Electrooxidation of glyphosate herbicide at different DSA® compositions: pH, concentration and supporting electrolyte effect

Electrochimica Acta 54 (2009) 2039–2045

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Electrooxidation of glyphosate herbicide at different DSA® compositions: pH, concentration and supporting electrolyte effect S. Aquino Neto 1 , A.R. de Andrade ∗ Departamento de Química, Faculdade de Filosofia Ciências e Letras de Ribeirão Preto, Universidade de São Paulo, 14040-901 Ribeirão Preto, SP, Brazil

a r t i c l e

i n f o

Article history: Received 29 April 2008 Received in revised form 3 July 2008 Accepted 10 July 2008 Available online 22 July 2008 Keywords: Glyphosate herbicide Electrochemical oxidation DSA® electrodes Electrolysis Wastewater

a b s t r a c t This paper has investigated the electrochemical oxidation of glyphosate herbicide (GH) on RuO2 and IrO2 dimensionally stable anode (DSA® ) electrodes. Electrolysis was achieved under galvanostatic control as a function of pH, GH concentration, supporting electrolyte, and current density. The influence of the oxide composition on GH degradation seems to be significant in the absence of chloride; Ti/Ir0.30 Sn0.70 O2 is the best electrode material to oxidize GH. GH oxidation is favored at low pH values. The use of chloride medium increases the oxidizing power and the influence of the oxide composition is meaningless. At 30 mA cm−2 and 4 h of electrolysis, complete GH removal from the electrolyzed solution has been obtained. In chloride medium, application of 50 mA cm−2 leads to virtually total mineralization (release of phosphate ions = 91%) for all the evaluated oxide materials. © 2008 Elsevier Ltd. All rights reserved.

1. Introduction Remediation of water polluted by toxic organic compounds such as herbicides, dyes, pesticides, pharmaceuticals, detergents and many other highly toxic compounds, even at very low concentration, has been the subject of many investigations [1–5]. These organic pollutants are responsible for a lot of environmental damage when they accumulate in the environment, both in landfields and water [6]. Many investigations are currently seeking for alternatives in order to degrade these toxic organic compounds. Conventional methods such as chemical and biological treatments are no longer efficient once many compounds are resistant to wastewater treatment. Some approaches to the treatment of toxic organic materials, which will remove or convert these pollutants to biodegradable compounds, are available. Electrochemical [7], electro-Fenton [8,9], ozonation [10], Fenton and photo-Fenton [11], photocatalysis, and heterogeneous photocatalysis [12] are some of the processes that have been frequently proposed for the treatment of organic pollutants by many researchers. Electrochemical processes are always claimed to be attractive alternatives to the treatment of organic pollutants due to their versatility, easier operation, effectiveness, and lower cost [13]. The efficiency of the electrochemical process depends on many factors,

∗ Corresponding author. Tel.: +55 16 3602 3725; fax: +55 16 3633 8151. E-mail address: [email protected] (A.R. de Andrade). 1 ISE member. 0013-4686/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2008.07.019

including the electrode material. Metallic oxide electrodes containing RuO2 and IrO2 have been widely used in environmental electrochemistry because of their mechanical resistance, low cost, and successful scale-up in the electrochemical industry. Besides leading to chloro-alkali production, dimensionally stable anode (DSA® ) electrodes are also a good alternative to the oxidation of various organic compounds [14]. Another material which is a good candidate for the electrochemical oxidation and destruction of organic compounds is the boron-doped diamond (BDD), because of its generally large potential window [15,16]. Electrochemical mineralization of organic compounds can occur directly at anodes through the generation of physically adsorbed hydroxyl radicals (Eq. (1)), and the oxidation of organics pollutants, R, may result in fully oxidized reaction products such as CO2 (Eq. (2)) [17]. MOx + H2 O → MOx (• OH) + H+ + e−

(1)

R + MOx (• OH) → CO2 + inorganic ions + MOx + H+ + e−

(2)

The oxygen evolution reaction (OER) is an undesirable side reaction responsible for lowering the current efficiency of the organic compound oxidation in the electrochemical process. The mechanism proposed [18] for this reaction involves the discharge of water molecules at the metal oxide surface (Eq. (1)). Depending of the characteristic of the anode material, the oxygen evolution proceeds via two different pathways: oxidation of weakly adsorbed hydroxyl radicals (Eq. (3)) or by the formation of the higher oxide followed

