Review on the physical chemistry of iodine transformations in the oceans

The transformation between iodate ( IO 3 − ), the thermodynamically stable form of iodine, and iodide (I-), the kinetically stable form of iodine, has received much attention because these species are often dependent on the oxygen concentration, which ranges from saturation to non-detectable in the ocean. As suboxic conditions in the ocean’s major oxygen minimum zones indicate that IO 3 − is minimal or non-detectable, the incorporation of IO 3 − into carbonate minerals has been used as a redox proxy to determine the O2 state of the ocean. Here, I look at the one and two electron transfers between iodine species with a variety of oxidants and reductants to show thermodynamics of these transformations. The IO 3 − to IO 2 − conversion is shown to be the controlling step in the reduction reaction sequence due to thermodynamic considerations. As IO 3 − reduction to IO 2 − is more favorable than NO 3 − reduction to NO 2 − at oceanic pH values, there is no need for nitrate reductase for IO 3 − reduction as other reductants (e.g. Fe2+, Mn2+) and dissimilatory IO 3 − reduction by microbes during organic matter decomposition can affect the transformation. Unfortunately, there is a dearth of information on the kinetics of reductants with IO 3 − ; thus, the thermodynamic calculations suggest avenues for research. Conversely, there is significant information on the kinetics of I- oxidation with various oxygen species. In the environment, I- oxidation is the controlling step for oxidation. The oxidants that can lead to IO 3 − are reactive oxygen species with O3 and •OH being the most potent as well as sedimentary oxidized Mn, which occurs at lower pH than ocean waters. Recent work has shown that iodide oxidizing bacteria can also form IO 3 − . I- oxidation is more facile at the sea surface microlayer and in the atmosphere due to O3.


Introduction
The thermodynamically favorable form of iodine in seawater is iodate (IO − 3 ). However, iodide (I -) is present in oxic, suboxic and anoxic waters. The one electron transfer reaction of Iwith molecular oxygen, 3 O 2 , to form the iodine atom (I•) and superoxide (O 2 -) is thermodynamically unfavorable as is the reaction of two Iwith 3 O 2 to form I 2 and H 2 O 2 Luther, 2011). Thus, other oxidants are required to initiate abiotic iodide oxidation, and Iis a known sink for O 3 . Biotic iodide oxidation has received much interest with one report showing conversion of iodide to iodate (Hughes et al., 2021). Iodate reduction can occur with common reductants (e.g., sulfide, Fe 2+ ), and various organisms that decompose organic matter using iodate as the electron acceptor. Nitrate reductase and dimethyl sulfoxide reductase enzymes from these and planktonic microbes are considered important mediators for biotic iodate reduction (e.g., Hung et al., 2005;Amachi, 2008). Thus, there has been extensive interest in the chemistry of these two iodine species and the possible intermediates that form during their 6-electron redox interconversion ever since the element, iodine, was first discovered as I 2 during the study of brown kelp algae of the Laminariales (kelps/seaweeds) by Courtois in the early 1800s (Wong, 1991;Küpper et al., 2008;Küpper et al., 2011). Possible chemical species that form during the IO − 3 ↔ I − interconversion are given in eqn. (1a, b). The loss of an O atom is equivalent to a two-electron transfer resulting in the reduction of the iodine from +5 in IO − 3 to +3 for IO − 2 (HOIO) to +1 for HOI (IO -) and to -1 for I -. The acid-base species in parentheses are minor species at seawater pH as the pK a values for HOI and HOIO are 10.7 (K a = 2 x 10 -11 ) and 4.49 (K a = 3.2 x 10 -5 ), respectively.
Equation 1b shows the interconversion between HOI and Ias HOI undergoes one-electron transfer to I 2 followed by another oneelectron reduction per I atom to I -.
This work considers the thermodynamics of these transformations during the reduction of IO − 3 , which occurs in anoxic systems, during organic matter decomposition and by phytoplankton, as well as the oxidation of I -, which occurs by the direct oxidation of iodide by iodide oxidizing bacteria, oxidized metals and reactive oxygen species (ROS) that are produced by certain microbes, (macro)algae and abiotic processes including photochemistry. In a previous work (Luther, 2011), the chemistry and thermodynamics of chloride, bromide and iodide oxidation were compared; however, I(+3) species (HOIO, iodous acid, and IO − 2 , iodite) and stepwise iodate reduction were not considered. Here, stepwise reactions of the iodine species in equations 1a and 1b with environmentally important reactants (including transient ROS species) are considered for both the oxidation of Iand the reduction of IO − 3 to affect their interconversion. The kinetics of these stepwise reactions are also considered. Kinetic data for the first step(s) in iodide oxidation are available, but less kinetic information is available for iodate reduction. Table 1 gives several equations for redox half-reactions that include the pH dependence for the reaction considered. These are pϵ(pH) relationships based on the balanced chemical equations and the thermodynamics of each chemical species. The basic mathematical approach has been fully developed in standard textbooks (Stumm and Morgan, 1996; and used in previous publications (Luther, 2010;Luther, 2011). Aqueous thermodynamic data to calculate the pϵ(pH) or logK(pH) relationships in Table 1 (at 25°C and 1 atm) are from Stumm and Morgan (1996) and other sources (Bard et al., 1985;Stanbury, 1989). The value used for the Gibbs free energy for Fe 2+ (-90.53 kJ/mole) is that discussed in Rickard and Luther (2007). Values of the free energy for HOIO and IO − 2 are from Schmitz (2008). The calculated pϵ value from each half-reaction is given as a function of pH as in the examples in Table 1, and these half reactions can be used for simple calculations of the pE values of full reactions (see next section). When H + or OHis not in a balanced equation for a half-reaction, there is no pH dependence on the half-reaction. The pe calculated is termed pϵ(pH) which provides a log K for each halfreaction at a given pH. Concentration dependence for the other reactants are not considered in the calculation; thus, these are considered standard state calculations. When concentration dependence is considered, the calculated pϵ value can vary as in the following example for the O 2 /H 2 O couple (O1 in Table 1).

