A Bifunctional Electroactive Ti₄O₇-Based Membrane System for Highly Efficient Ammonia Decontamination

Abstract

Herein, an electroactive filtration system, consisting of a Ti₄O₇ anode and a Pd-Cu co-modified nickel foam cathode, was developed and applied for the decontamination of ammonia from water. When assisted with an external electrical field, ClO• was generated on the surface of the Ti₄O₇ anode, which then reacted selectively with ammonia to generate N₂. The anodic byproduct, NO₃⁻, could also be reduced efficiently at the functional cathode to produce N₂ as well. Electron paramagnetic resonance technique and radical scavenging tests synergistically verified the essential role of ClO• during the highly efficient ammonia conversion process. Relative to conventional batch systems, the developed flow-through design demonstrated enhanced ammonia conversion kinetics, thanks to the convection-enhanced mass transport. The proposed technology also showed desirable stability across a wide environmental matrix. This work provides new insights for the development of advanced and affordable continuous-flow systems towards effective decontamination of ammonia.

1. Introduction

Excessive ammonia in the water environment usually causes eutrophication of water bodies and poses toxic effect to aquatic organisms [1,2]. Therefore, reducing the ammonia emission has been one of the primary objectives of current wastewater treatment processes. Among these available remediation processes, electrochemical oxidation was considered as a promising approach, due to its environmental friendliness, high efficiency, and easy controllability. The electrochemical processes have demonstrated enhanced oxidation power for the degradation of various toxic and refractory organic contaminants. For example, a Bi-doped SnO₂/Ti₄O₇ composite electrode was proven to be effective for the complete mineralization of atrazine and clothianidin within a few seconds at an applied potential of 3.5 V vs. SHE [3]. In an electrochemical system, both the electrode materials and reactor configuration play essential roles in determining the overall system performance (Table S1).

In terms of the electrode materials, several advanced types have been reported, such as boron-doped diamond (BDD), doped SnO₂, and carbon nanotubes (CNT) [4]. The BDD electrode possesses a wide potential window and high anodic stability, but suffers from high manufacturing cost [5]. The doped SnO₂ electrode demonstrated good performance and acceptable cost, but with limited service life [6]. While the CNT electrode possesses superior conductivity and a high specific surface area, the potential release of nanoscale CNT powders may cause secondary pollution [7]. Alternatively, sub-stoichiometric titanium oxide materials have attracted much recent interest from the scientific community for electrocatalytic applications due to their high conductivity, low cost, strong corrosion resistance, and excellent stability [8]. Among the family of sub-stoichiometric titanium oxide materials, Ti₄O₇ exhibits the highest conductivity, up to 1500 S cm⁻¹, and excellent resistance to harsh chemical conditions [3,9].

Regarding the reactor design, traditional electrochemical processes usually adopt flat plate electrodes and operate in flow-by mode. However, such a flow configuration usually leads to a large hydrodynamic diffuse boundary layer of > 100 μm [10]. Extensive experimental studies and modeling data suggest that an oxidation reaction only happens within a narrow zone of < 1 μm at the electrode/solution interface [11]. To address this issue, an effective solution, combining electrochemical oxidation with membrane filtration, was proposed, which can be operated in an alternative flow-through configuration. Convection flow perpendicular to the "active" membrane improves the mass transport of reactants towards the active sites by restricting the diffusive distance to the pore radius of the membrane electrode. For example, Zaky and Chaplin previously reported that a Ti₄O₇-based electroactive membrane showed a 10-fold increase in mass transport rate when compared with that obtained in traditional flow-by mode [8]. Meanwhile, such flow-through design is attractive for practical engineering applications, due to its simple controllability and high operability. However, there are only very limited reports available on the decontamination of ammonia from water bodies using a flow-through system. It can be envisaged that by adopting a flow-through design with high-performance electrode materials, rapid and effective ammonia nitrogen removal would be achieved.