2040

S. Aquino Neto, A.R. de Andrade / Electrochimica Acta 54 (2009) 2039–2045

by oxygen evolution (Eqs. (4) and (5)). +

MOx (• OH) → MOx + 12 O2 + H + e−

(3)

MOx (• OH) → MOx+1 + H+ + e−

(4)

MOx+1 →

1 O 2 2

+ MOx

electrochemical characterization of these anodes are given elsewhere [29–33]. Two spiraled platinized platinum wires (15 cm), placed parallel to each other, were used as counter electrodes. All potentials are referred to the saturated calomel electrode (SCE).

(5)

Frequently, electrochemical remediation needs more powerful oxidizing conditions as well as the electrogeneration of highly oxidizing species such as OH• , Cl2 and O3 takes place. This can be obtained by changing the electrode material (SnO2 , PbO2 , BDD) or the supporting electrolyte (NaCl). The growth of agriculture has led to an increase in the demand for agrochemicals. In recent years, the intensive use of herbicides has increased environmental concern, partly because of the adverse effects of these chemicals on soil and mainly aquatic microorganisms [19]. Glyphosate [N-(phosphonomethyl) glycine] is a highly effective broad-spectrum, post-emergence, non-selective herbicide that is widely employed in agriculture worldwide [20]. Glyphosate herbicide (GH) is highly soluble in water (12 g L−1 at 25 ◦ C), and it is used in more than 30 cultures to control annual weeds, mainly in sugarcane and soybean plantations. An increase in the use of GH is expected due to the increase in the development of transgenic plants tolerant to this compound [21]. GH has low toxicity; it is quickly biodegraded to its main metabolite aminomethylphosphonic acid (AMPA); which is more toxic and persistent than the original herbicide [22]. The low toxicity of this herbicide to human and animals makes it one of the main reasons for its wide use worldwide. However, exposure of non-target aquatic organisms to this herbicide is a concern of many ecotoxicologists. Several in vivo and in vitro studies on animals have revealed the mutagenic and carcinogenic effect of GH [23,24] as well as its impact on the environment and aquatic life [25]. The half-life of the commercial formulations is relatively long, about 7–70 days [26]. According to resolution 375 of CONAMA (National Brazilian Environmental Council), the maximum value allowed for GH in sweet water is 0.280 mg L−1 , also classified as “extremely persistent” by the US Environmental Protection Agency. The present work investigates the performance of a set of RuO2 and IrO2 based electrodes on GH electrodegradation. The aim of this study was to find out which were the most efficient electrolysis conditions for this reaction by evaluating different parameters such as supporting electrolyte, pH, and current density. 2. Experimental 2.1. Electrolytic system and electrodes The electrochemical measurements were conducted in an open system, using a three-compartment electrolytic cell consisting of: (i) a main body (50 mL of solution), (ii) two smaller compartments containing the counter electrodes, which were isolated from the main body by coarse glass frits. The electrolyses experiments were conducted in the galvanostatic mode, under magnetic stirring. Electrochemical experiments (cyclic voltammetry and galvanostatic electrolyses) were performed using a potentiostat/galvanostat Autolab, mode SPGSTAT30. All experiments were performed at 25 ± 1 ◦ C. The working anodes were 2 cm2 large and they were prepared by thermal decomposition. The precursor mixtures were applied on both sides of the pre-treated Ti support by brushing, as described in previous works [27–31]. The following nominal compositions were evaluated: Ti/(RuO2 )0.70 (Ta2 O5 )0.30 , Ti/Ru0.30 Ti0.70 O2 , Ti/Ru0.30 Sn0.70 O2 , Ti/Ru0.30 Pb0.70 O2 , and Ti/Ir0.30 Sn0.70 O2 . Details about the preparation, methodologies, and the physical and

