Chlorination of Iodide-Containing Waters in the Presence of CuO

Oct 14, 2014 - (26) The BET surface area was determined by a Micromeritics Tristar II to be ...... Ding , G.; Zhang , X. A picture of polar iodinated ...
0 downloads 0 Views 832KB Size
Article pubs.acs.org/est

Chlorination of Iodide-Containing Waters in the Presence of CuO: Formation of Periodate Chao Liu,† Elisabeth Salhi,‡ Jean-Philippe Croué,*,†,⊥ and Urs von Gunten*,‡,§ †

Water Desalination and Reuse Center, King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia ‡ Eawag, Swiss Federal Institute of Aquatic Science and Technology, Ueberlandstrasse 133, CH-8600 Dübendorf, Switzerland ⊥ Curtin Water Quality Research Centre, Department of Chemistry, Curtin University, Perth, WA-6845, Australia § School of Architecture, Civil and Environmental Engineering (ENAC), Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland S Supporting Information *

ABSTRACT: It has been shown previously that the disproportionation of halogen-containing oxidants (e.g., HOCl, HOBr, and ClO2) is enhanced by a CuO-catalyzed process. In this study, the transformation of iodine during chlorination in the presence of CuO was investigated. There is no significant enhancement of the disproportionation of hypoiodous acid (HOI) in the presence of CuO. The formation rate of iodate (IO3−) in the CuO−HOCl−I− system significantly increased when compared to homogeneous solutions, which was ascribed to the activation of HOCl by CuO enhancing its reactivity toward HOI. In this reaction system, iodate formation rates increase with increasing CuO (0−0.5 g L−1) and bromide (0−2 μM) doses and with decreasing pH (9.6−6.6). Iodate does not adsorb to the CuO surfaces used in this study. Nevertheless, iodate concentrations decreased after a maximum was reached in the CuO−HOCl−I−(−Br−) systems. Similarly, the iodate concentrations decrease as a function of time in the CuO−HOCl−IO3− or CuO−HOBr−IO3− system, and the rates increase with decreasing pH (9.6−6.6) due to the enhanced reactivity of HOCl or HOBr in the presence of CuO. It could be demonstrated that iodate is oxidized to periodate by a CuO-activated hypohalous acid, which is adsorbed on the CuO surface. No periodate could be measured in filtered solutions because it was mainly adsorbed to CuO. The adsorbed periodate was identified by scanning electron microscopy plus energy dispersive spectroscopy and X-ray photoelectron spectroscopy.



INTRODUCTION

iodine during oxidative water treatment conditions of fresh waters at near neutral pH.10 HOI can slowly disproportionate to I− and IO3−. This reaction pathway is not important for typical drinking water conditions.10 During oxidative water treatment HOI can be further oxidized to IO3− or react with dissolved organic matter (DOM) to form iodinated disinfection byproducts (IDBPs).8,11−13 I-DBPs became a particular concern because it has been discovered that they are more toxic (e.g., cytotoxicity, genotoxicity, and developmental toxicity), as a group, than their brominated or chlorinated analogues, which are regulated in drinking water.7,14−17 Iodate is nontoxic and quite redox-stable, and is therefore the desired sink for iodine in drinking waters.18



Iodine is present in natural waters in the form of iodide (I ), iodate (IO3−) and organo-iodine. In fresh water, I− is the main species, while in seawater IO3− predominates.1 Among these iodine species, organo-iodine is generally of minor importance.2,3 The total iodine concentrations are highly variable, with typically higher concentrations in seawater (median of ca. 60 μg L−1) than in fresh surface water (typically 0.5−20 μg L−1).1,4−6 Under extreme circumstances, for example, groundwater adjacent to halide rocks or oil field brines, the total iodine can be ≥100 μg L−1.1 A survey over 23 water supplies in the U.S. and Canada showed that I− in water resources ranged from 0.4 to 104 μg L−1 (median of 10.3 μg L−1).7 In oxidative water treatment processes, naturally occurring I− can be quickly oxidized to hypoiodous acid (HOI). For typical oxidant concentrations applied in drinking water, the half-life times for I− are in the order of milliseconds for ozonation and chlorination, while in the minute range for chloramination.8,9 HOI in equilibrium with OI− (pKa = 10.4) is the major form of © XXXX American Chemical Society

Received: July 2, 2014 Revised: October 14, 2014 Accepted: October 14, 2014

A

dx.doi.org/10.1021/es5032079 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

