Formation dynamics of inorganic iodine species during UV-based advanced oxidation of iopamidol and iohexol and their correlation with iodinated disinfection by-product yields†

Hojoong Ji, Jaehyeong Park, Seonyoung An, Seo-Yeong Choi and Jong Kwon Choe*
Department of Civil and Environmental Engineering and Institute of Construction and Environmental Engineering, Seoul National University, 35-402, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Republic of Korea. E-mail: jkchoe@snu.ac.kr; Fax: +82 2 873 2684; Tel: +82 2 880 2278

First published on 4th July 2025

---

---

Water impact Treatment of iodinated contaminants in water should focus not only on their degradation but also on the fate of released iodine to prevent formation of iodinated disinfection by-products. This study compared iodine species formation during treatment of iohexol and iopamidol across different UV-AOPs. UV with NaOCl and PMS fully oxidized iodide to iodate minimizing iodinated disinfection by-products while others did not.

1. Introduction

Iopamidol and iohexol are two most widely used iodinated contrast media (ICM) in medical X-ray imaging, and these ICM comprise 10–80% of the mass loading of pharmaceuticals in hospital and domestic wastewater.^1,2 ICM persistence leads to their wide occurrence in various aquatic systems, including the marine environment, surface water, and drinking water.^3–5 Long-term ICM exposure poses adverse health effects (e.g., thyroid dysfunction)^6,7 and their presence and persistence in aquatic systems are problematic as major precursors of iodinated disinfection by-products (I-DBPs). Chlorination and chloramination of ICM are known to produce I-DBPs during drinking water treatment processes,⁸ and ICM and I-DBP concentrations were positively correlated in an investigation of surface water and drinking water treatment plants.⁹ Although their concentrations are typically lower than those of chlorinated and brominated DBPs in drinking water, I-DBPs are known to be more cytotoxic and genotoxic than their chlorinated/brominated analogs.¹⁰ For instance, in an in vitro chronic cytotoxicity study, the LC₅₀ values of I-DBPs were 10–1000 times lower than those of other halogenated DBPs.¹¹ Thus, control of I-DBP formation during water treatment is deemed to be necessary.

Conventional wastewater treatment plants (WWTPs) employing biological and physicochemical processes are not effective in removing ICM. For instance, almost no ICM removal was reported in an activated sludge-based WWTP.¹² Moreover, diatrizoate and iopromide were not removed from hospital wastewater in lab-scale studies using an activated sludge bioreactor and membrane bioreactor.^1,13 Additionally, physicochemical processes such as ozonation or granular activated carbon adsorption were not overly effective in removing ICM, as 30–90% removal was attained with ozonation in a pilot scale study, and 70% removal was reported with granular activated carbon filtration combined with ozonation in a full scale study.^14,15

The use of advanced oxidation processes (AOPs), which generate reactive radicals that react with pollutants, has shown mixed results for ICM removal efficiency in water. The removal efficiency of iohexol, one of the most commonly used ICM, reached less than 40% during 7 min of reaction in UV/H₂O₂ treatment and less than 50% in Fe₃O₄/O₃ treatment with contact times of 10 min.^16,17 By contrast, UV-based AOPs (UV AOPs) that employ sulfate- or chlorine-based radical sources have shown better effectiveness in removing ICM. Iohexol was successfully degraded (>99%) within 1.5 and 5 min in UV/S₂O₈²⁻ (with a UV fluence of 0.225 mW cm⁻²) and UV/Cl₂ (3.02 mW cm⁻²) processes, respectively.^16,18 Similarly, iopamidol was effectively removed via UV/S₂O₈²⁻ and UV/Cl₂ treatment (0.128–0.133 mW cm⁻²), attaining more than 70% removal within 1 min and 99% removal within 2.0 min, respectively.^19,20

Previous studies have also investigated the formation of I-DBPs during chlorination of UV AOP-treated ICM. For example, I-HAAs are known by-products of UV/Cl₂ and UV/H₂O₂ treatment of ICM mixtures.²¹ Zhao et al. also reported I-THM and I-HAA formation during post-chlorination of UV AOP-treated iopamidol solutions, comparing UV only, UV/H₂O₂, UV/PDS and UV/NaOCl processes.¹⁹ Tian et al. suggested that iopamidol releases I⁻ during UV irradiation and I⁻ would likely be the primary contributor to I-DBP formation during post-chlorination.²² Reactive iodine species (RIS), such as HOI, I₂, and I₃⁻, are also known to generate I-DBPs due to their reactivity toward phenolic structures in natural organic matter (NOM).²³ While many studies including review papers suggested that I⁻ and RIS generated during UV AOP treatment of ICM are linked to the I-DBP formation, no studies so far have quantified these inorganic iodine species (i.e., I⁻, RIS, IO₃⁻) and compared their formation dynamics during the degradation of ICM via different AOPs to identify whether there is a direct correlation between these iodine species and formation of I-DBPs. Ye et al. suggested in their review paper the importance of measuring inorganic iodine species during ICM treatment as a future research area to clearly identify their contribution to the formation of I-DBPs.²⁴

Therefore, this study attempted to systematically compare multiple UV AOPs including mixed-oxidant processes using the same oxidant concentration and solution conditions. The aims of this study were 1) to compare the ICM degradation and inorganic iodide species formation characteristics across different UV AOPs and 2) to assess I-DBP formation during UV-AOP treatment in relation to RIS concentrations under an identical UV light source for unbiased comparison of multiple UV AOPs. Iohexol and iopamidol were chosen in this study because these two compounds are the most widely used ICM in clinical practices.

