English

• Nitrate pollution in surface water and underground aquifer has been increasingly serious as the massive production of nitrate wastewater in chemical fertilizers, metallurgy and explosives industries [1,2]. The increased exposure of nitrate (NO₃⁻) in aquatic ecosystems threat the human health and ecological balances [3-5]. Several technologies including ion exchange, biological denitrification, membrane filtration and electrodialysis have been developed to remove NO₃⁻ [6-9]. Electrocatalytic nitrate reduction (NO₃RR) to ammonia (NH₃) driven by renewable electricity resources represents a promising alternative, owing to green feature, mild reaction conditions and high reactivity [10-12]. More intriguingly, the produced NH₃-N could be recovered when coupled the electrochemical system with a membrane technology.

  The catalyst is the core of the NO₃RR technology, which determines the kinetics, selectivity and energy consumption of the NO₃⁻ conversion to NH₃. TiO₂, as a typical transition metal oxide, is considered as a promising NO₃RR electrocatalyst owing to its abundance, high chemical stability and superior selectivity toward NH₃ in product [13-15]. Zhang et al. [16] and Niu et al. [17] reported the successful synthesis of TiO₂ applies to NO₃RR, and the NH₃ selectivity reaching to 87.1% and 81.9%, respectively. They further demonstrated that the O_V site on the TiO₂ surface was the reactive center for NO₃RR, and the exposed Ti³⁺ at the O_V site had a large affinity to both the NO₃⁻ and N-intermediates, which facilitated the proton-coupled electron transfer to reactants at electrode surface, and guaranteed the selective formation of NH₃-N [18]. Albeit the above advancements, there is still space for improvement in the NO₃RR performance of TiO₂ due to the following two aspects. Firstly, the current NO₃RR kinetics suffers from an inferior mass transfer as the NO₃⁻ in the wastewater generally has a low concentration [19,20]. Furthermore, the NO₃⁻ with a negative charge bear a repulsive force from cathode owing to the electronic field, which makes it more difficult to approach the cathode [21-23]. On the other hand, the TiO₂, as a semiconductor, is poor in conductivity, which is actually unfavorable for the electrochemical system which urgently requires fast electron transfer [24-26,15].

  Herein, we reported a facile approach to the surface O_V-enriched and urchin-like TiO₂ microparticles (U-TiO₂), which were used for NO₃RR. The NO₃RR performance, including NO₃⁻-N removal efficiency and kinetics, mass/specific activity, product distribution, durability, faradaic current efficiency and reaction pathway on U-TiO₂ were then evaluated under a combined electrochemical analyses and in-situ spectrometric study. To identify the reactive center and the contribution of NO₃⁻-N mass diffusion, O_V content, the NO₃RR performance of U-TiO₂ was also compared to a solid amorphous TiO₂ (A-TiO₂) that had close size but more surface O_V, and a solid sphere. Finally, the impacts of NO₃⁻ concentration, coexisting anions, and dissolved organic organisms in nitrate-laden water on the NO₃RR performance were also performed to assess its application in nitrogen remediation technique.

