Electrochemical Removal Of Selenate From Aqueous Solutions

Abstract

Removal of selenate from solution is investigated in batch electrochemical systems using reactive iron anodes and copper plate cathode in a bicarbonate medium. Iron anodes produce ferrous hydroxide, which is a major factor in the removal of selenate from solution. Iron anodes also generate a significant decrease in the oxidation-reduction potential (ORP) of the solution because it prevents generation of oxygen gas at the anode by electrolysis. The removal rates varied from 45.1 to 97.4%, depending on current density and selenate concentration. The transformation of selenate by the process is modeled based on a heterogeneous reaction coupled with electrochemical generation of ferrous and hydroxide. The rates are optimized at lower initial concentrations, higher electrical currents, and the presence of anions. Presence of dissolved oxygen does not cause any significant effects the removal of selenate.

1. Introduction

Selenium (Se) is an essential trace element for animals and humans. It plays an important role in glutathione peroxidase, which protects cell membrane from damage caused by the peroxidation of lipids [1]. At low concentrations, its deficiency is known to cause an increase in heart and liver disease and may be associated with an increased risk of cancer [2, 3]. At high concentrations, selenium is toxic, can cause skin disease, gastrointestinal disturbance, damage to the central nervous system, and birth defects manifested in waterfowls [3]. Therefore, US EPA set the standard of selenium as 0.01 mg Se/L in drinking water. Selenium can be released into the environment from mining activities, fossil fuel combustion, oil refining, and agricultural irrigation. Selenium exists in the environment in four oxidation states (+VI, +IV, 0, and –II) and in several organic forms. Oxyaionic forms, selenate (Se^VIO₄²⁻) and selenite (Se^IVO₃²⁻) are the common forms in oxidized systems. Elemental selenium (Se(0)) and selenides (Se(-II)) exist in reducing zones and unweathered mineral formations. Selenate is mobile in groundwater and surface water because it has low adsorption and precipitation characteristics [4]. Selenite may be adsorbed onto various solid surfaces such as metal oxyhydroxides, clays, and organic matters.

Selenium, mainly selenate and selenite, can be transformed into Se(0) by several methods including biological reduction [5], abiotic reduction by green rust [2, 6, 7], zero-valent iron (ZVI) [8–14], and photocatalysts [3]. Among these reduction methods, ZVI has received more attention because it can be applied in permeable reactive barriers for in-situ remediation of Se in groundwater. However, the reduction of Se is highly dependent on the surface characteristics of iron and dissolved oxygen concentration in groundwater. Surface coating by iron oxide significantly decreases the reactivity, and ZVI does not transform Se under limited concentrations of dissolved oxygen [13–15]. Recently, electrochemical methods have been used to transform contaminants and overcome the uncontrolled reactivity of ZVI.

In an electrochemical system using iron anode, oxidation of iron anode prevents water oxidation and formation of oxygen, and produces ferrous and ferric ions leading to the formation of ferrous and ferric hydroxides[16]. Electrochemical reduction of selenate (selenate to selenite or Se(0)) and oxidation of ferrous or ferrous hydroxide could occur as follows:

$$Se{O}_4^{2-}+2e^{-}+3H^{+}\to HSe{O}_3^{2-}+H_{2}O{E}^0=1.060V$$ (1)

$$HSe{O}_3^{-}+4e^{-}+5H^{+}\to S{e}^0+3H_{2}O{E}^0=0.903V$$ (2)

$$\begin{array}{ccc}F{e}^{2+}\to F{e}^{3+}+e^{-}& & E^0=-0.771V\end{array}$$ (3)

$$Fe{(OH)}_2+O{H}^{-}\to Fe{(OH)}_3+e^{-}{E}^0=0.54V$$ (4)

The oxidation of ferrous to ferric (Eq.3) might not be coupled with the reduction of selenate or selenite (Eq.1 or Eq.2) based on the standard reduction potential.

Although electro-coagulation is reported to remove selenate from wastewater together with other chemicals [17], there is little understanding of the potential of selenate removal from groundwater, and the mechanisms of transformation using two iron-anodes redox system.

The aim of this study is to investigate the role of sacrificial iron anode and compare its performance with inert anodes for electrochemical reduction of selenate in groundwater. The basic transformation mechanisms of selenate are assessed in a mixed electrolyte system under different testing parameters, including initial concentration, electric current, and electrolyte type. The impact of various parameters on the electrochemical reduction of selenate in mixed electrolytes using iron and inert anodes are evaluated.