Calculations of aqueous redox potentials from half-reactions
Using the balanced half reaction and the Gibbs free energy of formation of each species at 25°C, the Gibbs free energy of the reaction and the equilibrium constant are calculated.
On expanding, eqn. 2b results.   Table 1) where concentration is used for O 2(aq) . Note that the above equations show a 1.50 log unit change for an O 2 concentration range from 1 μM to unity activity (O1a) so the calculations could vary an order of magnitude or more in either direction when concentration dependence is included. However, comparisons can be more easily made when combining different half-reactions at a given pH. This permits an assessment of which combined half-reactions are thermodynamically favorable and thus more likely to occur in a given environmental setting.

Coupling half-reactions
As an example of coupling two half reactions to determine whether a reaction is favorable, I use the data in Table 1 for the reduction of IO − 3 (Io5b) by NO − 2 (N1) in equation 3. At a pH of 7, the pe red values for IO − 3 and NO − 3 are 7.91 and 7.28, respectively. As NO − 2 is the reductant, it is oxidized; thus, the sign for pe red (7.28) is reversed to become pe oxid (-7.28).
For this work, Table 1 lists the pϵ(pH) values for Mn, Fe, oxygen, nitrogen, sulfur and iodine species for the relevant iodine redox reactions considered. Dissolved Fe(II) and Mn(II) are primarily hexaaquo species until the pH is > 7, where hydroxo complexes start to become important. As most reactions occur via one and twoelectron transfers, the calculations will permit assessment of a thermodynamically unfavorable step along a reaction coordinate of six-electrons as in the reduction of iodate to iodide and the oxidation of iodide to iodate. From surface waters to decomposition zones, seawater pH values range from 8 down to 7; thus, the following discussion will emphasize this pH range.
3 Results and discussion: Iodate reduction 3.1 Iodate and iodide speciation at different seawater oxygen conditions In the oxic environment, the oxidizing condition of the environment or pe is set by the 4-electron transfer reaction of the O 2(aq) /H 2 O couple [reaction O1 in Table 1]. At a pH of 8, temperature 25 0 C and a salinity of 35, 100% O 2(aq) saturation is 211 mM, which gives a pe of 12.58 ( Figure 1). As the IO − 3 =I − couple has a pe of 10.56 at pH = 8, IO − 3 is the thermodynamically favored iodine species. Entering the pe value for a given [O 2(aq) ] into equation Io6 allows the determination of the iodide to iodate ratio and the actual concentration of each assuming a total iodine concentration of 450-470 nM (Elderfield and Truesdale, 1980). Figure 1 shows the iodate and iodide concentrations are equivalent at a pe of 10.56. The vertical lines indicate the environmental pe for [O 2(aq) ] of 1, 10, 100 nM, and 1, 50 and 211 mM. As oxygen minimum zones (OMZ) of the Arabian Sea and the equatorial Pacific Ocean have [O 2(aq) ] concentrations in the 1-100 nM range (Revsbech et al., 2009;Lehner et al., 2015), calculations show that IO − 3 is the thermodynamically preferred iodine species even at 1 nM O 2(aq) , which gives a pe of 11.25 for the O 2(aq) / H 2 O couple. However, Iis the dominant iodine species detected in OMZ waters (Wong and Brewer, 1977;Luther and Campbell, 1991;Rue et al., 1997;Farrenkopf and Luther, 2002;Cutter et al., 2018). At [O 2(aq) ] concentrations ≤ 1 mM, IO − 3 , NO − 3 and Mn 2+ concentrations are now similar or higher in concentration and should determine the pe of the water.
As most reactions occur by 1-or 2-electron transfers, Figure 2 shows the redox sequence for two electron transfer redox couples NO − 3 =NO − 2 (N1), MnO 2 /Mn 2+ (Mn1) and the one-electron redox couple Fe(OH) 3 /Fe 2+ (Fe3) over a wide range of pH. The redox sequence at pH = 8 is as expected for the N, Mn and Fe systems. As NO − 2 (up to 12 mM) and Mn 2+ (up to 8 mM) are formed at OMZ oxicanoxic transition zones (e.g., Arabian Sea, Black Sea, Equatiorial Pacific, see Lewis and Luther, 2000;Trouwborst et al., 2006;Cutter et al., 2018, respectively) and are in higher concentration than O 2(aq) , the NO − 3 ! NO − 2 (N1) and MnO 2 ! Mn 2+ (Mn1) couples can be chosen to set the environmental pe. At pH = 8, the NO − 3 to NO − 2 pe is 6.15 and the MnO 2 to Mn 2+ pe is 4.80. At pH = 7, the NO − 3 to NO − 2 pe is 7.28 and the MnO 2 to Mn 2+ pe is 6.80. At these pe values, O 2(aq) is below 1 nM, and Iis now the thermodynamically favored iodine species when comparing these data with the IO − 3 =I − couple (pe of 10.56 at pH = 8). Figure 2 also shows that the IO − 3 =IO − 2 couple (Io5b) should be the first step in the reaction sequence of iodate to iodide (eqn. 1a). As the pe of the IO − 3 =IO − 2 couple has a more positive pe value than the N, Mn and Fe couples in Figure 2, IO − 3 reduction is more favorable than these couples even though it is very close to the NO − 3 =NO − 2 couple. Thus, IO − 3 is predicted to reduce before NO − 3 , and biological activity (e.g., nitrate reductase activity) is not necessary to reduce IO − 3 (see section IO − 3 and Iconcentrations for a total iodine concentration of 450 nM at different environmental pe values assuming that O 2(aq) reduction to H 2 O sets the pe value of ocean waters.