2.2. Chemicals GH (secondary standard, 95.5%) was provided by Fersol Indústria e Comércio LTDA. All other chemicals were analytical reagent grade and were used without further purification. H2 SO4 and NaOH were employed to adjust the pH values of the solutions. In all experiments, the ionic strength was kept constant ( = 1.5) by adjusting the Na2 SO4 and NaCl concentrations. Solutions were prepared with high-purity water from a Millipore Milli-Q system, and pH measurements were carried out with a pH electrode coupled to a Qualxtron model 8010 pH meter. 2.3. Analyses The chemical structure of GH does not have a chromophore group, so spectroscopic determinations have to be performed after its derivatization reaction. In this study, the GH decays after the electrolyses were followed in two different ways. The first derivatization consisted in the reaction of ninhydrin in the presence of the Na2 MoO4 catalyst at 100 ◦ C, which produced the Ruhemann’s purple product with a maximum absorption at 570 nm [34]. GH degradation was also followed by nitrosation reaction in acid media, which produced a UV spectrophotometrically active compound at 243 nm [35]. As shown in Eq. (6), GH total oxidation produces PO4 3− ions as one of the final degradation products. C3 H8 O5 NP + 8H2 O → 3CO2 + PO4 3− + NO3 − + 24H+ + 20e−

(6)

3−

release rate is a good indication of complete GH The PO4 degradation. PO4 3− determination was performed colorimetrically by the molybdenum blue method, according to the standard method [36]. The detection limit was 0.01 mg L−1 . UV–vis spectra were recorded using a Varian model Cary 50 Conc spectrophotometer. Total organic carbon (TOC) in solutions was determined with a Shimadzu TOC-VCPN . All results are based on the measurement of triplicate samples. AOX analyses were performed using a multi X 2000 model instrument (Analytika Jena, Germany). Residual chlorine and hypochlorite were removed by addition of sodium sulphite before the analyses. Briefly, the measurement involves the adsorption of organochlorine on activated carbon, mineralization of organically bound halogen by combustion at 950 ◦ C, and determination of chloride by micro-coulometric titration. The current density effect was evaluated by determination of the instantaneous current efficiency (ICE) during the electrochemical oxidation. Considering that during electrochemical incineration two parallel reactions (organic compound oxidation and OER) take place, the ICE is defined as the current fraction used for the organic oxidation. The ICE was calculated considering the values of chemical oxygen demand (COD) of the wastewater before and after the electrolysis, using the relation: ICE =

FV [(COD)t − (COD)t+t ] 8I t

(7)

where F is the Faraday constant (C mol−1 ), V is the volume of the electrolyte (m3 ), I is the applied current (A) and (COD)t and (COD)t+t are the chemical oxygen demand (g O2 m−3 ) at time t and t + t (s), respectively.

S. Aquino Neto, A.R. de Andrade / Electrochimica Acta 54 (2009) 2039–2045

Fig. 1. Cyclic voltammograms as a function of anode composition.  = 50 mV s−1 ,  = 1.5 (Na2 SO4 , pH 3).

2041

Fig. 2. Representative cyclic voltammograms at Ti/Ru0.3 Pb0.7 O2 electrode: (—) support electrolyte and (- - -) GH, 6 mmol L−1 .  = 50 mV s−1 ,  = 1.5, SE = Na2 SO4 , pH 3.

3. Results and discussion oxidation lowers the electrochemical degradation efficiency significantly.

3.1. Voltammetric characterization of the anodes Fig. 1 shows the cyclic voltammograms in Na2 SO4 of all the anodes investigated in this work. The voltammetric behavior observed for these electrodes is the one typical of thermally prepared oxide coatings. A detailed description can be obtained elsewhere [27–31]. Briefly, all voltammograms display the onset of the OER at potentials higher than 1.0 V vs. SCE. A comparison of the voltammograms shows that changes in the oxide composition promote a slight shift in the onset of OER, being Ti/(RuO2 )0.70 (Ta2 O5 )0.30 , Ti/Ru0.30 Pb0.70 O2 , and Ti/Ru0.30 Sn0.70 O2 the most active materials for OER. The electrochemically active area of the electrodes (number of active sites) was estimated by integrating the anodic charge (qa ) in the voltammetric curves in the double layer domain (E = 0.2–1.0 V vs. SCE) [37]. The electrochemical active area shows a great dependence on the valve metal (Table 1). For the same amount of active oxide (RuO2 /IrO2 ), the introduction of tin oxide promotes an increase in the number of active sites because of a better distribution of microcrystallites [38]. As expected, increasing ruthenium loadings (70% atomic) raise the active area. The preliminary stage of this investigation consisted in selecting a suitable electrode material where the highest rate of GH electrocatalytic degradation would occur. Cyclic voltammograms for all the electrode compositions in both the absence and presence of GH were conducted. Fig. 2 shows representative cyclic voltammograms obtained in the absence and in the presence of GH. Electrochemical characterization shows that GH is not electroactive in the potential window of the investigated oxide anodes. Its oxidation is hindered by the OER, which is the main difficulty inherent to the electrochemical organic compound degradation. In other words, competition between OER and organic