2. Materials and Methods

All reagents used in this study are ACS or higher grade. A stock solution of 79.0 mg Se/L selenate was prepared by dissolving Na₂SeO₄ (ACS grade from Johnson Matthey Co., USA) in deionized water. The electrochemical experiments were conducted in a 500 mL round flask with three extensions, two on the side to install electrodes and one in the center for sampling. The electrode materials include cast iron (10×15×150 mm³), copper plate (0.2×15×150 mm³), and mixed metal oxide (MMO) mesh (0.2×15×150 mm³). The MMO electrode consists of titanium coated with IrO₂ and Ta₂O₅. The distance between anode and cathode is 9.4 cm, and the mixture was agitated using magnetic stirrer at 100 rpm. A power supply model E3612A (Agilent Co. Ltd. USA) was used. The oxidation-reduction potential (ORP) and pH of the solution were measured by pH and ORP meters and micro-electrodes (Microelectronics Inc., USA). Samples were collected during testing for chemical analysis. The samples were filtered using a syringe with 0.45-µm filter and selenate and selenite were analyzed by an ion chromatograph (IC, DIONEX IC-5000, USA) using 35 mM KOH eluent under a flow rate of 1.0 mL/min. Dissolved and precipitated ferrous or ferric concentrations were analyzed using the 1,5-diphenoltroline method.

A summary of experiments and testing variables is presented in Table 1. The primary experiments were conducted under 90 mA and using 10 mM NaHCO₃ electrolyte with a cast iron anode and a copper plate cathode. Other variables evaluated include currents of 30, 60 and 120 mA and use of sulfate or carbonate solutions instead of bicarbonate solution as a background electrolyte. To assess the effect of oxygen, additional experiments were conducted with continuous nitrogen gas purging to maintain anoxic solution. To evaluate the removal mechanisms of selenate, two additional electrolysis experiments were conducted without selenate in the solution for one and two hours; then the current was turned off, the electrodes were removed, and the desired amount of selenate was added to the electrolyte.

Table 1.

Summary of experimental variable and results

Initial Se(VI) Conc. (mM)	Applied Current (mA)	Electrolyte (10 mM)	N2 gas purging	Average voltage (V)	Rate constant (k')	Regression Coefficient (r2)
0.1	90	NaHCO3	No	25.4	1.223	0.970
0.2	90	NaHCO3	No	27.6	0.853	0.997
0.4	90	NaHCO3	No	25.5	0.470	0.996
0.6	90	NaHCO3	No	23.5	0.291	0.990
1.2	90	NaHCO3	No	20.1	0.206	0.945
0.2	30	NaHCO3	No	10.5	0.350	0.970
0.2	60	NaHCO3	No	19.3	0.393	0.955
0.2	120	NaHCO3	No	35.6	1.106	0.981
0.2	90	Na2CO3	No	14.6	0.330	0.980
0.2	90	Na2SO4	No	13.7	0.766	0.673
1.0	90	Na2CO3	No	10.3	0.239	0.863
1.0	90	Na2SO4	No	9.1	0.358	0.988
0.2	90	NaHCO3	Yes	22.5	0.736	0.977
0.4	90	NaHCO3	Yes	25.1	0.538	0.991

To analyze the adsorbed selenate and selenite onto iron precipitates, samples collected from the reactor were centrifuged at 6000 rpm for 10 min, and the supernatant was discarded. A solution of 0.1 M NaOH was added, mixed for 2 hours, centrifuged, and the supernatant was analyzed [18].

3. Results and Discussions

3.1 Removal mechanism

Electrochemical transformation and removal of selenate by MMO and iron anodes in a mixed electrolyte of 15.8 mg Se/L concentration of selenate were evaluated. In the sodium bicarbonate electrolyte, electrolysis by MMO anode did not cause any significant removal of selenate from the solution (Fig. 1a). This indicates that selenate transformation by direct cathodic reduction is either limited under the testing conditions or that generation of anodic oxidation products (oxygen gas and protons) will limit any potential cathodic reduction of selenate in a mixed electrolyte. Development of high positive ORP shows that oxidizing condition will prevail (Fig.1b).

Figure 1.

(A) Removal of selenate and (B) change in redox potential in batch electrolysis cell using iron and inert anodes.