2. Materials and methods

2.1. Chemicals and reagents

Iohexol (≥99%), iopamidol (≥99%), peroxymonosulfate (PMS, KHSO₅·0.5KHSO₄·0.5K₂SO₄, OXONE®, monopersulfate compound), potassium iodide (KI, ≥99%), sodium acetate anhydrous (≥99.0%), sodium tetraborate decahydrate (≥99.5%), sodium thiosulfate (≥99%), 2,6-dibromophenol (≥99%), EPA 501/601 THM calibration mix, EPA 552.2 HAA calibration mix, and iodoform (≥99%) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Hydrogen peroxide (H₂O₂, 35% solution in water), acetic acid (99.8%), potassium phosphate monobasic (≥99%), and potassium phosphate dibasic anhydrous (≥99%) were obtained from Daejung Chemicals & Metals (Gyeonggi, South Korea). Peroxydisulfate (PDS, potassium persulfate, ≥99%), sodium hydroxide beads/pellets (≥99.99%), and iodine solution (I₂, 1.0 N in water) were purchased from Alfa Aesar (Ward Hill, MA, USA). Sodium hypochlorite solution (NaOCl; 5.65–6%) and methyl tert-butyl ether (MtBE, ≥99.0%) were purchased from Fisher Chemical (Pittsburgh, PA, USA). Suwannee River NOM (Cat. no. 2R101N) was obtained from International Humic Substances Society (Denver, CO, USA). Iodinated trihalomethane (I-THM) and iodinated haloacetic acid (I-HAA) standards, including dichloroiodomethane (≥95%), chlorodiiodomethane (≥95%), chloroiodoacetic acid (≥90%), and diiodoacetic acid (≥90%), were purchased from CanSyn Chemical Corporation (Toronto, ON, Canada). Stock solutions for diatrizoate, iohexol, and iopamidol were prepared with deionized water (≥18.2 MΩ cm) from a Milli-Q water purification system (Millipore, Billerica, MA, USA). These stock solutions were maintained at a concentration of 5 g L⁻¹ and stored in the dark at 4 °C prior to their use.

2.2. Degradation of ICM in UV AOPs

The reactor consisted of an acrylic water jacket, a glass reaction cylinder, a quartz tube UV lamp (TUV 15 W SLV/6; Philips®, Amsterdam, The Netherlands), and a thermostatic pump. The UV lamp was turned on for 30 min pre-reaction. The UV fluence rate at 254 nm was measured using atrazine degradation during UV photolysis of 5 μM atrazine solution (500 mL) maintained at pH 7.0 with 10 mM phosphate buffer. The UV fluence rate at 254 nm was calculated to be 2.88 ± 0.03 mJ cm⁻² s⁻¹ based on the quantum yield and molar absorption coefficients provided by Canonica et al.²⁵

A glass reaction tube with a 500 mL reaction volume containing a single ICM compound (10 ppm, 12.2 μM iohexol or 12.9 μM iopamidol), oxidants, and 10 mM phosphate buffer at pH 7.0 was irradiated under UV light. The initial ICM concentration was 10 ppm for the degradation kinetics and inorganic iodine speciation experiments and 50 ppm for the THM and HAA analyses. The temperature of the water-jacketed glass reaction tube was maintained at 20 °C and the solutions were continuously stirred at 700 rpm. During the reaction, 1 mL aliquots were periodically withdrawn and immediately transferred to 2 mL amber glass vials containing sodium thiosulfate or phenol to quench the oxidants and RIS. These samples were directly analyzed or diluted for optimal analysis.

2.3. Analytical methods

The iohexol and iopamidol concentrations in solution were determined using single quadrupole liquid chromatography-mass spectrometry (LC-MS) (6120 Single Quad LC/MS; Agilent Technologies, Santa Clara, CA, USA). Samples for ICM analysis were quenched with 1.0 mM sodium thiosulfate to minimize unwanted reactions between analytes and oxidants.

The iodide (I⁻) and iodate (IO₃⁻) concentrations were determined via ion chromatography (Dionex® ICS-1100) equipped with a Dionex® AG-23/AS-23 column, Dionex® DRS 600 suppressor, and Dionex® 7693 autosampler (all Thermo Fisher Scientific). Samples were prepared by adding 1.0 mM phenol as an RIS quencher to prevent further RIS reactions from forming I⁻ or IO₃⁻. The addition of phenol has no influence on I⁻ or IO₃⁻ that has already formed.

RIS were quenched with excessive 2,6-dibromophenol (2.5 mM) to form 4-iodo-2,6-dibromophenol via electrophilic substitution.²⁶ Molecular iodine (I₂) and hypoiodous acid (HOI) were able to produce 4-iodo-2,6-dibromophenol in an equal stoichiometric ratio. HOI was prepared by adding excessive KI to HOCl solution. Newly formed 4-iodo-2,6-dibromophenol was quantified by HPLC-UV (YL® 9100, Young Lin Instrument Co., Ltd., Anyang-Si, South Korea).