3.2. Electrolyses as a function of the anode composition Both spectrophotometric methods carried out after the derivatizations reactions (ninhydrin and nitrosation) give exactly the same degradation rate; for this reason, the results presented here are related only to the determination using the nitrosation reaction. Fig. 3 shows the GH degradation profile and ICE as a function of the anode composition when the current density is maintained at 50 mA cm−2 in Na2 SO4 . After 4 h of electrolysis, the influence of the oxide composition on the degradation rate is very significant. For the most efficient electrode compositions, Ti/Ir0.30 Sn0.70 O2 and Ti/Ru0.30 Ti0.70 O2 , GH removal is about 32% and 24%, respectively. For the less efficient compositions, Ti/Ru0.30 Sn0.70 O2 , Ti/(RuO2 )0.70 (Ta2 O5 )0.30 and Ti/Ru0.30 Pb0.70 O2 , only 12%, 8% and 6% of the starting material (1000 mg L−1 ) is oxi-

Table 1 Anodic charge (qa ) obtained for a set of mixed RuO2 and IrO2 oxides electrodes in Na2 SO4 medium ( = 1.5, pH 3) Electrode composition

qa (mC cm−2 )

Ti/(RuO2 )0.70 (Ta2 O5 )0.30 Ti/Ru0.30 Sn0.70 O2 Ti/Ir0.30 Sn0.70 O2 Ti/Ru0.30 Ti0.70 O2 Ti/Ru0.30 Pb0.70 O2

32.1 29.7 26.6 19.8 12.6

Fig. 3. Electrolysis efficiency in function of the electrode composition. () ICE, ( ) release of phosphate ions, () TOC removal, and ( ) GH removal from spectrophotometry data. t = 4 h, iap = 50 mA cm−2 ,  = 1.5 (Na2 SO4 , pH 3). GH initial concentration: 1000 mg L−1 .

2042

S. Aquino Neto, A.R. de Andrade / Electrochimica Acta 54 (2009) 2039–2045

Fig. 4. Proposed mechanism for GH degradation. (1) GH, (2) sarcosine, and (3) AMPA.

dized, respectively. The mineralization rate for these anodes in sodium sulphate solutions was very low. TOC removal is about 24% for the most efficient anode, Ti/Ir0.30 Sn0.70 O2 . The ICE data in these conditions was also very small (<5%), indicating that OER is an important side reaction to be considered in the electrochemical process. Analysis of Fig. 3 shows that the behavior of GH oxidation for the different anodes remains in the ICE profile; the highest efficiency performance was also found at Ti/Ir0.30 Sn0.70 O2 . There is excellent agreement between the results concerning mineralization obtained from the two analyses, release of phosphate ions, and TOC (Fig. 3). However, the degradation values obtained from these analyses are lower than those achieved with the spectrophotometric methods. Considering that the three methods are sensitive and reliable, we can infer that some byproducts other than CO2 and PO4 3− are also being formed. Fig. 4 shows the degradation mechanism proposed for GH degradation. According to a previous investigation using microorganisms, the main intermediates are AMPA and sarcosine (n-methylglycine) [39]. Formation of these species should explain the behavior observed during the electrolysis; i.e., the gap between the results of GH removal and the mineralization results is a clear indication of the formation of less oxidizable species. To obtain useful criteria to judge the real efficiency of different electrode materials, one must eliminate the contributions due to changes in the surface area. An adequate approach is to divide GH removal by the faradaic voltammetric charge (qa ). Fig. 5 shows the results obtained for the normalized and non-normalized oxidation data as a function of the electrode nominal composition. This figure shows that the contribution due surface area is quite important for the overall degradation efficiency. When the area effect is removed, Ti/Ru0.30 Ti0.70 O2 anode has the same efficiency than Ti/Ir0.30 Sn0.70 O2 , indicating that this material has a higher intrinsic efficiency than the latter one. The same behavior is observed for Ti/Ru0.30 Pb0.70 O2 anode. Results from GH degradation as a function of the electrolysis time showed that the influence of the electrode composition on the GH degradation remains exactly the same even after 12 h of electrolysis. This analysis also confirms that GH elimination is still small even after long-term electrolysis. The best GH removal value obtained after 12 h of electrolysis is 65% for Ti/Ir0.30 Sn0.70 O2 . TOC measurements for this electrode show that only 43% of the carbon is mineralized, which implies that 22% of byproducts were formed. The low oxidation rates obtained for the evaluated anodes can be explained by the general mechanism of organic compound oxidation [40]. The reaction of organic compounds is in continuous competition with the OER on the anode side, thus significantly decreasing the rate of organic compound oxidation. Briefly, the oxidation power of the anode is directly related to the overpotential for oxygen evolution. For DSA® -like anodes, the • OH radicals bind strongly to the surface, eventually leading to the indirect oxidation

