The deposition of ozone to seawater is known to be controlled by a variety of physical and chemical processes. At low wind speeds chemical loss is comparatively more important than loss due to physical processes. We have determined experimentally the relationship between ozone deposition velocity and concentration of iodide and dissolved organic matter in water buffered at seawater pH (8.0). The concentrations of both species used in this study are representative of those encountered in coastal and oceanic systems. We show that dissolved organic matter and iodide contribute to a similar degree to the chemical enhancement of ozone deposition to surface waters.
 Deposition to surface seawater is a significant mechanism for loss of atmospheric ozone, accounting for approximately one third of the global annual ozone deposition of 600–1000 Tg O3 yr−1 [Fairall et al., 2007; Ganzeveld et al., 2009].The deposition velocity (the ratio of flux to air-sea concentration difference) of ozone measured directly at sea shows large variations (0.01–0.12 cm s−1) [Gallagher et al., 2001b; Heikes et al., 1996; Lenschow et al., 1982]. The limited number of micrometeorological field measurements carried out to date inhibits the parameterisation and quantification of the various processes contributing to dry deposition of ozone. However, it is clear that wind speed and sea surface turbulence play a major role in ozone uptake by seawater at medium to high wind speeds [Gallagher et al., 2001a, 2001b], but chemical losses become comparatively more significant at low wind speeds [Chang et al., 2004; Garland et al., 1980; Oh et al., 2008]. Chang et al.  reviewed the importance of a number of chemical species in affecting ozone chemical loss at the sea surface, and concluded that iodide was the only species capable of explaining a substantial ozone loss at low wind speeds. These authors also suggested that there is insufficient evidence for a substantial contribution of surfactant films to ozone loss, because chemical species with a high enough concentration, combined with a sufficient reactivity towards ozone, have yet to be identified in surface seawater. As an example, Chang et al. showed how the contribution of organic compounds such as ethene and propene was negligible because of the combination of sub-nanomolar seawater concentrations and low reactivity with ozone. However,McKay et al.  demonstrated that there was a correlation between ozone deposition velocity and concentration of natural surfactants in seawater. Recently, Clifford et al.  suggested that chlorophyll (and similar macrocyclic compounds) represents a major chemical sink for ozone at the sea surface.
 Marine dissolved organic matter (DOM) is composed of a complex mixture of compounds, encompassing a range of molecular weights (from a few tens to tens of thousands a.m.u.), functional groups (for example phenolic, carboxylic) and degrees of saturation (unsaturated fatty chains, aromatic rings) [Buffle, 1988]. Whereas coastal DOM is predominantly composed of humic-type material of riverine origin, oceanic DOM is mostly generated in situ by biological activity [Buffle, 1988]. Apart from some relatively small molecules such as specific sugars, carbohydrates, aminoacids, lipids, organic acids and photosynthetic pigments (all present at sub-μM or sub-nM concentrations and accounting for 4–10% of the dissolved organic carbon (DOC) pool of marine DOM), most marine DOM is still uncharacterised at the molecular level, due to the complexity and low concentrations of the individual species [Benner, 2002]. However, the combined concentration of organic compounds in seawater is in the range of 40–80 μM DOC [Hansell et al., 2009]. Compounds containing carboxylic and phenolic groups and/or aromatic rings are likely to be reactive towards ozone, due to the presence of double bonds [von Gunten, 2003]. Although it is impossible to assess the effect of individual organic compounds, their reactivity towards ozone is likely to span a range of values, depending on their individual concentrations and reactivities. Here, we compare the effect of DOM as a whole to that of iodide on ozone uptake by water. We also assess the combined effect of iodide and DOM together on ozone uptake.
