Nitrate (NO
3-) is an ubiquitous pollutant in environment, originating from the extensive use of nitrogenous fertilizers, inadequate discharge of nitrate-laden domestic/industrial wastewater, and NO
x deposition. It poses serious threats to ecosystems and human health (
e.g., the methemoglobinemia) [
1]. Conventional physical-chemical treatments like ion exchange, electrodialysis and reverse osmosis can effectively separate NO
3- but produce concentrated solutions that require post-treatment steps. Biological denitrification is time-consuming and significantly affected by surrounding conditions, such as dissolved oxygen and temperature [
2]. Recently, electrochemical nitrate reduction (ENRR) has emerged as a highly competitive alternative owing to its mild reaction conditions and sustainability [
3,
4]. More intriguingly, when the product is ammonia (NH
3), the ENRR process utilizing nitrate pollutants as raw materials can be a promising alternative to the energy- and carbon-intensive Haber-Bosch process for NH
3 synthesis [
5,
6].
To accelerate the eight-electron reduction of NO
3--N and improve NH
3 selectivity within ENRR, various catalysts have been developed, including transition metals (Pt, Pd, Cu, Co, Ti, Fe and their alloys) [
7,
8], oxides (CuO, CoO and TiO
2) [
9-
11] and single- or di-atomic catalysts [
12,
13]. Among them, TiO
2 has emerged as a preferred ENRR catalyst owing to its substantial chemical stability, abundance and high NH
3-N selectivity [
14,
15]. For instance, Jia
et al. ever developed a TiO
2 nanotube-based powder catalyst that achieved a relatively high Faradaic efficiency (FE%) of 85.0% and a high NH
3 selectivity 87.1% in ENRR [
16]. Ren
et al. ever designed a series of oxygen vacancy (O
v)-bearing TiO
2-x nanoarray catalysts with anatase (A-TiO
2), rutile (R-TiO
2) and mixed-phase (A + R) crystal structures, respectively [
17]. They demonstrated that the A-TiO
2-x afforded the highest NH
3 yield of 122 mg h
-1 cm
-2 and a FE% of 91.1%. Wei
et al. developed a TiO
2 nanoparticle catalyst featuring a unique anatase-rutile heterojunction and O
vs, which enabled efficient ENRR with a FE% of 78.0% and NH
3 selectivity of 81.9% [
18]. Zhang
et al. also prepared a defective TiO
2 nanotube array catalyst on Ti mesh for ENRR, which delivered a high NH
3 selectivity of 91% and FE% of 85% [
19]. Besides demonstrating robust ENRR efficiency under various reaction conditions, these studies reached a consensus that the O
vs on the TiO
2 surface were the active sites for ENRR [
20,
21]. They could capture and stabilize the N-intermediates by having their oxygen or nitrogen ends occupy the O
v site [
22]. Along with strong adsorption, the N-O bonds of the N-intermediates were significantly activated, and their migration and inter-coupling over the catalyst surface were inhibited, resulting in a low energy barrier and high NH
3 selectivity for ENRR [
16].
It is well documented that TiO
2 has two crystalline forms: A-TiO
2 and R-TiO
2, which generally coexist. Given that the surface structures of A-TiO
2 and R-TiO
2 differ, it can be inferred that the chemical structure of the O
v at their surfaces as well as their roles in ENRR should differ accordingly. Although some studies have explored the dependence of the ENRR on crystalline composition, the interplay between the crystalline phase and surface O
v in ENRR remains underexplored [
18,
23,
24]. Herein we reported a facile approach to the defective and crystalline phase-mixed TiO
2 by calcinating the titanium nitride (TiN) powders under an air atmosphere [
25]. Both the phase composition and surface O
v number within TiO
2 was controllable by tuning the calcination temperature (450–750 ℃). The ENRR performances of these TiO
2 catalysts in terms of mass activity, NH
3-N selectivity, FE% and tolerance to impurities were then evaluated and compared to that of the commercially-available pure A-TiO
2 and R-TiO
2 phase, respectively. Finally, controlled experiments,
in-situ spectrometric tests and theoretical calculations were combined to probe into the synergistic effects of the crystalline phase and O
v on ENRR [
26].