of organics via formation/decomposition of a higher valence state oxide [16]. Analysis of Fig. 3 shows that Ti/(RuO2 )0.70 (Ta2 O5 )0.30 , Ti/Ru0.30 Sn0.70 O2 and Ti/Ru0.30 Pb0.70 O2 are more active for OER; therefore, they display low activity toward organic degradation. On the other hand, Ti/Ir0.30 Sn0.70 O2 , and Ti/Ru0.30 Ti0.70 O2 anodes lead to the highest degradation rates because of their higher overpotential for O2 evolution. 3.3. pH and concentration effects In an attempt to increase the degradation rate, pH and concentration effects were investigated. In order to determine the effect of the initial pH on the rate of GH degradation, electrolysis experiments were performed using the most active anode material (Ti/Ir0.30 Sn0.70 O2 ) operating at a current density of 50 mA cm−2 . The pH effect was investigated in the range 2–11 for fixed GH concentration (1000 mg L−1 ) and ionic strength ( = 1.5). The effect of the initial GH concentration was also evaluated by using initial concentrations ranging from 50 to 1000 mg L−1 . Fig. 6 shows that after 4 h of electrolysis, GH electrochemical oxidation is higher in acidic medium. A similar behavior for GH degradation at low pH values has been reported elsewhere [41]. Some works have shown that the OER is kinetically favored in alkaline medium [42]. Therefore, lower pH values diminish the oxygen evolution reaction in favor of organic compound oxidation. We also

Fig. 5. GH removal (black) and normalized GH removal (gray) in function of the electrode composition. t = 4 h, Iap = 50 mA cm−2 ,  = 1.5 (Na2 SO4 , pH 3). GH initial concentration: 1000 mg L−1 .

S. Aquino Neto, A.R. de Andrade / Electrochimica Acta 54 (2009) 2039–2045

Fig. 6. Electrolysis efficiency in function of pH for Ti/Ir0.30 Sn0.70 O2 . () GH removal, () TOC removal, and (䊉) release of phosphate ions. t = 4 h, iap = 50 mA cm−2 ,  = 1.5 (Na2 SO4 ).

observed that the degradation rate of GH increases with the bulk concentration. 3.4. Effect of the supporting electrolyte Another possible approach to increasing the degradation rate in DSA® electrodes is to investigate the supporting electrolyte effect. NaCl is one of the most attractive media in the field of electrocatalysis owing to its straightforward impact on electrochemical technology. The advantage of this process is the formation of powerful oxidizing species like chlorine radical, hypochlorous acid (HClO− ), and hypochlorite ion (ClO− ) during the electrolysis. These oxidizing species are responsible for promoting faster organic compound degradation. Oxide electrodes such as DSA® anodes are very active for Cl2 evolution. The chlorine evolution reaction (ClER) mechanism has been described elsewhere [43]. 2Cl−  Cl2(el) + 2e−

(8a)

Cl2(el)  Cl2(sol)

(8b)

The subsequent reaction of Cl2(sol) in solution results in the formation of some reactive species such as HClO and ClO− , as follows: Cl2(sol) + H2 O → HClO + Cl− + H+ +

HClO  H + ClO

−

(9) (10)

The major disadvantage of using NaCl as supporting electrolyte is the possible formation of organochlorine compounds (RCl) during the electrolysis. The formation of these persistent compounds along with electrolysis in chloride medium has been investigated elsewhere. However, investigations have shown that once the R Cl compound is formed, it is quickly consumed before the end of the electrolysis [44]. In order to investigate the influence of the initial concentration of chloride ion on the indirect GH oxidation via chlorine generation, the Ti/Ir0.30 Sn0.70 O2 anode was employed in the electrolyses of GH 1000 mg L−1 for 4 h. These experiments were performed at an initial pH value of 3 and a current density of 50 mA cm−2 . The effect of the initial chloride concentration on the oxidation rate was investigated over the range 200–3500 mg L−1 . Fig. 7 presents the GH removal and the phosphate ions release at different sodium chloride concentrations.