 The apparatus for studying the deposition velocity of ozone as a function of iodide and dissolved organic matter is shown in Figure 1. All experiments were carried out in the dark. A 2-L round glass cell was modified to accommodate inlets and outlets for gases and a stopcock for addition of reactants to the cell. The water and air in the cell were mixed with a home-made polytetrafluoroethylene (PTFE) stirrer, which consisted of a motor-operated rod inserted in a PTFE stopper. Two o-rings inserted in the stopper ensured that the vessel was air-tight. Hydrocarbon-free nitrogen and ultra-pure oxygen were mixed in a ratio of approximately 4:1 before passing through a home-built quartz cell, where ozone was generated with a UVP Pen-Ray mercury lamp. The ozone concentration was controlled by adjusting the oxygen flow rate, and was measured by a Thermoelectron 49C analyser, connected at the exit of the cell. The total gas flow rate through the cell (approximately 1.3 L min−1) was determined by the requirements of the ozone analyser.
 The glass cell and the stirrer were acid-washed for 24 hours (10% HCl AnalaR, BDH) and thoroughly dried prior to each experiment. At the beginning of the experiment the gas stream flowed directly to the ozone analyser through a cell bypass (Figure 1) until a constant O3reading was obtained. We used typical ozone concentrations for the experiments of 40 ppb and 250 ppb. Meanwhile, HPLC-grade water (BDH) was buffered with NaHCO3 at pH 8.0 and 800–1000 mL of the buffered solution were added to the cell. The stirrer was then switched on. The ozonised air flow was then diverted to the cell and left for ∼30–60 minutes, allowing sufficient time for the ozone to equilibrate with the water and vessel walls. Once the ozone concentration reached a constant value, in each experiment different amounts of iodide or/and DOM were added to the water. Because the solution was stirred, the reactants quickly mixed into the bulk and to the surface of the water, as indicated by the rapid drawdown of ozone after the injection (Figure 2). The ozone loss rate was then calculated by monitoring the decrease in ozone concentration with time. The addition of iodide and/or DOM was made in such a way as to perturb the system as little as possible. Thus, 40 mL water were drawn from the cell with an acid-washed, gas-tight glass syringe; microliter amounts of a 10−3M potassium iodide standard (BDH AristaR) were added to the syringe and the water re-injected into the vessel. The resulting drop in ozone concentration was the sole result of the addition of iodide. For the DOM additions, Suwannee River Natural Organic Material (NOM), sourced from the International Humic Substances Society (IHSS), was weighed and dissolved in 10 mL of water drawn from the cell prior to each injection. The choice of this material is discussed in detail below. We also carried out a set of “mixed” experiments where DOM was added to the water together with 150 nM iodide. This iodide concentration was chosen because it is representative of average iodide concentrations in surface seawater [Truesdale et al., 2000]. Initial experiments were carried out in artificial seawater as a background but because of the variable iodide concentration (present as contaminant in the NaCl) in solution, the data were less reproducible (data not shown here).
 The elemental composition of dry Suwannee River NOM is: 48.8% C; 3.9% H; 39.7% O; 1.02% N; 0.60% S; 0.02% P; 7.0% inorganic residue (Total 101.0%). The percentage carbon distribution obtained by solution state 13C NMR was: 8% carbonyl, 20% carboxyl, 23% aromatic, 7% acetal, 15% heteroaliphatic and 27% aliphatic [Thorn et al., 1989]. It was decided to use the IHSS Suwannee River NOM because there is no commercially available certified marine organic material; moreover, this material is at least representative of the NOM found in coastal systems affected by riverine inputs. Only ∼8% of the C in oceanic surface DOM is present in unsaturated (C = C and C = O) structures [Benner, 2002]. Since Suwannee river NOM is significantly more unsaturated, this may lead to a lower level of reactivity of marine DOM towards ozone. However, we consider the Suwannee NOM to better represent marine DOM than would an individual model organic compound, since marine DOM is composed (like the Suwannee NOM) of a wide range of compounds of varying composition and reactivity.