THM and HAA samples were prepared via liquid–liquid extraction from aqueous solutions to MtBE, based on EPA methods 551.1 and 552.2, respectively.^27,28 THM and HAA formation potentials (THMFP and HAAFP, respectively) were calculated by measuring the changes in THM or HAAME concentrations in aqueous solution prior to and after 3-day chlorination (5 mg L⁻¹) using gas chromatography-mass spectrometry (GC-MS) (Agilent® 5977B).

The details of the analytical methods are provided in Text A.S1–S3 in Appendix A.†

3. Results and discussion

3.1. Degradation kinetics of ICM in UV AOPs with oxidants

Fig. 1 shows the values of the observed degradation rate constants (k_obs) for iohexol and iopamidol after UV AOP treatment, including the UV-only, UV/H₂O₂, UV/PMS, UV/PDS, and UV/NaOCl processes as well as those utilizing UV irradiation with dual oxidants (i.e., PMS/PDS, PMS/H₂O₂, PDS/H₂O₂, PMS/NaOCl, PDS/NaOCl, H₂O₂/NaOCl). For single-oxidant UV AOPs, 1 mM of each oxidant was applied. For dual-oxidant UV AOPs, 0.5 mM of each oxidant (total 1 mM) was used to evaluate the potential synergistic or antagonistic interactions between mixed oxidants. Solutions containing ICM, oxidants and 1 mM phosphate buffer were exposed to 254 nm UV at a fluence rate of 2.88 ± 0.03 mJ cm⁻² s⁻¹. All kinetic experiments were performed in triplicate. ICM degradation in UV AOPs followed the pseudo first-order kinetics, and k_obs values were obtained with R² values > 0.98 (Fig. A.S1 and S2 in Appendix A†).

The k_obs values for iohexol were in the range of 0.25–1.0 min⁻¹, in which the slowest iohexol degradation was observed in the UV-only and UV/H₂O₂ + NaOCl processes while the fastest degradation rates were observed in the PMS and PDS-based processes. Compared to the UV-only process, the reactivity was increased by 58% when 1 mM H₂O₂ was added to enhance the formation of ·OH radicals. The UV/PMS, PDS and PMS + PDS processes generated ·SO₄⁻ radicals, which increased the reactivity by 3 to 3.5-fold, compared to that achieved in the UV-only process. Previous studies also report 3 to 7-fold enhancement in iohexol/iopamidol degradation in persulfate-based UV AOPs than the UV only process.^19,29

Whereas there was a negligible difference in kinetics between UV/H₂O₂ and UV/NaOCl processes for iohexol degradation, significant enhancement in reactivity for iopamidol degradation was observed in the UV/NaOCl system, with 14 times faster reactivity than that achieved in the UV-only process. The radicals formed in UV/NaOCl treatment (e.g., ·OH, ·Cl, ·OCl, and ·Cl₂⁻) are known to collaboratively induce electrophilic substitution in aromatic rings to form C–Cl bonds,³⁰ among which ·OCl radicals, in particular, have higher and more selective reactivity than other radicals with certain aromatic organics, such as diatrizoate (one of the ICM) and trimethoprim whose structures are similar to iopamidol.^31,32 The use of NaOCl alone was not able to degrade iohexol nor iopamidol (Fig. A.S1 and S2†), but UV/NaOCl showed specificity towards iopamidol (Fig. 1b). Rapid iopamidol degradation via the UV/NaOCl, UV/NaOCl + PMS, and UV/NaOCl + PDS processes was consistent with previously reported results claiming that the ·OCl radical selectively reacts with iopamidol.²⁰ However, such rapid degradation was not observed in UV/NaOCl treatment of iohexol. Unlike the k_obs values for iopamidol degradation, the k_obs for iohexol degradation in the UV/NaOCl process was 65% higher than that in the UV-only process and was comparable to that in other UV/oxidant processes. Wang et al. reported that the ·OH radical, rather than the ·OCl radical, plays a dominant role in iohexol degradation in the UV/Cl₂ process, which is identical to the UV/NaOCl process.¹⁸ Similar tendencies of more-than-5-fold reactivity enhancement for iopamidol and none for iohexol in UV/chlorine processes were also reported in previous studies.^18–20,30 On the other hand, significantly slower reactivity with ICM was also reported when NH₂Cl was used instead of NaOCl or Cl₂ in a UV system,³³ suggesting that ·OCl radicals likely play a key role in fast ICM degradation in the UV/NaOCl system. This kind of reactivity enhancement by ·OCl radicals also applies to diatrizoate and iopromide,^33,34 which bears a secondary amine (–NH-R) as a sidechain of the aromatic ring as in iopamidol. Since iohexol does not have a secondary amine as one of its sidechains (Fig. A.S3†), we suppose that these structural differences of ICM are somehow related to the availability of C–I bonds for ·OCl radicals.