The defective and phase-mixed TiO
2 with N doping (N-TiO
2) was prepared by calcining the TiN at an air atmosphere within a specific temperature range of 450–750 ℃. The representative Transmission electron microscope (TEM) image of N-TiO
2 calcined at 650 ℃ (denoted as N-TiO
2–650) is presented in
Fig. 1a, which shows irregular nanosheets with a mean lateral size of 20–30 nm. Corresponding High resolution transmission electron microscope (HRTEM) image in
Fig. 1b reveals that the nanosheets display two distinct lattice fringes with the spacing of 0.350 and 0.322 nm, corresponding to the (101) plane of A-TiO
2 and the (110) plane of R-TiO
2. These results suggest the presence of a phase mixture in the N-TiO
2–650. The presence of N in N-TiO
2–650 is confirmed by solid elemental analysis, and its content is only 0.17 wt%. The Energy dispersive spectrometer (EDS) elemental mapping results in
Fig. 1c reveal the uniform distribution of Ti, O and N on N-TiO
2–650. The XRD patterns of pure A-TiO
2, pure R-TiO
2 and N-TiO
2–650 in
Fig. 1d demonstrate that the N-TiO
2–650 exhibits a mixed crystalline phase of rutile and anatase, which is in good agreement with the HRTEM results. The relative content of A-TiO
2 to R-TiO
2 is estimated to be 49.3/50.7 on basis of their peak intensity in the X-Ray diffraction (XRD) pattern. X-ray photoelectron spectroscopy (XPS) analyses on N-TiO
2–650, A-TiO
2 and R-TiO
2 are further conducted to investigate the electronic states of Ti and O in samples. The Ti 2p XPS spectra in
Fig. 1e displays the Ti
4+ species in three samples. No Ti
3+ was detected possibly owing to its low content. The O 1s XPS spectra in
Fig. 1f confirm the presence of O
v on the surface of all the three samples by the characteristic peak at 531.4 eV, corresponding to the Ti site that is coordination unsaturated. The surface O
v content can be estimated by the peak area ratio of O
v to (O
L + O
v) (O
L denotes lattice oxygen). The N-TiO
2–650 is indicated to have the largest number of O
v (37.8%), followed by R-TiO
2 (21.2%) and A-TiO
2 (18.1%). The high O
v content can be partially attributed to the unique preparation process of N-TiO
2 via TiN oxidation.
Fig. 1g presents the Gibbs free energy for O
v generation on A-TiO
2 (101) and R-TiO
2 (110), which demonstrates that the O
v is more easily formed on R-TiO
2 than on A-TiO
2 (2.89
vs. 3.57 eV). We can speculate from these results that more O
v distribute on R-TiO
2 surface in N-TiO
2.
Fig. 1h shows the UV–vis-DRS spectra of three samples. The light absorption edge of N-TiO
2–650 is between those of A-TiO
2 and R-TiO
2. Correspondingly, the band gap increases from 3.08 eV of R-TiO
2, to 3.18 eV of N-TiO
2–650 and 3.39 eV of A-TiO
2 (
Fig. 1i). These results confirm the presence of mixed phases in N-TiO
2–650.
The phase composition and the O
v number in N-TiO
2 can be controlled by tuning the calcination temperature. The XRD patterns in
Fig. 2a show that the R-TiO
2 is gradually accumulated in N-TiO
2 along with the increase in calcination temperature.
Fig. 2b summarizes the relative amount of A-TiO
2 to R-TiO
2 as well as the N content in N-TiO
2. As the calcination temperature increases, the proportion of R-TiO
2 rises, while the N content decreases from 0.49% to 0.05%. This occurs because the N atoms in TiN are gradually replaced by O atoms during the oxidation process. Generally, more complete oxidation at higher temperatures results in lower N contents. The UV–vis-DRS spectra in Fig. S1 (Supporting information) reveal that the band gaps of these N-TiO
2 follow a descending order of A-TiO
2 > N-TiO
2–450 > N-TiO
2–550 > N-TiO
2–650 > N-TiO
2–750 > R-TiO
2. Furthermore, N-TiO
2–450 and N-TiO
2–550 exhibit distinct absorption peaks in the visible light range of 450–500 nm, which can be attributed to the formation of intermediate energy levels induced by N doping. The O 1s XPS spectra in
Fig. 2c show that as the temperature increases, the O
v content decreases from 43.2% to 35.2%. The same variation trend of the O
v content within this samples were observed in the EPR results present in
Fig. 2d. This reduction in O
v number should be associated with the reduced N doping observed at higher calcination temperatures.