FIGURE 2
Two electron transfer redox couples for O 3 (O4), N (N1), Mn (Mn1) and I (Io2a, Io4b, Io5b), and one electron transfer redox couple for Fe (Fe3). The oxidized species is always above the line and the reduced below the line as in the O 2 /H 2 O and the H 2 O/H 2 couples, which dictate the water stability field.
3.2). Because of the strong pH dependence for the Mn and Fe couples, they cross the IO − 3 =IO − 2 and NO − 3 =NO − 2 couples at lower pH, which have similar slopes. Thus, NO − 2 is predicted to reduce IO − 3 to IO − 2 (eqn. 3). Interestingly, Mn 2+ and Fe 2+ should be poorer reductants than NO − 2 for conversion of IO − 3 to IO − 2 at a pH< 6 and pH< 1, respectively, but are more favorable to reduce IO − 3 than NO − 2 above those pH values (see sections 3.2 and 3.3).
Although the IO − 3 to Iconversion occurs at higher pe, it is a 6electron transfer (IO6), which is not a facile process. Thus, the intermediates (IO − 2 and HOI) will dictate the reactivity sequence via a combination of thermodynamic and kinetic considerations.
As shown in Figure 2, IO − 2 reduction to HOI and HOI reduction to Iare also more favorable at higher pe values than the IO − 3 to Icouple. At a pH = 8, the IO − 2 to HOI couple has a pe value of 12.06 corresponding to 2 mM O 2 (see Figure 1). Similarly, the HOI to Icouple has a pe value of 12.66 corresponding to 250 mM O 2 . At a pH of 7, both couples have pe values greater than 13 indicating that, even at O 2 saturation, Iis the dominant species predicted when these intermediates form. At a pH = 7.5 (that is found in many OMZ waters), both IO − 2 to HOI and HOI to Icouples have pe values greater than 12.6; also indicating that at O 2 saturation, Iis the dominant species predicted. Thus, the intermediates IO − 2 and HOI are not predicted to be stable in marine waters; thus, the conversion of IO − 3 to IO − 2 is a key step. Interestingly, Hardisty et al. (2021) found in situ IO − 3 reduction in the oxycline where [O 2(aq) ] was 11 mM, but not at [O 2 (aq) ]< 2 mM. Lastly, the O 3 to O 2 + H 2 O couple is highly oxidizing indicating that all iodine couples should lead to IO − 3 formation. O 3 reactions will be discussed in more detail below (sections 4.2, 4.7).
In the next sections (3.2 -3.5), the thermodynamics for the conversion of iodate to iodide via the intermediates outlined in equations 1a and 1b by environmental reductants are considered to show what step, if any, in the reduction of iodate to iodide may be unfavorable over a wide range of pH. Iodate reduction is well known in the marine environment (e.g., Wong and Brewer, 1977;Wong et al., 1985;Luther and Campbell, 1991;Rue et al., 1997;Farrenkopf and Luther, 2002;Cutter et al., 2018) and occurs via chemical reductants like sulfide (Zhang and Whitfield, 1986) and via microbes like Shewanella putrfaciens (Farrenkopf et al., 1997) and Shewanella oneidensis (Mok et al., 2018) during dissimilatory reduction coupled with decomposition (oxidation) of organic matter as well as phytoplankton mediated processes (e.g., Chance et al., 2007).

Iodate reduction by NO
has not yet been shown to be a reductant for IO − 3 in aqueous lab studies (eqn. 3), HNO 2 is a reductant for MnO 2 (Luther and Popp, 2002) and Mn(III)-pyrophosphate (Luther et al., 2021). Figure 3A shows the thermodynamic calculations for the stepwise conversion of IO − 3 to Iby NO − 2 reduction (NO − 2 oxidizes to NO − 3 ). All 2-electron transfer reactions, which involve O atom loss for iodine, are favorable over the pH range. For seawater pH (7-8), the least favorable reaction is the IO − 3 to IO − 2 reaction whereas the IO − 2 to HOI and HOI to Ireactions are more favorable. Thus, the IO − 3 to IO − 2 conversion appears to be the controlling step in the reaction sequence. The 1-electron transfer reaction of HOI to I 2 is the most favorable, but the second 1-electron transfer reaction of I 2 to Iis only favorable at pH > 4. Thus, reduction of IO − 3 to Iby NO − 2 is predicted via 1electron or 2-electron transfer reactions at seawater pH values. The data plotted in Figure 2 indicate that once IO − 2 forms there is no thermodynamic barrier to Iformation.
The IO − 3 reaction with NO − 2 has been reported to produce I 2 in ice by Kim et al. (2019), but not in solution. The pH in the ice was 3 where HNO 2 and H 2 ONO + exist and are the likely reductants. Thus, polar areas may be locales for IO − 3 reduction. At seawater pH, the reaction seems to be hindered by kinetics in the transition state as each reactant (IO − 3 and NO − 2 ) is an anion, which will repel each other.