 The ozone loss rate (Jozone) was calculated from the variation in ozone concentration with time (i.e., the slope of the linear portion of a plot of ln (C/C0) against time), following the addition of reactants (iodide/ DOM/mixed iodide – DOM) to the cell. In all experiments Jozone represents the variation in ozone concentration due solely to the addition of the reactants to the vessel. Ozone deposition velocity was calculated from the experimental ozone loss rate using the following equation:
where vd is the deposition velocity of ozone to the water (cm s−1), Jozone is the loss rate of ozone (s−1), V is the volume of the headspace (cm3) and S is the surface area of the water in the vessel (cm2). Errors in individual vd values were estimated from the standard deviation of Jozone derived from linear regression and were generally less than 10%, but as high as 17% in a few cases. It should be made clear that the changes in vd values shown here represent the deposition velocity of ozone caused solely by the addition of specific amounts of reactant to the vessel, and exclude any physical effects and we use the term “chemical deposition velocity” (vdc) to identify these hereafter. Therefore, our calculated values of vdc are significantly lower than ozone deposition velocities measured directly at sea, because they account only for the effect of the added reactant(s). Although stirring the solution enhances the ozone surface transfer by increasing surface turbulence, it affects all stages of all the experiments to the same extent and so should not affect νdc.
 The results for ozone uptake by iodide, DOM, and mixed iodide (150 nM) - DOM in HPLC-grade water at a starting ozone concentration of 40 ppb and 250 ppb are shown inFigures 3 and 4, respectively. The slope and intercept values for all experiments in Figures 3 and 4 are shown in Table 1. A higher degree of scatter is present in the experiments carried out at 40 ppb ozone: here, the relatively small decrease in ozone concentration during the experiments (compared to those at 250 ppb) was more difficult to measure against the background ozone signal. This made the calculations at 40 ppb less precise than those at 250 ppb.
Table 1. Linear Regression Values and Correlation Coefficients for Ozone Deposition Velocity (cm s−1) Versus Iodide Concentration (nM), DOM Concentration (mg L−1), and DOM Concentration (mg L−1) at 150 nM Iodide Constant Concentration
DOM + I−(150 nM)
40 ppb O3
2.76 (±0.27) × 10−5
1.78 (±0.24) × 10−3
1.14 (±0.33) × 10−3
2.71 (±0.58) × 10−3
4.77 (±0.91) × 10−3
1.02 (±0.13) × 10−2
250 ppb O3
2.09(±0.08) × 10−5
1.38 (±0.14) × 10−3
1.12 (±0.14) × 10−3
1.54 (±0.19) × 10−3
2.96 (±0.50) × 10−3
5.78 (±0.50) × 10−3
 The deposition velocity of ozone to water increases with increasing iodide and DOM concentrations in a similar linear fashion (Figures 3 and 4). In this experimental system, varying iodide concentrations between typical oceanic concentrations of 50 nM and 200 nM caused ozone deposition velocity (vdc) to vary between 0.005 and 0.010 cm s−1. This enhancement of ozone deposition velocity occurred solely as a result of increased ozone-iodide reactions due to increasing iodide concentrations. We observed similar values of deposition velocity at DOM concentrations of 1–2 mg L−1, which correspond to 40–80 μM DOC and are representative of average oceanic DOC concentrations [Hansell et al., 2009] . Thus, it appears that iodide and DOM contribute to a similar extent to ozone deposition velocity in HPLC water.
 The combined contribution of DOM and 150 nM iodide to ozone deposition is shown in Figures 3c and 4c. At typical oceanic concentrations of 150 nM iodide and 80 μM DOC (corresponding to 2 mg L−1 DOM for the DOM used in this study), the combined effect of both reactants resulted in a (chemical) deposition velocity of 0.012 ± 0.001 cm s−1 for the experimental system used in this study, i.e., ∼80% of the value expected from the combination of the deposition velocities measured for iodide and DOM individually. Based on the results obtained for the two components individually, it seems likely that they contributed more or less equally to ozone deposition in the mixed system. It is not possible from our experiments to determine whether the reduced ozone deposition of the mixed system is due to suppression of deposition by one or other component, or by suppression of both deposition mechanisms.
