figs. 6 and S1) shows that the aerosol samples collected in 31 March-4 April, 4-7 April and 11-14 April were directly influenced by westerly and southwesterly air masses that passed over not only the high 129 I contaminated North Sea, Irish Sea and English Channel, but also above Sellafield and La Hague NRPs. However, 129 I concentration and 129 I / 127 I ratio in aerosol collected in 4-7 April were 2.6-fold and 1.7-fold higher than that aerosols collected in 31 March-4 April and 1.6-fold and 2.1 fold higher than that in 7-11 April, respectively. This might be attributed to the fact that most of air masses during 4-7 April mainly passed over both Sellafield and La Hague as well as the high 129 I contaminated North Sea, Irish sea, and English Channel, while air masses during 31 March-4 April passed over only Sellafield, south part of the highly 129 I contaminated of the North Sea and low 129 I level North Atlantic Ocean, and air masses during 7-11 April passed over Sellafield, north part of the highly contaminated North Sea and low 129 I level Atlantic Ocean. However, the 129 I directly atmospheric releases from the European NRPs and secondary emission from the 129 I contaminated sea cannot be isolated and identified based on the trajectories analysis of these three samples, because the air masses of these three aerosols passed over both reprocessing plant(s) and high contaminated seas. For the aerosol collected in 14-18 April, the air masses passed over only the north part of the highly contaminated North Sea and low 129 I level North Atlantic Ocean (Fig. 6c), the 129 I / 127 I ratio in this aerosol is comparable with the samples collected in 11-14 April and 31 March-4 April ((4-5) × 10 −7 ), and the 129 I concentration is higher than that in 31 March-4 April sample by a factor of 1.5. This result suggests that secondary emission of marine discharged 129 I in the contaminated seawater (the English Channel, North Sea, Kattegat and the Norwegian costal water) is the major source of 129 I in the westerly wind-influenced aerosols in Denmark in 2011 rather than the directly gaseous release of the two European NRPs. A relative lower 129 I concentra-tion (2.6 × 10 6 atoms m −3 ) and 129 I / 127 I ratio (2.8 × 10 −7 ) were observed in the aerosol collected in 7-11 April compared to the other four aerosol samples influenced by westerly wind. This might be attributed to the fact that the air masses during this period are mainly from the northwest and mainly passed over to the lower 129 I North Atlantic Ocean and small area of the northern North Sea (Alfimov et al., 2004a). The 129 I / 127 I atomic ratios in the studied aerosols ((1.8-8.6) × 10 −7 ) fell well in the range of 129 I / 127 I ratios in the North Sea surface water in 2005 and 2009 ((1-10)) × 10 −7 ) Christl et al., 2015). This also supports the above conclusion that the main source of 129 I in the aerosols in Denmark in 2011 is re-emission of marine discharged 129 I in the surface water of the North Sea, Irish Sea, and Kattegat.

Atmos
There was a debate for a long time on either atmospheric releases or re-emission of marine discharged 129 I from the two European NRPs as the dominant source of 129 I in the European atmosphere (Ernst et al., 2003;Hou et al., 2009a;Jabbar et al., 2011Jabbar et al., , 2012Lopez-Gutierrez et al., 2004;Michel et al., 2012;Persson et al., 2007;Reithmeier et al., 2010;Schnabel et al., 2001). High 129 I concentrations of (3-445) × 10 8 atoms L −1 in rain water collected at Zürich, Switzerland in 1994-1997, corresponding to estimated average 129 I / 127 I ratios of (3.2-4.0) × 10 −7 , have been reported. Based on the estimated annual re-emission of 129 I from the contaminated European water (1.37 kg) in the 1990s and combined atmospheric releases of 129 I of about 6.5 kg in 1991-1996 from the two European NRPs (Schnabel et al., 2001), as well as the relative positive correlation of 129 I concentration in the aerosols from Vienna, Austria with the integrated monthly atmospheric releases of 129 I from Sellafield NRP (Jabbar et al., 2011), the direct atmospheric release of 129 I was proposed as the dominant source of 129 I in the atmosphere in central Europe (Schnabel et al., 2001;Jabbar et al., 2011).