Biological iodate reduction
The marine literature has many reports on the uptake of IO − 3 (with or without NO − 3 ) by phytoplankton with the iodine released as I -(e.g., Elderfield and Truesdale, 1980;Wong, 2001;Wong et al., 2002;Chance et al., 2007;Bluhm et al., 2010). As a result of this iodate uptake, NO − 3 reductase was presumed by some researchers to be a key process for IO − 3 reduction to I -. Also, Bluhm et al. (2010) and Carrano et al. (2020) showed that algal senescence enhanced Irelease, and Hepach et al. (2020) showed that there is a considerable lag between IO − 3 uptake and Irelease due to senescence. NO − 3 reductase appears to reduce IO − 3 in some phytoplankton (Hung et al., 2005). However, de la Cuesta and Manley (2009) showed that Ican be up taken by phytoplankton, and that different phytoplankton uptake Iwhereas other phytoplankton uptake IO − 3 . Thus, there is no need for nitrate reductase for IO − 3 reduction as Ican be up taken by some phytoplankton rather than form from IO − 3 reduction. Moreover, Waite and Truesdale (2003) showed that nitrate reductase was not important for IO − 3 reduction by Isochrysis galbana. The latter study is consistent with the thermodynamics of the reduction IO − 3 to IO − 2 being more favorable than the reduction NO − 3 to NO − 2 . Furthermore, under anaerobic conditions, dissimilatory IO − 3 reduction occurs without nitrate reductase for the denitrifying bacterium, Pseudomonas stutzeri, (Amachi et al., 2007;Amachi, 2008). Reyes-Umana et al. (2022) and Yamazaki et al. (2020) showed that iodate reductase is in the periplasmic space of Pseudomonas sp SCT. Also, Mok et al. (2018) showed that dissimilatory IO − 3 reduction by Shewanella oneidensis does not involve nitrate reductase. Recently, Shin et al. (2022) showed that Shewanella oneidensis requires extracellular dimethylsulfoxide (DMSO) reductase involving a molybdenum enzyme center for IO − 3 reduction. Guo et al. (2022) studied bacterial genomes in a variety of environments and documented that Shewanella oneidensis are ubiquitous in all fresh and marine waters; they concluded that IO − 3 reduction is a major biogeochemical process. Thus, nitrate reductase (also an O atom transfer reaction) is not a requirement for bacterial IO − 3 reduction to I -.
The interconversion of dimethylsulfoxide with dimethylsulfide during dissimilatory IO − 3 reduction is another 2-electron O-atom transfer reaction. Moreover, the reactions of DMS to reduce HOI, IO − 2 and IO − 3 are thermodynamically favorable ( Figure 3B, DMSO reduction is in S3, Table 1 and occurs at a lower pe than NO − 3 and IO − 3 reduction). The reaction of DMS with HOI has been suggested by Müller et al. (2021) to be a sink for DMS based on the rapid reaction of DMS with HOBr. Figure 4 shows the thermodynamics for the stepwise conversion of IO − 3 to Iby reduction with Mn 2+ and Fe 2+ . Concentrations of Mn 2+ and Fe 2+ range from several nM to mM in OMZs (e.g., Trouwborst et al., 2006;Moffett and German, 2020) and in suboxic porewaters (e.g., Oldham et al., 2019;Owings et al., 2021) to mM in waters emanating from hydrothermal vents (e.g., Estes et al., 2022); in these cases, Mn 2+ and Fe 2+ are normally higher in concentration than the total iodine concentration. For the 2-electron transfer reactions with Mn 2+ , only the IO − 2 to HOI reaction is favorable over the entire pH range. The IO − 3 to IO − 2 reaction is favorable only at pH > 6 whereas the other reactions are favorable at pH > 3. Thus, the IO − 3 to IO − 2 conversion is the controlling step in the reaction sequence when Mn 2+ is the reductant. Using high resolution porewater profiles of Iand Mn 2+ obtained by voltammetric microelectrodes, Anschutz et al. (2000) showed that a Imaximum occurred at the depth where upward diffusing Mn(II) was being removed and proposed that Iformed by the reaction of IO − 3 with Mn 2+ under suboxic conditions. The reaction has not been investigated in laboratory studies.

Iodate reduction by Mn 2+ and Fe 2+
For the reaction sequence with Fe 2+ , all iodine species reductions are favorable over the entire pH range except for the I 2 to Ireaction, which is favorable at pH > 2.5. Thus, there is no thermodynamic inhibition to IO − 3 reduction to Iby Fe 2+ , and this abiotic reaction at a pH of 7 was reported to be 92% complete after 2 hours using initial concentrations of 2 mM Fe 2+ and 0.1 mM IO − 3 (Councell et al., 1997

Iodate reduction by sulfide
In sulfidic waters and porewaters, IO − 3 does not exist as sulfide reacts readily with it (Zhang and Whitfield, 1986), and S(0) forms as the initial sulfur product. Figure 5 shows the thermodynamics for the stepwise conversion of IO − 3 to IO − 2 and to HOI by sulfide where S(0) forms as an intermediate leading to S 8 . As the Gibbs free energy of formation for HSOH is unknown, HSOH could not be evaluated as an  intermediate, which on continued oxidation would form SO 2− 4 . The reaction of sulfide with I 2 and HOI is well known as the iodometric titration, so calculations were not performed. The only unfavorable iodine reduction reactions are the 1-electron reductions that lead to the formation of the HS radical (HS• or HS rad). The conversion of IO − 2 to HOI is more favorable as it has the larger DlogK reaction values. Again, the IO − 3 to IO − 2 conversion is the least favorable and likely controlling step in this reaction sequence.
3.6 Iodate reduction by NH 4 + Figure 6 shows the thermodynamics for the stepwise conversion of IO − 3 to IO − 2 by NH + 4 where hydrazine (N 2 H 4 ) and hydroxylamine (NH 2 OH) as well as their protonated forms could form as the first N intermediates. The thermodynamic calculations for these 2 electron transfers indicate that these reactions are not favorable. However, the reaction of the intermediates, if they could form by other processes, with IO − 3 to form N 2 is very favorable. Thus, the IO − 3 to IO − 2 conversion is the controlling step in the reaction sequence with NH + 4 .
4 Results and discussion: Iodide oxidation

