Mixed-metal Semiconductor Anodes for Electrochemical Water Splitting and Reactive Chlorine Species Generation: Implications for Electrochemical Wastewater Treatment

A procedure for the preparation of semiconductor anodes using mixed-metal oxides bound together and protected with a TiO 2 nanoglue has been developed and tested in terms of the relative efficiencies of the oxygen evolution (OER), the reactive chlorine species evolution (RCS), and the hydrogen evolution (HER) reactions. The composition of the first anode is a Ti metal substrate coated with IrTaO x and overcoated with TiO 2 (P 25) that was mixed with TiO 2 nanogel, while the second anode consists of a Ti metal substrate coated with IrTaO x and an over-coating layer of La-doped sodium tantalate, NaTaO 3 :La. The experimental efficiencies for water splitting ranged from 62.4% to 67.5% for H 2 evolution and 40.6% to 60.0% for O 2 evolution. The corresponding over-potentials for the Ti/IrTa-TiO 2 and Ti/IrTa-NaTaO 3 :La anodes coupled with stainless steel cathodes of the same dimensions were determined to be 437 mV and 367 mV for the OER, respectively, and 239 mV and 205 mV for RCS, respectively. The preparation procedure described herein should allow for easier production of large-surface area anodes at lower costs than standard methods.


Introduction
We have recently developed prototype electrochemical reactor systems for the treatment of human or domestic wastewater that are powered by either AC or DC power sources [1][2][3][4].In these electrochemical treatment systems, human wastewater is oxidized at an array of semiconductor anodes based on the generation rate of Reactive Chlorine Species (RCS), which is a key active compound in the process of wastewater treatment.Water oxidation and Cl ´oxidation to free chlorine (Cl 2 ) have been known to compete against each other during electrocatalysis due to sharing the active sites for both processes, while H 2 is produced at matched metal cathodes.Hydrogen generation via electrochemical water splitting is generally environmental friendly and reliable for the large-scale hydrogen production that would be required for a hydrogen economy [5][6][7].The cathodic generation of molecular hydrogen is effectively limited by the corresponding anodic oxidation of water.In order to improve the efficiency of water splitting, it is essential to reduce the anodic overpotential for the oxygen evolution reaction (OER) by using optimized electron transfer catalysts.If a chloride ion is present in the electrolyte or wastewater undergoing treatment, then the oxidation of Cl ´to Cl 2 takes place at lower applied potentials in light of the overpotential for the OER [4].
A significant effort combining both theoretical and experimental work has been made over the past decade in order to identify new materials that could lower the overpotential of the OER [8].RuO 2 and IrO 2 base anodes are referred to as dimensionally stable anodes (DSA), which are often used both for the electrochemical generation of reactive chlorine species (RCS) and the OER at relatively low overpotentials [9].However, the high cost of the iridium oxide precursor reagents coupled with the relatively poor long-term stability of IrO 2 has essentially impeded the large-scale commercial application of DSAs.
The process of attachment of nanoparticulate metal oxide semiconductors (e.g., IrO 2 , RuO 2 , Ta 2 O 5 , SnO 2 , Bi 2 O 3 ) to base-metal surfaces is critical for producing efficient anodes and, in some cases, functionalized cathodes.Anatase (TiO 2 ) nanogels have been reported to be a sticky inter-particle binding agent or nanoglue that results in improved electrochemical performance and higher current efficiencies for base-metal supported semiconductor anodes [10].