The use of mixed oxidants did not exert synergistic effects prominently on ICM degradation. For instance, the k_obs values in the UV AOP utilizing 0.5 mM PMS and 0.5 mM PDS were 0.95 ± 0.02 for iohexol and 0.83 ± 0.10 min⁻¹ for iopamidol, which were within 15% difference to the values of these separate UV/PMS and UV/PDS processes (0.82 and 0.83 for iohexol and 0.75 and 0.80 for iopamidolmin⁻¹, respectively). Among the mixed-oxidant combinations for iopamidol, UV AOPs utilizing NaOCl consistently showed high reactivity (3.9–4.2 min⁻¹), except for the UV/H₂O₂ + NaOCl process (0.31 min⁻¹). The presence of H₂O₂ could lead to ·OCl coupling with ·OH to form reactive singlet (¹O₂) or triplet oxygen (³O₂), resulting in much lower or no reactivity toward ICM.³⁵

3.2. Formation of inorganic iodine species

The kinetic profiles for the formation of I⁻, RIS, and IO₃⁻ during iohexol and iopamidol treatment via UV AOPs were examined, as shown in Fig. 2, A.S4 and S5.† For the UV-only process during iohexol degradation, I⁻ was released dominantly along with RIS, and no IO₃⁻ formation was observed within 30 min (Fig. 2a), and for iopamidol degradation, only I⁻ was observed with no RIS nor IO₃⁻ (Fig. 2b). This finding suggested that the UV-only process could degrade ICM and release I⁻, but it did not have the sufficient ability to further oxidize I⁻ into IO₃⁻. For the UV/H₂O₂ process, gradual increases in I⁻ and RIS concentrations were observed over 30 min, with the molar ratio of RIS to I⁻ being less than approximately 1:6. For the UV/PDS process, the I⁻ and RIS concentrations increased for the first 8 min, and then decreased, followed by a subsequent increase in IO₃⁻ concentration. It is likely that I⁻ and RIS were further oxidized to IO₃⁻ during the process. For the UV/PMS and UV/NaOCl processes, rapid formation of IO₃⁻ was observed, whereas little to no formation of I⁻ and a rapid decrease of a low amount of RIS were observed over time. While their speciation patterns differed, the UV AOPs released 20.1–38.0 μM total inorganic iodine within 30 min, which was equivalent to approximately 50–98% of the maximum iodine yield from both ICM. The remaining iodine likely existed in the form of organic iodine moieties formed during ICM degradation.

The formation dynamics and characteristics of inorganic iodine species across different oxidants were further examined using results from dual-oxidant processes (Fig. A.S4 and S5†). Overall, the use of H₂O₂ as a single oxidant or as a mixture resulted in the promotion of I⁻ or RIS formation and minimal IO₃⁻ formation. For example, the UV/PDS + H₂O₂ process yielded the largest amount of I⁻ (29.8 μM for IHX, 28.6 μM for IPM) and no formation of IO₃⁻ was observed. This phenomenon is different from UV with PDS only, which yielded only small concentrations of I⁻ (3.7 μM for IHX, 0 μM for IPM) and a high amount of IO₃⁻ (35.9 μM for IHX, 24.6 μM for IPM) at a reaction time of 30 min. Also, unlike UV/PMS in which preferential formation of IO₃⁻ over I⁻ or RIS was observed, the UV/PMS + H₂O₂ process showed a rapid increase in RIS concentrations during the first 8 minutes with gradual formation of I⁻ (accounting for 44–48% of iodine at t = 30 min) and only 32–35% IO₃⁻ yields. The most prominent contrast was observed for the UV/NaOCl + H₂O₂ process where only I⁻ formation was observed, as opposed to the UV/NaOCl process where only IO₃⁻ was present and no I⁻ nor RIS were formed during the entire reaction. For the UV/NaOCl + H₂O₂ process, the reaction between NaOCl and H₂O₂ may have occurred, eliminating each other to form singlet oxygen under ambient conditions.

To further understand the transformation behaviors of inorganic iodine species, UV-AOP treatment of I⁻ was conducted, as shown in Fig. A.S6,† which showed a similar pattern to UV-AOP treatment of ICM. UV irradiation without any oxidants did not oxidize I⁻ to RIS or IO₃⁻. The UV/H₂O₂ process produced <5 μM RIS and no further oxidation to IO₃⁻ was observed. The UV/PMS and UV/NaOCl processes oxidized I⁻ to IO₃⁻ without producing RIS within 30 min. In the UV/NaOCl process, a gradual decrease in I⁻ concentration and a subsequent increase in IO₃⁻ concentration were observed, with almost all I⁻ converted to IO₃⁻ within 30 min. In the UV/PMS process, almost instantaneous oxidation to IO₃⁻ was observed. In the absence of UV irradiation, PMS could quickly oxidize I⁻ into IO₃⁻, but RIS were still present in solution (the presence of I₂/I₃⁻ was visually detectable by the yellow color), indicating that further oxidation to IO₃⁻ in this system requires both UV irradiation and PMS. Interestingly, the UV/PDS process rapidly converted I⁻ to RIS with a maximum of 45 μM within 4 min, but then the RIS concentration slowly decreased until 30 min, again turning into I⁻. The transformation behavior of inorganic iodine species such as I⁻ (Fig. A.S6†) and that of ICM (Fig. 2) in UV AOPs indicated that the ICM degradation rate and oxidative transformation rates of iodine species influence the overall behavior of inorganic iodine species formation during ICM degradation.