Iodide oxidation by NO 3 -, MnO 2 and Fe(OH) 3
Figures 3-6 showed the reduction of IO − 3 with reductants. All values of DlogK reaction > 0 indicate a favorable reaction and all values of DlogK reaction < 0 indicate an unfavorable reaction. These figures can be used to discuss the reverse reaction of Ioxidation with oxidants. For reverse reactions, when a DlogK reaction < 0, then Ioxidation is favorable, but when DlogK reaction > 0, Ioxidation is unfavorable.
In Figure 3, NO − 3 is not an oxidant for I -(reverse of the NO − 2 and I 2 reaction) except for the formation of I 2 at a pH< 4.
In Figure 4, MnO 2 oxidizes Ito HOI (reverse of the Mn 2+ and HOI reaction) at a pH< 3 and Ito I 2 (reverse of the Mn 2+ and I 2 reaction) at a pH< 5. A couple of laboratory studies showed Ioxidation with synthetic birnessite (d-MnO 2 ). First, Fox et al. (2009) showed that I 2 was produced over the pH range 4.50 -6.25, and that IO − 3 formed in smaller amounts. The kinetics of the reaction were slower at higher pH by 1.5 log units (> 30-fold) and were slower when smaller amounts of MnO 2 were added (Table 2). Allard et al. (2009) investigated the same reactants to a pH of 7.5 and found I 2 and IO − 3 as products; above pH = 7 the reaction is very slow. Iodate was found mainly in lower pH waters. Both I 2 and IO − 3 adsorb to the birnessite surface. Similar results have been found over the pH range 4-6 for Mn(III) solids (Szlamkowicz et al., 2022). These MnO x reactions with Iare much slower that the reactions with reactive oxygen species (Table 2). Nevertheless, these are important as Elderfield (1987a, 1987b) showed that the conversion of iodide to iodate occurred in marine sediments. Figure 4 also shows that Ioxidation by Fe(OH) 3 to I 2 (reverse of the Fe 2+ and I 2 reaction) should occur only at a pH< 2.5. The Ito HOI conversion (reverse of the Fe 2+ and HOI reaction) is favorable at pH ≤ 0.5. Figure 7 shows the thermodynamics of Ioxidation to I 2 by oxygen species. Figure 7A shows that the one-electron process for Ioxidation with 3 O 2 is thermodynamically unfavorable over all pH whereas Figure 7B shows that the two-electron process is favorable at a pH< 3. Figure 7A shows that the successive 1-electron oxidations of Iwhere superoxide (O − 2 ) is reduced to hydrogen peroxide (H 2 O 2 ), which is reduced to hydroxyl radical (•OH). Only •OH is thermodynamically favorable over the pH range considered. O − 2 and H 2 O 2 show favorable reactions at pH< 9 and pH< 6, respectively.

Iodide oxidation by oxygen species
By contrast, Figure 7B shows that the reactions of Iwith the 2electron oxidants H 2 O 2 , 1 O 2 and O 3 are all thermodynamically favorable. The likely reaction pathway is the loss of 2-electrons to produce I + , which then reacts with Ito form I 2 . Note that H 2 O 2 reacts to form H 2 O not •OH in Figure 7B. The 2-electron reaction with O 3 ( Figure 7B) is more favorable than the 1-electron reaction ( Figure 7A). Wong and Zhang (2008) showed that H 2 O 2 oxidizes Iin artificial seawater from pH 7-9, which is consistent with Figure 7B. However, Ioxidation does not lead to iodate. In fact, Ireforms. They proposed that I 2 formed and was reduced back to I -, but they did not provide a mechanism. The reverse reaction of I 2 with •OH (DlogK reaction < 0 in B A

FIGURE 5
Thermodynamics for the 2-electron transfer reductions of (A) IO − 3 (Io5b) and (B) IO − 2 (Io4b) by sulfide species (S1, S2, S4, S5). The vertical line represents the pK a1 value for H 2 S. Data above the horizontal line at DlogK (DlogK reaction ) = 0 indicate a favorable reaction and data below the horizontal line indicate an unfavorable reaction. the plot) is favorable to reform H 2 O 2 and Iat pH > 6 whereas the reverse reaction of H 2 O 2 with I 2 to reform O − 2 and Iis favorable at a pH > 9 (DlogK reaction < 0 in the plot). These thermodynamic data indicate that H 2 O 2 can form I 2 in a 2-electron transfer ( Figure 7B) and then reduce I 2 to Iin a 1-electron transfer ( Figure 7A).
At seawater pH, superoxide, O − 2 , can oxidize Ito I 2 and the reaction occurs with a rate constant of 10 8 M -1 s -1 (Bielski et al., 1985; Table 2). Because I 2 is a good electron acceptor, the subsequent reaction of O − 2 with I 2 leads to I − 2 (Schwarz and Bielski, 1986). As to be discussed in section 4.3, I 2 reacts with organic matter to form organoiodine compounds. Extracellular O − 2 is generated by Roseobacter sp. AzwK-3b (Li et al., 2014) and results in the oxidation of Mn 2+ to Mn (III,IV) oxides. However, Li et al. (2014) found that O − 2 also oxidized I -. Considering that extracellular O − 2 formation is a widespread phenomenon among marine and terrestrial bacteria, this could represent an important first step in the pathway for iodide oxidation in some environments. The Mn oxides formed by Roseobacter sp. AzwK-3b are not the oxidant as MnO 2 kinetics is slower (Table 2).
To obtain IO − 3 , further oxidation of I 2 to HOI must occur, and •OH is one candidate with a rate constant of 1.2 x 10 10 (Buxton et al., 1988; Table 2). Also, O 3 has a rate constant of 1.2 x 10 9 (Liu et al., 2001). I 2 is a prominent intermediate in Ioxidation yet HOI is needed to form IO − 3 . HOI can form directly from Iand I 2 oxidation or from hydrolysis of I 2 (reverse of eqn. 5), which is fast at basic pH (Wong, 1991). Figure 8A shows that of the successive 1-electron oxidants (starting from O 2 ) for I 2 oxidation, only •OH is thermodynamically favorable over all pH to form HOI whereas O 3 is favorable at pH > 6, and O − 2 is favorable at pH< 6. H 2 O 2 as a 1-electron oxidant cannot oxidize I 2 to form HOI, but H 2 O 2 can reduce HOI to I 2 (reverse of the O − 2 and I 2 reaction). Figure 8B indicates that, as 2-electron oxidants, H 2 O 2 and O 3 oxidation can lead to HOI formation. Comparing Dlog K values in Figures 7, 8 indicates that oxidation of I 2 to HOI is less favorable than the oxidation of Ito I 2 .
These data also indicate why the comproportionation reaction of HOI with Ito form I 2 can occur (eqn. 5, Carpenter et al., 2013).
Although disproportionation of HOI to IO − 3 and I -(eqn. 6) is fast in strongly basic solution, it is not detectable at seawater pH (Wong, 1991). Figure 9 shows the successive 2-electron oxidation reactions of I -, HOI and IO − 2 with 3 O 2 , 1 O 2 , H 2 O 2 and O 3 . 3 O 2 cannot affect the oxidation at any pH. Figure 9 shows that O 3 oxidation reactions with I -, HOI and IO − 2 are favorable; thus, O 3 can affect the complete oxidation of Ito IO − 3 . Also, the H 2 O 2 oxidation reactions of I -, HOI and IO − 2 are favorable and can lead to IO − 3 formation; however, the kinetics of H 2 O 2 oxidation can be slow. Haloperoxidase enzymes from organisms enhance the kinetics (Butler and Sandy, 2009) as does the reaction of H 2 O 2 with carboxylic acids secreted by microbes to form peroxy carboxylic acids, which in turn oxidize Ito I 2 (Li et al., 2012). The reactive oxygen species 1 O 2 can oxidize Iat pH< 10, oxidize HOI at pH > 5, and IO − 2 over all pH. Thus, 1 O 2 can be an oxidant of Ito IO − 3 at seawater pH. These data indicate that HOI oxidation leads to IO − 3 formation. Interestingly, DlogK values in Figure 9A show that the thermodynamics of Ioxidation by the 2-electron oxidants O 3 and H 2 O 2 to form HOI is slightly less favorable than I 2 formation Thermodynamics for the reduction of IO − 3 to IO − 2 (IO5b) by NH + 4 to hydrazine species (N3a, N3b), hydroxylamine species (N4a, N4b) and by N 2 H + 5 (N5) and NH 2 OH (N6). The vertical lines represent the pK a values for NH 3 OH + (5.82) and N 2 H + 5 (7.93), respectively. Data above the horizontal line at DlogK (DlogK reaction ) = 0 indicate a favorable reaction and data below the horizontal line indicate an unfavorable reaction.  Figure 7B). Conversely, thermodynamics of Ioxidation by H 2 O 2 as a 2-electron oxidant to form HOI ( Figure 9A) is more favorable than I 2 formation ( Figure 7A). As shown in Figure 10, a potentially potent oxidant for Iis N 2 O, which is an O atom transfer oxidant like O 3 . However, the N 2 O concentration in seawater is minor, but the largest reported values are 90 and 250 nmol kg -1 for the OMZs of the Arabian Sea (Freing et al., 2012) and the Eastern Tropical North Pacific (Damgaard et al., 2020), respectively. These values are smaller than the total iodine concentration in seawater. The N 2 O concentration in the atmosphere is 335 ppbv (August 2022, https://www.n2olevels.org), which is equivalent to 0.0331 Pa or 7.8 nM dissolved in surface seawater (salinity of 35) at 20 0 C using the solubility data from Weiss and Price (1980).