Anode Characterization
The sequentially layered anode assembly can be visualized in a series of SEM micrographs showing the surface morphologies as determined by EDS analyses for each layer in Figure 1.The surface sampling depth was 1-2 µm for the determination of the elemental composition.The SEM micrograph of the mixed-metal oxide layer of IrO 2 /Ta 2 O 5 on the Ti base-metal plate has a wave-like surface morphology in contrast to the typically observed cracked-film morphology for IrO 2 deposits applied to base-metal substrates [11].The initially deposited layer consists of Ir, O, Ti, and Ta, while the seal coating layer made from Ni(NO 3 ) 3 ¨6H 2 O contains NiO with minor amounts of Ni(OH) 2 as determined by XPS analysis.The electron probe microanalysis of the Ti-IrTa-NaTaO 3 :La composite anode shows that the outermost surface of the anode consists of NaTaO 3 :La mixed with nanoparticulate TiO 2 in distinctly different size domains (Figure 2).The magnified SEM image shown in Figure 2d allows us to distinguish the location of TiO 2 nanoparticles that are located between the larger NaTaO 3 :La particles.The TiO 2 nanoglue serves as an inter-particle binding agent via interlocking surface-hydroxyl groups of hydrated TiO 2 with the surface hydroxyl groups of hydrated NaTaO 3 .These interlocking surface hydroxyl groups condense during elevated temperature dehydration to form > M-O-M < bonds during the annealing steps at 100 ˝C [10].The FTIR spectra shown in the Figure 3 can be used to compare the combination of NaTaO3:La and TiO2 (P25) in the particle phase (Figure 3a) to the electrode surface when prepared with TiO2 nanoglue as shown in Figure 3b.The IR spectrum of the TiO2 nanoglue composite is dominated by surface hydroxyl groups that are characterized by a strong > OH stretch centered near 3100 cm −1 .A broader peak normally observed at 535 cm −1 that is characteristic of skeletal > Ti-O-Ti < vibrations was not detected in the dehydrated TiO2 nanoparticles.Peaks at 1400 and 1500 cm −1 are attributed to C-O vibrational modes that originate from either residual titanium isopropoxide or from acetic acid, which was added to adjust pH.The TiO2 nanogel was used during the formation of the electrode outer layers of the Ti-IrTa-TiO2 and Ti-IrTa-NaTaO3:La electrodes in order to facilitate the interparticle binding of TiO2 (P25) and NaTaO3:La; however, the nanoglue was still detectable after annealing as measured by the OH-vibrational peak at 3100 cm −1 .The XRD pattern of Ti-IrTa-NaTaO3:La is consistent with a combination of anatase TiO2 that originated from amorphous TiO2 nanoglue and NaTaO3:La.In contrast, a mixed phase of anatase and rutile TiO2 is evident in the XRD pattern for Ti-IrTa-TiO2 (Figure 3b).
SEM micrographs clearly show a nanostepped surface morphology of NaTaO3:La due to a change of the surface structure of NaTaO3 due to doping with lanthanum, while the surface of undoped NaTaO3 appears to be relatively flat (Figure S1) [12].NiO deposited on to NaTaO3 doped with La has been reported to be an efficient water-splitting photocatalyst.The evolution of O2 is reported to take place on the grooves of the nanostep, while H2 is catalytically evolved on the ultrafine NiO particles.The loading of NiO nanoparticle coupled with the doping of La into the structure of NaTaO3 increases the lifetime of the trapped electrons and holes, which lead to higher yields of H2 and O2 under continuous illumination [12].The FTIR spectra shown in the Figure 3 can be used to compare the combination of NaTaO 3 :La and TiO 2 (P 25 ) in the particle phase (Figure 3a) to the electrode surface when prepared with TiO 2 nanoglue as shown in Figure 3b.The IR spectrum of the TiO 2 nanoglue composite is dominated by surface hydroxyl groups that are characterized by a strong > OH stretch centered near 3100 cm ´1.A broader peak normally observed at 535 cm ´1 that is characteristic of skeletal > Ti-O-Ti < vibrations was not detected in the dehydrated TiO 2 nanoparticles.Peaks at 1400 and 1500 cm ´1 are attributed to C-O vibrational modes that originate from either residual titanium isopropoxide or from acetic acid, which was added to adjust pH.The TiO 2 nanogel was used during the formation of the electrode outer layers of the Ti-IrTa-TiO 2 and Ti-IrTa-NaTaO 3 :La electrodes in order to facilitate the inter-particle binding of TiO 2 (P 25 ) and NaTaO 3 :La; however, the nanoglue was still detectable after annealing as measured by the OH-vibrational peak at 3100 cm ´1.The XRD pattern of Ti-IrTa-NaTaO 3 :La is consistent with a combination of anatase TiO 2 that originated from amorphous TiO 2 nanoglue and NaTaO 3 :La.In contrast, a mixed phase of anatase and rutile TiO 2 is evident in the XRD pattern for Ti-IrTa-TiO 2 (Figure 3b).
SEM micrographs clearly show a nanostepped surface morphology of NaTaO 3 :La due to a change of the surface structure of NaTaO 3 due to doping with lanthanum, while the surface of un-doped NaTaO 3 appears to be relatively flat (Figure S1) [12].NiO deposited on to NaTaO 3 doped with La has been reported to be an efficient water-splitting photocatalyst.The evolution of O 2 is reported to take place on the grooves of the nanostep, while H 2 is catalytically evolved on the ultrafine NiO particles.The loading of NiO nanoparticle coupled with the doping of La into the structure of NaTaO 3 increases the lifetime of the trapped electrons and holes, which lead to higher yields of H 2 and O 2 under continuous illumination [12].
Given the favorable photosynthetic water-splitting properties of NaTaO 3 :La upon irradiation, we decided to utilize it for the outer layer of our metal oxide semiconductor anode.The synthesized TiO 2 nanogel is characterized by various analytical techniques: TEM image of TiO 2 nanoglue indicated around 8 nm of the particle size; XRD pattern confirmed an anatase phase of crystalline structure;surface hydroxyl groups were verified by FTIR; and the band-gap energy for the nanoglue was determined to be 3.13 eV (Figure S2).Given the favorable photosynthetic water-splitting properties of NaTaO3:La upon irradiation, we decided to utilize it for the outer layer of our metal oxide semiconductor anode.The synthesized TiO2 nanogel is characterized by various analytical techniques: TEM image of TiO2 nanoglue indicated around 8 nm of the particle size; XRD pattern confirmed an anatase phase of crystalline structure; surface hydroxyl groups were verified by FTIR; and the band-gap energy for the nanoglue was determined to be 3.13 eV (Figure S2).