It is commonly known that I⁻ is easily oxidized to RIS (commonly referred to as free iodine or HOI) with mild oxidants such as NH₂Cl, whereas the oxidation of RIS to IO₃⁻ is much slower under the same conditions.²³ However, RIS can be rapidly converted to IO₃⁻ with strong oxidants such as ozone,³⁶ ferrate,³⁷ or activated persulfate.³⁸ IO₃⁻ is considered to be an ideal iodine sink due to its thermodynamic stability and non-toxic characteristics.^39,40 In this regard, among all of the investigated UV AOPs, the UV/NaOCl and UV/PMS processes were the most efficient in forming IO₃⁻, followed by the UV/PDS process, whereas the UV-only and UV/H₂O₂ processes had limited ability to oxidize iodine species to IO₃⁻.

3.3. THM and HAA formation and their correlation with RIS concentration

ICM may serve as the iodine source as well as organic precursors for DBP formation during UV-AOP treatment; therefore, the formation of iodinated and non-iodinated THMs and HAAs during UV-AOP treatment of ICM was investigated in the absence (Fig. A.S7 and S8†) and presence of 5 mg L⁻¹ NOM (Fig. A.S9a and b†). The initial ICM concentrations were increased by 5-fold to 50 ppm to acquire sufficient amounts of I-THMs and I-HAAs after the UV AOP treatment, and oxidant concentrations were doubled to 2 mM to prevent oxidant depletion throughout the 30 minutes of UV AOP reaction. In the absence of NOM, minimal formation (0–0.9 μM) of I-THMs (i.e., CHClI₂, CHI₃) was observed after 30 min of UV-AOP treatment, accounting for only 0–1.4% of the initial ICM molar concentrations (60.9 μM iohexol, 64.3 μM iopamidol). No I-THMs were observed in the UV/PMS and UV/PDS processes, while the UV-only and UV/H₂O₂ processes showed higher yields. Iohexol treatment in UV/NaOCl also yielded I-THMs, which might correlate to the lack of C–I bond specificity in degradation kinetics, and slower total inorganic iodine release in comparison with that of iopamidol. To further investigate the possibility of additional formation of iodinated and non-iodinated disinfection by-products after UV-AOP treatment, solutions treated with UV AOPs for 30 min were subsequently chlorinated with 5 mg L⁻¹ NaOCl for 3 days, and the formation potentials of iodinated and non-iodinated THMs and HAAs were quantified (herein referred to as 3-day THMFPs and HAAFPs, respectively) for each UV AOP. For all UV AOPs, except for the UV-only process, the 3-day total iodinated THMFPs in the absence of NOM were below the LOD (Fig. A.S7 and S8†). For the UV-only process, the 3-day iodinated THMFPs (i.e., CHClI₂ and CHI₃) and non-iodinated THMFPs (i.e., CHCl₃) reached 4.6 and 1.3 μM respectively for iohexol, and 4.6 and 0.8 μM respectively for iopamidol. These iodinated and non-iodinated THMFPs account for 7.2–7.4% and 1.2–2.2% of the initial ICM concentration, respectively.

In the presence of 5 mg L⁻¹ NOM, 30 min UV-AOP treatment resulted in the formation of iodinated and non-iodinated THMs and HAAs, ranging from 0.5 to 83 μM, accounting for 0.8–136.6% of the initial ICM molar concentration (Fig. A.S9a and b†). Higher concentrations of THMs and HAAs were found particularly in the UV AOPs where NaOCl was used, with maximum THM + HAA concentrations of 83 μM for iohexol and 79.2 μM for iopamidol, respectively. In contrast, although they were lower amounts than non-iodinated THMs and HAAs, higher concentrations of I-THMs and I-HAAs were generally found in the UV AOPs where PDS was used, with levels ranging from 6.2 to 13.8 μM for iohexol and 0.3 to 7.2 μM for iopamidol. The iodine in these I-THMs and I-HAAs accounted for 0.1–7.6% of the iodine that originally constituted ICM. Similar results for UV/PDS treatment were previously reported.^19,41 In contrast, the UV AOPs where NaOCl was used showed no formation of I-THMs and I-HAAs.

While 30 min UV AOP treatments transformed iodine into I-THMs and I-HAAs with yields less than 7.6%, subsequent chlorination resulted in higher production of I-THMs and I-HAAs (Fig. A.S9† and 3). The 3-day THMFPs and HAAFPs for iohexol and iopamidol ranged from 3.7 to 190.8 μM, which are equivalent to 0.1–97.2% of total iodine in solution. The majority of THMFPs and HAAFPs accounted for the non-iodinated THMs and HAAs; however, considerable amounts of I-THMs and I-HAAs were formed in certain processes, sometimes taking up to more than 50% of total THMs and HAAs (Fig. A.S9†). To further visualize the formation behavior of I-THMs and I-HAAs, the total concentrations of iodine atoms in THMs and HAAs remaining after post-chlorination are shown in Fig. 3. UV only and processes using PDS or H₂O₂ led to high yields of I-THMFPs and I-HAAFPs ranging from 2.9–19.4 μM for iohexol and 4.7–64 μM for iopamidol, which account for 7.4–97.2% of iodine in the solution, most likely due to the higher formation of I⁻ and RIS. This was particularly evident in the iopamidol treatment with the UV/PDS process, in which 64 μM I-THMs and I-HAAs were formed among the 95 μM total THMs and HAAs, and 97.2% of iodine was converted to THMs and HAAs during post-chlorination. By contrast, the UV/PMS, UV/NaOCl and UV/NaOCl + PMS (processes without the use of PDS or H₂O₂) treatments for iohexol and iopamidol led to lower iodinated THMFPs and HAAFPs. The UV AOP treatment of I⁻ was also conducted as a control group (Fig. A.S9c†) and an identical pattern was observed, where I-THMFPs and I-HAAFPs were clearly identified only in the UV-only process or processes where PDS or H₂O₂ was used.