An investigation on 129 I in rain water (1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005) over Germany has shown a high but slightly decreasing gradient of 129 I concentrations and 129 I / 127 I ratios from coastal to inland areas (Krupp and Aumann, 1999;Szidat et al., 2000;Michel et al., 2012). A high 129 I level ( 129 I / 127 I ratio of (2-8) × 10 −7 ) was also measured in air samples (gas and particle associated iodine) collected at an island in the North Sea (Foehr) in April 2002, which is similar to the 129 I / 127 I ratio measured in the seawater surrounding the Island (Ernst et al., 2003;Michel et al., 2012). This 129 I level in the aerosol is also comparable with those in precipitation in inland Germany. A significant correlation of 129 I with marine-derived chlorine (R 2 = 0.62) and bromine (R 2 = 0.62) were measured in precipitation in Denmark in -2006(Hou et al., 2009a. All these investigations suggested that 129 I in the European atmosphere, at least in central and northern Europe, mainly originates from the re-emission of marine discharged 129 I in the European seawater (Krupp and Aumann, 1999; Atmos. Chem. Phys., 16, 1971-1985, 2016 www.atmos-chem-phys.net/16/1971/2016/ Szidat et al., 2000;Michel et al., 2012;Ernst et al., 2003;Hou et al., 2009a;Persson et al., 2007). The reported 129 I / 127 I ratios in the atmosphere (aerosol and precipitation) in central and northern Europe remain constantly high from about (3-4) × 10 −7 at Zurich in 1994-1997 (Schnabel et al., 2001), (3-14) × 10 −7 in 1988-1995 (Krupp and Aumann, 1999;Bachhuber and Bunzl, 1992) and (  , if the contribution from another European NRP at Marcoule (France), which was operated in 1960-1997, is included, the atmospheric releases of 129 I were 9-17 kg yr −1 in 1980-1997(Reithmeier et al., 2006. The insignificant influence of the remarkably reduced atmospheric releases of 129 I from the European NRPs since 2002 on the 129 I / 127 I ratios in the atmosphere in central and northern Europe confirms that the atmospheric releases 129 I from the two European NRPs at La Hague and Sellafield is not the dominant source of 129 I in the European atmosphere, at least after 2002. However, this does not conflict with the suggestion that 129 I in the atmosphere in Zürich in 1994-1997 and Vienna in 2001 mainly originated from the atmospheric releases of the European NRPs due to the relative high atmospheric releases of 129 I from the European NRPs in 1980-2002, and relatively lower marine discharges of 129 I from the two European NRPs before 1990. The marine discharges of 129 I from the two European NRPs has significantly increased from less than 60 kg yr −1 before 1990 to more than 100 kg yr −1 in 1995 and keeping relative constant at about 250 kg yr −1 from 1998 until now (Reithmeier et al., 2006). The 129 I discharged to the English Channel from the La Hague NRP and to the Irish Sea from the Sellafield NRP has dispersed to a large area in the North Sea, Kattegat, Norwegian Sea, and further to the Arctic by sea currents, and mainly remains in the surface water (less than 100 m) in this area. Different from the atmospherically released 129 I which is quickly dispersed in the atmosphere and deposited on the land in a short time (a few weeks/months), the marine discharged 129 I is integrated in the marine system with much less and slow removal from the water body to the sediment as particle associated form and to the atmosphere in gaseous forms. The huge amount and constantly increased inventory of 129 I in the marine system has made the marine discharged 129 I the major source of 129 I in the present environment, therefore, the continuous re-emission of 129 I from the contaminated seawater to the atmosphere makes the marine discharged 129 I the dominated source in the atmosphere in Europe, and probably all over the world after 2002. The very well matched trend of 129 I / 127 I ratio in the atmosphere in central and northern Europe with the marine discharges of 129 I from the two European NRPs ) and constantly high 129 I level in the precipitation in the Northern Hemisphere outside Europe (e.g. in Japan before 2011 and China) also confirmed this assumption (Toyama et al., 2013;Zhang et al., 2011). In addition, the measurement of 129 I in the aerosol collected in the high altitude Alps sites in 2001 and back trajectories analysis of air masses reached to these locations have also confirmed that re-emission of marine discharged 129 I from the contaminated European seawater was the major source of 129 I in these aerosols (Jabbar et al., 2012).