Electrochemical Characterization
The mixed metal oxide anodes were characterized by cyclic voltammetry (CV).The anodes were paired with stainless steel (SS) cathodes that were separated by 5 mm.The CV measurements were carried out in either a 50 mM NaCl or a 50 mM K2SO4 solution at a fixed scanning rate of 20 mV s −1 over a range of applied potentials ranging from 0 to 2.0 V.In Figure 4, the resulting CV plots are shown for (a) IrO2/Ta2O5 coated on to a Ti base-metal plate (designated as Ti-IrTa), (b) a sealing coat of NiO on Ti-IrTa (designated as Ti-IrTa-Ni), (c) an over-coating layer of TiO2 on Ti-IrTa (designated as Ti-IrTa-TiO2), and (d) an over-coating layer of NaTaO3:La on Ti-IrTa (designated as Ti-IrTa-NaTaO3).The Reactive Chlorine Species (RCS) and oxygen onset potentials (OER) were determined from the intersection of the tangents between the baseline and the current signals.The observed onset potentials occurred close to 1.2 V for the production of RCS and 1.25 V for the evolution of oxygen, OER.The corresponding overpotential for the RCS was determined to be 239 mV for the Ti-IrTa-TiO2 anode and 205 mV for Ti-IrTa-NaTaO3:La anode, respectively (Figure 4c).

Electrochemical Characterization
The mixed metal oxide anodes were characterized by cyclic voltammetry (CV).The anodes were paired with stainless steel (SS) cathodes that were separated by 5 mm.The CV measurements were carried out in either a 50 mM NaCl or a 50 mM K 2 SO 4 solution at a fixed scanning rate of 20 mV s ´1 over a range of applied potentials ranging from 0 to 2.0 V.In Figure 4, the resulting CV plots are shown for (a) IrO 2 /Ta 2 O 5 coated on to a Ti base-metal plate (designated as Ti-IrTa), (b) a sealing coat of NiO on Ti-IrTa (designated as Ti-IrTa-Ni), (c) an over-coating layer of TiO 2 on Ti-IrTa (designated as Ti-IrTa-TiO 2 ), and (d) an over-coating layer of NaTaO 3 :La on Ti-IrTa (designated as Ti-IrTa-NaTaO 3 ).The Reactive Chlorine Species (RCS) and oxygen onset potentials (OER) were determined from the intersection of the tangents between the baseline and the current signals.The observed onset potentials occurred close to 1.2 V for the production of RCS and 1.25 V for the evolution of oxygen, OER.The corresponding overpotential for the RCS was determined to be 239 mV for the Ti-IrTa-TiO 2 anode and 205 mV for Ti-IrTa-NaTaO 3 :La anode, respectively (Figure 4c).
The CV profiles observed in the 50 mM K 2 SO 4 electrolyte solutions had steeper polarization curves compared to the comparative measurement in NaCl solutions except for the NiO-coated electrode (Figure 4b).The CV curves obtained in the K 2 SO 4 electrode solutions indicate that only the OER is taking place.In this case, the overpotentials observed for the OER on Ti-IrTaO x -TiO 2 and on Ti-IrTaO x -NaTaO 3 :La were determined to be 437 mV and 367 mV, respectively.The measured overpotentials are comparable to previously reported values for IrO 2 of 300 mV [13].The anode with a sealing coat of NiO on the Ti-IrTaO x electrode had a relatively gentle J-V curve (i.e., slope) and a lower overpotential of 212 mV.This may be attributed to surface corrosion during electrochemical water oxidation [9].In comparison, anodes prepared with an outer layer of NaTaO 3 :La and TiO 2 (P 25 ) bound together with TiO 2 nanoglue had steeper J-V curves with a cathodic shift in the onset potential for both electrolyte solutions that was close to 100 mV.The CV profiles observed in the 50 mM K2SO4 electrolyte solutions had steeper polarization curves compared to the comparative measurement in NaCl solutions except for the NiO-coated electrode (Figure 4b).The CV curves obtained in the K2SO4 electrode solutions indicate that only the OER is taking place.In this case, the overpotentials observed for the OER on Ti-IrTaOx-TiO2 and on Ti-IrTaOx-NaTaO3:La were determined to be 437 mV and 367 mV, respectively.The measured overpotentials are comparable to previously reported values for IrO2 of 300 mV [13].The anode with a sealing coat of NiO on the Ti-IrTaOx electrode had a relatively gentle J-V curve (i.e., slope) and a lower overpotential of 212 mV.This may be attributed to surface corrosion during electrochemical water oxidation [9].In comparison, anodes prepared with an outer layer of NaTaO3:La and TiO2 (P25) bound together with TiO2 nanoglue had steeper J-V curves with a cathodic shift in the onset potential for both electrolyte solutions that was close to 100 mV.