We have previously developed a flow-by Ti₄O₇–based electrochemical system that enables ammonia conversion under an electric field [12]. However, the poor mass transport in that design significantly limits the potential for large-scale application of this promising technology. Herein, we further developed a flow-through system using the Ti₄O₇ electrode as a functional anode, as well as a membrane material. When assisted by an external electric field, we hypothesized that the flow-through Ti₄O₇ electrode enables consistent production of highly reactive OH•, which further reacts with free chlorine to generate ClO•; these ClO• can react quickly with electron-rich groups to selectively convert ammonia to N₂ [13]. Meanwhile, a Pd-Cu co-modified nickel foam (Pd-Cu/NF) cathode enables the reduction of anodic byproduct—nitrate—to N₂ again. At both electrodes, the dominant product is N₂. The influence of several key operating factors (e.g., anode potential, Cl⁻ concentration, and solution pH) on the ammonia decontamination performance were systematically investigated. The electron paramagnetic resonance (EPR) technique and scavenger experiments both confirmed the essential role of ClO• within the proposed system. This work is dedicated to providing an effective and affordable design for the decontamination of aqueous pollutants, based on the promising sub-stoichiometric titanium oxide materials.

2. Results and Discussion

2.1. Characterizations

The morphology of Ti₄O₇ electrodes was characterized by field emission scanning electron microscopy (FESEM). Figure S1a reveals that the macroscale Ti₄O₇ particles were interconnected and formed macroporous structure. This led to a BET surface area of 2.8 ± 0.4 m² g⁻¹ for the Ti₄O₇ anode [14]. X-ray diffraction (XRD) patterns of the electrode materials, in accordance with that of standard spectra of Ti₄O₇, with characteristic peaks centered at 20.8°, 26.2°, 29.5°, 31.6°, 34.0°, 36.2°, 40.5°, 53.1°, 55.0°, 63.7°, and 66.3° (indicated by red points) (Figure S1b). Other Magnéli phases were not observed in the XRD pattern, indicating a high-purity Magnéli phase Ti₄O₇ electrode has been successfully prepared during the high-temperature fabrication process. The presence of Magnéli phase Ti₄O₇ as the main component ensures an excellent electrical conductivity because Ti₄O₇ exhibits excellent conductivity, compared with other sub-stoichiometric titanium oxides [15]. The X-ray photoelectron spectroscopy (XPS) survey pattern of the Ti₄O₇ electrode confirms the presence of C 1s, O 1s, and Ti 2p with specific ratio of 21.91, 54.92, and 23.17%, respectively. The Ti 2p characteristic peaks centered at 458.7 eV (2p_3/2) and 464.3 eV (2p_1/2) can be assigned to Ti⁴⁺, while peaks located at 456.2 eV (2p_3/2) and 461.9 eV (2p_1/2) can be assigned to Ti³⁺ (Figure S1d) [16,17]. As displayed in Figure S2, the C 1s XPS spectra can be divided into three peaks centered at 284.5 eV, 285.9 eV, and 288.3 eV, corresponding to C=C, C–O, and C=O, respectively. While the high resolution O 1s spectrum were deconvoluted into two peaks centered at 529.7 eV and 531.1 eV, respectively, corresponding to Ti–O bond and –OH group [18].

Compared to original gray NF, the color of the Pd-Cu/NF changed to black after the loading of Cu and Pd (Figure S3), although the macropore size and surface morphology did not change significantly (Figure S4) [6]. Such macroporous dimensions of the Pd-Cu/NF cathode only generated a BET surface area of 1.3 ± 0.1 m² g⁻¹. The presence of Pd and Cu in NF can also be verified by the XPS technique, as well. The bonding energy of the spin-orbit coupling of metal Pd⁰ can be observed at 340.5 eV and 335.2 eV, respectively. Another pair of Pd²⁺ signals were observed at 342.2 eV and 337.0 eV (Figure S4c). The successful deposition of Cu onto the NF can be verified by the Cu 2p XPS spectra. As displayed in Figure S4d, the signals located at 931.8 eV and 953.0 eV were assignable to the emissions from Cu 2p 3/2 and Cu2p 1/2 levels, which corresponded to the typical binding energy of Cu⁰ [19]. Other binding energies present at 934.4 eV for Cu 2p_3/2 and 954.0 eV for Cu 2p_1/2 were ascribed to Cu²⁺ [6,20]. Besides these peaks, the presence of shake-up satellite peaks within the region of 940–945 eV (Cu 2p_3/2) and 959–965 eV (Cu 2p_1/2) also indicated the presence of Cu²⁺ species on the cathode surface. A similar phenomenon was also previously documented [21].