2043

Fig. 7. Effect of different Cl− concentration for Ti/Ir0.30 Sn0.70 O2 electrode on the electrodegradation of GH. (䊉) GH removal and () release of phosphate ions. t = 4 h, iap = 50 mA cm−2 ,  = 1.5 (Na2 SO4 + NaCl, pH 3).

Analysis of Fig. 7 shows that increasing initial chloride ion concentrations lead to a significant increase in the rate of GH removal and organic mineralization. Even at a very low NaCl concentration (220 mg L−1 ), there is an increase of 42% PO4 3− release and 53% in GH removal. For the best chloride concentration (2662 mg L−1 ), 91% of PO4 3− release removal is obtained. The removal efficiency decreases at chloride concentrations higher than 3000 mg L−1 . The decrease in the GH oxidation can be explained considering that the increase in Cl2 evolution is enhanced towards the formation of more oxidizable species (HClO, ClO− ) [43]. Another possibility is the formation of chlorate (ClO3 − ), which is favored at high current density and high chloride concentration [45]. Again, the results show a gap between PO4 3− release and GH removal, which could be related to the formation of chlorinated compounds. In order to investigate this possibility, AOX analysis of the electrolyte solution was performed at the end of the electrolysis. Values of AOX concentration were found ranging from 0.15 to 1.5 mg L−1 . Considering that the starting material concentration is very high (1000 mg L−1 ), the R Cl values are very small to explain the difference observed in Fig. 7. In conclusion, formation of less oxidizable species like AMPA and sarcosine (Fig. 4) might be responsible for this behavior. In order to investigate the influence of the electrode material on organic compound oxidation in Cl-medium, GH electrolyses (1000 mg L−1 ) on different DSA® compositions were performed. At a fixed chloride concentration (2662 mg L−1 ), 50 mA cm−2 was applied for 4 h. In chloride medium, there is no significant influence of the anode composition on the oxidation rate (Fig. 8). In fact, all the investigated electrodes have similar efficiency concerning GH degradation. In all cases, virtually complete GH elimination is observed and the mineralization rate (phosphate release) reaches almost 91%. Results for the normalized and non-normalized oxidation data as a function of the electrode nominal composition was also evaluated in chloride medium. From Fig. 9 we can depict that when the area effect is eliminated, the Ti/Ru0.30 Pb0.70 O2 efficiency is better compared with the other materials. This indicates that the intrinsic efficiency of this material for producing oxidizing species in the presence of chloride is better compared with the other ones. 3.5. Current density effect Fig. 10 shows the effect of current density on GH oxidation for the Ti/Ir0.30 Sn0.70 O2 electrode. Considering data of GH removal

2044

S. Aquino Neto, A.R. de Andrade / Electrochimica Acta 54 (2009) 2039–2045

Fig. 8. Effect of support electrolyte on the degradation of GH on different electrode compositions. () 2662 mg L−1 of Cl− and () 0 mg L−1 of Cl− . t = 4 h, iap = 50 mA cm−2 ,  = 1.5 (Na2 SO4 + NaCl, pH 3).

and release of phosphate ions, the increase in the current density provides a mild increase in the oxidation performance in both investigated supporting electrolytes (Fig. 10A). In the absence of chloride, even when the oxidation performance is increased at high current density (100 mA cm−2 ), the maximum degradation values are lower than 50%. In chloride medium, the production of chlorine/oxidizing species reaches a saturation value at a current density higher than 30 mA cm−2 . In this condition, an almost complete mineralization (91% of phosphate ions released) is achieved at low current density values (I ≥ 30 mA cm−2 ). The ICE data displayed in Fig. 10B shows that, for both supporting electrolytes, that the increase in the current density favors the undesirable side reactions (OER and ClER), i.e., a considerable decrease in the ICE value is observed. Comparing the ICE values in both supporting electrolytes (SE) one can observe that the results in Na2 SO4 are much lower than 100%. Considering that we applied a current density corresponding to the anode potential in the region of water discharge, the OER plays an important role of the total current consumed. To maximize the performance of the