Electrocatalysis
Given the current level of interest in photo-electrochemical water splitting, chlorine production, and electrochemical wastewater treatment [14][15][16] for small scale application or for distributed water treatment systems, alternatives to the platinum group metals or boron-doped diamond electrodes are needed to lower the costs of electrochemical treatment [17][18][19][20][21][22][23][24][25][26].Alternative electroactive materials are needed in order to lower the cost of production of semiconductor anodes for larger-scale practical applications.A major limitation of DSA systems that are currently in use is the dependence on the use of IrO2 or RuO2 as primary ohmic contact material.For example, the typical precursor reagents used, IrCl3 and RuCl3, are quite expensive with the cost of IrCl3 at three times the cost of RuCl3.
Zaradjanin et al. [27] reported on the critical importance of the chemical composition of the semiconductor anode materials in terms of improving interfacial electron transfer rates and for

Electrocatalysis
Given the current level of interest in photo-electrochemical water splitting, chlorine production, and electrochemical wastewater treatment [14][15][16] for small scale application or for distributed water treatment systems, alternatives to the platinum group metals or boron-doped diamond electrodes are needed to lower the costs of electrochemical treatment [17][18][19][20][21][22][23][24][25][26].Alternative electroactive materials are needed in order to lower the cost of production of semiconductor anodes for larger-scale practical applications.A major limitation of DSA systems that are currently in use is the dependence on the use of IrO 2 or RuO 2 as primary ohmic contact material.For example, the typical precursor reagents used, IrCl 3 and RuCl 3 , are quite expensive with the cost of IrCl 3 at three times the cost of RuCl 3 .
Zaradjanin et al. [27] reported on the critical importance of the chemical composition of the semiconductor anode materials in terms of improving interfacial electron transfer rates and for optimization of electrochemical efficiencies.In the case of reactions taking place in an NaCl electrolyte, there is major competition between the OER and RCS reactions for the same active sites on the anode surfaces [18,28].

Water Oxidation
Metal oxide semiconductor electrodes are often used for electrochemical water splitting into O 2 and H 2 in alkaline, acidic, or neutral solutions in spite of their relatively large bandgap energies coupled with valence band edges close to ´3.0 eV.The half-reactions for water splitting and corresponding overall reaction are given in Equations ( 1)-(3): The Ti-IrTa-TiO 2 and Ti-IrTa-NaTaO 3 :La anodes were tested for their water splitting capability in 50 mM K 2 SO 4 alone to avoid a direct composition of reactive chlorine species (RCS) from chloride at the electrode surfaces.The OER reaction mechanism on the metal oxide anodes can be considered to take place according to the sequence of reactions outlined by Comninellis [29].The key reactions in the sulfate electrolyte solutions involve surface bound hydroxyl groups, >MOH, bound or adsorbed water >MOH 2 + , or a hydroxide ion within the near surface electrical double layer, which depends on the electrolyte solution pH and the pH zpc of the active metal oxide at the solid-solution interface [17,30].
If the applied potential in a sulfate electrolyte solution is kept below +2.0 V, then the oxidation of SO 4 2´t o SO 4 ´and S 2 O 8 2´s hould not be a competitive process.Figure 4b shows the cyclic voltammetry (CV) profile of the anodes, Ti-IrTa-TiO 2 and Ti-IrTa-NaTaO 3 :La, in a 50 mM K 2 SO 4 electrolyte solution at the circum-neutral pH of 5.8.The Gibbs free energy (∆G) for water splitting at room temperature is 237.16 kJ/mol at a potential of 1.23 V at pH 0. The observed onset potential for the Ti-IrTa-NaTaO 3 anode coupled with stainless steel cathode is about 1.255 V, which is giving an overpotential of 367 mV at circum-neutral pH in 50 mM K 2 SO 4 .The rate of H 2 and O 2 generation was determined for the Ti-IrTa-NaTaO 3 and Ti-IrTa-TiO 2 anodes using 6 cm 2 active surfaces coupled with stainless steel cathodes at a separation distance of 5 mm as a function of the applied anodic potential, which varied from 1.35 to 2.0 V. Figure 5 shows the time-dependent evolution of H 2 and O 2 as a function of time as measured simultaneously, and O 2 generation rate and current density, which increases with an increase in applied potential.The measured mole ratio of evolved H 2 to O 2 was not found to be 2 to 1 while the generation rate of H 2 and O 2 decreased as a function of time.
The less than 2 to 1 stoichiometric ratio in the reactor headspace may be due to a number of factors including different gas solubility between O 2 and H 2 in water, which depends on the relative partial pressures in headspace of the reactor.The O 2 solubility is 25 times that of H 2 at 25 ˝C and a total pressure of 1.0 atm.In a well-mixed reactor that is lacking a salt bridge or proton-exchange membrane separator, oxygen can be simultaneous oxidized and reduced.Figure 6 shows the H 2 and O 2 evolution rates along with the current densities and current efficiencies.The generation rates of H 2 and O 2 were found to increase linearly with an increase in current density, as given in Table 1.The faradaic efficiencies ranged from 62.4% to 67.5% for H 2 and from 40.6% to 60% for O 2 , respectively.At these efficiencies the Ti-IrTa-TiO 2 and Ti-IrTa-NaTaO 3 anodes, if coupled with more efficient cathodes, could be used for H 2 generation [31].