2.2. Ammonia Conversion

The catalytic performances of the Ti₄O₇ electrodes were evaluated using ammonia oxidation as a model reaction. As shown in Figure S5, an evident promotion was observed for ammonia conversion in the flow-through system, with effluent ammonia concentration decreased from 30 to 0.7 mg·L⁻¹. However, for batch and single-pass systems, the ammonia concentration decreased to 15.1 mg·L⁻¹ and 26.3 mg·L⁻¹, respectively, under similar conditions. The improved performance in the flow-through mode can be due to the convection-enhanced mass transport. The removal kinetics was the lowest in single-pass filtration mode, because the liquid residence time within the filtration device was rather limited (5 sec), which caused insufficient contact between ammonia molecules and the electrode. Additionally, in the batch mode, the mass transport of ammonia was only dominated by diffusion and, consequently, only the surface active sites contributed to the ammonia conversion. This demonstrates the advantage of a flow-through design compared to the flow-by mode in conventional batch system.

2.3. Impacts of Key Operational Parameters

2.3.1. Effect of Anode Potential

To study the effect of the anode potential on ammonia oxidation, different anode potentials were exerted. As displayed in Figure 1a, when the anode potential increased from 1.9 V to 2.8 V vs. Ag/AgCl, the ammonia conversion efficiency increased from 43.2% to 85.9%. This indicates that a higher potential was beneficial for ammonia conversion. However, further increasing the anode potential to 3.1 V vs. Ag/AgCl contributed negatively to the ammonia oxidation kinetics, due to the occurrence of other side reactions, such as oxygen evolution reaction, which decreased the overall current efficiency [22]. Hence, 2.8 V vs. Ag/AgCl was selected as optimal for subsequent studies.

2.3.2. Effect of [Cl⁻]

As shown in Figure 1b, when the [Cl⁻] concentration increased from 0.08 M to 0.14 M, the ammonia conversion increases sharply from 59.6% to 97.8%. This may be due to the increase of chloride ions producing more radicals (e.g., Cl• and ClO•) and contributing to the radical-mediated ammonia oxidation reaction [22]. However, further increasing [Cl⁻] concentration to 0.16 M caused a slight decrease in the ammonia oxidation efficiency. Other side reactions may occur in the presence of excess [Cl⁻] (Equations (1) and (2)), which consumes as-produced radicals and, hence, reduces the overall ammonia conversion efficiency [22].

$$\mathrm{Cl}•+{\mathrm{Cl}}^{-}\to {\mathrm{Cl}}_2^{•-}$$

(1)

$${\mathrm{Cl}}_2^{•-}+{\mathrm{H}}_2\mathrm{O}\to \mathrm{HOCl}+{\mathrm{Cl}}^{-}+{\mathrm{H}}^{+}+{\mathrm{e}}^{-}$$

(2)

2.3.3. Effect of Solution pH

The effect of initial solution pH on the ammonia removal was studied and the results are displayed in Figure 1c. The removal efficiency of ammonia after 3 h was 73.2% at pH 3 and 81.9% at pH 5. This may be attributed to the poor stability of hypochlorous acid under acidic conditions [23]. However, the conversion efficiency of ammonia significantly improved under alkaline conditions. At pH 9 and 11, the ammonia conversion rate after 1.5 h was 97.4% and 100%, respectively. Ammonia in the aqueous solution may be present in the non-ionized form (NH₃) or the ionized form (NH₄⁺) (Equation (3)) [24]. In these two forms, the non-ionized ammonia is more susceptible to oxidation [25]. According to Candido et al., NH₃ is the dominant species in alkaline media, with a pH greater than 10. When pH < 7, NH₄⁺ is a dominant species. In the medium pH, the two substances are in equilibrium [26]. Besides, alkaline conditions also promoted the cycling of Cl⁻ and enhanced indirect oxidation [23]. Considering the ammonia loss caused by stripping ammonia in the solution under the pH of 11 [27], we chose pH 9 for subsequent investigations.