As mentioned in section 3.2, the UV/NaOCl and UV/PMS processes showed low I⁻ and RIS concentrations along with high IO₃⁻ concentrations during ICM degradation, resulting in minimal formation of I-THMs and I-HAAs. However, ICM degradation by UV/H₂O₂, UV/PDS, or mixed processes containing PDS or H₂O₂ yielded relatively high formation of I⁻ or RIS, which resulted in higher yields of I-THMs and I-HAAs during post-chlorination. For the UV-only process, ICM degradation formed I⁻ and RIS without IO₃⁻, in which I⁻ likely reacted with chlorine (i.e., hypochlorous acid) to form RIS and with aromatic rings in ICM and NOM to form I-THMs and I-HAAs.^24,42,43 For the UV AOPs utilizing dual oxidants (Fig. A.S9a and b†), similar trends were observed to those with the use of single oxidants. Those utilizing NaOCl yielded lower formation of I-THMs and I-HAAs and higher formation of THMs and HAAs, while those utilizing H₂O₂, PDS, and UV-only resulted in higher formation of I-THMs and I-HAAs and lower formation of THMs and HAAs.

Furthermore, the correlations between inorganic iodine species formation after 8 min single/mixed-oxidant UV-AOP treatment and I-THM/I-HAA concentrations post-chlorination were investigated, as shown in Fig. 4. A reaction time of 8 min was selected since most ICM had decomposed at this point and a sufficient amount of inorganic iodine species had formed. Unpaired t-tests were performed for each compound to statistically verify the correlation. Positive correlations between I⁻/RIS and I-THM/I-HAA concentrations were observed for both iohexol (R² = 0.5590, t = 3.954) and iopamidol (R² = 0.3021, t = 2.815). This implies that the treatment with UV AOPs that favors the formation of I⁻ or RIS will likely result in more formation of I-DBPs during subsequent chlorination of UV AOP-treated water.

Based on these observations, we found that regardless of the difference in degradation kinetics or degradation mechanisms among UV AOPs with different oxidants, iodine speciation is the dominant correlation factor for formation of I-DBPs. Our findings suggest that ensuring minimal formation of I⁻ and RIS during the treatment of ICM can directly minimize the formation of I-DBPs. It is also important to prevent the reduction of IO₃⁻ to I⁻ or RIS before chlorination. It is reported that photolytic degradation of IO₃⁻ to I⁻ resulted in I-THM formation.^44–46 It is also noteworthy to mention that I-DBPs can also undergo degradation in some UV AOPs, such as UV-only, UV/NaOCl and UV/H₂O₂ processes.^47,48

4. Conclusions

This study comparatively investigated the degradation efficiency, iodine species formation, and I-THM/I-HAA formation of iohexol and iopamidol for single and dual-oxidant UV AOPs utilizing NaOCl, PMS, PDS, and H₂O₂. This is the first study to carefully examine the distribution and formation dynamics of iodine species during UV AOP treatment of ICM. UV AOPs utilizing NaOCl showed the highest activity for degrading iopamidol, seemingly due to selective reactions, while H₂O₂ may have hindered this selectivity. UV AOPs utilizing NaOCl or PMS (in specific, also without the mixed use of PDS or H₂O₂) could efficiently transform iodide released from degradation of ICM into IO₃⁻, which is a desirable iodine sink, whereas UV AOPs utilizing PDS or H₂O₂ as major oxidants could not fully oxidize iodine species. Consequently, these processes (i.e., UV/PDS, UV/H₂O₂) yielded RIS, which are less oxidized forms than IO₃⁻ and precursors of I-DBPs. The results of the I-THM and I-HAA analyses roughly correlated with these findings, in that higher yields of I-THMs and I-HAAs were generally found in UV AOPs utilizing PDS or H₂O₂ compared to those with NaOCl or PMS, and the I-THM and I-HAA formation became more prominent during post-chlorination. Overall, the findings from this study suggest that the treatment of iodine-containing contaminants such as ICM via UV AOPs needs to consider not only the efficacy of these processes for degradation of the parent contaminants but also the fate of iodine species released from these compounds and the formation of I-DBPs. For treatment of ICM via UV AOPs, UV/NaOCl and UV/PMS are more preferred options in terms of both degradation efficiency of ICM and formation of I-DBPs. Furthermore, it is important to keep iodine species in the form of IO₃⁻ to minimize and prevent I-DBP formation; future studies should focus on identifying strategies to convert iodine species to IO₃⁻ and prevent IO₃⁻ from reducing back to other iodine species. The findings from this study contribute to minimizing the impact of ICM-contaminated hospital and medical imaging wastewater on receiving water bodies by identifying practical and sustainable AOP technologies that can efficiently remove ICM contaminants and prevent the formation of toxic by-products.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF-RS-2023-00208545). It is also supported by the BK21 PLUS research program of NRF and the Institute of Engineering Research at Seoul National University.