 Literature data suggest that the main species controlling the deposition velocity of ozone to seawater is iodide [Chang et al., 2004], with organic species contributing to a much lesser extent, because of the lower kinetic rate constants (compared to iodide), combined with low (sub-nanomolar) concentrations. For example, the second-order kinetic rate constants for the reaction of ozone with alkenes and with iodide are 2–8 × 105 M−1 s−1 and 2 × 109 M−1 s−1, respectively [Dowideit and von Sonntag, 1998; Garland et al., 1980; see Chang et al., 2004, Table 1]. Because alkenes are present at sub-nanomolar concentrations, and iodide concentrations are of the order of 50–200 nM in seawater, the contribution of iodide to the ozone chemical loss rate is ∼106times higher than that of alkenes. Although alkenes are certainly representative of non-methane hydrocarbons produced in seawater, these gases represent only a minimal fraction of the overall content of organic material in seawater. The marine DOM pool in seawater, present at concentrations of 40–80 μM DOC [Hansell et al., 2009], is composed of a wide variety of compounds. A high but variable fraction of these comprises unsaturated and aromatic moieties, all potential sinks for atmospheric ozone. Although the contribution of these individual chemical species to ozone uptake is low (because of their low concentrations), their combined effect is shown here to be significant, and comparable to that of iodide. The more “empirical” approach followed in this study considers the contribution of DOM as a whole, rather than the contribution of individual chemical species present at sub-nM concentrations. This likely explains why our results indicate a more significant role for organic material in ozone deposition than estimated previously. However, it should be noted that because of the higher degree of unsaturation in Suwannee River DOM with respect to marine DOM, our results might overestimate the effect of organic matter compared to that of iodide. The approach used in this study does not allow determination of the reactivities of individual compounds. The value of 0.012 cm s−1 determined for chemical ozone deposition velocity for average ambient concentrations of iodide (150 nM) and DOM (80 μM DOC) in this study lies in the lower part of the range of deposition velocities measured at sea (0.01–0.1 cm s−1 [Fairall et al., 2007]). However it should be noted that whereas we measure the sole effect of the addition of specific chemical species to water, the overall deposition velocities measured at sea include a combination of chemical, as well as physical (e.g., wind speed) effects. On the other hand, values of deposition velocity determined in our experiments at values of 3–6 mg L−1 DOM (which correspond to 120–250 μM DOC, typical of coastal waters [Cauwet, 2002, and references therein]) compare well with those determined using a similar setup by McKay et al.  for UK coastal sites. Our values also compare well with the range (10−5 to 10−3 m s−1) estimated by Clifford et al.  for the contribution of chlorophyll a to ozone deposition velocity to oceanic surfaces, but are significantly lower than those estimated for humic acid films (0.4–0.9 cm s−1 [D'Anna et al., 2009]).
 Whereas up to the recent past the effect of organic material on ozone deposition to water was considered negligible compared to that of iodide [Chang et al., 2004], this study, together with the recent studies carried out by Clifford et al.  and Reeser et al. suggest that the importance of organic material as a sink for ozone to water surfaces should be reconsidered. Here, we show that dissolved organic material is an ozone sink approximately as important as iodide. Of course our results depend on the type of organic material chosen for this study, and further experiments with marine natural organic material are necessary to assess the real extent of ozone deposition to seawater. Further studies should also be carried out to assess the effect of light on ozone uptake by organic material in seawater, not only by specific organic molecules, but also by natural marine dissolved organic matter as a whole. In addition it would be beneficial to conduct the experiments in a set-up that enabled better determination of the kinetics of the reactions occurring both in terms of single reactants (e.g., iodide, specific organic molecules) as well as mixtures and combinations of species (e.g., natural organic material from various sources).
 Financial support for this work was provided by the Natural Environment Research Council (NERC) grant NE/C511572/1. We thank two anonymous reviewers who helped to improve the quality of the paper.
 The Editor thanks two anonymous reviewers for their assistance in evaluating this paper.