$${\mathrm{Cl}}_2^{•-}+{\mathrm{H}}_2\mathrm{O}\to \mathrm{HOCl}+{\mathrm{Cl}}^{-}+{\mathrm{H}}^{+}+{\mathrm{e}}^{-}$$

(3)

2.3.4. Effect of Flow Rate

The effect of flow rate on the ammonia oxidation is available in Figure S6. When the flow rate increased from 0.5 to 1.5 mL·min⁻¹, the ammonia conversion efficiency increased from 73.7% to 97.4%. However, further increasing flow rate to 2.0 mL·min⁻¹ caused a slight decrease in ammonia oxidation performance. This can be explained by the fact that a higher flow rate may shorten the residence time within the filtration system and cause incomplete ammonia reaction.

2.4. Ammonia Decontamination Mechanism

EPR and radical quenching experiments were carried out to identify the main active substances for ammonia conversion. As shown in Figure 2a, the signal of DMPO–ClO• was detected in the presence of [Cl⁻] [28,29]. A four-line characteristic DMPO−OH• signal with 1:2:2:1 intensity was observed, and other faint peaks may correspond to Cl• [30]. Similar results were previously reported by Zhang et al. [29]. This confirmed the generation of ClO•, OH• and Cl• within the present electrochemical flow-through system. Furthermore, radical quenching tests were performed using nitrobenzene (NB), tert-butanol (tBuOH), and HCO₃⁻ as scavengers (Figure 2b). NB reacts rapidly with OH•, while tBuOH has high affinity to HO•, Cl•, and ClO• [22,31]. In addition, HCO₃⁻ can be served as a scavenger for HO• and Cl•, while its reaction with ClO• is negligible [32]. In comparison with the control experiments, free of any scavengers, tBuOH showed complete inhibition to the conversion of ammonia, indicating the importance of radicals in ammonia conversion. In contrast, the efficiency of ammonia conversion was only reduced by 15% after the introduction of NB, implying that OH• was not the major radical. We also found that the addition of NaHCO₃ did not completely inhibit the conversion of ammonia, suggesting that ClO• was the main oxidant. In addition, we conducted another control experiment using Na₂SO₄ instead of NaCl. As expected, only the DMPO−OH• signal (Figure 2a) and negligible ammonia conversion (< 8%) were observed (Figure S7), further verifying the essential role of [Cl⁻].

Based on the above discussion, a plausible mechanism of ammonia conversion was proposed and illustrated in Figure 3. On the surface of the Ti₄O₇ anode, water and chloride ions were effectively oxidized under high potential (e.g., 2.8 V vs. Ag/AgCl) to generate OH• and free chlorine (e.g., HClO or ClO⁻), respectively. The ClO• can be produced by the instantaneous adduct of OH• with free chlorine (Equations (4) and (5)) [33,34]. In particular, the ClO• can react selectively with NH₄⁺ (k_ClO• = 3.1 × 10⁹ M⁻¹ s⁻¹), compared with that of OH• (k_OH• = 8.9 × 10⁷ M⁻¹ s⁻¹) and Cl• (k_Cl• = 1.1 × 10⁹ M⁻¹ s⁻¹) [29].

$$\mathrm{OH}•+\mathrm{HClO}\to \mathrm{ClO}•+{\mathrm{H}}_2\mathrm{O}$$

(4)

$$\mathrm{OH}•+{\mathrm{OCl}}^{-}\to \mathrm{ClO}•+{\mathrm{OH}}^{-}$$

(5)

These ClO• radicals would extract H from the NH₄⁺ molecules to form the N-centered radicals (e.g., •NHCl) [34], while the N-centered radicals would be unstable [29]. Once formed, these •NHCl would undergo self-decay to produce N₂ (Equations (6)–(8)) [35]. In addition, NO₃⁻ may be produced during this process, resulting in incomplete denitrification (Equations (9) and (10)).

