Comparative Analysis of Photocatalytic and Electrochemical Degradation of 4-Ethylphenol in Saline Conditions.

We evaluated electrochemical degradation (ECD) and photocatalytic degradation (PCD) technologies for saline water purification, with a focus on rate comparison and formation and degradation of chlorinated aromatic intermediates using the same non-chlorinated parent compound, 4-ethylphenol (4EP). At 15 mA·cm-2, and in the absence of chloride (0.6 mol·L-1 NaNO3 was used as supporting electrolyte), ECD resulted in an apparent zero-order rate of 30 μmol L-1·h-1, whereas rates of ∼300 μmol L-1·h-1 and ∼3750 μmol L-1·h-1 were computed for low (0.03 mol·L-1) and high (0.6 mol·L-1) NaCl concentration, respectively. For PCD, initial rates of ∼330 μmol L-1·h-1 and 205 μmol L-1·h-1 were found for low and high NaCl concentrations, at a photocatalyst (TiO2) concentration of 0.5 g·L-1, and illumination at λmax ≈ 375 nm, with an intensity ∼0.32 mW·cm-2. In the chlorine mediated ECD approach, significant quantities of free chlorine (hypochlorite, Cl2) and chlorinated hydrocarbons were formed in solution, while photocatalytic degradation did not show the formation of free chlorine, nor chlorine-containing intermediates, and resulted in better removal of non-purgeable hydrocarbons than ECD. The origin of the minimal formation of free chlorine and chlorinated compounds in photocatalytic degradation is discussed based on photoelectrochemical results and existing literature, and explained by a chloride-mediated surface-charge recombination mechanism.

Introduction

Water pollution is one of the greatest challenges of modern society.[1,2] The treatment of polluted water to enable its reuse is essential,[3] while efficient removal of (recalcitrant) organic pollutants, present in low concentrations, requires development of innovative technology.[4−7] Particularly difficult is treatment of industrial effluents containing high amounts of sodium chloride, which impedes biological treatment due to negative effects of the NaCl on the microbial flora such as plasmolysis.[8,9]

Advanced oxidation processes (AOP) such as photocatalytic degradation (PCD) or electrochemical degradation (ECD) are promising methodologies for the removal of (recalcitrant) organics in saline conditions. Highly reactive oxygen species (ROS), e.g., hydroxyl radicals (•OH) are generated oxidatively (anodically) in both methodologies.[10] Sodium chloride, ubiquitous in many industrial wastewaters,[8] is known to further enhance degradation rates in ECD, due to (anodic) chloride oxidation and the consecutive formation of reactive chlorine species (RCS: e.g., Cl•, Cl2, or HOCl).[11−15] However, the major disadvantage of ECD in saline media is the formation of chlorinated byproducts.[12,16−20] These byproducts typically induce high water toxicity, and thus deteriorate the environment.[21−23]

For TiO2-based photocatalytic degradation, chloride is usually considered to be an inhibitor. Several mechanisms have been proposed to explain the inhibition: scavenging of holes or OH-radicals by chloride ions;[24−27] blocking of active surface sites by chloride ions;[25,28−32] formation of an inorganic salt layer;[33] or chloride acting as surface-charge-recombination-center for photogenerated charge carriers.[34] Often, more than one mechanism is used to explain the inhibiting effect.[35−40] In addition, aggregation of TiO2 particles has been reported to reduce the number of absorbed photons influencing degradation rates.[41] While the inhibiting effect of chloride, using TiO2 as photocatalyst, has been discussed frequently in the literature, detailed studies regarding the formation of reactive chlorine species (RCS), such as hypochlorite, and chlorinated intermediates in aqueous solution (starting from a nonchlorinated parent molecule) are rare, and the obtained results seem to be inconsistent. For example, negligible amounts of organochloride compounds have been reported during phenol degradation in high saline media [50 g·L–1 NaCl],[38,42,43] which appears to be in disagreement with detection of several chlorinated intermediates arising from the photocatalytic degradation of the azo dye Acid Orange 7 at moderate salinity [5.8–11.8 g·L–1].[39]

The role of chloride ions in TiO2-based PCD of organic pollutants is thus not fully understood[44,45] and requires further studies.[46] In particular, the formation of toxic byproducts needs attention, to reveal the environmental impact of AOPs.[4,20,47−52]

Although some comparative studies between different AOPs, or a combination of technologies have been reported,[53−62] to the best of our knowledge a comparative study specifically addressing the concentration dependent formation of chlorinated intermediates during ECD, PCD, and photoelectrochemical degradation (PECD), starting from a nonchlorinated parent compound in saline solutions, has not yet been performed.