The low 129 I concentrations, (11-13) × 10 5 atoms m −3 , were observed in the aerosol samples collected in 18-26 April and 26 April-2 May 2011. Back trajectory analysis shows that in this period the air masses at the sampling site were mainly transported by easterly or northwesterly winds, i.e. from the European continent (Fig. S1). Terrestrial emissions of iodine occur through vegetation and terrestrial microorganisms (Bewers and Haysom, 1974). The relatively lower 129 I / 127 I ratios in the terrestrial samples in Europe (10 −8 -10 −7 ) (Osterc and Stibilj, 2013;Ezerinsk et al., 2016) compared to that in the contaminated European seawater (10 −7 -10 −6 ) have been reported. This is reflected in these two aerosol samples by their relatively low 129 I concentrations. An elevated 210 Pb level (249-253 µBq m −3 ) (Table 1) for this period is also consistent with a continental origin ( 210 Pb in the air is a decay product of 222 Rn which is mainly released from the soil in the continental area). However, it should be mentioned that the 129 I level in the European soil is still much higher than that in the uncontaminated area such as in China and Chile ( 129 I / 127 I ratio of 10 −10 -10 −9 ) (Zhang et al., 2011;Daraoui et al., 2012); this is attributed to the high 129 I level in the atmosphere in Europe. Consequently, the 129 I level in the aerosols with air masses from terrestrial area is lower than that from the contaminated European seas, but it is still much higher than that in other places with less contamination. It should be mentioned that the conclusion of re-emission of the NRPs marine discharged 129 I in the contaminated seas as the dominant 129 I source in the atmosphere in Europe and other locations in the Northern Hemisphere suggested in this work does not exclude the contribution of the direct atmospheric releases of 129 I from the NRPs. In the local area surrounding the NRPs, the atmospheric releases might become the dominant source.
The highest 129 I concentration (97.0 × 10 5 atoms m −3 ) found in 8-15 December 2014 might be related to the increased 129 I re-emission from the North Sea since 2011, where the air masses passed over (Fig. S2), as well as probably increased liquid releases of NRPS, which needs further investigations. (2)

Species of 129 I and 127 I in aerosols
WSI is virtually pure iodide in the aerosols investigated, with iodate and water-soluble organic iodine accounting for less than 3 % of total iodine, and these are only measurable in two samples. Iodate was once considered to be the only WSI species in aerosol (Vogt et al., 1999). This was supported by earlier field observations demonstrating that iodate was dominant in size-segregated aerosols from the tropical Atlantic (Wimschneider and Heumann, 1995). However, this iodate-dominant feature was not found in other aerosol samples, e.g. in the northwest Atlantic Ocean and in tropical atmospheric aerosols (Baker, 2004(Baker, , 2005. In these cases, iodide was the dominant water soluble iodine species in the aerosols, as observed in this study. Significant amounts of soluble organic iodine, accounting for 83-97 % of WSI, has been reported in aerosols collected at the Mace Head atmospheric research station on the west coast of Ireland (Gilfedder et al., 2008). Water-soluble organic iodine accounting for 4-75 % of WSI were also measured in aerosols collected from a cruise from the UK to the Falkland Islands in 2003 (Baker, 2005). This suggests that the proportion of soluble organic iodine in aerosols varies regionally and depends on particular aerosol sources and formation processes. Some of this variability might also be related to the analytical methods employed for speciation analysis . It is not clear how iodide is formed in the atmosphere, in an oxidizing environment containing oxygen and ozone. Early models predicted a negligible iodide concentration in particle phases based on the assumption that the iodide in aerosols only originates from the low levels of gaseous HI (McFiggans et al., 2000;Vogt et al., 1999). This work in combination with the previous reports (Baker, 2004; suggests that there must be other primary pathways that contribute to iodide formation at the observed levels. It is generally accepted that iodine atoms are photochemically produced by photolysis of gaseous iodinated compounds, and oxidized by ozone to form reactive iodine oxides (Carpenter, 2003;Saiz-Lopez et al., 2012;Vogt et al., 1999). The for-mation of iodide from iodine atoms and other reactive iodine compounds must rely on electron-donors that are capable of reducing high valence iodine species to iodide. One possibility is the involvement of sulfur compounds (Chatfield and Crutzen, 1990). Possible reaction pathways are given in Table 3. Gaseous SO 2 can be formed in nature by reactions of dimethyl sulfate (DMS) with hydroxide and nitrate. Human activity is a major source of atmospheric SO 2 , globally about three times as much SO 2 as natural processes (Galloway, 1995). By associating with H 2 O, these reactions produce HSO − 3 and SO 2− 3 (Eqs. 1 and 2 in Table 3). Native iodine and other reactive species (not shown) can be reduced to I − on gas-aerosol interfaces (Eq. 3 in Table 3). Other iodine species in aerosols can be also reduced by reductive sulfur compounds to form iodide (Eq. 4 in Table 3). The electron-donors are not limited to sulfur compounds either, for example, nitrogen in the form of gaseous ammonia (NH 3 → NO 2 / NO 3 ) (McFiggans et al., 2000;Saiz-Lopez and Plane, 2004) and elemental mercury (Hg 0 → HgO / HgX, where X is a halogen, I − , Br − or Cl − ) (Lindberg et al., 2002;Simpson et al., 2007) are also candidates responsible for iodide formation.