References

1. L. Kovalova, H. Siegrist, H. Singer, A. Wittmer and C. S. McArdell, Environ. Sci. Technol., 2012, 46, 1536–1545 CrossRef CAS PubMed .
2. T. Azuma, K. Otomo, M. Kunitou, M. Shimizu, K. Hosomaru, S. Mikata, M. Ishida, K. Hisamatsu, A. Yunoki, Y. Mino and T. Hayashi, Sci. Total Environ., 2019, 657, 476–484 CrossRef CAS PubMed .
3. L. J. Fono and D. L. Sedlak, Water Res., 2007, 41, 1580–1586 CrossRef CAS .
4. E. Hapeshi, A. Lambrianides, P. Koutsoftas, E. Kastanos, C. Michael and D. Fatta-Kassinos, Environ. Sci. Pollut. Res., 2013, 20, 3592–3606 CrossRef CAS .
5. A. Mendoza, J. Aceña, S. Pérez, M. López de Alda, D. Barceló, A. Gil and Y. Valcárcel, Environ. Res., 2015, 140, 225–241 CrossRef CAS PubMed .
6. E. Kornelius, J. Y. Chiou, Y. S. Yang, C. H. Peng, Y. R. Lai and C. N. Huang, J. Clin. Endocrinol. Metab., 2015, 100, 3372–3379 CrossRef CAS .
7. R. Kubicki, J. Grohmann, K. G. Kunz, B. Stiller, K. O. Schwab and N. Van Der Werf-Grohmann, J. Pediatr. Endocrinol. Metab., 2020, 33, 1409–1415 CrossRef CAS .
8. S. E. Duirk, C. Lindell, C. C. Cornelison, J. Kormos, T. A. Ternes, M. Attene-Ramos, J. Osiol, E. D. Wagner, M. J. Plewa and S. D. Richardson, Environ. Sci. Technol., 2011, 45, 6845–6854 CrossRef CAS .
9. Z. Xu, X. Li, X. Hu and D. Yin, Chemosphere, 2017, 184, 253–260 CrossRef CAS PubMed .
10. M. J. Plewa, E. D. Wagner, P. Jazwierska, S. D. Richardson, P. H. Chen and A. B. McKague, Environ. Sci. Technol., 2004, 38, 62–68 CrossRef CAS .
11. H. Dong, Z. Qiang and S. D. Richardson, Acc. Chem. Res., 2019, 52, 896–905 CrossRef CAS .
12. T. A. Ternes and R. Hirsch, Environ. Sci. Technol., 2000, 34, 2741–2748 CrossRef CAS .
13. M. Escolà Casas, R. K. Chhetri, G. Ooi, K. M. S. Hansen, K. Litty, M. Christensson, C. Kragelund, H. R. Andersen and K. Bester, Sci. Total Environ., 2015, 530–531, 383–392 CrossRef .
14. T. A. Ternes, J. Stüber, N. Herrmann, D. McDowell, A. Ried, M. Kampmann and B. Teiser, Water Res., 2003, 37, 1976–1982 CrossRef CAS .
15. W. Seitz, J. Q. Jiang, W. H. Weber, B. J. Lloyd, M. Maier and D. Maier, Environ. Chem., 2006, 3, 35–39 CrossRef CAS .
16. C. Y. Hu, Y. Z. Hou, Y. L. Lin, Y. G. Deng, S. J. Hua, Y. F. Du, C. W. Chen and C. H. Wu, Chemosphere, 2019, 229, 602–610 CrossRef CAS .
17. C. Y. Hu, Y. Y. Zhu, B. Xu, T. Y. Zhang, Y. L. Lin, C. Xiong, Q. B. Wang, D. D. Huang and L. Xu, Sep. Purif. Technol., 2022, 289, 120810 CrossRef CAS .
18. Z. Wang, Y. L. Lin, B. Xu, S. J. Xia, T. Y. Zhang and N. Y. Gao, Chem. Eng. J., 2016, 283, 1090–1096 CrossRef CAS .
19. X. Zhao, J. Jiang, S. Pang, C. Guan, J. Li, Z. Wang, J. Ma and C. Luo, Chemosphere, 2019, 221, 270–277 CrossRef CAS PubMed .
20. F. X. Tian, W. K. Ye, B. Xu, X. J. Hu, S. X. Ma, F. Lai, Y. Q. Gao, H. B. Xing, W. H. Xia and B. Wang, Chem. Eng. J., 2020, 398, 125570 CrossRef CAS .
21. I. J. Lopez-Prieto, K. D. Daniels, S. Wu and S. A. Snyder, J. Environ. Chem. Eng., 2021, 9, 105760 CrossRef CAS .
22. F. X. Tian, B. Xu, Y. L. Lin, C. Y. Hu, T. Y. Zhang and N. Y. Gao, Water Res., 2014, 58, 198–208 CrossRef CAS .
23. D. B. Jones, A. Saglam, A. Triger, H. Song and T. Karanfil, Environ. Sci. Technol., 2011, 45, 10429–10437 CrossRef CAS PubMed .
24. T. Ye, T.-Y. Zhang, F.-X. Tian and B. Xu, Water Res., 2021, 206, 117755 CrossRef CAS .