In this study the influence of NaCl on the degradation rate and mechanism of the model compound 4-ethylphenol (4EP) is reported. Two salt concentrations, corresponding to brackish (1.75 g·L–1) and seawater salt concentrations (35 g·L–1), covering a relevant concentration range of industrial effluents,[8] have been studied. TiO2 particle suspensions or TiO2 thin films illuminated by UV light were used for PCD and PECD, respectively. A platinized Ti (Ti/Pt) anode was utilized for ECD, and the performance compared to a Boron-Doped-Diamond (BDD) electrode.

The obtained results clearly demonstrate that the rate of degradation of 4EP in ECD and PECD is mainly the result of chemical chlorination, since free chlorine and various chlorinated hydrocarbons were detected. In contrast, the formation of chlorinated compounds was significantly smaller in PCD, and the overall removal effective, despite an inhibiting effect of chloride.

Materials and Methods

Chemical and Reagents

The following chemicals were used in this study: titanium dioxide (Hombikat UV 100, Sachtleben (Venator)), 4-ethylphenol (Sigma-Aldrich 99%), sodium chloride (Sigma-Aldrich ≥99%), sodium nitrate (Sigma-Aldrich ≥99.0%), demineralized water (Merck Milli-Q system, resistivity >18 MΩ·cm), water (LC–MS grade, Biosolve BV), acetonitrile (LC–MS grade, Biosolve BV), formic acid (98–100%, LC–MS grade, Merck), 1-(4-hydroxylphenyl)ethanol (for synthesis, see the Supporting Information, SI), 2-(4-hydroxylphenyl)ethanol (Sigma-Aldrich 98%), 4-hydroxybenzaldehyde (Sigma-Aldrich 98%), 4-hydroxy-acetophenone (Sigma-Aldrich 99%), 4-ethylresorcinol (Alfa Aesar 98%), 4-Ethylcatechol (Sigma-Aldrich 95%), and 2-chloro-4-ethylphenol (AKos GmbH, > 95%), 2,6-dichloro-4-ethylphenol (AKos GmbH, > 90%).

Degradation Experiments

If not otherwise stated, then all experiments were performed in solutions containing NaCl or NaNO3 at various concentrations (0.03 mol·L–1 or 0.6 mol·L–1). Sodium nitrate (NaNO3) was used for comparison, since NaNO3 is supposed to be an “inert” water additive in oxidative processes.

Photocatalytic Degradation

Pretreatment of TiO2 photocatalyst, via annealing at 600 °C for 4 h (named in the following H600), was performed to achieve an improved photocatalytic performance (optimum between adsorbed surface OH groups and available holes at the surface), as reported in earlier work by our group.[63] The presence of the anatase phase was confirmed by XRD analysis (see XRD pattern in Figure S1).

4EP belongs to the group of alkylphenols, which have been found in process water from the petrochemical industry,[65] or as degradation product from alkylphenol ethoxylates (used in the chemical industry).[66] These compounds are suspected to induce endocrine disruptive effects, and thus pose a health risk.[67] In this study 4EP was used as model compound offering several experimental advantages, for example a moderate toxicity and an adequate water solubility. Moreover, the structure consists of three functional groups (benzene ring, alcohol group, alkyl chain) representing important functional groups of more complex compounds found in wastewater[5,6] and is suitable to study the selectivity in reactions with chlorine.[68,69]

PCD experiments were performed as described in the following: 4EP stock solution (approximately 50 mg·L–1) was presaturated with air for 20 min. The catalyst H600 [25(± 1) mg], the required amount of salt and 50 mL stock solution were mixed in a quartz glass beaker, covered with a quartz glass lid and stirred (350 rpm) in the dark for 30 min before the first sample was taken (0 h measurement). UV-irradiation of the suspension was achieved using a custom-made closed box reactor[64] equipped with eight Philips UV lamps (TL-D 18W BLB, λmax ≈ 375 nm, intensity ≈ 0.32 mW·cm–2). Any influence of UV light absorption by the salt solutions on the PCD rate was excluded (Figure S2).[33] Liquid samples were taken after 0.5 h, 1 h, 2 h, 3 h, 4 and 5 h of continuous illumination. Each time approximately 1.5 mL solution was taken. Afterward, the samples were filtered using a Phenex RC membrane filter (0.2 μm; Phenomenex), and analyzed by HPLC-UV and LC–MS.