Fig. 3a exhibits the Linear sweep voltammetry (LSV) curves of the A-TiO
2, R-TiO
2 and N-TiO
2–650 in the presence and absence of NO
3--N, respectively. In compared to A-TiO
2 and R-TiO
2, N-TiO
2–650 displays a smaller current response to hydrogen evolution in the absence of NO
3--N while sharper current increase in the presence of NO
3--N, implying a more effective ENRR on the phase-mixed N-TiO
2–650. The ENRR performances of A-TiO
2, R-TiO
2 and N-TiO
2–650 are further evaluated by potential-controlled batch tests at −0.80 V
vs. RHE. Plotting the concentrations of NO
3--N, NO
2--N, NH
3-N and total N versus reaction time in
Fig. 3b show that all the three catalysts are effective in converting NO
3--N to NH
3-N with minimal residue of NO
2--N at the end of reaction. Noted that the total nitrogen number in the solution remains almost constant among the three catalysts, indicating that minimal other products, such as N
2 and N
2O, are formed. As expected, N-TiO
2–650 affords the highest conversion efficiency of 91.8% within 6 h, followed by A-TiO
2 (72.9%) and R-TiO
2 (42.1%). To explore more about the phase-dependent ENRR activity of TiO
2, all the N-TiO
2-
x (
x = 450, 550, 650 and 750) are subjected to potential-controlled batch ENRR tests.
Fig. 3c shows that the removal kinetics of NO
3--N vary in a volcano-like trend with the
x value. The N-TiO
2–650 delivers the fastest NO
3- removal, followed by N-TiO
2–550, N-TiO
2–450 and N-TiO
2–750. The product distributions at the end of reaction are presented in
Fig. 3d. As observed, the NO
3- are almost completely converted to NH
3-N with minimal formation of NO
2--N, resulting in the NH
3-N selectivity exceeding 98% for all the four samples.
Fig. 3e summarizes the mass activities and FE% of A-TiO
2, R-TiO
2 and N-TiO
2-
x. As observed, the N-TiO
2-
x outperform both A-TiO
2 and R-TiO
2 in terms of mass activity and FE%. Among the N-TiO
2-x samples, N-TiO
2–650 affords both the highest mass activity (22.2 mg
N h
-1 g
cat.−1) and FE% (79.4%).
Given the high ENRR performance, N-TiO
2–650 is considered a promising catalyst for NH
3 synthesis using NO
3--N pollutants as raw materials. Durability is a key descriptor of catalytic performance, and the durability of N-TiO
2–650 is evaluated through cycling batch ENRR experiments. The results in
Fig. 3f show that the conversion efficiency of NO
3--N to NH
3-N remains above 90% over 5 cycles, with the NO
2--N yield staying below 0.5 mg/L. Accordingly, N-TiO
2–650 is expected to be durable in long-term ENRR. To evaluate the versatility of N-TiO
2–650 for various application scenarios, its ENRR performance is reassessed under different working potentials and in the presence of impurity ions and dissolved organics that coexists with NO
3--N in wastewater.
Fig. 4a depicts the time-course
C/
C0 for NO
3--N during ENRR on N-TiO
2–650 under different working potentials ranging from −0.60 V to −0.90 V. It is observed that ENRR does not begin until the potential is more negative than −0.70 V, with a more negative potential leading to accelerated NO
3--N conversion.
Fig. 4b shows that less NO
2- remains and more NH
3-N is formed at more negative potentials, indicating increased NH
3-N selectivity under these conditions. It is noted that the N mass balance is not achieved and the FE% for ENRR decreases to 39.2% at −0.90 V, compared to 79.4% at −0.80 V. These phenomena may result from the predominance of HER at this negative potential, which carries some N species away from the solution while simultaneously wasting electrons.
Fig. 4c exhibits the time-course
C/
C0 for NO
3--N during ENRR on N-TiO
2–650 under in the presence of Cl
- (5 mmol/L), CO
32- (1 mmol/L), Mg
2+(1 mmol/L), Ca
2+ (2 mmol/L) and HA (3.3 mg/L). Their concentrations are set to approximate those in natural water body [
3].
Fig. 4c shows that the impurities, with the exception of Cl
- and natural organic materials (NOMs), exert minimal effects on the ENRR kinetics. In the presence of 5 mmol/L Cl
- and 3.3 mg humic acid (representative of NOMs), the conversion efficiency at the end of reaction drops to 71.5% and 85.9%, respectively, compared to 91.3% in the control experiment.
Fig. 4d demonstrate that except for humic acid, the other impurities show limited effects on the NH
3 selectivity. The negative effects of Cl
- and humic acid can be attributed to their potential to occupy active sites and alter the structure of the electric double layer at the electrode/solution interface [
20,
27]. These results demonstrate the high tolerance of N-TiO
2–650 to impurities, but further efforts are required to maintain its high performance under more complex conditions in wastewater.