We note that the percentage of WSI 129 I and 127 I in marine-sourced aerosol from the North Sea is relatively lower than that in the continental aerosol, as compared to the European continent-sourced aerosols (Figs. 4 and 5). This is consistent with the findings drawn from an iodine speciation study of coastal aerosol samples from England (Baker et al., 2001), where the concentrations of total water-soluble iodine from continental aerosols were significantly higher than those from marine aerosols.
A large proportion of 129 I and 127 I in our aerosol samples were NaOH-soluble, which is consistent with the results of aerosol from Tsukuba, Japan, collected shortly after the Fukushima nuclear accident in March 2011 . Abundant NaOH-soluble 129 I (32-44 % of total 129 I) in Fukushima-derived aerosols was attributed to coarse vegetation-related organic particles concentrated during spring. The measured NaOH-soluble iodine (NSI) fractions of 129 I and 127 I during the entire sampling period in the Atmos. Chem. Phys., 16, 1971-1985, 2016 www.atmos-chem-phys.net/16/1971/2016/ spring of 2011 are similar. This indicates that NSI is relatively stable and less affected by the source and pathways of air masses than WSI. NaOH leaching is often used to extract organic substance in fractionation analysis of soil and sediment (Englund et al., 2010a;Hou et al., 2003) based on the high solubility of organic matter, such as humic substances, as well as on nucleophilic substitution and decomposition of organic matter. Organic compounds are important contributors to aerosols, such as lipidic, saccharides, and proteinaceous materials (O'Dowd et al., 2004;Quinn et al., 2014). A significant portion of atmospheric aerosols was found to be humic-like substances (HULIS), named for their strong structural similarity to humic and fulvic acids (Havers et al., 1998). Most of these organic compounds are water-soluble, but a significant water-insoluble fraction of the HULIS material is hydrophobic and acidic in character, and can be dissolved in an alkaline solvent, like NaOH (Feczko et al., 2007;Havers et al., 1998). On the other hand, the hydroxide anion can also initiate a nucleophilic substitution or elimination of iodine-containing organic compounds, which releases iodine from the organic substances. NaOH-soluble iodine is therefore suggested to be likely associated with organic substances in aerosols. RII in aerosols has received less attention than WSI (Gilfedder et al., 2010;Tsukada et al., 1987). The early report on water-insoluble iodine fraction in aerosol particles showed that water-insoluble iodine accounted for 27-58 % of total iodine bound in aerosols from Tokyo, Japan, collected in 1983-1984(Tsukada et al., 1987. Another similar result of 17-53 % of total iodine as insoluble species was reported for aerosols from the west coast of Ireland in 2007, and from a ship transect from China to Antarctica in 2005-2006(Gilfedder et al., 2010. Taking the alkaline-leachable iodine in aerosols into account, these results are very consistent with our observations from Risø (Fig. 5). The residual insoluble 129 I fractions were reported to be 4-23 % of total 129 I in Fukushima-derived aerosol particles , less than the proportion in the aerosols collected in Denmark. This discrepancy might reveal the different formation processes of RII species for the NRPs-derived 129 I in this study as compared to those from Fukushima. A significant difference is the timing of the 129 I releases. NRPs have releasing 129 I into the European environment for about 50 years, allowing 129 I to follow geochemical pathways on timescales ranging from days to decades. In contrast, RII in Fukushimaderived aerosols had only days to react with their environment prior to sampling, 15-22 March 2011.