$${\mathrm{NH}}_4^{+}+\mathrm{ClO}•\to {\mathrm{NH}}_2\mathrm{Cl}+\mathrm{OH}•+{\mathrm{H}}^{+}$$

(6)

$${\mathrm{NH}}_2\mathrm{Cl}+\mathrm{ClO}•\to •\mathrm{NHCl}+\mathrm{HClO}$$

(7)

$$•\mathrm{NHCl}\stackrel{•\mathrm{NHCl}}{\to }{\mathrm{N}}_2+2{\mathrm{H}}^{+}+2{\mathrm{Cl}}^{-}$$

(8)

$${\mathrm{NH}}_4^{+}+8\mathrm{Cl}•+3{\mathrm{H}}_2\mathrm{O}\to {\mathrm{NO}}_3^{-}+8{\mathrm{Cl}}^{-}+10{\mathrm{H}}^{+}$$

(9)

$${\mathrm{NH}}_4^{+}+4\mathrm{HClO}\to {\mathrm{NO}}_3^{-}+{\mathrm{H}}_2\mathrm{O}+6{\mathrm{H}}^{+}+4{\mathrm{Cl}}^{-}$$

(10)

2.5. Nitrate Reduction on the Pd-Cu/NF Cathode

NO₃⁻ was difficult to attach onto the cathode surface, due to the like-charge repulsion by the negatively-charged cathode. Moreover, these NO₃⁻ can be converted back to NH₄⁺ as well. To alleviate this issue, a Pd-Cu/NF, with three-dimensional network and excellent electrical conductivity, was used as the cathode material [2,36,37]. It is of note that the coating of Pd and Cu is indispensable. These adsorbed NO₃⁻ anions were reduced to NO₂⁻ at the Cu sites and further reduced to N₂ at the Pd sites of the cathode [30]. Results suggest that a similar ammonia removal kinetics was observed using a Cu-NF or a Pd-NF cathode, and the NO₃⁻ generated was 15.9 and 17.1 mg·L⁻¹, respectively. However, the integration of Pd-Cu sites evidently boosted the kinetics of NO₃⁻ reduction to N₂ and inhibited the conversion of NH₄⁺ to NO₃⁻, resulting in only 5.1 mg·L⁻¹ NO₃⁻, while the presence of Pd or Cu-alone only led to limited N₂ yield (Figure 3). For the NF cathode, the ammonia conversion rate was rather slower, while the NO₃⁻ reduction rate is similar to the Pd/NF or the Cu/NF cathode.

2.6. Stability Evaluation of the Ti₄O₇ Electrode

The stability of the anode is important for practical applications. Hence, the stability of the Ti₄O₇ electrode was evaluated in consecutive cycles for ammonia conversion under the conditions of potential 2.8 V vs. Ag/AgCl, [Cl⁻] of 0.14 M, and pH 9. Figure 4a shows that only marginal declines in ammonia removal efficiency were experienced in the three cycles, indicating the outstanding stability of the Ti₄O₇ electrode in catalytic oxidation. In addition, the ammonia conversion was further examined under actual water matrixes. To do this, the 30 mg·L⁻¹ ammonia-spiked lake water and the tap water were challenged with the filtration system under optimal conditions. Results suggest that a similar ammonia oxidation trend was obtained in tap water samples when compared to that in DI water conditions (Figure 4b). Moreover, a > 91.7% ammonia conversion was achieved when challenging with the ammonia-spiked lake water (TOC = 96.3 mg·L⁻¹), which was considered as a more complex matrix compared to tap or DI water. Moreover, the energy consumption of electrochemical filtration was calculated at a typical total cell potential of 3.6 V (corresponds to a potential of 2.8 V vs. Ag/AgCl) to be 2.31 kWh·kg⁻¹ COD for the proposed technology. These values are comparable to or even lower than current state-of-the-art electrochemical oxidation processes, with a power consumption of 5–100 kWh·kg⁻¹ COD [5]. In addition, the energy per volume is only 0.4 kWh·m⁻³, comparable to the most advanced electrochemical processes with energy consumption in the range of 0.1–40 kWh·m⁻³ [38]. This indicates that the electroactive filtration system can be served as a promising unit for ammonia decontamination.