To probe into the synergistic effects of phase composition and O
v on ENRR,
in-situ test is first performed at the potentials of −0.70, −0.80 and −0.90 V, respectively, to identify the reaction pathway on N-TiO
2–650. K
15NO
3 is used as the source of nitrate to eliminate interference from the nitrogen reduction reaction for N-intermediate identification. The spectra in
Fig. 5a demonstrate the occurrences of NO, N, N
2, NH, NH
2 and NH
3, while N
2O and NH
2OH are absent [
12]. When coupling the results with the detected NO
2- and NH
3 in solution, the ENRR is considered to proceed
via the mass diffusion of NO
3- from bulk solution to catalyst surface, forming adsorbed
*NO
3. Subsequently, the
*NO
3 evolves to
*NO
2,
*NO and
*N through stepwise dehydrogenation reactions, and
*N further converts to
*NH,
*NH
2 and
*NH
3 through stepwise hydrogenation reactions (
Fig. 5a) [
28]. These spectra also suggest that the hydrogenation of
*N predominates over the coupling of
*N-
*N to N
2, consistent with the high NH
3 selectivity achieved by N-TiO
2–650 [
29]. The resolution of strong peaks assigned to H
2 suggests that side hydrogen evolution is inevitable on N-TiO
2–650, resulting in an overall FE% below 100%. Furthermore, the signals for all the N-intermediates are stronger under more negative potentials, verifying the boosted ENRR at these conditions.
According to Singh
et al.'s results, the adsorption strengths of N-intermediates on the catalyst can be described by the adsorption strength of a single N or O atom on the catalyst [
30]. In this context, the adsorption strengths of the single O and N atom on pristine A-, R-TiO
2, as well as them with surface O
v sites were compared by Density functional theory (DFT) calculations, respectively [
12,
31]. The A-TiO
2 (110) and R-TiO
2 (101) facets are modeled separately for investigation, and the O
v site is created by removing one O atom from the facet. The results in
Fig. 5b show that the adsorption strengths of the single N and O atom on the O
v sites of both TiO
2 phases are much larger than that on pristine surface. It is suggested that the O
v sites are more likely to be the active sites for ENRR, consistent with the results reported in literature [
20]. Furthermore, compared to R-TiO
2 (110), the O
v site on A-TiO
2 (101) binds more strongly with N-intermediates, implying that the activity of the O
v site within ENRR differs between A-TiO
2 and R-TiO
2.
The thermodynamics for ENRR on the O
v sites of A-TiO
2 (101) and R-TiO
2 (110) were further investigated by DFT calculations to gain more insight to the role of O
v [
32]. The Gibbs free energy profiles in
Fig. 5c show that the overall ENRR process on the O
v sites of both A-TiO
2 (101) and R-TiO
2 (110) is thermodynamically spontaneous, owing to the negative value for
*NO
3 conversion to
*NH
3 [
32]. On the O
v site of R-TiO
2 (110), only the dehydrogenation step of
*NOH to
*N requires additional energy input. In contrast, on the O
v site of A-TiO
2 (101), there are energy barriers for two steps: the dehydrogenation conversion of
*NOH to
*N and the hydrogenation conversion of
*NH
2 to
*NH
3. Additionally, the energy barrier for ENRR on the O
v site of R-TiO
2 (110) (0.40 eV) is smaller than that on the O
v site of A-TiO
2 (101) (0.44 and 0.56 eV), demonstrating that the O
v on R-TiO
2 is more active for NO
3- conversion. Besides reactivity, the renewability of the O
v site, associated with the desorption kinetics of NH
3 from it, is critical for maintaining its activity [
33]. Herein the energy barriers for NH
3 desorption from the O
v sites of A-TiO
2 and R-TiO
2 are investigated by DFT calculations [
34]. The Gibbs energy profiles in
Fig. 5d reveal that NH
3 desorption is thermodynamically unfavorable on both TiO
2 phases due to the strong affinity of the O
v sites with NH
3. Nevertheless, there is a lower energy barrier for NH
3 desorption from the R-TiO
2 (110), indicating that the O
v site on R-TiO
2 can be more easily renewed compared to that on A-TiO
2 [
35].