Electrochemical Degradation

ECD of 4EP was performed in a custom-made single-compartment cell (see scheme in Figure S3). A platinized Ti plate electrode or BDD electrode (Magneto special anodes B.V; geometric surface area 3.14 cm2) was used to study the electrochemical degradation of 4EP. A Pt mesh and a Ag/AgCl electrode (3 M NaCl, BASi) were used as a counter electrode and as a reference electrode, respectively. The reactor was filled with 70 mL 4EP stock solution (presaturated with air). Degradation experiments were carried out under galvanostatic conditions at 15 mA·cm–2 (VersaSTAT4 potentiostat, PAR or biologic SP-300/VMP3). The electrolyte was constantly stirred during the experiments. Sampling was performed as described for PCD. For high NaCl concentration, samples were also taken after 1 min, 2.5 min, 5 min, 10 min, 15 min, 20 min, and 30 min. The cell was not purged with air, and the cathodic reaction is thus likely the formation of hydrogen. The surface area of the Pt gauze was significantly larger than the applied anode, and therefore the cathodic reaction was not limiting the degradation of hydrocarbons (see Figure S3).

To evaluate reactivity of 4EP with reactive chlorine species, 9.5 mL of 4EP stock solution [50 mg·L–1] was mixed with sodium hypochlorite solution (0.5 mL, equal to approximately 20 mg·L–1 free chlorine) and treated for 15 min. Filtration was followed by LC–MS analysis.

Photoelectrochemical Degradation

PECD experiments were carried out in a custom-made Teflon based PEC-cell. The reactor was equipped with a thin film of TiO2 on a Ti-electrode (for the preparation see SI; geometric surface area 3.14 cm2 [TiO2/Ti]) or H600 coated on FTO ([H600/FTO], 2.54 cm2]) and a Pt-mesh counter electrode. The reactor was filled with the 4EP stock solution containing 0.03 mol·L–1 NaCl, and pretreated by purging air. Experiments were carried out under potentiostatic conditions at 0.1 V vs Ag/AgCl reference electrode (3 M NaCl, BASi) using a potentiostat (VersaStat3, PAR or biologic SP-300). The electrolyte was constantly stirred and illuminated with a 365 nm LED [for TiO2/Ti experiments] or 375 nm LED [for H600/FTO experiments] (Roithner Lasertechnik GmbH, APG2C1-365/375-E, 200 mW output power) placed in front of the glass window of the PEC-cell.

Analytical Methods

HPLC analysis was performed by a ThermoFisher Scientific Dionex Ultimate 3000 HPLC system equipped with a UV detector and a reversed phase Luna Omega polar C18 column (3 μm, 150 × 2.1 mm2, protected by a polar C18 security guard column (both Phenomenex)). The sample injection volume was 5 μL. The oven temperature was set to 30 °C. A water ([A]; 0.1% formic acid) - acetonitrile (0.1% formic acid) gradient system was used at a flow rate of 0.200 mL/min and a total run time of 60 min. (0 min 100% A, 1 min 100% A, 30 min 50% A, 35 min 50% A, 45 min 5% A, 50 min 5% A, 55 min 100% A, and 60 min 100% A). For detection, UV absorption was measured at 276 nm. For the quantification, an external standard calibration was applied (for more information see SI Figure S4 and Tables S1–S3).

For the identification of intermediates, a Bruker amaZon SL ion trap using electrospray ionization (conditions: capillary 4500 V, end plate offset 500 V, 15 psi nebulizer gas pressure (N2), 8 L/min dry gas flow (N2), 200 °C dry gas temperature, scan between 50 to 1000 m/z) was coupled to the HPLC system. A divert-valve was used for salt separation. Ionization was performed in both positive and negative mode under alternating conditions using automatic fragmentation for data-dependent acquisition of both MS and LC MS data.

Free chlorine ((FC); total concentration of Cl2 + HOCl + OCl–) measurements were performed in adaption of the EPA-DPD method 330.5, using a Hanna Instruments free chlorine test kit (HI93701) and photometer (HI83099) for a concentration range of 0.00 to 2.50 mg·L–1 [accuracy ±0.03 mg·L–1]. For higher free chlorine concentrations a Hanna Instruments test kit (HI95771) and photometer (HI96771) were used [accuracy ±2 mg·L–1].

