Iron(III)-induced photooxidation of arsenite in the presence of carboxylic acids and phenols as model compounds of natural organic matter

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

proportion. Fe(III) can react with catechol to form a complex to produce Fe(II) and 107 semiquinone radicals. This is a pH-dependent process, which is related to changes in UV-A irradiation (366 nm), the rate coefficient k exp was 1.27 × 10 -2 min -1 (250 122 nM/100 min initial rate was ca. 5.08 x 10 -3 μM min -1 ). Such rate is negligible in our 123 system. And they also fail to study the transformation mechanism of As(III) from the

163
The property parameters of model compounds of NOM used in this study are listed and UVB irradiation are shown in Fig. 1 and 2. Moreover, the experimental values of 218 k obs-UVA and k obs-UVB are listed in Table S2. The pH was set to pH 7 for all investigated 219 systems, as it is an average pH of natural waters and the efficiency of photochemical 220 oxidation of As(III) induced by CFH at pH 7 was high. As shown in Fig. 1 and 2  Fe + As + EDTA Fe + As + Succinic acid Fe + As + Maleic acid Fe + As + Oxalic acid Fe + As a Fe + As + Citric acid Fe + As + Malic acid Fe + As + Salicylic acid Fe + As + Lactic acid Fe + As b C/C 0

Time (min)
Fe + As + Catechol Fe + As + 4-Nitrocatechol Fe + As c As Fe + As + Gallic acid  Fe + As + EDTA Fe + As + Succinic acid Fe + As + Maleic acid Fe + As + Oxalic acid Fe + As a Fe + As + Citric acid. Fe + As + Malic acid Fe + As + Salicylic acid Fe + As + Lactic acid Based on the data presented in Fig. 1 and Table S2, the values of k obs-UVA in the 266 presence of various NOM were determined and classified into three groups. The first 267 group (group I), with k obs-UVA of 0.026-0.039 min -1 (relative standard deviation, RSD 268 15%), included lactic, maleic, salicylic, oxalic, succinic, citric, and malic acids as well 269 as EDTA (Figs. 1a and 1b). Moreover, the second group (group II) with k obs-UVA 270 established at 0.0087-0.0099 min -1 (RSD 9%) included 4-nitrocatechol and catechol 271 (Fig. 1c). Lastly, the third group (group III), with a k obs-UVA value of 0.0028 min -1 , 272 involved gallic acid (Fig. 1d). The relative minimum inter-group variances of k obs-UVA 273 between each group were determined at 62% (between group I and II) and 68% 274 (between group II and III). Analogously, the values of k obs-UVB in the presence of 275 various NOM were also classified into three groups. The value of k obs-UVB for the first 276 group was determined at 0.081-0.109 min -1 (Fig. 2a and 2b, RSD 10%), while for the 277 second group it was 0.018-0.024 min -1 (Fig. 2c, RSD 20%). Additionally k obs-UVB for 278 the third group was established at 0.0062 min -1 (Fig. 2d). The relative inter-group 279 minimum variances of k obs-UVB between each group were 70% (between group I and II) 280 and 66% (between group II and III). The classification of the NOM into 3 groups 281 based on the value of k obs was valid, since the relative minimum intergroup variance 282 was sufficiently large (>3RSD of intragroup data,  Overall, the classification in terms of k obs correlates with the present functional 285 groups. Group I includes polycarboxylic acids (Figs. 1a and 2a) or hydroxyl 286 carboxylic acids (Fig. 1b and 2b), group II contains polyphenols without carboxyl 287 moieties ( Fig. 1c and 2c), while group III consists of carboxylic polyphenols (Fig. 1d   288 and 2d). However, such classification is not sufficiently precise, particularly for NOM

299
The k obs-UVB value is noticeably larger than the corresponding k obs-UVA value for each 300 evaluated NOM. Analogous results were obtained for the system without NOM, 301 where the ratio of k obs-UVB /k obs-UVA was established at 2.82. The quantum yields of the 302 photolysis of the Fe(III)-As(III) complex were estimated to be close ( = 1.4,

303
Text S1 and Fig. S2). Hence, the difference between k obs-UVB and k obs-UVA in the 304 absence of NOM can be attributed to the intensity of the radiation and its absorption 305 by the Fe(III)-As(III) complex, which is not wavelength-dependent. Additionally, to 306 confirm whether the effect of the NOM structure on k obs is dependent on the 307 wavelength, the correlation analysis between k obs-UVA and k obs-UVB in the presence of 308 NOM was carried out, as shown in Fig. 3. Linear fitting analysis demonstrated that 309 both in the presence and in the absence of NOM, k obs-UVB was approximately three 310 times higher than k obs-UVA , and the correlation was significant (r = 0.996, p < 0.001).

311
This result confirmed the hypothesis that the effect of NOM at low concentrations is 312 not dependent on the wavelength band of UVB and UVA, but on the structure.  respectively, and all the data were within the 95% prediction interval. It is noteworthy 373 that the logK Fe-NOM parameter accounted for a large proportion in the fitting formula.

374
This meant that among the chemical properties of NOM, the most important factor 375 affecting k obs was the coordination equilibrium constant between NOM and iron. In 376 addition, some of the two-parameter regressions were also significant, e.g., molecular 377 weight and pKa 1 . Thus, these simpler models may also be utilized to estimate k obs 378 when logK Fe-NOM data are not available (Fig. S4).