ROS in seawater
Reactive oxygen species exist in marine waters, but at low concentrations. O 3 penetrates a few micrometers through the water-air interface at surface iodide concentrations (Carpenter et al., 2013). Powers and Miller (2014) showed that solar-induced processes with organic matter in freshwater and seawater are a major source of ROS (as O − 2 , H 2 O 2 , and •OH) with the inventory and production rates for H 2 O 2 in surface seawater being highest of the ROS. Also, Sutherland et al. (2020) report that dark, extracellular O − 2 production is prolific among marine heterotrophic bacteria, cyanobacteria, and eukaryotes. In surface ocean waters, the concentration of H 2 O 2 ranges from 20 -80 nM (Yuan and Shiller, 2001), biological O − 2 production gives a total concentration of~0.07 to 0.30 nM (Sutherland et al., 2020), •OH concentration is~10 -18 M (Mopper and Zhou, 1990), and 1 O 2 concentration ranges from 10 -13 to 10 -14 M (Sunday et al., 2020). However, these ROS concentrations are typically smaller than Iconcentrations, which range from 10 to 200 nM (Chance et al., 2019). Thus, Ioxidation in seawater samples should be difficult to observe experimentally. Hardisty et al. (2020) tracked the addition of stable isotopes of iodide in sample incubations and report the rate of Ioxidation to be 118-189 nM yr -1 , which is similar to rates reported by mass balance approaches (Campos et al., 1996a;Truesdale et al., 2001;Žic and Branica, 2006;Žic et al., 2008). Hardisty et al. (2020) report that the product is likely HOI that results in the formation of organic-iodine compounds (see section 4.6) which on decomposition can release I -.
As the surface concentrations of ROS are smaller than the Iconcentration, the question is how does Iget oxidized to IO − 3 in seawater? Microbial processes and the oxidation of I species in the atmosphere by ROS are likely candidates. These are now discussed.

Iodide oxidation in brown kelp
Brown kelp are the strongest accumulators of iodine as Iamong living organisms (up to 100 mM, Küpper et al., 2008). The element iodine was discovered by the formation of I 2 during exposure of brown kelp to concentrated sulfuric acid, which oxidized Ito I 2 . Kelp releases Ion the thallus surface and in the apoplast when undergoing oxidative stress during the partial emersion of the brown kelp forest at low tide; e.g., by exposure to high irradiance, desiccation, and atmospheric O 3 . Kelp contain vanadium haloperoxidases (Colin et al., 2003;Küpper et al., 2008) that enhance Ioxidation by H 2 O 2 . Whereas the nonenzymatic reaction of Iwith H 2 O 2 is slow, the reactions with O 3 , O − 2 , 1 O 2 , and •OH are very fast (> 10 8 M -1 s -1 , Table 2); they are also faster than MnO 2 oxidation of I -. Küpper et al. (2008) consider Ias the simplest antioxidant known. Hughes et al. (2021) report that IO − 3 production occurs in cultures of the ammonia-oxidizing bacteria Nitrosomonas sp. and Nitrosococcus oceani supplied with I -, but not in cultures of three different nitrite oxidizing bacteria. Information on the enzymes mediating the oxidation were not studied. Nevertheless, NH + 4 oxidation via nitrification occurs via NH 2 OH formation which is a 2-electron reaction. Further reaction of NH 2 OH via metalloenzymes (e.g., Mo and W oxidases that transfer O atoms) leads to NO − 2 (a 4electron transfer) and NO − 3 . Ilikely goes through the intermediates IO − 2 and HOI to form IO − 3 , but these intermediates are reactive and not detectable by present analytical methods unlike NO − 2 . Consistent with the Hughes et al. (2021) report, Elderfield (1987a, 1987b) showed that the conversion of iodide to iodate occurred in marine sediments. Microbial intervention is likely, but reaction of Iwith oxidized Mn is possible depending on the pH. Amachi and Iino (2022) reviewed the genus Iodidimonas, which was originally found in brines, but was also cultured from seawater enriched with I -. I 2 is the first oxidation product. Iodidimonas contains the iodide oxidizing enzyme (IOX), which is an extracellular protein that contains multicopper oxidases. Iodidimonas requires O 2 , not H 2 O 2 , as the electron acceptor. Other oxidants are required to oxidize I 2 to IO − 3 with O 3 and •OH being the most effective. Thermodynamics for the reaction of 3 O 2 (O2), 1 O 2 (O5), H 2 O 2 (O3), and O 3 (O4) as 2-electron oxidants with (A) Ito form HOI (Io2), (B) HOI to form IO − 2 (Io4b) and (C) IO − 2 to form IO − 3 (Io5b). The vertical lines represent the pK a value of 4.49 for HO 2 I dissociation to IO − 2 . Data above the horizontal line at DlogK (DlogK reaction ) = 0 indicate a favorable reaction and data below the horizontal line indicate an unfavorable reaction.