Non-purgeable organic carbon (NPOC) content was determined with a SHIMADZU total organic carbon analyzer, TOC-L CPH/CPN (see SI for more information)

Results and Discussion

Electrochemical Degradation (ECD) of 4EP

Degradation curves of 4EP obtained for ECD at various salt conditions (NaNO3 or NaCl containing electrolytes), at constant current density of j = 15 mA·cm–2, are shown in Figure . A linear degradation behavior was obtained for both NaNO3 concentrations (Figure a), and zero-order kinetic constants can be estimated to be ∼20 μmol L–1·h–1 and 30 μmol L–1·h–1 for the low or high salt concentrations, respectively (voltage–time curves are shown in Figure S5 and quantified numbers are presented in Table S4).

Figure 1

Electrochemical degradation of 4EP at various salt conditions: a) 0.03 mol·L–1 NaNO3 (light blue), 0.6 mol·L–1 NaNO3 (dark blue); and b) 0.03 mol·L–1 NaCl (light green), 0.6 mol·L–1 NaCl (dark green). ECD was performed at a Ti/Pt working electrode at constant current density of 15 mA·cm–2.

In contrast to NaNO3, which cannot be oxidized to reactive radicals, the degradation of 4EP in chloride-containing media proceeded much faster (Figure b). Again assuming linear regression (zero-order), constants of ∼300 μmol L–1·h–1 and ∼3750 μmol L–1·h–1 can be computed for low and high NaCl concentration, respectively. The fast 4EP degradation in the NaCl experiments can be explained by chlorine mediated electrochemical oxidation, where free chlorine (i.e., hypochlorite) and chlorine radicals support the degradation of organic pollutants.[11−15] In agreement with this hypothesis, the measured concentrations of FC amounted to 0.95 mg·L–1 (after 5 h at 0.03 mol·L–1 NaCl) and 111 mg·L–1 (after 30 min at 0.6 mol·L–1).

Theoretically an electric charge of 110 Coulomb is required for full mineralization of 4EP, corresponding to approximately 40 min of continuous treatment at the applied current density of j = 15 mA.cm–2.

The presence of RCS was found to induce the formation of undesired chlorinated organic intermediates and byproducts (RCl), in agreement with literature.[12,16−20] LC–MS analysis revealed the formation of 2-chloro-4-ethylphenol (2C4EP) and 2,6-dichloro-4-ethylphenol (26DC4EP), see Figure . The same intermediates were also formed during chemical chlorination of 4EP with sodium hypochlorite solution (20 mg·L–1 FC), showing that conversion of 4EP in saline conditions resembles chemical chlorination. The chlorination of the aromatic ring at the ortho position is directed by the adjacent OH-group and is a typical reaction of phenolic compounds with electrophilic oxidants.[70,71] Furthermore, the identified intermediates are in good agreement with other literature reports on phenol and bisphenol conversion, using degradation at BDD electrodes at low salt concentrations[16,18,19] and the chemical chlorination of methylphenols.[72]

Figure 2

Reaction path of 4EP with RCS (mainly hypochlorite) during ECD at 15 mA·cm–2 at a Ti/Pt electrode or BDD electrode in NaCl media (0.03 mol·L–1 NaCl and 0.6 mol·L–1 NaCl). 2-Chloro-4-ethylphenol (2C4EP) and 2,6-dichloro-4-ethylphenol (26DC4EP) were identified as the main aromatic intermediates.

In order to understand the subsequent degradation process of the two intermediates, the concentration vs time profiles were determined (Figure ). Obviously, 4EP is converted into 2C4EP almost instantaneously. For the low NaCl concentration (Figure a) a maximum in 2C4EP concentration is obtained after 1–1.5 h of treatment time, with 2C4EP already exceeding the concentration of 4EP. The measured concentration of 149 μmol L–1 [23.3 mg·L–1] of 2C4EP corresponds to a molar fraction of 34% relative to the initial molar concentration of 4EP (for more information see also Figure S6).

Figure 3

Formation and degradation of 2C4EP (red) and 26DC4EP (blue) formed during ECD of 4EP (black) in 0.03 mol·L–1 NaCl (a and c) and 0.6 mol·L–1 NaCl solution (b and d) using Ti/Pt (a and b) or BDD electrode (c and d).