The oxidation of HOI and OI− to IO3− by ozone is fast (kO3+HOI = 3.6 × 104 M−1 s−1; kO3+OI‑ = 1.6 × 106 M−1 s−1), whereas for typical chlorine concentrations in drinking waters, the kinetics for IO3− formation is in the range of hours, depending on the pH (kHOCl+HOI = 8.2 M−1 s−1, and kOCl−+HOI = 52 M−1 s−1).8 The reaction between HOCl/OCl− and HOI to iodite (IO2−) is the rate-limiting step, with a faster further oxidation of IO2− to IO3−. In contrast, monochloramine does not oxidize HOI.8 Therefore, the probability of the formation of iodoform (a typical I-DBP) is higher for chloramination than for ozonation or chlorination.11,19 However, chlorination can produce higher levels of brominated iodo-trihalomethanes.20,21 Bromide (Br−), which usually is simultaneously present with I−, can be oxidized by HOCl to hypobromous acid (HOBr).22 During chlorination, iodate formation is significantly accelerated in the presence of Br−, because HOBr reacts faster with HOI than HOCl.23 Copper pipes are widely used in municipal distribution systems and household plumbing. Cupric oxide (CuO) is one of the major corrosion products of copper pipes.24,25 Recently, we have shown that CuO can catalyze HOCl or HOBr disproportionation to produce chloride (Cl−) and chlorate (ClO3−) or Br− and bromate (BrO3−), respectively.26,27 The latter pathway (eq 1) may lead to elevated BrO 3 − concentrations during chlorination of bromide-containing waters.

spectrophotometer (Cary 100).30 Concentrations of IO4− in the CuO−IO4− systems were also analyzed by the DPD method (same as for residual HOX), a calibration curve (Figure S1, Supporting Information (SI)) in the range of 0−40 μM showed a good linearity (R2 > 0.99). Bromate and iodate were quantified by a Dionex 3000 reagent free ion chromatograph (IC) through an AS9 column with a postcolumn reaction with UV/vis detection of I3− at 288 nm.31,32 The quantification limits for BrO3− and IO3− were 2.5 μg L−1 and the standard deviation was ±10%. Elemental analyses of the reacted CuO layer were performed by scanning electron microscopy plus energy dispersive spectroscopy (SEM-EDS) on a scanning electron microscope (Quanta 200D, FEI) according to our previous method.27 X-ray photoelectron spectroscopy (XPS) analyses were carried out in a Kratos Axis Ultra DLD spectrometer equipped with a monochromatic Al Ka X-ray source (hν = 1486.6 eV) operating at 150 W, a multichannel plate and delay line detector under a vacuum of ∼10−9 mbar. The survey and highresolution spectra were collected at fixed analyzer pass energies of 160 and 20 eV, respectively. Experimental Setup and Procedures. All experiments were conducted in the dark and under continuous agitation using a magnetic stirrer. Reactions were initiated by the injection of an aliquot of a HOCl, HOBr, or HOI stock solution, respectively, to the buffered solutions containing CuO in the presence or absence of I− or IO3− at room temperature (22 ± 1 °C). The CuO doses ranged from 0.05 to 0.5 g L−1, similar to our previous study to be able to investigate the reaction kinetics in a reasonable time frame.26 The pH was adjusted to 6.6, 7.6, 8.6, 9.6, and 10.6 with HNO3 or NaOH solutions in the presence of a 2.5 mM tetraborate buffer. The pH changes were less than 0.2 pH unit during the course of the reaction. Bromide was added when required (concentrations ranged from 0.5 to 2.0 μM, that is, 40−160 μg L−1). Samples were withdrawn at preselected time points, filtered within 15 s (insignificant considering reaction times of several hours) through a 0.45-μm syringe filter (surfactant-free cellulose acetate membrane). The filter was pretreated with HOCl or HOBr solutions (1 mg L−1) and then rinsed with deionized water to avoid a potential oxidant demand. The filtered samples were analyzed for residual oxidant concentrations by the DPD method. For IC analyses, the samples were quenched immediately with sulfite. To characterize the reacted CuO particles, 100 mL of CuO (1 g L−1) suspensions containing HOCl or IO3− with an initial concentration of 8.6 and 10 mM, respectively, were filtered with a 0.45 μm glass fiber membrane. Similarly, a CuO (1 g L−1) suspension containing IO4− (1 mM) after a 4 h reaction time was also filtered. Thereafter, the CuO layer on the membrane was rinsed with 1500 mL of deionized water. The obtained CuO layer was then characterized by SEM-EDS and XPS. Model calculations to simulate the complex reaction systems with HOCl, Br−, I−, HOBr and HOI in various combinations were performed using the program Kintecus.33

CuO

3HOBr ⎯⎯⎯→ BrO−3 + 2Br− + 3H+

(1)

It was proposed that CuO can enhance the reactivity of halogen-containing oxidants (e.g., HOCl, HOBr, and ClO2), thereby accelerating the disproportionation.26−28 However, it is unknown if the reactivity of the CuO-activated species toward iodine is enhanced. This might significantly affect the distribution between inorganic iodine species and I-DBPs. The objective of this study was to investigate the fate of iodine during chlorination of iodide-containing water in the presence of CuO. The influence of drinking water quality parameters such as CuO concentration, Br− concentration, and pH on the formation kinetics of BrO3− and IO3− as well as the potential formation of periodate (IO4−) was investigated. Finally, reaction mechanisms are proposed to explain our observations.