  Fig. 1a illustrated the synthetic route to urchin-like TiO₂ catalyst (U-TiO₂). In this route, amorphous TiO₂ particles (A-TiO₂) were firstly prepared as precursors via the hydrolysis of TTIP in an NH₃ aqueous solution. They further evolved into U-TiO₂ when subjected to a hydrothermal treatment at 110 ℃ in the presence of F⁻ and PVP, which served as the etching and protecting agents, respectively [27]. In compared to the electrodeposition and electroshock methods, the developed method was more readily to succeed in controlling the morphology and structure of the TiO₂ particles [28,29]. The SEM images in Figs. S1 and S2 (Supporting information) and the TEM images in Figs. 1b and c revealed that the A-TiO₂ particle was a microsphere with smooth surfaces and a mean size of 600 nm, whereas the U-TiO₂ particles displayed unique urchin-like structures with particle size growing to approximately 630 nm. The HRTEM image in Fig. 1d revealed that the U-TiO₂ particle was composed of small crystalline domains with distinct lattice spacing of 0.352 nm, corresponding to the (101) planes of anatase phase [30,31]. The amorphous-anatase conversion during the hydrothermal process was also confirmed by the XRD patterns in Fig. 1h, where the characteristic diffraction peaks of anatase phase were discerned on U-TiO₂ (JCPDS No. 21–1272). Because the control test led to a relatively poor crystallinity in TiO₂ (denoted as H-TiO₂, TEM image shown in Fig. S3 in Supporting information), the substantial improvement in crystallinity from A-TiO₂ to U-TiO₂ would be also attributed to the presence of F⁻ and PVP [27]. The EDS elemental mapping results in Figs. 1e-g demonstrated the uniform distribution of Ti and O on U-TiO₂ and the absence of ionic impurities, such as F⁻. The N₂ adsorption-desorption isotherm curves in Fig. S4 (Supporting information) revealed the mesoporous structures of the three TiO₂ samples. U-TiO₂ features a type-IV isotherm plot with a H3 hysteresis loop, while the A-TiO₂ and H-TiO₂ displayed type-I isotherm plots. On basis of the plot, the Brunauer–Emmett–Teller (BET) surface and pore size in these TiO₂ samples were estimated. Interestingly, the BET surface area of them followed a decreasing order of H-TiO₂ (586.3 m²/g) ≈ A-TiO₂ (489.9 m²/g) > U-TiO₂ (257.7 m²/g), whereas their mean pore size exhibited a reverse order of U-TiO₂ (5.7 nm) > A-TiO₂ (2.2 nm) ≈ H-TiO₂ (2.5 nm). The much smaller BET surface area for urchin-like structure could be attributed to the enlarged pore size [31]. Albeit the decreased physical surface area, the large pore size, urchin-like structure would benefit for the exposure of active sites, the mass diffusion of dilute reactants and the electron transfer. Indeed, the ECSA (Fig. S5 in Supporting information) of U-TiO₂ reached 92.1 m²/g, much higher than that of H-TiO₂ (74.2 m²/g) and A-TiO₂ (53.6 m²/g). The presence of oxygen vacancy (O_V) in A-TiO₂, H-TiO₂ and U-TiO₂ was evidenced by the signal at g = 2.004 in ESR spectra and the O 1s XPS peak at 531.4 eV (Figs. 1i and j) [32-34]. A-TiO₂ should have a larger number of O_V due to the displayed stronger ESR signal and the stronger XPS peak for O_V. It was suggested that the hydrothermal process reduced the number of O_V, possibly due to the improved crystallinity.

  Figure 1

  

  Figure 1.  (a) Schematic illustration for the synthesis of U-TiO₂, TEM images of (b) A-TiO₂ and (c) U-TiO₂, (d) HR-TEM image of U-TiO₂, (e-g) EDS elemental mapping images of U-TiO₂, (h) XRD patterns, (i) O 1s XPS spectra and (j) ESR spectra of as-synthesized A-TiO₂, H-TiO₂ and U-TiO₂.

  DownLoad: Full-Size Img PowerPoint

  To estimate the NO₃RR activity of catalysts, LSV tests in the electrolyte solution with and without NO₃⁻-N were conducted for A-TiO₂, H-TiO₂ and U-TiO_2, respectively. The dash LSV curves in Fig. 2a, obtained in the NO₃⁻-N-free solution, demonstrated a decreasing order of U-TiO₂ > A-TiO₂ > H-TiO₂ in the proton transfer kinetics at their surfaces according to the position of onset potential as well as the current density under the potential more negative than onset potential. After the addition of NO₃⁻-N, the onset potential for electrochemical reaction (the solid LSV curves) shifted negatively by around 0.30 V in compared to that in NO₃⁻-N-free solution for all the three catalysts. The abnormal phenomena could be ascribed to the fact of the hydrogen adsorption was hindered by the strong adsorption of NO₃⁻ on active sites, and the current response, assigned to NO₃⁻-N reduction, occurred under a more negative polarization potential. Furthermore, both the values of onset potential and current density for NO₃⁻-N reduction, as observed in the solid LSV curves, followed a decreasing order of U-TiO₂ > A-TiO₂ > H-TiO₂, suggesting the superior NO₃RR activity of U-TiO₂. To evaluate the performances of catalysts in removing NO₃⁻-N, potential-control batch NO₃RR tests were conducted in an Ar-saturated 50 mmol/L Na₂SO₄ solution with 22.5 mg/L NO₃⁻-N at −0.60 V vs. RHE (the same below). Fig. 2b displayed the reaction time-courses of C/C₀ (C denoted the NO₃⁻-N concentration) during NO₃RR on A-TiO₂, H-TiO₂ and U-TiO₂ respectively. As observed, all the three catalysts were active to remove the NO₃⁻-N, but obviously, the U-TiO₂ was the best with a NO₃⁻-N removal efficiency of 98.5% in 240 min of reaction, followed by the A-TiO₂ (62.3%) and H-TiO₂ (60.9%). Fig. 2c plotted the reaction time-dependent concentrations of NO₃⁻-N, NO₂⁻-N and NH₃-N as well as their total in the three reaction systems. As observed in the three reaction systems, the amount of NH₃-N kept continuous rising with the consumption of NO₃⁻-N. Intriguingly, the NO₂⁻-N, which was one common intermediate resolved during the conversion of NO₃⁻-N to NH₃-N, was evidenced in a very low yield (≤0.4 mg/L) during NO₃RR. As the NO₂⁻-N was one more toxic species than NO₃⁻-N, and its concentration was restricted under 1.0 mg/L by Standards for drinking water quality (GB5749–2022). The well control in NO₂⁻-N made the NO₃RR on our TiO₂-based catalysts a very safe process for practical environmental remediation of nitrate-contaminated water. On the other hand, the low NO₂⁻-N yield as well as the nearly constant total N content in solution indicated the high product selectivity towards NH₃-N. For example, the NH₃-N selectivity on U-TiO₂ reached 98.1%, which was the highest among the one reported in literatures (Table S1 in Supporting information).