Protons are essential reactants for ENRR, and the kinetics of their transfer from H
2O molecules to reactants on catalyst surface are closely associated with the denitrification kinetics [
7,
36]. To assess the efficacy of proton transfer kinetics, kinetic isotope experiments (KIE) are performed for ENRR on N-TiO
2–650. A KIE value obtained within these experiments serves as a key descriptor of the proton transfer kinetics from H
2O dissociation. It is defined as the ratio of NO
3- removal kinetics or NH
3 production kinetics in H
2O versus D
2O media, respectively [
37].
Fig. 5e plots the concentrations of NO
3--N and NH
3-N against the electrolysis time in the two media, respectively. Compared to H
2O media, both NO
3- removal and NH
3 production become much slower in D
2O media.
Figs. 5f and
g demonstrate that both NO
3- removal and NH
3 production follow pseudo-first-order kinetics, with the KIE value calculated as
kapp, H2O/
kapp, D2O reaching 2.77 for NO
3- removal and 3.45 for NH
3 production (
kapp refers to apparent reaction rate constant). These results demonstrate that the ENRR on N-TiO
2 largely depends on the proton transfer kinetics [
38]. To compare the capability of O
v sites on A-TiO
2 and R-TiO
2 in proton adsorption, the Gibbs free energy profiles for proton adsorption on these O
v sites were investigated by DFT calculations.
Fig. 5h exhibits that the H adsorption is spontaneous on the O
v site of both A- and R-TiO
2, owing to the negative value of Δ
G [
39]. Furthermore, the O
v site on the A-TiO
2 surface has a greater affinity for protons than that on R-TiO
2. This suggests that the defective A-TiO
2 is more active in proton transfer, making ENRR less dependent on proton transfer.
From the above, we can conclude that the O
vs on the TiO
2 surface are the active sites for ENRR, but their roles differ between A-TiO
2 and R-TiO
2. Specifically, the O
vs on the R-TiO
2 surface are more active for ENRR and renew more easily during ENRR, while those on A-TiO
2 surface perform better in proton adsorption. Accordingly, a dual-site mechanism involving the O
v site on R-TiO
2 surface and the O
v site on A-TiO
2 surface can be expected on the defective and phase-mixed N-TiO
2, with the synergy of the dual O
v sites contributing to significantly enhanced ENRR performance (
Fig. 5i) [
40]. Based on the distinct functions of the O
vs on R-TiO
2 and A-TiO
2 surfaces, it also becomes clear why the ENRR efficiency shows a volcano-like variation with the catalyst varying from N-TiO
2–450 to N-TiO
2–750. As schematically described in
Fig. 5j, as the catalyst varies from N-TiO
2–450 to N-TiO
2–750, the content of R-TiO
2 in the catalyst rises while the O
v number decreases. Correspondingly, the overall efficacy of O
v in proton transfer decreases owing to the decreased number of O
vs distributed on the A-TiO
2 surface. Nevertheless, the mean activity and renewability of them improve as an increased proportion of O
vs are distributed on the R-TiO
2 surface. The tug-of war of the two reverse varying trends leads to a peak ENRR efficiency on N-TiO
2–650 with a moderate R-TiO
2 content of 49.3%.
Herein the synergy between phase composition and Ov on TiO2 for ENRR is investigated using phase-mixed and defective N-TiO2 as platform catalysts. The N-TiO2 are prepared by calcining TiN powders in an air atmosphere with the phase composition and Ov number controlled by calcination temperature. The ENRR performance of these N-TiO2 samples surpasses that of pure A- and R-TiO2, showing a volcano-like trend with respect to R-TiO2 content and Ov number in the catalyst. Mechanistic studies reveal that Ovs are the active sites, but their functions differ between A-TiO2 and R-TiO2. Specifically, the Ovs on R-TiO2 are more active in NO3- conversion and renew more easily during reaction, while those on A-TiO2 are more efficient in proton adsorption. The synergy between Ovs on R-TiO2 and A-TiO2 leads to higher ENRR efficacy on N-TiO2. On the other hand, when the catalyst varies from N-TiO2–450 to −750, the overall efficacy of Ovs in proton transfer decreases owing to the decreased proportion of Ovs distributed on A-TiO2. Nevertheless, the mean activity and renewability of Ovs improve as an increased proportion of them are distributed on R-TiO2. The tug-of-war between the two opposing trends leads to a peak ENRR performance on N-TiO2–650 with a moderate R-TiO2 content of 49.3%. This work highlights the synergy between crystalline phase and surface defects in heterogeneous catalysis, suggesting that optimizing phase composition and enriching surface defects could enhance the ENRR on TiO2.
PDF
Full
61
35
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