26DC4EP is already formed during accumulation of 2C4EP indicating a continuous consecutive transformation of 4EP into 2C4EP and 26DC4EP (Figure ). Again these results are in agreement with the chemical chlorination of phenol[70] and recent studies on phenol degradation at a BDD electrode in solutions of low salinity.[16] A concentration maximum of 45 μmol L–1 [8.6 mg·L–1] in 26DC4EP, corresponding to a molar fraction of 10% relative to the initial molar concentration of 4EP, was obtained after 2 h. It took as long as 4 h of ECD at low NaCl concentration, to remove both intermediates completely from solution (below the limit of detection (LOD), see Table S3 for more information).

Similar concentration profiles were obtained for the high salt conditions, however within shorter reaction times (Figure b). Already after 5 min, 2C4EP reached its maximum concentration (90 μmol L–1 [14.1 mg·L–1], corresponding to 21 mol % of the initial 4EP concentration). 26DC4EP reached its maximum concentration after 10 min (30 μmol L–1 [5.7 mg·L–1], corresponding to 7 mol % of the initial 4EP concentration). After 20 min of treatment both intermediates were removed from solution. For comparison, the performance of a BDD electrode is shown in Figure c and d) for the low and high NaCl concentration. While small changes in the kinetic curves of the degradation of 4EP and formation of chlorinated intermediates are present, generally both electrodes perform similarly under the studied conditions.

The results suggest that significant amounts of the same mono- and dichlorinated intermediates were formed and built-up in the solution for both NaCl concentrations and both electrode types. Although the apparent removal of the starting compound 4EP is fast (compared to NaNO3) a large fraction of 4EP is directly converted into chlorinated compounds in a consecutive reaction path (Figure ). Interestingly for the higher salt concentration less 2C4EP and 26DC4EP were measured, however the higher FC concentration led to a faster removal rate which most likely suppresses more accumulation of these intermediates.

In addition to the two identified intermediates various other chlorinated compounds were detected by MS. While identification of their structure exceeds the scope of this work, it is interesting to note that the degradation of these intermediates is significantly delayed (Figure S7). The degradation pattern indicates that these compounds are obtained by additional reactions of the two main intermediates, 2C4EP and 26DC4EP. Hydroxylation[72,73] and subsequent ring opening[73,74] might cause the formation of several other intermediates. In addition, dimerization has been proposed for chemical chlorination of methylphenols leading to polychlorinated methylphenoxymethylphenols,[72] and the formation of polychlorinated phenol oligomers was observed during ECD of phenol and 2-chlorophenol.[74,75] Therefore, it is likely that the detected intermediates are congeners of the reported polychlorinated phenol oligomers.

The formation of small chlorinated end products such as chloromethanes[17,76] and chlorinated acetic acids[17] has been reported for ECD in saline water, and formation likely occurs 4–5 h after the start of our experiment. In order to analyze the residual organic content after oxidation of 4EP, non-purgeable organic carbon (NPOC) measurements were performed. The NPOC values (Figure ) clearly indicate that large amounts, between 64% to 86% of the initial organic carbon, were still present. Even after extended treatment times of 5 h (high salinity) only 23% of organic carbon was effectively removed. These nonvolatile compounds are most probably low molecular weight organic compounds such as organic acids, which have been detected after aromatic ring cleavage during ECD in NaCl solution[74] and might also consist of chlorinated derivates such as chlorinated organic acids.[17]

Figure 4

Non-purgeable organic carbon (NPOC) measurements of treated waters after ECD of 4EP at constant current conditions (15 mA·cm–2) at a Ti/Pt or BDD electrode in NaCl media. NPOC values were measured after 5 h ECD in 0.03 mol·L–1 NaCl media and after 30 min and 5 h for 0.6 mol·L–1 NaCl.

The experimental observations highlight a major bottleneck of electrochemical AOP techniques. Although ECD in saline conditions appeared to be efficient, especially compared to inert salt conditions, undesirable (poly)chlorinated aromatic intermediates were formed for both electrodes that require long treatment times. Moreover, the degradation of 4EP and (chlorinated) intermediates on a Ti/Pt electrode led only to a breakdown of the aromatic compounds, and the total removal rate of organic carbon was quite poor. Only for BDD at low salt concentration slightly higher NPOC removal was achieved, most likely due to an increased contribution of generated OH radicals. Still, 64% of the NPOC remained after 5 h treatment and significant amounts of RCl were formed (Figure ), leading to the conclusion that for both electrodes chlorination is the main degradation pathway under the studied conditions, which is in agreement with a recent report of high faradaic efficiency for chloride oxidation on BDD electrodes.[77]