25. S. Canonica, L. Meunier and U. von Gunten, Water Res., 2008, 42, 121–128 CrossRef CAS .
26. X. Zhao, E. Salhi, H. Liu, J. Ma and U. Von Gunten, Environ. Sci. Technol., 2016, 50, 4358–4365 CrossRef CAS .
27. U.S. EPA., Method 551.1: Determination of Chlorination Disinfection Byproducts, Chlorinated Solvents, and Halogenated Pesticides/Herbicides in Drinking Water by Liquid-Liquid Extraction and Gas Chromatography With Electron-Capture Detection, 1995.
28. U.S. EPA., Method 552.2: Determination of Haloacetic Acids and Dalapon in Drinking Water by Liquid-Liquid Extraction, Derivatization and Gas Chromatography with Electron Capture Detection, 1995.
29. C. C. Mo, F. X. Tian, B. Xu, F. Lai, Y. Q. Gao, Y. Ma, Z. H. Feng, Z. J. Yao, D. S. Bi and X. J. Hu, J. Water Process Eng., 2024, 67, 106199 CrossRef .
30. X. Kong, J. Jiang, J. Ma, Y. Yang and S. Pang, Chemosphere, 2018, 193, 655–663 CrossRef CAS .
31. Z. Wu, J. Fang, Y. Xiang, C. Shang, X. Li, F. Meng and X. Yang, Water Res., 2016, 104, 272–282 CrossRef CAS .
32. C. Y. Hu, Y. Z. Hou, Y. L. Lin, A. P. Li and Y. G. Deng, Chem. Eng. J., 2019, 360, 1003–1010 CrossRef CAS .
33. Y. Wu, S. Zhu, W. Zhang, L. Bu and S. Zhou, Chem. Eng. J., 2019, 375, 121972 CrossRef CAS .
34. Y. Cha, T. K. Kim, J. Lee, T. Kim, A. J. Hong and K. D. Zoh, J. Hazard. Mater., 2022, 437, 129371 CrossRef CAS PubMed .
35. R. Castagna, J. P. Eiserich, M. S. Budamagunta, P. Stipa, C. E. Cross, E. Proietti, J. C. Voss and L. Greci, Atmos. Environ., 2008, 42, 6551–6554 CrossRef CAS .
36. S. Allard, C. E. Nottle, A. Chan, C. Joll and U. von Gunten, Water Res., 2013, 47, 1953–1960 CrossRef CAS .
37. J. Shin, U. Von Gunten, D. A. Reckhow, S. Allard and Y. Lee, Environ. Sci. Technol., 2018, 52, 7458–7467 CrossRef CAS PubMed .
38. L. Wang, Y. Ji, J. Lu, X. Yin, Q. Zhou and D. Kong, Chemosphere, 2017, 181, 400–408 CrossRef CAS .
39. U. Von Gunten, Water Res., 2003, 37, 1469–1487 CrossRef CAS .
40. J. Criquet, S. Allard, E. Salhi, C. Joll, A. Heitz and U. Von Gunten, Environ. Sci. Technol., 2012, 46, 7350–7357 CrossRef CAS PubMed .
41. C. Y. Hu, Y. Z. Hou, Y. L. Lin, Y. G. Deng, S. J. Hua, Y. F. Du, C. W. Chen and C. H. Wu, Chem. Eng. J., 2020, 379, 122403 CrossRef CAS .
42. Y. Xia, Y. L. Lin, B. Xu, C. Y. Hu, Z. C. Gao, Y. L. Tang, W. H. Chu, T. C. Cao and N. Y. Gao, Water Res., 2018, 147, 101–111 CrossRef CAS PubMed .
43. T. Bond, E. H. Goslan, S. A. Parsons and B. Jefferson, Environ. Technol. Rev., 2012, 1, 93–113 CrossRef CAS .
44. L. Z. Tang, Y. L. Lin, B. Xu, Y. Xia, T. Y. Zhang, C. Y. Hu, Y. L. Tang, T. C. Cao, Q. M. Xian and N. Y. Gao, Water Res., 2021, 193, 116851 CrossRef CAS PubMed .
45. F. X. Tian, X. J. Hu, B. Xu, T. Y. Zhang and Y. Q. Gao, J. Hazard. Mater., 2017, 326, 138–144 CrossRef CAS PubMed .
46. F. X. Tian, S. X. Ma, B. Xu, X. J. Hu, H. B. Xing, J. Liu, J. Wang, Y. Y. Li, B. Wang and X. Jiang, Chemosphere, 2019, 221, 292–300 CrossRef CAS .
47. A. Q. Wang, Y. L. Lin, B. Xu, C. Y. Hu, S. J. Xia, T. Y. Zhang, W. H. Chu and N. Y. Gao, Chem. Eng. J., 2017, 323, 312–319 CrossRef CAS .
48. Y. Xiao, L. Zhang, J. Yue, R. D. Webster and T. T. Lim, Water Res., 2015, 75, 259–269 CrossRef CAS .

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ew00393h

This journal is © The Royal Society of Chemistry 2025