  Figure 2

  

  Figure 2.  Comparisons of (a) LSV in NO₃⁻-N-contained and NO₃⁻-N-free Na₂SO₄ electrolyte, (b) time-course C/C₀, (c) evolution in production distribution, (d) pseudo-first-order rate equation curves, (e) specific and mass activities between different cathodes in NO₃RR, (f) durability tests of U-TiO₂. Reaction conditions: [initial NO₃⁻-N] = 22.5 mg/L; cathodic potential = −0.60 V; electrolysis time = 240 min.

  DownLoad: Full-Size Img PowerPoint

  The kinetics study revealed that the NO₃RR reaction on A-TiO₂, H-TiO₂ and U-TiO₂ all obeyed the pseudo-first-order mode (lnC/C₀ = −k_app t) (Fig. 2d), suggesting that the NO₃RR was primarily limited by the mass diffusion of reactants. NO₃RR on U-TiO₂ was the fastest with a k_app = 9.18E-03 min⁻¹, followed by that on A-TiO₂ (5.75E-03 min⁻¹) and H-TiO₂ (3.82E-03 min⁻¹). On basis of the k_app, the mass (MA) and specific activities (SA) of the three catalysts were calculated. As observed in Fig. 2e, U-TiO₂ displayed the highest MA of 1.15 min⁻¹ g_cat⁻¹, followed by A-TiO₂ (0.72 min⁻¹ g_cat⁻¹) and H-TiO₂ (0.48 min⁻¹ g_cat⁻¹). However, the SA of the catalysts followed a different order of A-TiO₂ (2.56 min⁻¹ m_cat⁻²) > U-TiO₂ (2.37 min⁻¹ m_cat⁻²) > H-TiO₂ (1.23 min⁻¹ m_cat⁻²). Given the lower ECSA of A-TiO₂, the higher SA but lower MA on A-TiO₂ suggested that the A-TiO₂ might have a higher density of active centers at surface than U-TiO₂. The underlying mechanism was discussed below. The faradaic current efficiency (F.E.%) of the three catalysts for NH₃-N production was also calculated. As shown in Fig. S6 (Supporting information), the U-TiO₂ delivered the largest one of 79.0%, followed by A-TiO₂ (59.7%) and H-TiO₂ (51.1%). The lower F.E.% than 100% in the three systems could be ascribed to the side hydrogen evolution reaction [35]. These results demonstrated that the NO₃RR on U-TiO₂ was more energy-efficient, and also safer due to the release of less explosive H₂. As a critical descriptor for efficient catalyst, the durability of U-TiO₂ for NO₃RR was also evaluated through repeated batch NO₃RR tests. The results in Fig. 2f revealed that U-TiO₂ could retain its high performance after at least 5 consecutive cycles of reactions. The TEM image of the used U-TiO₂ in Fig. S7 (Supporting information) showed that the catalyst kept its urchin-like structure as well as the exposure of the (101) facet of anatase. Furthermore, the O 1s XPS spectra in Fig. S8 (Supporting information) evidenced the preservation of O_V on the used U-TiO₂, suggesting that the O_V could be in-situ regenerated under the reductive potential of NO₃RR.