Photocatalytic Degradation (PCD) of 4EP

The photocatalytic degradation of 4EP at various salt conditions is shown in Figure . Using a pseudo first order approximation, implying surface mediated conversion is dominating, rate constants are estimated to be 1.13 h–1 in the absence of salt, 0.8 h–1 in the presence of NaNO3 (high and low concentration), and 0.8 h–1 and 0.5 h–1 for low and high concentration of NaCl. To compare the rates to ECD, initial rates can be calculated (using the initial concentration of 410 μmol L–1 [50 mg·L–1]) to be ∼330 μmol L–1·h–1 or ∼205 μmol L–1·h–1 for the low or high NaCl concentration, respectively. At low NaCl concentration, the rates of ECD and PCD are comparable, but in contrast to ECD, addition of NaCl to high concentration, had a negative effect on the degradation rates of the parent compound 4EP, leading to an order of magnitude difference under these conditions (∼3750 μmol L–1·h–1 vs 205 μmol L–1·h–1 for respectively ECD and PCD). The inhibiting effect of NaCl is in agreement with observations reported for PCD of phenol.[38] Since the inhibitation at 0.6 mol·L–1 NaCl is clearly more significant than at 0.6 mol·L–1 NaNO3, the inhibition is mainly attributed to the presence of Cl–.[24]

Figure 5

Photocatalytic degradation of 4EP at various salt conditions: no salt (black curve), 0.03 mol·L–1 NaNO3 (light blue), 0.6 mol·L–1 NaNO3 (dark blue), 0.03 mol·L–1 NaCl (light green), and 0.6 mol·L–1 NaCl (dark green). PCD was performed under 375 nm illumination in a TiO2 (H600) slurry [0.5 g·L–1].

While the exact reason for the negative effect of anions is still under debate, blocking of active surface sites[25,28−32] appears unlikely as the main explanation. Considering a point of zero charge (PZC) for TiO2 between 5 to 6,[78] the TiO2 particle surface charge in this study is neutral as measurements were performed at circumneutral conditions. Moreover, under illumination the buildup of a negative surface charge has been observed for TiO2 particles (pH > isoelectric point)[79,80] hindering competitive adsorption or blocking of active surface sites by chloride anions. In the case of chloride adsorption on the TiO2 particles (pH < PZC, for example possible due to a pH drop during PCD because of dissolved CO2 produced during degradation; see also Figure S8) the oxidation of chloride ions, by either photogenerated valence band holes and/or hydroxyl radicals becomes likely.[24−27,35−37,39,40,57] As a consequence RCS formation would occur, supporting the 4EP removal as observed in ECD. However, our results indicate that the degradation pathways of 4EP during PCD and ECD are different and RCS likely do not contribute to PCD. Thus, just blocking of active sites by Cl– cannot fully explain the observations. In order to verify this conclusion we analyzed the intermediates formed during (PCD).

Unlike in the case of ECD, chlorinated intermediates were identified only in extremely low quantities, and HPLC analysis showed the following aromatic intermediates in significantly larger quantities: (1) 1-(4-hydroxyphenyl)ethanol, (2) 4-hydroxybenzaldehyde, (3) 4-hydroxyacetophenone, (4) 4-ethylresorcinol, and (5) 4-ethylcatechol (Figure a; more information SI Table S1). Thus, the model compound 4EP is oxidized at least at three carbon atoms during PCD, which is typical for nonselective radical reactions (i.e., by •OH). The formation of the isomer 2-(4-hydroxyphenyl)ethanol could not be revealed due to overlapping retention times with (1) (SI Table S1). However, the increased stability of secondary alkyl radicals favors the formation of intermediate (1) and its subsequent oxidation product (3). Therefore, the formation of 2-(4-Hydroxyphenyl)ethanol is of minor contribution during PCD of 4EP. The same intermediates were detected for all PCD conditions. The nonselective radical reaction pathway is also in good agreement with other reports, e.g., for PCD of ethylbenzene,[81] phenol[37,42,82] or p-cresol,[83] confirming that RCS are insignificantly involved in the degradation path.