Experiments with iron anode (Fig. 1a) show significant removal of selenate from the electrolyte at rates that are dependent on the applied current. Almost complete removal is observed under 90 mA in 6 hours. The removal is associated with iron anode oxidation and release of ferrous ions into the solution. The standard electrode potential for selenate to selenite or selenite to Se(0) indicates that reduction may not occur due to oxidation of ferrous into ferric in the solution. The reducing potential of ferrous hydroxide may contribute to the transformation of selenate. However, the ORP value with iron anode tends to increase after 4 hours of operation (Fig. 1b). Initially, iron oxidation is the dominant anodic electrolysis reaction; however, formation of higher Fe(II) concentration and buildup of iron oxides on the anode inhibits the direct iron oxidation causing water oxidation to occur at the anode.

To verify the role of ferrous hydroxide, experiments were conducted using an electrolyte containing Fe²⁺ generated from electrolysis and OH⁻/CO₃²⁻. The system was prepared by applying an electric current through an electrolyte without selenate. After periods of 1 and 2 hours; the current was turned off, the electrodes were removed, and selenate was added. In this case, partial reduction of selenate occurred at rates that depend on the applied current and the duration of electricity (Fig. 1a); which control the total concentration of iron ions generated by electrolysis. The partial removal of selenate from the solution without any current application indicates that the removal of selenate is impacted by the electrolytic reaction. Although the total charge generated from 180 mA for 1 hour is equal to that generated from 90 mA for 2 hours, the removal rate of selenate was higher in the 180 mA experiment. The concentration of total (ferric+ferrous) and ferrous iron in the electrolyte was 8.31 mM and 7.47 mM, under 180 mA, and 6.75 mM and 5.43 mM, under 90 mA (Fig. 2b). This confirms that the removal of selenate is dependent on the concentration of ferrous precipitates generated by anodic oxidation of iron, and the removal of selenate is not directly proportional to the amount of applied current even though the removal increased with the applied current.

Figure 2.

(A) Removal of selenate by Fe(OH)₂ generated by electrochemical reaction and (B) Concentration of iron precipitates.

Generation of hydroxide ions at the cathode enhance the formation of Fe(OH)₂, a reducing agent for selenate and selenite, leading to improved performance [19], based on the following reduction reactions.

After 6 hours of reaction time (Fig. 2b), the concentration of ferrous precipitates decreases while the concentration of ferric precipitates increases. The ferrous precipitates were oxidized to ferric precipitates and selenate was reduced. Although only small concentrations of selenite, the product of selenate reduction, were measured in the solution and the concentration decreased slightly; the major concentration of reduced Se was in a solid precipitate form and not dissolved. A 0.1 M NaOH solution was used to extract adsorbed selenate and selenite, however, the desorbed selenium was negligible, and most precipitates were non-extractable by the 0.1 M NaOH (data not shown). Non-extractable fraction may be in the form of Se(0) [13] and Fe²⁺ and SeO₃²⁻ do not form any crystalline precipitates such as FeSeO₃Fe₂(SeO₃)₂and Fe₂(SeO₃) ₃·6H₂O [20]. Therefore, the non-extractable insoluble Se found in this study is most likely in the form of elemental selenium or selenide.

3.2 Initial concentration

The influence of initial selenate concentration on selenate removal is presented in Fig. 3. Initial selenate concentrations selected for testing are 7.9, 15.8, 31.6, 47.4, 94.8 mg Se/L in 10 mM sodium bicarbonate solution. Selenate concentration decreased continuously over time in all experiments under 90 mA. Selenate concentration decreased to less than 0.79 mg Se/L from initial concentrations of 7.9 and 15.8 mg Se/L within 6 hours. The removal efficiencies were 97.4% (3.97 mg) for 7.9 mg Se/L, 93.4 % (7.46 mg) for 15.8 mg Se/L, 77.1% (12.1 mg) for 31.6 mg Se/L, 59.0 % (13.8 mg) for 47.4 mg Se/L, and 45.1 % (21.0 mg) for 94.8 mg Se/L concentration, respectively.

Figure 3.

Effect of initial concentration on (A) Selenate removal in time and (B) pH changes in time. The experiments were carried out at 90 mA in 10 mM sodium bicarbonate solution.