Wastewater Treatment
There is a growing interest in the use of electrochemistry for wastewater treatment, especially in smaller-scale distributed systems.The oxidation of organic and reduced inorganic compounds on the surface of metal oxides anodes is initiated via the formation of surficial hydroxyl radicals and/or by direct electron transfer to surface-trapped holes [17,30].Chloride ions (Cl ´), which are normally present at variable levels in wastewater, undergo "indirect" oxidation by surface-bound hydroxyl radicals, leading to the production of reactive chlorine species (RCS) including free chlorine (Cl 2 , HOCl, ClO ´) [17,30] and chlorine radicals (Cl ´, Cl 2 ´) [2,16].The RCS generation rate increases with an increase in applied voltage (E a ) up to a limiting value.The Ti-IrTa-TiO 2 and Ti-IrTa-NaTaO 3 :La anodes were evaluated with respect to their free chlorine generation rates and electrochemical efficiencies when coupled with matched stainless steel cathodes.Figure 7 shows the observed generation rates for the RCS, current densities, and current efficiencies as a function of applied anodic potential over the range of +1.35 to +2.5 V during potentiostatic electrolysis (Table 2).The production of RCS is directly proportional to the electrode current density.The OER and RCS have similar standard redox potentials of 1.23 and 1.36 V at pH 0, respectively.On the surface of the mixed-metal oxide anodes, RCS evolution appears to be kinetically favored even though the OER is thermodynamically favored.The electrochemical oxidation of Cl ´on metal oxide anodes can be attributed to either oxidation by hydroxyl radical or by active lattice oxygens [17,32,33] as illustrated below: where k 8 = 4.3 ˆ10 9 M ´1 s ´1 for the homogeneous oxidation of Cl ´by OH.
In typical wastewater with relatively high chloride ion concentrations, there is an obvious competition between the RCS production and the OER on the anode surfaces.In the case of RuO2 or RuO2/TiO2 the MOx+1 surficial sites appear to be the principal reactive sites for chloride oxidation [34].In a typical sequence it initiates by an oxidation of Cl − to a surface bound OCl − (Equation ( 10)), the generation of reactive chlorine species (RCS) is developed as first order in concentration of Cl − with a pseudo-steady-state approximation on MOx(OCl − ).However, at circumneutral pH, the Cl2 reacts with water to form hypochlorous acid and hypochlorite, as described −  In the case of RuO 2 or RuO 2 /TiO 2 the MO x+1 surficial sites appear to be the principal reactive sites for chloride oxidation [34].In a typical sequence it initiates by an oxidation of Cl ´to a surface bound OCl ´(Equation (10)), the generation of reactive chlorine species (RCS) is developed as first order in concentration of Cl ´with a pseudo-steady-state approximation on MO x (OCl ´).However, at circumneutral pH, the Cl 2 reacts with water to form hypochlorous acid and hypochlorite, as described in Equations ( 14) and (11).The formation of chlorate, ClO 3 ´, which we expect to detect from a subsequent reaction that is led by OCl ´or HOCl as a primary reductant(Equations ( 15) and ( 16)).The stepwise sequence of reactions for the production of RCS can be written as follows: H + `-OCl é HOCl ppK aHOCl " 7.53q 2Cl é Cl 2 (13) Cl 2 `H2 O Ñ HOCl `Cl -`H+

.3. Urea Degradation
Urea is the most abundant nitrogen-containing compound in freshly discharged human waste.Upon hydrolysis, urea forms NH 3 and CO 2 , eventually leading to the formation of chloramines, NHCl 2 , NH 2 Cl, and NCl 3 .Urea decomposition and subsequent chlorination using mixed metal oxide anodes [35][36][37] has been reported.Urea degradation using the alternative Ti-IrTa-TiO 2 and Ti-IrTa-NaTaO 3 :La anodes has been investigated as part of this study.During electrochemical oxidation, N 2 and CO 2 along with H 2 and O 2 were formed as gaseous products, while NH 4 + , NO 3 ´, NH 2 Cl, NHCl 2 , and NCl 3 were formed as reactive intermediates and products in the liquid phase.Simultaneous analysis of the gas and liquid phases was performed with a quadrupole mass spectrometer (QMS) for the headspace gas analysis and a DX-2000 ion chromatographic system was used to quantify ions in the aqueous phase [3].Free reactive chlorine (HOCl or OCl ´, Figure 7) is a primary oxidant during urea degradation at relatively fast reaction rates.Figure 8 shows the concentration vs. time profiles of NH 4 + and NO 3 formed during potentiostatic electrolysis at an applied voltage of 2.25 V of a solution of urea at an initial concentration of 41.6 mM with an electrolyte concentration of 50 mM Cl ´in a reaction volume of 70 mL.During a typical reaction, the free molecular chlorine (Cl 2 ) was below the detection limit, while the majority of total chlorine was found to be RCS plus an array of inorganic chloramines [38].The observed concentration vs. time profiles of aqueous products are similar to previous reports [3,38].This suggests that urea degradation takes place via reactions between urea and electrochemically-generated reactive chlorine (i.e., RCS) rather than by a direct oxidation pathway on the electrode surface, which normally takes place via direct electron transfer from nitrogen in urea to the active sites on the anode surface [1].The first step during the oxidation of urea by chlorine has been reported to proceed slowly and thus it is the rate-determining step in the overall electrochemical oxidation and degradation of urea [38][39][40].Based on previously reported results, RCS are formed on the anode surface, which then leads to formation of urea chlorinate to tetrachloro-urea that is subsequently oxidized to CO 2 along with the formation of chloramines.
The reaction sequence shown below in Equations ( 17) to (26)   The reversible hydrolysis of chloramines generates NH4 + /NH3 depending on pH and the [RCS] [39].Trichloramine oxidation by RCS also leads to NO3 − formation [1,38]; however, the yields are quite low for the electrochemical oxidation of urea initiated by RCS.The observed time profiles ofNH4 + and NO3 − shown in Figure 8a,b are consistent with previous reports for urea degradation on a bismuth-doped TiO2 electrode (BiOx/TiO2); in addition, the NH4 + ion in liquid phase increased The reversible hydrolysis of chloramines generates NH 4 + /NH 3 depending on pH and the [RCS] [39].Trichloramine oxidation by RCS also leads to NO 3 ´formation [1,38]; however, the yields are quite low for the electrochemical oxidation of urea initiated by RCS.The observed time profiles ofNH 4 + and NO 3 ´shown in Figure 8a,b are consistent with previous reports for urea degradation on a bismuth-doped TiO 2 electrode (BiO x /TiO 2 ); in addition, the NH 4 + ion in liquid phase increased proportionally along with an increased applied anodic potential, as shown in Figure 8c.The total nitrogen decay profile could be used as a surrogate for urea decomposition, since TN in the aqueous phase includes [NO The TN removal rate, in the case of the Ti-IrTa-TiO 2 anode, was 31% after 20 min of electrolysis at an applied potential of 3.25 V.This level of reduction is higher than we had previously reported for a multi-metal anode with an overcoat of Bi-doped TiO 2 [2].The rate of urea degradation based on the TN decay was ~1.0 mM min ´1 with the Ti-IrTa-NaTaO 3 :La anode.In comparison, the rate of urea degradation or decomposition was found to be 1.3 times faster with the Ti-IrTa-TiO 2 anode than with the Ti-IrTa-NaTaO 3 :La anode.The time profiles for the generated N 2 and CO 2 during electrolysis of urea degradation are shown in Figure S3.From the time-dependent concentration profiles of Figure 8, it is clear that the [NH 4 + ] increased slowly compared to nitrate.NCl 3, which is a likely intermediate leading to the formation of NO 3 ´via the reaction of NH 2 Cl and free chlorine (HOCl) [38], was at pH 6.5.On the other hand, NH 4 + might be formed through a reversible hydrolysis of the inorganic chloramine (NH 2 Cl).The formation of CO 2 is generated via a different pathway from N 2 generation, as described in Equations ( 18) and ( 19) above, and it is indicated that the CO 2 formation is slower than the N 2 generation, as shown in the time profiles of Figure S3.