MATERIALS AND METHODS Reagents. All chemical solutions were prepared from reagent grade chemicals or stock solutions using deionized water (Barnstead Nanopure (Skan); 18.2 mΩ cm). A sodium hypochlorite (NaOCl) solution was used as the source of chlorine (Sigma-Aldrich). Cupric oxide particles were prepared according to a previously published method.26 The BET surface area was determined by a Micromeritics Tristar II to be 33.7 m2 g−1. The pHpzc (pH at which CuO particles have a zero charge) was determined to be 8.6.29 HOBr solutions were prepared by reaction of NaOCl with Br− according to a previously described method.26 HOI solutions were prepared freshly by the reaction of NaOCl with equimolar I− concentrations. Analytical Methods. Residual oxidant stands for the sum of [HOCl], [HOBr], and [HOI] (where [HOX] = concentrations of HOX and OX−, X = Cl, Br, I; and [HOX]0 is the initial concentration). Oxidant concentrations were analyzed spectrophotometrically by the N,N-diethyl-p-phenylenediamine (DPD) method at 515 nm on a Varian UV−visible



RESULTS AND DISCUSSION Fate of HOI in the CuO−HOI System. Figure S2 (SI) shows a plot of 1/[HOI] vs time in the presence or absence of CuO at various pH values. The HOI decrease follows secondorder kinetics in the pH range 8.6 to 10.6. The calculated second-order rate constants (k) for HOI decrease in the presence of CuO are 8.6, 26.0, and 16.0 M−1 s−1 for pH 8.6, 9.6, B

dx.doi.org/10.1021/es5032079 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

leads to an increasing oxidant consumption rate, which can be ascribed to the previously described CuO-catalyzed reactions of HOCl.26 A pseudo first order kinetic model was previously used to fit the oxidant decay for chlorination of bromidecontaining waters in the presence of CuO.26,27 Figure S3a (SI) indicates that the oxidant decrease also follows pseudo first order kinetics (R2 > 0.97), with pseudo first order rate constants (k′) of 8 × 10−5, 2.5 × 10−4 and 4.0 × 10−4 s−1 for CuO doses of 0.05, 0.2, and 0.5 g L−1, respectively. If the oxidant decrease in absence of CuO is fitted with pseudo first order kinetics, the k′ value is 7 × 10−5 s−1, indicating that the presence of CuO significantly enhances the oxidant consumption rate. Along with the fast oxidant loss, iodate formation was enhanced by the presence of CuO, and increasing CuO doses enhanced the iodate formation rate (Figure 1b and Figure S3b, SI). More than 90% of the initial I− was transformed to IO3− after reaction times of 45, 60, and 90 min for CuO doses of 0.5, 0.2, and 0.05 g L−1, respectively. These results confirm our previous observations that the reactivities of hypohalous acids are enhanced by CuO. In the presence of 0.05 g L−1 CuO, after the maximum iodate formation at about 120 min, the iodate concentration decreased as a function of the reaction time. However, in the presence of 0.2 and 0.5 g L−1 CuO, the iodate concentrations remained stable. Potential reasons for the significant iodate loss in the presence of 0.05 g/L CuO include (i) adsorption of IO3− and (ii) formation of other iodine species. Figure S4 (SI) shows that the iodate concentration is constant in a CuO−IO3− system, even for long reaction times (a few days), in the presence of either 0.1 or 0.2 g L−1 CuO (without or with pretreatment with 30 μM HOBr). This indicates that there is no significant adsorption of iodate to CuO, which is similar to our previous findings for chlorite, chlorate and bromate.26,28 It is hypothesized that IO3− is further oxidized to other iodine species, probably IO4−. This hypothesis is further supported by the fact that in the presence of 0.2 and 0.5 g L−1 CuO, the available residual oxidant at the maximum IO3− production (i.e., after 60 min) is significantly lower than that in the presence of 0.05 g L−1 CuO. Therefore, the residual oxidant concentration might be decisive for a further transformation of IO3− to IO4−. Iodate Formation in the CuO−HOCl−I− System: Effect of pH. The IO3− formation during chlorination of iodidecontaining waters in the presence of CuO was further studied for pH values ranging from 6.6 to 9.6 (Figure 2). The residual oxidant decreased rapidly at any of the selected pH values due to the interaction between HOCl and CuO as well as the production of IO3− from the reaction between HOCl and iodine. The oxidant decay rates increase with decreasing pH. Applying pseudo first order kinetics to interpret the oxidant decay curves results in k′ values increasing from 1.0 × 10−4 to 8.8 × 10−4 s−1 for pH values decreasing from 9.6 to 6.6 (Figure S5, SI). This trend is similar to our previous study involving the CuO−HOCl and CuO−HOCl−Br− systems.26 The formation of IO3− at various pH values showed a complex pattern. At pHs 6.6 and 7.6, the IO3− concentrations increased as a function of the reaction time up to 20 and 30 min, corresponding to IO3− yields of 77% and 100%, respectively. Thereafter, the IO3− concentrations decreased due to the further transformation of IO3− to other iodine species (e.g., IO4−). In contrast, iodate concentrations