  Previous researches had verified that the O_V sites with exposed Ti³⁺ species were the primary reactive centers of TiO₂-based catalysts for NO₃RR [16]. To probe into the role of O_V in our system, the NO₃RR performances of U-TiO₂ and A-TiO₂ before and after the quench of their surface O_V through immersing the cathodes in 10.0 mmol/L PO₄³⁻ solution prior to reaction were all examined [36]. As presented in Fig. 3a, after being treated in 10.0 mmol/L PO₄³⁻, the O_V/O_L ratio on U-TiO₂ was indeed reduced from initial 0.18 to 0.13, owing to the occupation of O_V by PO₄³⁻ via Ti³⁺-PO₄³⁻ complexation [37]. Correspondingly, the NO₃⁻-N removal efficiency for the O_V-quenched U-TiO₂ dropped by 27.8%, in compared to that of the fresh U-TiO₂ (Fig. 3b). All these clearly verified the critical role of O_V as reactive centers on our TiO₂ catalyst. Fig. 3b also showed that the NO₃⁻-N removal efficiency on the O_V-quenched A-TiO₂ was reduced by a larger extent of 53.2% in compared to that on fresh A-TiO₂. The displayed more significant efficacy of O_V on A-TiO₂ indicated a larger density of O_V at its surface, which was consistent with the XPS results in Fig. 1i. Given the critical role of O_V in NO₃RR, the large density of surface O_V on A-TiO₂ would rationalize why A-TiO₂ had a larger SA than U-TiO₂ (Fig. 2e).

  Figure 3

  

  Figure 3.  (a) O 1s XPS spectra of U-TiO₂ after immersion in 10.0 mmol/L PO₄³⁻ solution. (b) NO₃RR performance of A-TiO₂ and U-TiO₂ before and after submerged in the PO₄³⁻ solution. (c) Nyquist plots of the samples. (d) Schematic illustration of the potential-dependent NO₃RR mechanism. (e, f) DEMS results of the NO₃RR on U-TiO₂.

  DownLoad: Full-Size Img PowerPoint

  Though A-TiO₂ had a larger number of O_V on surface, its MA was still lower than that of U-TiO₂. This suggested that besides the number of active sites (i.e., O_V), the NO₃RR performance on U-TiO₂ depended on other factors, such as the charge transfer at catalyst/water interfaces. The Nyquist plots in Fig. 3c revealed an increasing order of U-TiO₂ < H-TiO₂ < A-TiO₂ in charge transfer resistance at catalyst/water interfaces [38]. Intriguingly, this order was exactly opposite to that of their MA, indicating that the NO₃RR performance was primarily associated with the charge transfer rate at catalyst surface. Generally, the semiconductor with more defects or in amorphous structure usually displayed higher electronic conductivity. The U-TiO₂, with high crystallinity and less defects, would have inferior electronic conductivity to A-TiO₂. Accordingly, the low charge resistance at U-TiO₂/water interfaces should benefit from the urchin-like hierarchical structure, which provided abundant channels/voids for mass diffusion and also exposed more reactive centers for charge transfer [27,39,40]. All these demonstrated that the superior NO₃RR performance on U-TiO₂ originated from the reproducible O_V site (reactive center) as well as the urchin-like structure (Fig. 3d). Notably, the urchin-like structure took the leading contribution in our system.

  To get more knowledge about the NO₃RR on U-TiO₂, the NO₃RR pathway was explored with in-situ DEMS test, which could resolve the reaction intermediates during the conversion of NO₃⁻-N to NH₃-N, even when the intermediates were of a short lifetime. As demonstrated in Figs. 3e and f, the m/z signals for 2, 14, 15, 16, 17, and 30 were discerned under a working potential of −0.60 V, which could be assigned to H₂, N, NH, NH₂, NH₃, and NO species, respectively. No m/z signals for N₂ and NO₂ were captured, indicating the inferior selectivity of NO₃⁻ conversion to N₂ on U-TiO₂. This was consistent with our experimental results that little N₂ was produced. On basis of the resolved species and the detected product in aqueous solution, we could propose a NO₃RR pathway of ^*NO₃⁻ → ^*NO₂⁻ → ^*NO → ^*N → ^*NH → ^*NH₂ → ^*NH₃ on U-TiO₂ surface (^* denotes the reactive sites on catalyst). Such a pathway of NO₃⁻-NH₃ conversion was also reported on many O_V-triggered NO₃RR systems, which, therefore, solidified that the O_V on TiO₂ was the primary reaction centers.