3. Materials and Methods

3.1. Chemicals and Materials

The Ti₄O₇ powder was provided by Kelai New Material Co., Ltd. (Hunan, China). Nickel foam (NF, 0.5 mm thickness) was supplied by Lizhiyuan Electronics Co., Ltd. (Shanxi, China). Sulfuric acid (H₂SO₄, 96%–98%), acetone (≥ 99.5%), hydrochloric acid (HCl, 36%–38%), sodium hydroxide (NaOH, ≥ 96%), copper(II) sulfate pentahydrate (CuSO₄·5H₂O, ≥ 99%), palladium (II) chloride (PdCl₂, ≥ 98%), ammonium sulfate ((NH₄)₂SO₄, ≥ 99%), sodium chloride (NaCl, ≥ 99.5%), and sodium sulfate (Na₂SO₄, ≥ 99%) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China).

3.2. Preparation of Electrodes

The Ti₄O₇ anodes were synthesized by mixing 1 g Ti₄O₇ powder with 2–3 wt % paraffin oil as binder. The powder and binder were mixed for 10 min. The mixture was transferred in module (D = 1.2 cm) and a pressure of ~ 50 kg was applied by a hydraulic tablet machine (Lixin experimental instrument, Shenzhen, China). The sheets were sintered in a tube furnace at N₂ atmosphere at 900 °C for 6 h. The fabrication of Pd-Cu/NF, Pd/NF, and Cu/NF electrodes followed a previously reported electrodeposition method [39].

3.3. Electrochemical Experiments

The ammonia degradation experiment was carried out in a customized flow-through electrochemical apparatus (Figure S8), including a Ti₄O₇ working electrode, a Pd-Cu/NF counter electrode, and a saturated Ag/AgCl reference electrode. In a typical experiment, 30 mL of 30 mg·L⁻¹ (NH₄)₂SO₄ solution and certain NaCl electrolyte were passed through the filtration casing and then returned. The flow rate was tuned by a Cole-Parmer ISM833C peristaltic pump (Vernon Hills, IL, USA). The applied voltage was provided by a Chenhua CHI 760E electrochemical workstation (Shanghai, China). Solution pH was adjusted by 0.1 M H₂SO₄ or NaOH. Ammonia concentration in the influent and effluent was measured by the standard Nessler reagent method. NO₃⁻ and NO₂⁻ concentration was determined by a Thermo Fisher DionexTM ICS-5000 ion chromatography (IC, Waltham, MA, USA).

3.4. Analytical Methods

A Hitachi S-4800 field emission scanning electron microscope (FESEM, Tokyo, Japan) and a Rigaku D/max-2550/PC X-ray diffraction (XRD, Austin, TX, USA) was applied to characterize surface morphologies and crystal phases of the electrode materials, respectively. X-ray photoelectron spectroscopy (XPS, Model No.: Thermo Fisher Scientific Escalab 250Xi, Waltham, MA, USA) was performed under high vacuum. Total organic carbon (TOC) was measured by a multi N/C 2100 TOC Analyzer (Analytik Jena AG, City, Germany). A MS5000 electron paramagnetic resonance (EPR) spectrometer (Freiberg Instruments Inc., Freiberg, Germany) was employed to examine produced reactive radicals. The Brunauer Emmett Teller (BET) surface area of the electrode materials was performed using an Anton Paar Autosorb iQ-c analyzer (Boynton Beach, FL, USA).

4. Conclusions

This work presents an effective design of a continuous-flow electrochemical system towards selective and efficient decontamination of ammonia from water. The proposed technology consists of a Ti₄O₇-based anode and a Pd-Cu/NF cathode; both are indispensable. Upon the application of an external electric field, the system enables a sustainable production of ClO•, which then reacts with ammonia selectively to produce N₂. Any anodic byproducts were further reduced to N₂ again at the cathode. Various advanced characterizations were employed to confirm the effectiveness of the system. This study extends the application of electrocatalytic technology and provides an affordable and feasible strategy for water remediation.