A pseudo first order reaction kinetic has been proposed for selenate transformation [14]. In this study, however, the selenate was transformed by ferrous hydroxide, and the reaction rate is influenced by the concentrations of both selenate and ferrous hydroxide,

$$\frac{d\left[Se{O}_4^{2-}\right]}{dt}=-k{\left[Se{O}_4^{2-}\right]}^{\alpha }{[Fe{(OH)}_2]}^{\beta }$$

The ferrous in the ferrous hydroxide is generated from the electro-oxidation of iron anode. The total mass of ferrous generated by anodic reaction is calculated by Faraday’s law [21, 22].

$$[F{e}^{2+}]=\frac{It}{zFV}\eta$$

where I the current (A), t the electrolysis time (s), z the number of electrons involved in the reaction, F is Faraday’s constant (95,485 As/mol), V the volume of solution (L), and η columbic efficiency. A fraction of the ferrous ions are oxidized into ferric, and the presence of dissolved form of ferrous is limited. Generally, the precipitation reaction rate is fast, and the concentration of ferrous hydroxide is dependent on the rate of anodic reaction.

$$\frac{d[Fe{(OH)}_2]}{dt}\approx \theta \frac{d[F{e}^{2+}]}{dt}=\theta \frac{I}{zFV}\eta$$

$$[Fe{(OH)}_2]=\frac{I\theta }{zFV}\eta t$$

where θ is the fraction of ferrous forms in the total iron.

To simplify, we assumed that η and θ are constant and α=β=1, and the concentration of selenate in the system can be estimated by:

$$\text{ln}\frac{{\left[Se{O}_4^{2-}\right]}_t}{{\left[Se{O}_4^{2-}\right]}_{t=0}}=-\frac{k}{2}\frac{I\theta }{zFV}\eta t^2=-k^{\text{'}}I{t}^2$$ (5)

Eq.(5) was found to best describe the data in Fig. 4a based on the regression coefficient observed from 0.673 to 0.997. The rate constant (k') and regression coefficients are summarized in Table 1. The reduction rate sharply decreases with the increase in the initial concentration. Although the removal rate at higher concentration was lower than that at lower concentration, the total amount of selenate removed by the process increase with increasing initial concentration. All experiments show gradual increase in pH because only a fraction of hydroxide ions generated from the cathodic reaction is neutralized. Although ferrous hydroxide will form and limit the increase in pH, selenate reduction by ferrous hydroxide will consume hydroxide ions according to Eq.(4).

Figure 4.

Influence of applied current on removal of selenate (A) under different currents and (B) as a function of applied electric charge.

The different removal rates of selenate under different currents again show that the transformation is not affected by direct cathodic reduction. In cathodic reduction, the reduction rate is highly dependent on the surface area of the cathode [23]. On the other hand, a mediated electrochemical reduction is less influenced by the surface area or the cathode material. The reduction rate normally increases with increasing initial concentration to a steady state value under cathodic reduction; while the reduction rate normally decreases with the initial concentration in the electrolytic reduction. When ZVI is used to remove selenate, the reduction rate of selenate is found to be highly dependent on the initial concentration and tend to decrease sharply with the increase in the initial concentration of selenate. This is because ZVI serves as an electron donor for the reduction of selenate and the total amount of electron available for selenate is limited by surface area of ZVI at the initial stage [14]. However, electrochemical oxidation of iron anode is induced by the applied current, which contributes to the continuous transformation of selenate. Additionally, the cathode materials used in this study (copper plate, MMO, cast iron) did not influence the reduction of selenate (data not shown).

3.3 Applied current

Batch experiments were conducted to determine the effect of the applied current using 15.8 mg Se/L of selenate electrolytes. The removal efficiency and the rate constant increased with increasing electric current (Fig.4a and Table 1). Assuming 100% efficiency, the theoretical mass of iron generated by anodic oxidation in 6 hours is 189.5 mg under 30 mA, 379.0 mg under 60 mA, 568.4 mg under 90 mA, and 757.9 mg under 120 mA, respectively. The mass of selenate removed from the solution was 2.25, 4.09, 7.46, and 7.76 mg, respectively, under 30, 60, 90, and 120 mA for 15.8 mg Se/L initial concentration of selenate in 500 mL solution. The amount increased 1.82 times, 3.32 times, and 3.45 times as the current increased from 30 mA to 60, 90 and 120 mA, respectively. This implies that there is an optimum value for the electric current and that increasing the current beyond the optimum value will decrease the efficiency of the process. The coulomb efficiency for iron oxidation may decrease with time and applied current. In this case, the removal efficiency of the selenate will decrease with increasing the current more than 90 mA.

The removal of selenate was calculated based on the electrical charge supplied under different currents (Fig.4b). The removal of selenate is proportional to the accumulated electrical charge regardless of applied current. The removal efficiency was linear with applied electrical charge until all selenate is removed. The fitted relationship between removal of selenate and electrical charge supplied follows the equation: removal (%) = 0.114 (accumulated electrical charge(C))^0.87 or removal (%) = 372.6 (accumulated electrical charge(C))/(6687 + accumulated electrical charge(C)).