  To assess the application of U-TiO₂-driven NO₃RR on different scenarios, the NO₃RR performances of U-TiO₂ were investigated under different NO₃⁻-N feeding concentrations. The results in Fig. 4a showed that with the feeding concentrations increasing from 22.5 mg/L to 90.0 mg/L, more NO₃⁻-N are remained after a reaction of 240 min. Corresponding NO₃⁻-N removal efficiency ((1-C/C₀) × 100%) was found to decrease from initial 98.5% to 46.1%. However, as observed in Fig. 4b, the NH₃ yield at the end of reaction increased with the feeding NO₃⁻-N concentration, indicating that more NO₃⁻-N was converted to NH₃-N when fed a higher concentration. These results supported the conclusion that the NO₃RR kinetics on U-TiO₂ under a NO₃⁻-N concentration range of 22.5~90.0 mg/L was limited by the mass diffusion of NO₃⁻. Generally, a higher NO₃⁻-N concentration would promote the collision frequency between NO₃⁻-N and catalyst, contributing to an accelerated NO₃RR [6]. Additionally, Fig. 4b also revealed that more NO₂⁻-N was produced with the increase in feeding NO₃⁻-N concentration. This might be ascribed to the fact that the NO₃⁻ would compete the reactive centers with NO₂⁻, some of which was squeezed off the reactive centers before being hydrogenated to NH₃. Given the large toxicity of NO₂⁻, this result reminded us to carefully optimize the reaction conditions when dealing with the wastewater with concentrated NO₃⁻-N. Fortunately according to our previous work, prolonging the reaction time could be an efficient way to control the residue of NO₂⁻ in solution [41]. Fig. 4c also compared the F.E.% for NH₃-N production on U-TiO₂ under different NO₃⁻-N concentrations. It was revealed that the F.E.% depended less on the feeding NO₃⁻-N concentration. This was unexpected as an enhanced NO₃⁻-NH₃ conversion was usually accompanied with a larger F.E.% [42]. The underlying mechanism will be explored in the future work.

  Figure 4

  

  Figure 4.  Effects of (a-c) initial NO₃⁻-N concentration on NO₃⁻-N removal, evolution in production distribution, MA and F.E.%. (d-f) The presence of coexisting ions on NO₃⁻-N removal, pseudo-first-order rate equation curves for NO₃RR, and evolution in production distribution over U-TiO₂. Reaction conditions: [initial NO₃⁻-N] = 22.5 mg/L for (d-f); cathodic potentiall = −0.60 V; electrolysis time = 240 min for all the tests.

  DownLoad: Full-Size Img PowerPoint

  As most of the industrial processes employed natural water as water resources, the resultant wastewater usually contained some impurity ions (e.g., Cl⁻, CO₃²⁻, Mg²⁺ and Ca²⁺) and natural organic materials (NOM), which might pose some effects on the electrochemical process. Herein their effects on the NO₃RR performance of U-TiO₂ were investigated under a potential of −0.60 V. Humic acid was introduced as the representative of NOM. The concentrations of the impurities were set according to the values of them in natural water, such as Cl⁻ (5.0 mmol/L), CO₃²⁻ (1.0 mmol/L), Mg²⁺ (1.0 mmol/L) and Ca²⁺ (2.0 mmol/L), and humic acid (3.3 mg/L) [43]. Fig. 4d revealed that the presence of Ca²⁺, Mg²⁺ and Cl⁻ promoted NO₃RR, due to the faster drop in C/C₀ in compared to that in the control experiment. On basis of the C/C₀ value at 150 min, we could tell that the Ca²⁺ was the most efficient species in promoting NO₃RR, followed by Mg²⁺ and Cl⁻. In contrast, the inclusion of CO₃^2– and humic acid slowed down the NO₃RR. The detrimental effects of CO₃^2– was more significantly as the C/C₀ at the end of reaction remained 75.4%. Fig. S9 (Supporting information) showed that the NO₃RR on U-TiO₂ still obeyed the pseudo-first-order reaction mode, indicating that the mass diffusion of NO₃⁻ continued to be the rate-determined step for NO₃RR even when the NO₃RR was boosted in a significant extent in the presence of Ca²⁺. Fig. 4e summarized the k_app, which followed a decreasing order of Ca²⁺ > Mg²⁺ > Cl⁻ > Control > humic acid > CO₃^2–. Fig. 4f compared the products of NO₃RR at 150 min in the presence of impurity ions and humic acid. As observed, the inclusion of Ca²⁺, Mg²⁺ and Cl⁻ produced more NH₃-N, while the presence of humic acid and CO₃^2– reduced yield of NH₃-N, both of which was consistent with the displayed efficacy of them on NO₃RR kinetics. All these suggested that the impurity species affects primarily on the kinetics of NO₃⁻-N conversion to NH₃-N rather than the reaction pathway. Fig. 4f also showed that the NO₂⁻-N yields in all the investigated systems were kept under a low level, even when the NO₃RR was hindered under a large extent in the presence of CO₃²⁻. The underlying mechanism was unclear, but this unique feature was appealing as our U-TiO₂ had a strong resistant to water quality in the aspect of the well control in NO₂⁻-N yield.