HOI and I 2 formation leads to organic iodine
Competing with the inorganic interconversion between iodide and iodate is the formation of organic iodine compounds. Formation of C-I bonds can occur during the reduction of IO − 3 and oxidation of I -. Complete reduction of IO − 3 to Idoes not need to occur intercellularly and can lead to HOI and I 2 formation as in Figure 4. The first step in Ioxidation also leads to I 2 and HOI. I 2 is neutral and adds to organic compounds such as olefins, which are not very reactive in seawater, whereas I + in HOI reacts with a-keto compounds and peptides through keto-enol isomerization (Truesdale and Luther, 1995). Both I 2 and HOI lead to volatile and nonvolatile organic-iodine (R-I) compounds with C-I or N-I bonds, and Harvey (1980) showed that N-iodo amides were the main organic iodine components in marine sediments. On decay of organic compounds, the C(N)-I bond breaks leading to Irelease, which mimics the senescence pathway outlined by Bluhm et al. (2010) and Hepach et al. (2020). Recently, Ooki et al. (2022) showed that CH 3 I and CH 3 CH 2 I formed in sediments from polar and subpolar seas and was related to increased phytodetritus at the seafloor after the spring bloom. Allard and Gallard (2013) showed that the oxidation of Iby birnessite in the presence of organic matter also led to CH 3 I over the pH range 4-5.
As total iodine in surface ocean waters is lower by a few percent compared to deep waters (Wong, 1991), the decomposition of organic-iodine leads to some Irelease, which may be oxidized to IO − 3 by ammonia-oxidizing bacteria (Hughes et al., 2021). This is similar to release and oxidation of NH + 4 to NO − 3 from particulate organic matter in deep waters that results in an increase of NO − 3 concentration with depth (recycled element profile). Deep waters contain mainly IO − 3 , so not much Iis released to the deep-water column by in situ water column processes, and most organic-iodine gets to the sediments where it is released as I - Elderfield, 1987a, Kennedy andElderfields 1987b;Luther et al., 1995). Elderfield (1987a, 1987b) and Shimmield and Pedersen (1990) report that the molar I/C ratio in planktonic organisms is 10 -4 whereas it is typically >10 -3 in sediments. Decomposition of sedimentary organic-I releases Ito porewaters and the overlying water column where it can be transported hundreds of kilometers offshore along isopycnal surfaces in OMZs (Farrenkopf and Luther, 2002;Cutter et al., 2018). 4.7 Surface seawater and atmospheric formation of IO 3 -, and iodine speciation in the atmosphere There is significant literature showing that coastal and oceanic regions are sources of iodine emissions to the atmosphere, and I note some important aspects of this air-sea connection. Ireacts with O 3 to form IO -, which at seawater pH forms HOI. Carpenter et al. (2013) showed that this reaction occurs in the first few micrometers below the air-water interface and that HOI is ten-fold greater than I 2 above the sea surface. HOI contributes 75% of the observed iodine oxide aerosol levels over the tropical Atlantic Ocean, and these iodine emissions to the atmosphere have increased 3-fold over the last century due to the increase in anthropogenic O 3 (Carpenter et al., 2021). O 3 reacts stepwise with this gaseous HOI (IO -) and gaseous IO − 2 to form IO − 3 , which can attach to aerosols.
Formation and release of gaseous I 2 from seawater to air permits photochemical breaking of the I-I bond to form gaseous I atoms, •I, which are reactive radicals. Similarly, release of volatile organic-iodine compounds leads to the homolytic cleavage of the C-I bond to form I•. O 3 reacts readily with I• to form gaseous IO• in the marine boundary layer (Whalley et al., 2010). Further stepwise oxidation of gaseous IO•/HOI leads to IO − 3 . In laboratory experiments using mass spectrometry detection, Teiwes et al. (2019) showed that hydrated iodide, I(H 2 O) -, reacts with gaseous O 3 to form IO − 2 directly without formation of gaseous HOI or IO -; thus, HIO 3 =IO − 3 can form in a twostep reaction sequence in the atmosphere.
Using mass spectrometry to evaluate atmospheric I x O y cluster and (nano)particle formation above seabed macroalgae, Sipilä et al. (2016) showed the stepwise formation of HIO 3 via HOI and IO•, which leads to (I 2 O 5 ) x clusters (x=2-5) containing HIO 3 that result in iodine rich aerosol particles. These data on the formation of I 2 O 5 aerosols agree with the exothermic DH reaction values of iodine oxide species reacting with O 3 and each other calculated using quantum mechanics (Kaltsoyannis and Plane, 2008). Sipilä et al. (2016) also showed that cluster formation increased as a burst at low tide indicating significant I 2 release from the macroalgae (and subsequent oxidation) as found by Küpper et al. (2008). Hydration of I 2 O 5 leads to two IO − 3 . In mass spectrometry laboratory studies, Martıń et al. (2022) showed that new iodine containing (nano) particles and IO − 3 also form in the presence of NO − 3 and provide DH reaction data for the gas phase reactions involved. Experiments using the CERN CLOUD (Cosmics Leaving Outdoor Droplets) chamber documented the formation of HIO 3 via iodooxy hypoiodite, IOIO, as an intermediate (Finkenzeller et al., 2022) and the fast growth of HIO 3 as (nano)particles (He et al., 2021).
In recent atmospheric campaigns, Koenig et al. (2020) showed that IO − 3 is the main iodine reservoir as it forms on aerosols in the stratosphere with iodine being responsible for 32% of the halogen induced O 3 loss. Cuevas et al. (2022) also showed that iodine can dominate (∼73%) the halogen-mediated lower stratospheric ozone loss during summer and early fall, when the heterogeneous reactivation of inorganic chlorine and bromine reservoirs is reduced.
The information in the preceding paragraphs along with the thermodynamic data from Martıń et al. (2022), Figure 2 (the half reaction for O 3 to O 2 and H 2 O) and Figure 9 predict that IO − 3 should be the dominant species in the atmosphere. Although reduction of IO − 3 is not predicted in an oxidizing atmosphere, analyses of rainwater (Campos et al., 1996b;Truesdale and Jones, 1996;Baker et al., 2001;Hou et al., 2009), aerosols (Gilfedder et al., 2008;Droste et al., 2021) and snow (Gilfedder et al., 2008) in the marine boundary layer indicate that aqueous iodide and iodate coexist. Hou et al. (2009) reviewed wet iodine speciation data and reported that IO − 3 predominates over Ifrom marine sources/air masses whereas Ipredominates from continental air masses.
There are several ways that I -(or reduced I) can form in rainwater and aerosols. The interconversion between IO − 3 and Iat the pH of wet deposition also leads to HOI and I 2 , which can react with organic material forming C-I bonds that can release I -(section 4.6). This material has been given the term soluble organically bound iodine and can be larger than the sum of the concentrations of IO − 3 and Iin aerosols (Gilfedder et al., 2008;Droste et al., 2021). Soluble organically bound iodine can form from release of natural organic iodine from land and sea (a primary source) or from the reaction of natural organic material with HOI or I 2 in the atmosphere (a secondary source). On photolysis of C-I, I• forms and reacts with O 3 , and on C-I reaction with nucleophiles, Iforms. During a study on the formation of cloud condensation nuclei, Huang et al. (2022) also showed that natural gaseous organic material in the marine boundary layer reacts with IO − 3 in aerosols resulting in gaseous I 2 , which can be reoxidized to IO − 3 (catalysis) or react to form organic-I compounds. Lastly, Cuevas et al. (2022) reported that photolysis of IO − 3 particles in the stratosphere at a wavelength of about 260 nm can lead to gaseous I• and O 2 during transport from the tropics to the Antarctic region. Thus, there are several pathways for reduction of IO − 3 in the atmosphere.