The origin of the RII fraction is not well understood at present. It is possible that part of the RII fraction is derived from suspended soil particles (Xu et al., 2013). It has been demonstrated that iodine can be associated with metal oxide (notably iron and manganese). A relatively large fraction of iodine (about 38 %) in soil and sediment has been observed in Fe / Mn oxides associated form (Hou et al., 2003). Our data show that RII fraction is as high as 67 % of total aerosol iodine. In addition to metal oxides associated iodine, speciation analysis of 129 I in soil shows that residual iodine after leaching with NaOH and weak acid accounts for less than 10 % of the total, and this component is assumed to be associated with minerals (Hou et al., 2003;Qiao et al., 2012). As stated above, the aerosols collected in early April 2011 were mainly marine-derived aerosols with relatively higher RII percentage than those continental-derived aerosols (Fig. 5). This might be attributed to the fact that some marine components facilitate the association of gaseous iodine with oxides and minerals.

Fukushima-derived 129 I signal in the European atmosphere
The Fukushima Dai-ichi nuclear power plant accident on 11 March 2011 released radioiodine to the atmosphere, primarily as 131 I and 129 I, which was mainly transported eastwards by prevailing westerly winds. Based on 129 I levels in the Fukushima offshore seawater, the released 129 I from this accident was estimated to be 1.2 kg . Radioactive iodine in the air dispersed across the Pacific Ocean, American continent and Atlantic Ocean, and some fraction arrived in the European continent after 1-2 weeks (Clemenza et al., 2012;Leon et al., 2011;Manolopoulou et al., 2011;Pittauerová et al., 2011). Anthropogenic 129 I has been reported from a variety of environmental samples in Japan, including soil, seawater, precipitation, and aerosols (Buesseler et al., 2012;Hou et al., 2013;Muramatsu et al., 2015;Xu et al., 2013Xu et al., , 2015. The level of 129 I in aerosols collected in Tsukuba, about 170 km from the Fukushima Dai-ichi NPP, reached 5 × 10 8 atoms m −3 . While the Fukushima-derived 129 I in environmental samples was less presented outside of Japan. 131 I in the aerosol samples col-lected at Risø, Denmark, 10 days after the Fukushima accident have been observed (Fig. 7) (Nielsen et al., 2011). The radioactivity of 131 I reached the peak on 24-30 March 2011 (763 µBq m −3 in aerosol), then fell to below detection limits for 131 I in the middle of May. Based on the measured 131 I radioactivity in the aerosol samples and an 129 I / 131 I atomic ratio of 16.0 ± 2.2 deduced from the aerosol samples collected at Tsukuba, Japan shortly after the Fukushima accident , the Fukushima-derived 129 I signal in Denmark can be reconstructed (Fig. 7). The highest 129 I concentration in the aerosols in Denmark from the Fukushima accident is estimated to be 6.3 × 10 4 atoms m −3 on 30-31 March 2011, which accounts for less than ∼ 6 % of total 129 I (1.1-9.7 × 10 6 atoms m −3 ) in Denmark when the Fukushima 131 I peak was measured. Considering the rapid decline of 129 I levels in aerosols and precipitation in Japan to nearly pre-accident levels within 2 years (Xu et al., 2013), the contribution of Fukushima-derived 129 I to the 129 I level and inventory in the Europe is now negligible in comparison to NRPs-derived 129 I in the European atmosphere.

Conclusions
Based on the analytical results on speciation analysis of 129 I and 127 I in aerosols collected in Denmark immediately after the Fukushima accident and the discussion above, the following conclusions can be drawn. Iodide is the dominant species (> 97 %) of the watersoluble iodine in the aerosols of this study, its formation might be related to atmospheric reductants, such as reductive SO 2 and disulfites. The most dominant species of iodine in aerosols are NSI and RII, accounting for up to 80 % of total iodine. NSI is likely predominantly bound to organic matter, such as HULIS, while RII might be associated with inorganic components, such as metal oxides.
Westerly-dominated air masses from the North Sea contribute to high 129 I concentrations in aerosol compared to easterly air masses from the northeastern European continent. Aerosol 129 I in Denmark is strongly influenced by the European nuclear reprocessing plants, in which secondary emission of 129 I from surface seawater in the North Sea, Norwegian coastal water, and Kattegat is a main contributor.
Fukushima-derived 129 I contributed less than ∼ 6 % to the total 129 I in the European atmosphere immediately after the Fukushima accident, a negligible quantity when compared to 129 I levels released from Sellafield and La Hague nuclear reprocessing plants.
Water-insoluble iodine (NSI and RII) in aerosols that is primarily associated with organic compounds is crucial for the investigation of geochemical cycling of iodine in the atmosphere due to its large fraction. This requires more intensive studies to extend our observations on both temporal and spatial scales.
The Supplement related to this article is available online at doi:10.5194/acp-16-1971-2016-supplement.