Experimental Details
3.1.Synthetic Procedures: La-Doped NaTaO 3 and TiO 2 Nanogel Lanthanum-doped NaTaO 3 powder (denoted as NaTaO 3 :La) was prepared in a solid-state reaction [11] Reagent grade starting materials, La 2 O 3 , Na 2 CO 3 , and Ta 2 O 5 , were mixed in a ratio Na:La:Ta = (1 ´x to x to 1).An excess amount of sodium (5 mol%) was added to the mixture in order to compensate for the volatilization during thermal processing.The reagent mixture was calcined in air at 1170 K for 1 h and 1420 K for 10 h with periodic breaks for grinding.Excess sodium was removed by water extraction after calcination.On average, the doping level was 2 mol% La into NaTaO 3 .TiO 2 was synthesized according to a procedure used by Li et al. [10].Titanium tetra-isopropoxide, Ti(OCH(CH 3 ) 2 ) 4 or (TTIP), was diluted with 2-propanol and then added drop by drop into Milli-Q water where the mass ratio was M H2O /M TTIP = 110 in the presence of acetic acid at pH = 2.After complete hydrolysis the suspension was heated at 90 ˝C for 4 h with vigorous stirring.TiO 2 particles were collected by centrifugation to yield a concentrated 15 wt.% gel.The resulting 15 wt.% TiO 2 nanogel is defined as a "nanoglue."

