The interfacial reactivity of arsenic species with green rust sulfate (GR SO4 )

11 Arsenic (As) contamination in groundwater is a significant health and environmental concern 12 worldwide because of its wide distribution and toxicity. The fate and mobility of As is greatly 13 influenced by its interaction with redox-active mineral phases, among which green rust (GR), an Fe II - 14 Fe III layered double hydroxide mineral, plays a crucial role. However, the controlling parameters of 15 As uptake by GR are not yet fully understood. To fill this gap, we determined the interfacial reactions 16 between GR sulfate (GR SO4 ) and aqueous inorganic As(III) and As(V) through batch adsorption 17 experiments, under environmentally-relevant groundwater conditions. Our data showed that, under 18 anoxic conditions, GR SO4 is a stable and effective mineral adsorbent for the removal of As(III) and 19 As(V). At an initial concentration of 10 mg L -1 , As(III) removal was higher at alkaline pH conditions 20 (~ 95% removal at pH 9) while As(V) was more efficiently removed at near-neutral conditions (> 21 99% at pH 7). The calculated maximum As adsorption capacities on GR SO4 were 160 mg g -1 (pH 8-9) 22 for As(III) and 105 mg g -1 (pH 7) for As(V). The presence of other common groundwater ions such as 23 Mg 2+ and PO 43- reduces the efficiency of As removal, especially at high ionic strengths. Long-term 24 batch adsorption experiments (up to 90 days) revealed that As-interacted GR SO4 remained stable, with 25 no mineral transformation or release of adsorbed As species. Overall, our work shows that GR SO4 is


1.
Introduction 119 Full details of all tested parameters [e.g., varying pH (7 to 9), adsorbent loading (solid to solution 120 ratio, S/L 2 to 6 g L -1 ), ionic strength (IS* 0.5 to 0.005 M), competing ions (Ca 2+ , Mg 2+ , PO 4 3-) and the Supporting Information (Text S1, Table S1).  shape and position in our sample and the theoretical spectrum (blue line in Fig.1b; Fig. S2a). This 160 revealed that the shape of the Fe L 3 -edge for the GR SO4 sample matched the linear reference fit for a 161 Fe(II)/Fe(III) ratio of 2, with minor differences. This is evidenced by the changes in shape and 162 position of the L 3 peak in the theoretical spectrum as the GR composition becomes more Fe(III)-rich.
ratio corresponding to 2.

171
The surface chemistry of the synthesized GR SO4 was analyzed by XPS and the wide scan

185
The iron chemistry of the synthesized GR SO4 was characterized by Mössbauer spectroscopy 186 which revealed two apparent doublets (Fig. S3), but with a certain line broadening of the outer is in agreement with the ratio of 2 from our EELS data (Fig. 1b, Fig S2), as well as literature data

219
Error bars represent standard deviations of triplicate experiments (< 5% relative). Note: IS* here is 220 defined as the ionic strength based on a 10x and 100x dilution from the initial 0.5 M IS of the GR SO4 221 suspension (further details, see in Supporting Information Text S1).

223
At all pH values tested, the As(V) removal efficiencies (Fig. 3a) were higher compared to 224 As(III). This is likely because of the higher adsorption affinity of the pentavalent species on iron 225 (oxyhydro)oxide surfaces. No significant differences in As(V) removal efficiencies between pH 7, 8 226 and 9 were observed (i.e. within analytical uncertainties < 2%). Although there were no significant efficiency (50.1 ± 1.5% at pH 7) increased by more than 30% at pH 8 (83.7 ± 0.9%) and another 10% 232 increase was measured at pH 9 (94.6 ± 0.1%). Such surface polymerization of As(III) complexes has 233 been previously suggested for GR Cl

246
With increased adsorbent loading from 2 to 4 g L -1 , the removal efficiency of As(III) also 247 increased by ~15% from 34.6 ± 2.7 to 50.1 ± 1.5% (Fig. 3b). This increase was caused by the larger 248 number of active surface sites available for As(III) complexes (Asere et al., 2017). However, with 249 further increase in loading to 6 g L -1 , the efficiency decreased to 39.2 ± 6.2%. In the case of As(V), no 250 significant differences (< 0.3% relative) in removal efficiencies were observed among the adsorbent 251 loadings tested (Fig. 3b).

252
The removal efficiencies for both As species decreased with increasing ionic strength, IS*

261
Fe 2+ , NH 4 + , Na + , Cland SO 4 2ions) have been shown to have little or no effect on As adsorption     were identified in XRD patterns of these long-term equilibrated and As-interacted samples (Fig. 5a).

312
The TEM images and SAED patterns (Fig. 5b) also showed that the GR SO4 particles in the 90-day   Table S7. Based on the fitting, the adsorption of As species 332 on GR SO4 is best described using the Langmuir model, indicating a homogenous monolayer binding of

333
As surface complexes at the solid/water interface (Leus et al., 2017). Using the Langmuir adsorption 334 model, we determined the maximum As adsorption capacities for both As species onto GR SO4 (Table   335 1). At alkaline pH, the maximum adsorption capacity of As(III) was 2.2 times higher than the value at  (Table S1).

345
The spatial distribution of the adsorbed As(III) on the GR particles, at an initial concentration 346 of 500 mg L -1 , was examined using HAADF-STEM imaging coupled with EDX mapping (Fig. 7).

347
The EDX elemental map (Fig. 7d) and associated intensity profile (Fig. 7g) show higher 348 concentrations of As can be found near the GR particle edges (ca. two times higher than the 001 GR 349 surface). In addition, the HAADF-STEM image (Fig. 7a)   Using the adsorption isotherm modelling data, we compared the calculated adsorption 379 capacities for As species on GR SO4 and with literature data for all described iron (oxyhydr)oxides, 380 oxyhydroxysulfates and sulfides, which have also been evaluated for their efficiency as mineral 381 substrate for the treatment of As contaminated groundwater resources (Table 1).

382
Our data show clearly that GR SO4 is among the most effective adsorbents among all the 383 phases listed in Table 1. This finding has important implications for the fate and mobility of As in 384 anoxic groundwaters where GR SO4 exists. To the best of our knowledge, this is the first study to report 385 the adsorption isotherms of As(III) and As(V) for GR SO4 , as well as the in-depth examination of 386 critical adsorption parameters for As removal. We have shown that at circum-neutral and slightly 387 alkaline pH conditions, GR SO4 can efficiently adsorb large amounts of As(III) and As(V), making

417
As for redox-active mineral adsorbents, arsenic can still be released from GR SO4 since its 418 sequestration is highly dependent on pH conditions and redox environment. Sudden changes in pH or 419 Eh of the system may cause potential release of surface immobilized As species back into the 420 groundwater either by dissolution or redox-change driven transformation of GR phases (Cundy et al., groundwater systems. The removal of As is also highly pH dependenthigh As(III) removal was 435 obtained at higher pH while As(V) removal was found to be more favourable at circum-neutral 436 conditions. GR SO4 exhibited fast As uptake rates at alkaline conditions.