and 10.6, respectively. The presence of CuO does not significantly affect the HOI disproportionation in contrast to HOCl, HOBr, or ClO2 for which disproportionation is significantly enhanced by CuO.26,28 For example, the k values for HOI disproportionation are 8.0 and 8.6 M−1 s−1 in absence and presence of CuO, respectively. These values are comparable to the previously reported apparent second order disproportionation rate constant of 8.8 ± 2.2 M−1 s−1 at pH 8.96 in the presence of 3 mM borate.34 The maximum rate for the noncatalyzed HOI disproportionation is expected to be at a pH equal to the pKa of HOI (10.4). In the presence of borate buffer, however, due to a buffer-catalyzed reaction between two molecules of HOI, the maximum apparent rate constant for HOI disproportionation is shifted to a pH close to 9.6.10 Iodate Formation in the CuO−HOCl−I− System: Effect of CuO Dose. The formation rate of iodate during chlorination was investigated to test whether this reaction is enhanced in the presence of CuO. Figure 1 shows oxidant decay and iodate formation for various CuO doses (0.05−0.5 g L−1) at pH 8.6.

Figure 1. Effect of CuO dose on (a) oxidant evolution and (b) iodate formation. Experimental conditions: [HOCl]0 ≅ 11 μM, [I−]0 = 1 μM, [CuO] = 0−0.5 g/L, pH 8.6, T = 22 °C. Symbols represent the experimental data; lines represent simulations from Model 1 (SI) in the absence of CuO.

In absence of CuO, HOCl reacts quickly with I−. The loss of residual oxidant is fast at the initial stage of the reaction (up to 90 min). After this period, the residual oxidant is more stable with a final total oxidant loss of approximately 3 μM, with the formation of 1 μM iodate after 120 min. This observation confirms the expected stoichiometry of 3 for the oxidation of I− to IO3−. In the absence of CuO, the experimental data of this study (Figure 1) is in good agreement with model calculations based on the rate constants reported previously (Model 1, SI). In the presence of CuO, the residual oxidant decays much faster than in the absence of CuO. Increasing the dose of CuO C

dx.doi.org/10.1021/es5032079 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

(Figure S6b, SI). The rate of the oxidant decrease was enhanced at any Br− dose, since the presence of Br− leads to the formation of HOBr, which has a higher reactivity than HOCl toward HOI or CuO. However, the increase of the consumption rate of the residual oxidant is not proportional to the Br− dose. The maximum IO3− concentration (>90% yield) was reached at 10, 15, and 30 min for Br− doses of 2, 0.5, and 0 μM, respectively. Thereafter, iodate concentrations decrease in these experiments (Figure S6c, SI). As expected, the presence of Br− enhances the oxidation of HOI to iodate but also its further transformation. Even though the iodate formation rate increases with increasing Br− concentrations in absence of CuO, the enhancement of iodate formation in the presence of CuO was decreasing (difference between the dashed and solid lines in Figure S6c, SI). Due to the fact that the reactions between HOBr/OBr− and HOI/OI− are already very fast (k ≥ 740 M−1 s−1 at pH 7),23 significant enhancement by CuO is not expected for higher Br− levels (e.g., [Br−]0= 2 μM). This indicates that the CuO catalysis is limited to relatively slow reactions (e.g., disproportionation of HOCl, HOBr, and ClO2, oxidation of HOI by HOCl). The decrease of IO3− after its maximum formation points toward a further transformation of IO3− to other iodine species (e.g., IO4−) in the CuO−HOCl−I−−Br− system. It is assumed that HOBr contributes to the further oxidation of IO3−. It should be noted that the BrO3− concentrations were less than 3 μg L−1 (i.e., 0.02 μM, corresponding to a bromate yield