  The detrimental effects of humic acid and CO₃^2– on O_V-driven catalysis had been documented in literature, which could attributed to the strong complexation effect between the carboxyl group and the exposed cation at the O_V site (i.e., the Ti³⁺), leading to the poisoning of the reactive centers [44,41]. The promoting effects of Ca²⁺ and Mg²⁺ and Cl⁻ were unexpected. In general, the NO₃⁻ with a negative charge suffered a repulsive force from cathode owing to the electronic field, which raised extra energy for them to approach electrode for NO₃RR. We speculated that the presence of Ca²⁺ and Mg²⁺ might alter the structure of double electric layer at electrode/solution interface by forming an instantaneous neutral ion pair, which, as a result, weakened the repulsive force between NO₃⁻ and the cathode [45]. In the future, more efforts will be devoted to unraveling the interactions between the impurity ions and the performance of the catalyst.

  Herein, we developed a defective urchin-like TiO₂ micro-particles catalyst for NO₃RR. When subjected to 22.5 mg/L of NO₃⁻-N, it could afford a mass activity of 1.15 min⁻¹ mg_catalyst⁻¹, a low yield of toxic NO₂⁻-N intermediate (≤0.4 mg/L) and an exceptional high NH₃-N selectivity of 98.1% under the working potential of −0.60 V, outperforming most of the reported catalysts. We also demonstrated that the O_V was the real reactive sites for NO₃RR, but rather than its content, the NO₃RR kinetics were more dependent on the urchin-like structure, which was beneficial for the interfacial electron transfer of TiO₂ and the mass diffusion of NO₃⁻-N around TiO₂. The in-situ DEMS test revealed that the NO₃RR on U-TiO₂ followed a pathway of ^*NO₃⁻ → ^*NO₂⁻ → ^*NO → ^*N→ ^*NH → ^*NH₂ → ^*NH₃. The U-TiO₂ could keep its robust performance under a wide NO₃⁻-N concentration range and in the presence of Ca²⁺, Cl⁻ and Mg²⁺. However, the humic acid and CO₃²⁻ in wastewater posed detrimental effects on NO₃RR.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  The authors acknowledge the financial support by National Natural Science Foundation of China (Nos. 22176019, 51978110), the Science and Technology Research Program of Chongqing Municipal Education Commission (Nos. KJQN201800829, KJQN201900837, KJZD-K202000802, KJQN201901527), Chongqing Research Student Science and Technology Innovation Project (No. CYS22724), Innovation and Entrepreneurship Training Plan for College Students (No. 202111799007).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108410.

  
  
• 1. [1]

     X. Zou, J. Xie, C. Wang et al., Chin. Chem. Lett. 34 (2023) 107908. doi: 10.1016/j.cclet.2022.107908

  2. [2]

     Z.Y. Wu, M. Karamad, X. Yong, et al., Nat. Commun. 12 (2021) 2870. doi: 10.1038/s41467-021-23115-x

  3. [3]

     T. Feng, F. Li, X. Hu, et al., Chin. Chem. Lett. 34 (2023) 107862. doi: 10.1016/j.cclet.2022.107862

  4. [4]

     J. Lim, C.Y. Liu, J. Park, et al., ACS Catal. 11 (2021) 7568–7577. doi: 10.1021/acscatal.1c01413

  5. [5]

     F. Ni, Y. Ma, J. Chen, et al., Chin. Chem. Lett. 32 (2021) 2073–2078. doi: 10.1016/j.cclet.2021.03.042

  6. [6]

     T. Zhu, Q. Chen, P. Liao, et al., Small 16 (2020) 2004526. doi: 10.1002/smll.202004526

  7. [7]

     Y. Wang, W. Zhou, R. Jia, et al., Angew. Chem. Int. Ed. 59 (2020) 5350–5354. doi: 10.1002/anie.201915992

  8. [8]