Figure 6

a) Chemical structures of the identified aromatic intermediates during PCD of 4EP observed for all studied salt conditions (no salt, 0.03 mol·L–1 NaNO3, 0.6 mol·L–1 NaNO3, 0.03 mol·L–1 NaCl, 0.6 mol·L–1 NaCl: (1) 1-(4-hydroxyphenyl)ethanol, (2) 4-hydroxybenzaldehyde, (3) 4-hydroxyacetophenone, (4) 4-ethylresorcinol, and (5) 4-ethylcatechol. b) Formation and degradation of intermediate (3) 4-hydroxyacetophenone during PCD of 4EP at various salt conditions: no salt (black curve), 0.03 mol·L–1 NaNO3 (light blue), 0.6 mol·L–1 NaNO3 (dark blue), 0.03 mol·L–1 NaCl (light green), and 0.6 mol·L–1 NaCl (dark green).

For the most abundant intermediate (3), its formation and degradation at different salt conditions is shown in Figure b. At high NaCl concentration a pronounced inhibition of formation and subsequent degradation was observed, whereas the influence of NaNO3 or low NaCl concentration is less. The observations are related to the 4EP removal, where the strongly inhibited 4EP degradation explains the delayed formation of (3) at high NaCl concentration. This is in agreement with reports for PCD of phenol.[38]

As 2C4EP and 26DC4EP were detected as main chlorinated intermediates in ECD the obtained chromatograms were thoroughly analyzed in PCD. Only for high salt concentration small amounts of 2C4EP could be measured and quantified (Figure ). Similarly to ECD measurements, formation of 2C4EP occurs immediately, slowly accumulates in the reactor, and its degradation is strongly retarded in agreement with intermediate (3) monitored during PCD at high saline conditions (see Figure b). If at all formed, then the concentration of 26DC4EP remained below the LOD.

Figure 7

Formation of the monochlorinated intermediate 2C4EP during PCD of 4EP in 0.6 mol·L–1 NaCl solution.

Again FC concentrations were determined. For low NaCl concentrations the measured values are below the limit of accuracy; for high NaCl concentrations only small amounts of FC could be detected [0.15 mg·L–1]. This shows that PCD is hardly mediated by FC species, in agreement with the low quantities of chlorinated intermediates.

In general the oxidation of chloride by scavenging OH radicals is not favored at the conditions of this study.[45,52,84] Zhang et al. concluded chloride being a poor hydroxyl radical scavenger at circumneutral to alkaline conditions, because the equilibrium of the reaction is favored for the reverse reaction to Cl– and •OH (>99.98%).[52] The absence of FC for smaller Cl– concentrations is also in agreement with findings reported by Krivec et al.,[32] who did not detect Cl• radicals. For higher NaCl concentrations the results match the observations by Azevedo et al.[38] who reported small amounts of 4-chlorophenol (not quantified) formed during phenol PCD at high salt concentrations. Anyway, the results indicate that chloride is barely oxidized during PCD, and formation of RCS is negligible.

As for the ECD experiments, the overall removal of organic carbon during the PCD treatment was estimated by determination of the NPOC content. Results obtained after 5 h of PCD in saline water are shown in Figure . In the absence of any salt NPOC was below the LOD. This is in agreement with the non-affected 4EP and intermediate removal (compared to salt containing solutions) and indicates a complete removal of the organic carbon within the duration of the PCD experiments. For the low and high NaCl concentration, residual NPOC remained in the solution in agreement with the delay in 4EP degradation. For example, although 4EP and the main intermediate (3) were removed after 5 h treatment in 0.03 mol·L–1 NaCl solutions, NPOC of 13% was measured. Since the UV chromatogram did not show any significant peaks after 5 h of PCD, the residual organic carbon is most likely due to compounds arising from ring-opening reactions, such as short chain aliphatic compounds and carboxylic acids.[82,85,86] At high NaCl concentrations, a residual NPOC of 46% remained. Calculated photonic efficiencies of PCD experiments for the various salt conditions are shown in Table S5 and an estimation of the energy consumption compared to ECD experiments is shown in Table S6. For the applied conditions the energy consumption of the PCD treatment was approximately between 1 and 2 orders of magnitude higher compared to the ECD; however, it is important to note that the applied reactors were not optimized for energy consumption, and the applied light sources have low efficiency for illumination of the photocatalytic slurry (less than 1%). Although ECD thus seems to appear beneficial in terms of consumed energy per kg removed non purgeable organic carbon (NPOC), degradation of 4EP by PCD leads to significantly smaller NPOC than ECD at comparable rate of 4EP conversion (low NaCl concentration). Moreover, the concentration of chlorinated intermediates formed during PCD is significantly lower (for low and high NaCl concentration) compared to ECD, thus photocatalysis appears to be the preferred method for application, in particular if illumination efficiency and reactor design are optimized.