3.4 Background electrolyte

Three different background solutions were used to evaluate the effect of electrolyte chemistry on the removal of selenate. The background electrolytes influenced the initial and temporal change of the solution pH (Fig. 5). NaHCO₃ solution had the lowest final pH compared to Na₂CO₃ and Na₂SO₄. The lower pH value in NaHCO₃ is due the buffering capacity of bicarbonate. Na₂SO₄ had the lowest initial pH, but sharply increased to a pH of 10.0 within the first hour (Fig.5c). The Na₂CO₃ electrolyte pH was maintained around 11.0 throughout testing. The redox potential decreased to −300 mV in Na₂SO₄ solution. For Na₂CO₃ solution, the ORP initially decreased sharply to −130 mV and gradually increased over time during testing (Fig. 5d).

Figure 5.

Influence of background electrolyte on removal of selenate: (A) selenate removal, (B) percent removal, (C) pH change, and (D) ORP change in time using different electrolytes.

The removal of selenate was comparable for both Na₂SO₄ and NaHCO₃ solutions; however, the removal rate decreased when Na₂CO₃ solution was used. For initial concentration of 15.8 mg Se/L mM of selenate, the percent of selenate removed after 6 hours was 86.5, 64.0, and 93.4% in Na₂SO₄, Na₂CO₃, and NaHCO₃, respectively (Fig.5a&b). Using a higher initial concentration of 79.0 mg Se/L, the percent of selenate removed was 66.7, 49.5, and 45.1% after 6 hours, and increased to 92.7, 73.4, and 67.0% after 10 hour for in Na₂SO₄, Na₂CO₃, and NaHCO₃, respectively. The difference in removal of selenate between the different background solutions is most likely due to the generation of different types of solid precipitates containing sulfate or carbonate.

3.5 Purging with nitrogen

The influence of dissolved oxygen on the process was evaluated by comparing the experiments with purging nitrogen or without purging. Purging is a general way to remove dissolved oxygen from aqueous solutions, which might decrease the direct oxidation of ferrous hydroxide to ferric hydroxide by dissolved oxygen. The influence of purging nitrogen on the transformation of selenate was negligible (Fig.6a & b). However, the dissolved oxygen increased the redox potential of the solution (Fig.6d), the lowest values of ORP were −27 to 11 mV without purging even though the value decreased significantly compared to the initial value (297 mV). Dissolved oxygen oxidizes ferrous to ferric and the ferric hydroxide precipitate consumes 50% more hydroxide than ferrous hydroxide. As a result, the un-purged system has lower solution pH (Fig.6c), and the pH was 8.8 and 10.0, respectively, after 6 hours. Purging the system with nitrogen gas remove dissolved oxygen making the system more reducing (Fig.6d), and the ORP was lowered to approximately – 140 mV. However, the complete removal of dissolved oxygen may induce ferrous hydroxide precipitates. The color of the solution changed from green to reddish because of ferric hydroxide precipitation. Other than these differences in the solution chemistry, the removal of selenate was almost similar in both cases (approximately 93%, Fig.6b) because the oxidation of ferrous hydroxide was coupled with selenate or selenite reduction. Studies report that the removal of selenate by ZVI under anoxic condition is limited because iron was not oxidized by the dissolved oxygen [14]. However, in this study, dissolved oxygen had negligible influence because oxidation of iron was induced by electricity. The result indicates that electrochemical removal of selenate by iron electrodes can occur regardless of oxic conditions.

Figure 6.

Influence of N₂ purging on removal of selenate: (A) selenate removal, (B) percent removal, (C) pH change, and (D) ORP change in time with or without N₂ purging.

4. Conclusions

Electrochemical removal of selenate using iron anode is investigated. The removal of selenate occurs in solution, possibly due to the formation of ferrous hydroxide precipitates. The selenite reduced from selenate was possibly reduced to Se(0) or Se(−II), non-extractable precipitates by ferrous hydroxide. The reduction rate of selenate increases with increasing the applied current and with decreasing the initial concentration of selenate. The amount of the reducing agent, ferrous hydroxide, is proportional to the applied current. Even though the pH and redox potential vary for different background electrolytes, selenate was removed from the solution but at different rates. The presence of dissolved oxygen would oxidize ferrous irons to ferric forms. The partial oxidation of ferrous could slightly reduce the amount of ferrous hydroxide, a reducing agent of selenate. This electrochemical system can be applied to both anoxic and oxic system to remove selenate. Even though the final concentration in this study is higher than the regulation level, the electrochemical system is effective in removing selenate and reducing its concentration in water system in the groundwater or mining areas.