Electrode Fabrication
The metal oxide hetero-junction anodes were prepared by sequential deposition as follows: Ti metal coupons (2 ˆ5 cm 2 ) were pretreated by sand blasting the surface; the sand-blasted Ti-metal coupons were then etched in a boiling 10 wt.% oxalic acid solution for 10 min.The first deposition layer of IrTaO x (i.e., anti-passivation ohmic contact) was prepared by coating with the precursor solution that contained 73 mM H 2 IrCl 6 with 27 mM TaCl 5 dissolved in a solution of ethanol and isopropanol at a 1:1 volume ratio.The deposited layer of IrTaO x was then annealed for 10 min at 525 ˝C.This procedure was repeated four to six times; the final over-coating layer was annealed for an additional hour at 525 ˝C.A protective over-coating layer of TiO 2 in the form of a viscous paste was then applied.The alcohol-TiO 2 coating paste was prepared by mixing 10% by weight Aeroxide P 25 TiO 2 with 15 wt.% of the TiO 2 nanogel; the mixture was then ultrasonically mixed with a 20 kHz sonication system for 1 h.The weight ratio of the P 25 TiO 2 to the nanogel TiO 2 was fixed at a ratio of 7:1.The NaTaO 3 :La overcoat was then prepared as a slurry by mixing 15 wt.% NaTaO 3 :La suspended in alcohol also containing 15 wt.% TiO 2 nanogel at a weight ratio of 3.75:1.The outer-layer slurry coating was then applied with a doctor blade using tape as a spacer on the IrTaOx layer.The sequential deposition process can be summarized in the following steps used for the preparation of anode "A." (1) A mixed suspension of IrO 2 /Ta 2 O 5 at the mole ratio of 73:27 is deposited to provide an anti-passivation layer; (2) a protective sealing coat of Ni(NO 3 ) 3 ¨6H 2 O is then applied; and (3) a slurry of TiO 2 (P 25 ) and TiO 2 nanogel is deposited at the weight ratio of 3.75:1.The preparation of anode "B" was prepared following a two-step sequence of (1) application of an initial coating on the Ti base-metal with a mixture of IrO 2 /Ta 2 O 5 at the mole ratio of 73:27 and then (2) deposition of a suspension slurry of NaTaO 3 :La and TiO 2 nanogel at the weight ratio of 2.85:1.The preparation sequence involves depositing an initial ohmic contact layer on to the base Ti metal, which is followed by a sealing coat layer that is then followed by an overcoating layer (or an overcoat).After each coating layer is deposited, the composite material is thermally annealed as described above.
The electrochemical setup consisted of a mixed-metal oxide semiconductor anode coupled with a stainless steel cathode a separation distance of 5 mm.The effective surface areas for both anodes and cathodes were 6 cm 2 (3 cm ˆ2 cm).The temperature-controlled electrochemical cells were connected to a potentiostat (SP-50, Bio-Logic, Grenoble, France).The applied anodic potential, Ea, was adjusted based on continuous monitoring of the response current (I) and cathodic potential (Ec).Electrochemical water splitting was carried out in electrolyte solutions of either 50 mM NaCl or 50 mM K 2 SO 4 in a cell with total volume of 105 mL; this volume included 35 mL of head space was used to measure the production of H 2 and O 2 during electrolysis.Before each experiment, the reactor was purged high purity N 2 for 45 min.The anodic and cathodic current efficiencies (Faradaic Efficiency, FE) for the production free reactive Cl 2 (e.g., Cl 2 , HOCl, and OCl -) and/or O 2 and H 2 were determined according to the expression FE (%) = [Cl 2 production rate (mol/s) ˆn ˆF/I] ˆ100, in which "n" is the number of electrons transferred for the production of Cl 2 , O 2 , and H 2 , respectively, I is the current (A), t is time (s) and F is the Faraday constant (96,485 C/mol).The free reactive chlorine concentrations were determined using a HACH standard DPD (N,N-diethyl-p-phenylenediamine) method and a HACH DR 900 colorimeter (HACH, Loveland, CO, USA).The method was calibrated using a SpecCheck Secondary Gel Standard Set for DPD Chlorine that has an analytical range from 0 to 6.5 mg/L as free reactive Cl 2 .

Instrumentations
X-ray photoelectron spectroscopy (XPS) analysis was conducted by using a surface science instrument M-probe spectrometer with a monochromatic 1486.6 eV Al Kα X-ray line source directed 35 ˝to the sample surface, which is controlled by ESCA25 capture software.UV-Vis diffuse reflectance spectrum was measured using a Shimadzu UV-2101PC (dual beam) (Shimadzu, Kyoto, Japan) equipped with an integration sphere attachment (Shimadzu ISR-260, Shimadzu, Kyoto, Japan), which is used for reflection and transmittance measurement of liquid and solids.IR spectra were obtained using a Nicolet iS50 FTIR spectrometer (Thermo Scientific Inc., Waltham, MA, USA) integrated a diamond accessory.XRD apparatus used is a PANalytical X-ray diffractometer (X'Pert Pro) (PANalytical, Westborough, UK) which is a closed system that is completely remote controlled via computer.The XRD data was obtained by automatic scanning of a given range of the angle, 20 ˝to 80 ˝.SEM images were collected using a ZEISS 1550VP Field Emission Scanning Electron Microscope (SEM) (Carl Zeiss, Jena, Germany) operating at 10 kV acceleration voltages.Elemental analysis was performed using an energy dispersive X-ray spectroscopy system (EDS, Carl Zeiss, Jena, Germany) integrated with SEM, with the electron beam voltage set at 15 KeV.The particle size of TiO 2 nanogel was estimated with a Transmission Electron Microscope (TEM) image.The analysis of the interest ions, NO 3 ´and NH 4 + ,

Figure 1 .
Figure 1.SEM micrographs and associated EDS data with the associated chemical composition for the sequential coating layers.(a) Top surface over-coating layer (b) Second coating layer of NiO applied as an optional layer, containing a minor amount of Ni(OH)2 as determined by XPS analysis.(c) Primary coating of IrTaOx on the surface of the Ti base-metal support.

Figure 1 .
Figure 1.SEM micrographs and associated EDS data with the associated chemical composition for the sequential coating layers.(a) Top surface over-coating layer (b) Second coating layer of NiO applied as an optional layer, containing a minor amount of Ni(OH) 2 as determined by XPS analysis.(c) Primary coating of IrTaO x on the surface of the Ti base-metal support.

Figure 2 .
Figure 2. SEM images of overcoated substrate surfaces coupled together with TiO2 nanogel: (a) NaTaO3; (b) NaTaO3:La mixed with 1.5% TiO2 nanogel; (c) additional TiO2 nanogel applied on the surface shown in Figure 2b for a total of 3 wt.% TiO2 nanogel; (d) higher resolution SEM image of the over-coating layer of NaTaO3:La mixed with 1.5% TiO2 nanogel inter-particle binding agent that highlights a nanogel binder location shown in the red box.