Figure 8

Non-purgeable organic carbon (NPOC) measurements of treated waters after 5 h PCD of 4EP in 0.03 mol·L–1 NaCl and 0.6 mol·L–1 NaCl.

Finally, photoelectrochemical degradation (PECD), a hybrid technology, was studied. The experiments were performed with a TiO2/Ti electrode at an applied anodic bias of 0.1 V vs Ag/AgCl irradiated with UVA light (see Figure S9 for more information). In recent reports Zanoni et al. showed already that oxidation of chloride to FC at TiO2 photoelectrodes occurs[87] and the generation of RCS has been proposed to assist the PECD of organic compounds.[46,57,61,88] Still, reports discussing the formation of intermediates are scarce.[50,89] After 15 min of irradiation at low NaCl concentration, 92% of 4EP was removed and an intensive formation of RCS was observed (as detected by FC [27 mg·L–1]). The HPLC analysis (Figure ) revealed significant amounts of the same intermediates, as observed during ECD (57 μmol L–1 [9 mg·L–1] 2C4EP and 89 μmol L–1 [17 mg·L–1] 26DC4EP). Interestingly, without additional bias (resembling PCD conditions) intermediate (3) was observed as main intermediate and the formation of chlorinated compounds was excluded (Figure S10). Thus, the 4EP reaction pathway can be easily modified by the applied bias. The blank experiment (PECD in dark at 0.1 V vs Ag/AgCl) did not lead to any photocurrent, RCS formation, or 4EP degradation (Figure S11), because the applied potential was below any relevant standard redox potential (e.g 1.36 V vs SHE for chlorine evolution or 2.8 V vs SHE for •OH generation). To verify the observations and exclude a particular effect of the applied TiO2, PECD experiments were conducted with the H600 photocatalyst coated on FTO glass, showing similar trends (see Table S7).

Figure 9

HPLC chromatogram obtained after 15 min PECD of 4EP in 0.03 mol·L–1 NaCl solution using a TiO2/Ti electrode with 0.1 V[vs Ag/AgCl] bias and 365 nm UV light. The majority of 4EP was removed and transformed into significant amounts of the monochlorinated 2C4EP and dichlorinated 26DC4EP, similar as observed during ECD experiments.

The effect of an applied anodic bias can be explained by the following: In a photocatalytic approach charge carrier (electrons and holes) are usually generated close to the particle surface and physical separation of reduction and oxidation sites is hardly achieved on the nanoscale. Thus, oxidation and reduction occur in close proximity and the recombination of oxidized chloride with photogenerated CB electrons is likely. This surface-charge recombination mechanism (quenching of photogenerated charge carriers by chloride ions)[34] is effectively suppressed when an anodic bias is applied, e.g., in PECD experiments. Upon biasing the electrode photogenerated conduction band (CB) electrons from the TiO2 surface are pulled toward the back contact and in contrast to PCD, reduction of Cl-radicals by available CB electrons is unfeasible. Therefore, pronounced RCS generation could be observed during PECD, leading to a distinctive formation of RCl. In addition, the Coulombic repulsion between the TiO2 surface and chloride ions[90] is lowered upon polarization of the TiO2 surface.[50] This facilitates the adsorption of Cl– on the positively charged TiO2 surface allowing for chloride oxidation (RCS formation) and subsequent formation of RCl.

Combining all information, it is evident that although not all of the organic carbon could be removed during PCD, the removal of organic carbon was more effective in the purely light-driven approach compared to ECD (under the conditions applied in this study). Especially, the significantly smaller formation of RCS and RCl intermediates during PCD is of importance. During ECD in saline media 4EP was mainly removed by conversion into (polychlorinated) compounds and only small amounts of 4EP could be completely mineralized. In contrast, during PCD only for high NaCl concentrations minor formation of monochlorinated 2C4EP could be detected. The measured 2C4EP concentration during ECD at 0.6 mol·L–1 was nearly 2 orders of magnitude higher than that detected after 5 h of PCD. The difference in the AOP techniques is likely governed by a surface-charge recombination mechanism in PCD, which suppresses FC and RCl formation. Overall, the formation of chlorinated intermediates in both AOP technologies is important for consideration of practical wastewater treatment applications.