     C.A. Clark, C.P. Reddy, H. Xu, et al., ACS Catal. 10 (2020) 494–509. doi: 10.1021/acscatal.9b03239

  9. [9]

     S. Xu, Y. Shi, Z. Wen, et al., Appl. Catal. B: Environ. 323 (2023) 122192. doi: 10.1016/j.apcatb.2022.122192

  10. [10]

      D. Chen, S. Zhang, X. Bu, et al., Nano Energy 98 (2022) 107338. doi: 10.1016/j.nanoen.2022.107338

  11. [11]

      J.N. Gao, N. Shi, Y.F. Li, et al., Environ. Sci. Technol. 56 (2022) 11602–11613. doi: 10.1021/acs.est.1c08442

  12. [12]

      J.N. Gao, N. Shi, X.B. Guo, et al., Environ. Sci. Technol. 55 (2021) 10684–10694. doi: 10.1021/acs.est.0c08552

  13. [13]

      X. Zhang, C. Wang, Y. Guo, et al., J. Mater. Chem. A 10 (2022) 6448–6453. doi: 10.1039/d2ta00661h

  14. [14]

      D.E. Kim, D. Pak, Chemosphere 228 (2019) 611–618. doi: 10.1016/j.chemosphere.2019.04.071

  15. [15]

      J. Chen, X. He, D. Zhao, et al., Green Chem. 24 (2022) 7913–7917. doi: 10.1039/d2gc03083g

  16. [16]

      R. Jia, Y. Wang, C. Wang, et al., ACS Catal. 10 (2020) 3533–3540. doi: 10.1021/acscatal.9b05260

  17. [17]

      Z. Wei, X. Niu, H. Yin, et al., Appl. Catal. A: Gen. 636 (2022) 118596. doi: 10.1016/j.apcata.2022.118596

  18. [18]

      H. Hirakawa, M. Hashimoto, Y. Shiraishi, et al., J. Am. Chem. Soc. 139 (2017) 10929–10936. doi: 10.1021/jacs.7b06634

  19. [19]

      W.J. Sun, H.Q. Ji, L.X. Li, et al., Angew. Chem. Int. Ed. 60 (2021) 22933–22939. doi: 10.1002/anie.202109785

  20. [20]

      Z. Song, Y. Liu, Y. Zhong, et al., Adv. Mater. 34 (2022) 2204306. doi: 10.1002/adma.202204306

  21. [21]

      X. Fu, J. Zhang, Y. Kang, Chem. Catal. 2 (2022) 2590–2613. doi: 10.1016/j.checat.2022.09.001

  22. [22]

      T.H. Jeon, Z.Y. Wu, F.Y. Chen, et al., J. Phys. Chem. C 126 (2022) 6982–6989. doi: 10.1021/acs.jpcc.1c10781

  23. [23]

      L. Li, C. Tang, X. Cui, et al., Angew. Chem. Int. Ed. 60 (2021) 14131–14137. doi: 10.1002/anie.202104394

  24. [24]

      C. Liang, P. Li, Y. Zhang, et al., J. Power Sources 372 (2017) 235–244. doi: 10.1016/j.jpowsour.2017.10.079

  25. [25]

      S. Gu, A.N. Marianov, H. Xu, et al., J. Mater. Sci. Technol. 80 (2021) 20–27. doi: 10.1016/j.jmst.2020.09.053

  26. [26]

      X. Lu, Y. Wang, J. Huang, et al., Chem. Eng. J. 434 (2022) 134648. doi: 10.1016/j.cej.2022.134648

  27. [27]

      J.H. Pan, X.Z. Wang, Q. Huang, et al., Adv. Funct. Mater. 24 (2014) 95–104. doi: 10.1002/adfm.201300946

  28. [28]

      J. Kim, H. Kim, G.H. Han, et al., Exploration 2 (2022) 20210077. doi: 10.1002/EXP.20210077

  29. [29]

      L. Du, H. Xiong, H. Lu, et al., Exploration, 2 (2022) 20210363. doi: 10.1002/EXP.20210363

  30. [30]

      T. Wu, H. Zhao, X. Zhu, et al., Adv. Mater. 32 (2020) 2000299. doi: 10.1002/adma.202000299

  31. [31]

      L. Deng, B. Chang, D. Shi, et al., Renew. Energ. 170 (2021) 858–865. doi: 10.1016/j.renene.2021.02.040

  32. [32]

      P.S. Murphin Kumar, V.K. Ponnusamy, K.R. Deepthi, et al., J. Mater. Chem. A 6 (2018) 23435–23444. doi: 10.1039/c8ta07380e