Figure 2 .
Figure 2. SEM images of overcoated substrate surfaces coupled together with TiO 2 nanogel: (a) NaTaO 3 ; (b) NaTaO 3 :La mixed with 1.5% TiO 2 nanogel; (c) additional TiO 2 nanogel applied on the surface shown in Figure 2b for a total of 3 wt.% TiO 2 nanogel; (d) higher resolution SEM image of the over-coating layer of NaTaO 3 :La mixed with 1.5% TiO 2 nanogel inter-particle binding agent that highlights a nanogel binder location shown in the red box.

Figure 5 .Figure 5 .
Figure 5.The generation rates for H2 and O2 were determined using 6 cm 2 active surface areas for specific test anode and cathode.The Ti-IrTa-NaTaO3 anode was paired with a stainless steel cathode at a separation distance of 5 mm.Evolution rates were measured as a function of the applied anodic potential over the range of 1.35 to 2.0 V vs. SHE in a 50 mM K2SO4 electrolyte:(a) evolved H2 and O2 as measured simultaneously as a function of time at the 1.5 applied anodic potential; (b) O2 evolution rate; and (c) current density as function of applied potential; with (d) current efficiency for O2 evolution.

Figure 5 .Figure 6 .
Figure 5.The generation rates for H2 and O2 were determined using 6 cm 2 active surface areas for specific test anode and cathode.The Ti-IrTa-NaTaO3 anode was paired with a stainless steel cathode at a separation distance of 5 mm.Evolution rates were measured as a function of the applied anodic potential over the range of 1.35 to 2.0 V vs. SHE in a 50 mM K2SO4 electrolyte:(a) evolved H2 and O2 as measured simultaneously as a function of time at the 1.5 applied anodic potential; (b) O2 evolution rate; and (c) current density as function of applied potential; with (d) current efficiency for O2 evolution.

Figure 7 .
Figure 7. Comparative performance of the Ti-IrTa-TiO2 (○) and Ti-IrTa-NaTaO3:La (□) anodes coupled with stainless steel cathodes in terms of (a) current efficiency, (b) reactive chlorine generation rates, and (c) current density as a function of applied anodic potential during potentiostatic electrolysis of 50 mM NaCl solutions using the same cell dimensions and surface areas.

Figure 7 .
Figure 7. Comparative performance of the Ti-IrTa-TiO 2 ( ) and Ti-IrTa-NaTaO 3 :La (˝) anodes coupled with stainless steel cathodes in terms of (a) current efficiency, (b) reactive chlorine generation rates, and (c) current density as a function of applied anodic potential during potentiostatic electrolysis of 50 mM NaCl solutions using the same cell dimensions and surface areas.

Figure 8 .
Figure 8.(a) [NH4 + ] and (b) [NO3 − ] versus time during the potentiostatic electrolysis of a solution containing 42 mM of urea and 50 mM NaCl in a volume of 70 mL for the Ti-IrTa-TiO2 and Ti-IrTa-NaTaO3:La anodes; (c) NH4 + production rate as a function of applied potential; (d) total chlorine production (ClDPD) at +2.25 V for Ti-IrTa-TiO2 and Ti-IrTa-NaTaO3:La; (e) [ N]T vs. time; (f) head-space gas composition at an anodic potential of +3.25 V as determined by online MS.

Figure 8 .
Figure 8.(a) [NH 4 + ] and (b) [NO 3 ´] versus time during the potentiostatic electrolysis of a solution containing 42 mM of urea and 50 mM NaCl in a volume of 70 mL for the Ti-IrTa-TiO 2 and Ti-IrTa-NaTaO 3 :La anodes; (c) NH 4 + production rate as a function of applied potential; (d) total chlorine production (Cl DPD ) at +2.25 V for Ti-IrTa-TiO 2 and Ti-IrTa-NaTaO 3 :La; (e) [ΣN] T vs. time; (f) head-space gas composition at an anodic potential of +3.25 V as determined by online MS.

Table 1 .
H 2 evolution rates and current efficiencies as a function of the applied anodic potential over the range from 1.375 to 2.0 V during electrolysis of water in 50 mM K 2 SO 4 .

Table 2 .
Measured rates of reactive chlorine generation, current efficiencies, and current densities as a function of applied anodic potential from 1.38 to 2.5 V NHE during potentiostatic electrolysis in 50 mM NaCl solutions.
has been proposed to account for the decomposition of urea during chlorination:2 pH 2 Nq 2 CO `pOCl -{HOCl{Cl -q Ñ H 2 NpCOqNHCl `H2 NpCOqNCl 2 `2 H 2 O 3 ´], [NH 4 + ], [NH x Cl 3´x ], although some of the chloramines react to form N 2 , as illustrated above via breakpoint chlorination.In general, the summation of the detected concentration of NH 4 + and NO 3 ´, and Cl DPD -containing nitrogen in the form of a chloramine (NH x Cl 3´x ), is much smaller than the degassed N 2 , N 2 O, NO, or NO 2 .