Fig. 10. Effect of current density for Ti/Ir0.30 Sn0.70 O2 electrode on the degradation of GH,  = 1.5 (Na2 SO4 + NaCl, pH 3), t = 4 h. (A) ( and ) GH removal and (䊉 and ) release of phosphate ions. Open symbol (absence of Cl− ), close symbol (2662 mg L−1 of Cl− ). (B) () ICE in absence of Cl− , ( ) ICE in 2662 mg L−1 of Cl− .

organic compound oxidation, one can work at low current density and change the supporting electrolyte. In this latter condition, the formed species are sufficiently strong to oxidize GH and improve the ICE performance, reaching 80% yield. 4. Conclusions The influence of the oxide composition on GH elimination seems to be significant in the absence of chloride. However, results obtained after 4 h of electrolysis indicate low ICE values, since only 32% of GH removal and 24% of TOC occur for the best anode (Ti/Ir0.30 Sn0.70 O2 ) at 50 mA cm−2 . This low oxidation rate is a consequence of the competition between OER and organic compound oxidation. The GH degradation is favored at low pH values. To maximize the ICE, one can work at low current density and change the supporting electrolyte. In chloride medium, almost total combustion of GH is achieved for all the studied anodes. Finally, all the metallic oxide electrodes containing RuO2 and IrO2 evaluated in this work display good performance showing the effectiveness of DSA® electrodes for the anodic mineralization of organic pollutants in chloride medium; perhaps opening the possibility of using this material under mild oxidation conditions for the treatment of organic waste in water treatment. Acknowledgements

Fig. 9. GH removal (black) and normalized GH removal (gray) in function of the electrode composition in chloride medium (2662 mg L−1 ). t = 4 h, Iap = 50 mA cm−2 ,  = 1.5 (Na2 SO4 + NaCl, pH 3). GH initial concentration: 1000 mg L−1 .

Financial support from FAPESP, and CNPq is gratefully acknowledged. Aquino Neto also acknowledges the MS-CAPES fellowship

S. Aquino Neto, A.R. de Andrade / Electrochimica Acta 54 (2009) 2039–2045

and the grant obtained from the University of São Paulo to participate in the 6th ISE Spring Meeting. Authors also acknowledge FERSOL Indústria e Comércio LTDA for kindles providing GH samples. References [1] E. Brillas, M.A. Banos, M. Skoumal, P.L. Cabot, J.A. Garrido, R.M. Rodríguez, Chemosphere 68 (2007) 199. [2] F.H. Oliveira, M.E. Osugi, F.M.M. Paschoal, D. Profeti, P. Olivi, M.V.B. Zanoni, J. Appl. Electrochem. 37 (2007) 583. [3] G.R.P. Malpass, D.W. Miwa, S.A.S. Machado, P. Olivi, A.J. Motheo, J. Hazard. Mater. 137 (2006) 565. [4] C.C. Jara, D. Fino, V. Specchia, G. Saracco, P. Spinelli, Appl. Catal. B: Environ. 70 (2007) 479. [5] E. Weiss, K. Groenen-Serrano, A. Savall, J. Appl. Electrochem. 37 (2007) 1337. [6] R.A. Relyea, Ecol. Appl. 15 (2005) 618. [7] C.R. Costa, C.M. Botta, E.L. Espindola, P. Olivi, J. Hazard. Mater. 153 (2008) 616. [8] I. Sirés, N. Oturan, M.A. Oturan, R.M. Rodríguez, J.A. Garrido, E. Brillas, Electrochim. Acta 52 (2007) 5493. [9] I. Sires, J.A. Garrido, R.M. Rodriguez, E. Brillas, N. Oturan, M.A. Oturan, Appl. Catal. B: Environ. 72 (2007) 382. [10] P. Canizares, R. Paz, C. Sáez, M.A. Rodrigo, J. Electrochem. Soc. 154 (2007) E165. [11] P. Raja, A. Bozzi, W.F. Jardim, G. Mascolo, R. Renganathan, J. Kiwi, Appl. Catal. B: Environ. 59 (2005) 249. [12] S.K. Kansal, M. Singh, D. Sud, J. Hazard. Mater. 141 (2007) 581. [13] D. Gandini, E. Mahè, P.A. Michaud, W. Haenni, A. Perret, C. Comninellis, J. Appl. Electrochem. 30 (2000) 1345. [14] S. Trasatti, Electrochim. Acta 45 (2000) 2377. [15] M. Panizza, A. Kapalka, C. Comninellis, Electrochim. Acta 53 (2008) 2289. [16] R.T.S. de Oliveira, G.R. Salazar-Banda, S.A.S. Machado, L.A. Avaca, Electroanalysis 20 (2008) 396. [17] A. Kapałka, G. Fóti, C. Comninellis, J. Appl. Electrochem. 38 (2008) 7. [18] S. Fierro, T. Nagel, H. Baltruschat, C. Comninellis, Electrochem. Commun. 9 (2007) 1969.