Conclusions
The reduction of IO − 3 to Iin solution is a facile process by biotic and abiotic reactions. The intermediates IO − 2 and HOI dictate the reactivity sequence via a combination of thermodynamic and kinetic considerations. The IO − 3 to IO − 2 conversion is the least favorable and likely controlling step in this reaction sequence, but there is no need for nitrate reductase for IO − 3 reduction based on numerous studies. The data from this study indicate that once IO − 2 forms there is no thermodynamic barrier to Iformation. Chemical reduction of all iodine species (not iodide) by sulfide, Fe 2+ and Mn 2+ are favorable at seawater and sedimentary pH values, but only sulfide has been studied in the laboratory at oceanic pH values. Dissimilatory IO − 3 reduction during organic matter decomposition seems to be a key process as the IO − 3 =IO − 2 couple is more favorable than the NO − 3 =NO − 2 couple. However, the oxidation of Iback to IO − 3 via 3 O 2 has a major thermodynamic barrier in solution, and the disproportionation of HOI at seawater pH values is not measurable. Thus, ROS, oxidized Mn and microbes are important for Ioxidation to IO − 3 due to favorable thermodynamics and kinetics (Table 2). Recent reports of microbial oxidation have not documented the entire six-electron oxidation in a stepwise manner so further work on this topic is necessary. Oxidation of Iby oxidized Mn is a pH dependent reaction and less likely at seawater pH values but could occur in sedimentary environments. The reactions of O 3 and •OH with iodine species (not IO − 3 ) are thermodynamically favorable over all pH. However, ROS are not normally in significant concentration in seawater to influence IO − 3 formation. Notable exceptions are for (1) sea surface microlayer, which adsorbs atmospheric O 3 , and (2) the reaction of Fe 2+ with O 2 that leads to Fenton chemistry with •OH production. Systems where Fenton chemistry can occur are at/near hydrothermal vents (Shaw et al., 2021), submarine groundwaters (Burns et al., 2010), and sediments or water columns where O 2 and Fe 2+ concentration profiles overlap including ancient earth (Chan et al., 2016). Iis a major sink for O 3 in the sea surface microlayer and the atmosphere. IO − 3 formation in the atmosphere and IO − 3 redeposition to surface seawater may be major iodine processes with the latter being similar to the deposition of trace metals from wet and dry deposition to the surface ocean (e.g., Chance et al., 2015;Meskhidze et al., 2019). Most atmospheric iodine originates from marine sources where Ioxidation to I 2 and homolytic cleavage of C-I bonds occurs; thus, gaseous iodine emissions from the ocean are reduced. IO − 3 forms from these sources during oxidation by O 3 in the atmosphere. An estimate of atmospheric deposition of IO − 3 to the ocean surface could be made by using the amount of IO − 3 in rainwater and aerosols that would be returned to the ocean surface, but more information on iodine speciation in rainwater and aerosols is needed as global spatial coverage appears limited. Despite major advances in iodine geochemistry over the last two decades, significant research is still needed on the processes that affect Ioxidation to IO − 3 in the atmosphere, seawater and ocean sediments.

Author contributions
The author confirms being the sole contributor of this work and has approved it for publication.