  33. [33]

      J. Wang, M. Xu, J. Zhao, et al., Appl. Catal. B: Environ. 237 (2018) 228–236. doi: 10.1016/j.apcatb.2018.05.085

  34. [34]

      L. Huang, D. Li, J. Liu, et al., Y. Feng, J. Hazard. Mater. 393 (2020) 122329. doi: 10.1016/j.jhazmat.2020.122329

  35. [35]

      J. Sun, W. Gao, H. Fei, et al., Appl. Catal. B: Environ. 301 (2022) 120829.6. doi: 10.1016/j.apcatb.2021.120829

  36. [36]

      X. Lv, K. Jiang, H. Wu, et al., ACS EST Water 2 (2022) 1451–1460. doi: 10.1021/acsestwater.2c00205

  37. [37]

      F. o. Ahimou, C.J.P. Boonaert, Y. Adriaensen, et al., J. Colloid Interf. Sci. 309 (2007) 49–55. doi: 10.1016/j.jcis.2007.01.055

  38. [38]

      P. Kuang, Y. Wang, B. Zhu, et al., Adv. Mater. 33 (2021) 2008599. doi: 10.1002/adma.202008599

  39. [39]

      H. Shin, H. -i. Kim, D.Y. Chung, et al., ACS Catal. 6 (2016) 3914–3920. doi: 10.1021/acscatal.6b00384

  40. [40]

      D.C. Nguyen, T.L. Luyen Doan, S. Prabhakaran, et al., Nano Energy 82 (2021) 105750. doi: 10.1016/j.nanoen.2021.105750

  41. [41]

      G. Jiang, M. Peng, L. Hu, et al., Chem. Eng. J. 435 (2022) 134853. doi: 10.1016/j.cej.2022.134853

  42. [42]

      X. Zhang, Y. Wang, C. Liu, et al., Chem. Eng. J. 403 (2021) 126269. doi: 10.1016/j.cej.2020.126269

  43. [43]

      B.P. Chaplin, E. Roundy, K.A. Guy, et al., Environ. Sci. Technol. 40 (2006) 3075–3081. doi: 10.1021/es0525298

  44. [44]

      Z. Mao, L. Liu, H.B. Yang, et al., Electrochim. Acta 391 (2021) 138886. doi: 10.1016/j.electacta.2021.138886

  45. [45]

      R.R. Nazmutdinov, D.V. Glukhov, G.A. Tsirlina, et al., J. Electroanal. Chem. 582 (2005) 118–129. doi: 10.1016/j.jelechem.2005.03.021

• 

  Figure 1  (a) Schematic illustration for the synthesis of U-TiO₂, TEM images of (b) A-TiO₂ and (c) U-TiO₂, (d) HR-TEM image of U-TiO₂, (e-g) EDS elemental mapping images of U-TiO₂, (h) XRD patterns, (i) O 1s XPS spectra and (j) ESR spectra of as-synthesized A-TiO₂, H-TiO₂ and U-TiO₂.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Comparisons of (a) LSV in NO₃⁻-N-contained and NO₃⁻-N-free Na₂SO₄ electrolyte, (b) time-course C/C₀, (c) evolution in production distribution, (d) pseudo-first-order rate equation curves, (e) specific and mass activities between different cathodes in NO₃RR, (f) durability tests of U-TiO₂. Reaction conditions: [initial NO₃⁻-N] = 22.5 mg/L; cathodic potential = −0.60 V; electrolysis time = 240 min.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) O 1s XPS spectra of U-TiO₂ after immersion in 10.0 mmol/L PO₄³⁻ solution. (b) NO₃RR performance of A-TiO₂ and U-TiO₂ before and after submerged in the PO₄³⁻ solution. (c) Nyquist plots of the samples. (d) Schematic illustration of the potential-dependent NO₃RR mechanism. (e, f) DEMS results of the NO₃RR on U-TiO₂.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Effects of (a-c) initial NO₃⁻-N concentration on NO₃⁻-N removal, evolution in production distribution, MA and F.E.%. (d-f) The presence of coexisting ions on NO₃⁻-N removal, pseudo-first-order rate equation curves for NO₃RR, and evolution in production distribution over U-TiO₂. Reaction conditions: [initial NO₃⁻-N] = 22.5 mg/L for (d-f); cathodic potentiall = −0.60 V; electrolysis time = 240 min for all the tests.

  下载: 全尺寸图片 幻灯片
•