2045

[19] M.A. Farah, B. Ateeq, M.N. Ali, R. Sabir, W. Ahmad, Chemosphere 55 (2004) 257. [20] O.P. de Amarante, T.C.R. dos Santos, N.M. Brito, M.L. Ribeiro, Quim. Nova 25 (2002) 589. [21] M.D.K. Owen, I.A. Zelaya, Pest Manag. Sci. 61 (2005) 301. [22] G.M. Williams, R. Kroes, I.C. Munro, Regul. Toxicol. Pharmacol. 31 (2000) 117. [23] C. Bolognesi, S. Bonatti, P. Degan, P.E. Gallerani, M. Peluso, R. Rabboni, P. Roggieri, A. Abbondandolo, J. Agric. Food Chem. 45 (1997) 1957. [24] V. Lin, V. Garry, J. Toxicol. Environ. Health, Part A 60 (2000) 423. [25] M.T.K. Tsui, L.M. Chu, Chemosphere 52 (2003) 1189. [26] J.P. Giesy, S. Dobson, K.R. Solomon, Rev. Environ. Contam. Toxicol. 167 (2000) 35. [27] R.D. Coteiro, F.S. Teruel, J. Ribeiro, A.R. De Andrade, J. Braz. Chem. Soc. 17 (2006) 771. [28] D.T. Cestarolli, A.R. De Andrade, J. Electrochem. Soc. 154 (2007) E25. [29] J. Ribeiro, A.R. De Andrade, J. Electrochem. Soc. 151 (2004) D106. [30] P.P.D. Alves, M. Spagnol, G. Tremiliosi, A.R. De Andrade, J. Braz. Chem. Soc. 15 (2004) 626. [31] J. Ribeiro, P.P.D. Alves, A.R. De Andrade, J. Mater. Sci. 42 (2007) 9293. [32] D.T. Cestarolli, A.R. De Andrade, Electrochim. Acta 48 (2003) 4137. [33] R.D. Coteiro, A.R. De Andrade, J. Appl. Electrochem. 37 (2007) 691. [34] B.L. Bhaskara, P. Nagaraja, Helv. Chim. Acta 89 (2006) 2686. [35] Food and agriculture organization of the United Nations, specification and evaluations for plant protection products, Glyphosate, 2000/2001. [36] American Public Health Association, American Water Works Association, Water Pollution Control Federation, Standard Methods for the Examination of Water and Wastewater, 20th ed., American Public Health Association, Washington, DC, U.S.A., 1998. [37] S. Trasatti, Electrochim. Acta 36 (1991) 225. [38] J.C. Forti, P. Olivi, A.R. De Andrade, Electrochim. Acta 47 (2001) 913. [39] R.E. Dick, J.P. Quinn, Appl. Microbiol. Biotech. 43 (1995) 545. [40] C. Comninellis, Electrochem. Acta 39 (1994) 1857. [41] C. Shifu, L. Yunzhang, Chemosphere 67 (2007) 1010. [42] N. Sato, G. Okamoto, Electrochim. Acta 10 (1965) 465. [43] S. Trasatti, Electrochim. Acta 32 (1987) 369. [44] C. Comninellis, A. Nerini, J. Appl. Electrochem. 25 (1995) 23. [45] A. Kraft, M. Stadelmann, M. Blaschke, D. Kreysig, B. Sandt, F. Schroder, J. Rennau, J. Appl. Electrochem. 29 (1999) 861.