Dual Active Sites Enhanced Plasmon‐Driven N₂ Photoreduction Over MoO_3−X Nanosheets Under Visible Excitation

ABSTRACT

Nitrogen photo‐fixation used solar energy as a prospective way, has received more attention. In this study, a series of Ru‐C/MoO_3‐x photocatalysts with enhanced localized surface plasmon resonance effects for N₂ photo‐fixation are developed, which not only construct abundant adsorption active sites but also promote the photocatalytic performances through reducing the N≡N bond dissociation activation energy. Notably, an additional LSPR peak at around 490 nm is observed in the UV–vis DRS spectra upon the introduction of the Ru element. Furthermore, increased N₂ adsorption capacity and more efficient carrier separation are obtained via the carbon decoration over the catalysts. Among these, the highest ammonia production rate over the Ru‐C co‐decorated photocatalyst achieves 864 µmol·g_cal ⁻¹·h⁻¹. Furthermore, a markedly apparent quantum efficiency exceeding 1.9% is achieved under visible light excitation at 420 nm, representing one of the highest values ever reported for photocatalytic N₂ fixation over metal oxide catalysts. Density functional theory calculations illustrate that the photocatalytic nitrogen reduction reaction follows the distal pathway, and the Ru/C composite activity sites in the samples can decrease the energy barrier in the rate‐controlling step to further enhance the N₂ photoreduction.

A novel Ru‐C/MoO_3‐x photocatalyst with an enhanced localized surface plasmon resonance (LSPR) effect for N₂ photo‐fixation is developed. The as‐prepared samples not only construct abundant adsorption active sites but also improve the N₂ photoreduction by reducing the N≡N bond dissociation activation energy.

1. Introduction

Nitrogen photo‐fixation driven by solar energy with N₂ or Air and water as the original materials has been recognized as a promising, environmental friendly, and economic strategy of ammonia synthesis [1, 2]. Breaking the stable N≡N triple bond is the core issue during the N₂ photoreduction reaction (pNRR) process [3, 4]. Enhancing N₂ molecules adsorption and activation onto the photocatalyst‐water interface is the necessary stage in the overall pNNR process for the heterogeneous catalytic reaction [5, 6, 7]. These need to develop an appropriate catalytic system to improve the N₂ adsorption and activation. Nowadays, various materials have been explored for the N₂ photo‐fixation; among them, metal oxides are the most significant kind of photocatalysts. However, the reported metal oxide catalysts have relatively low conversion rates for N₂ fixation. For example, transition metals doped W₁₈O₄₉ oxides exhibited limited photocatalytic N₂ fixation rates; among them, P‐Fe/ W₁₈O₄₉ has the highest rate of only 187.6 µmol·g_cal ⁻¹·h⁻¹ [8]. The pure TiO₂ exhibits no catalytic activity for the N₂ photoreduction, and Ni‐doped mesoporous TiO₂ shows a definite photocatalytic NH₃ production rate [9]. Furthermore, a novel Ce/S co‐doped TiO₂ photocatalyst has an enhanced photocatalytic nitrogen fixation rate and apparent quantum efficiency at 420 nm [10]. Widely used existing photocatalysts are generally subjected to charge recombination due to poor interfacial electron transfer and restricted surface diffusion.

As we know, photocatalysis can be significantly improved by generating a localized surface plasmon resonance (LSPR) effect and then generating hot carriers to accelerate the redox reactions of the catalytic system. Plasmon‐enhanced photocatalytic N₂ fixation has also recently served as a green, sustainable NH₃ production. Plasmonic metals such as Au, Ag, Ru, Cu, etc. can absorb near‐infrared light to induce the LSPR effect [11, 12, 13, 14]. Some efforts to more efficiently utilize the LSPR effect have been made. Wang et al. reports the combination of Au nanoclusters and MOFs catalysts enhances the proton supply to the ‐N‐N ^* sites, and further promotes N₂ photo‐fixation performances [15]. The hot carrier generated via the LSPR effect over Ag nanoparticles in the S‐type Ag/Ag₆Si₂O_7/Bi₁₂O₁₇Cl₂ heterojunction can easily overcome the Schottky barrier to promote the photocatalytic reaction [16]. The Ag/W₁₈O₄₉/PMoV heterostructure with improved hot‐electron concentration has exhibited the dual LSPR effect during the photocatalytic N₂ reduction process [17]. LSPR‐induced hot electrons formation and photothermal conversion over Cu/Cu₂O catalytic system for enhanced overall N₂ fixation from air is reported, and the production rate of NH₄ ⁺ have reached above 690 µmol·g_cal ⁻¹·h⁻¹ under air conditions, respectively [18]. N₂ photoreduction activity of traditional metal‐semiconductor plasmonic catalysts is significantly improved; however, a higher N₂ dissociation barrier leads to a lower product generation rate. Therefore, designing an efficient photocatalytic N₂ fixation catalyst should consider not only catalysts with excellent photoinduced redox ability but also provide necessary N₂ adsorption and activation sites.

Molybdenum oxide (MoO_3‐x) as a semiconductor has attracted much more attention in photocatalytic systems, and an efficient band structure can be obtained through introducing the surface oxygen vacancies (O_v) for the improved N₂ molecule adsorption activity [19, 20, 21]. Furthermore, the MoO_3‐x has exhibited a stronger LSPR effect in the visible region during the photocatalytic N₂ fixation process [22, 23, 24]. The asymmetric vacancies in MoO_3‐x effectively improve the N₂ adsorption activity [25]. The switchable oxygen vacancies improve the photocatalytic N₂ fixation performance significantly [26]. Low N₂ adsorption capacity and light response ability still limit their applications. Carbon‐based materials have attracted more attention for their excellent surface absorption activity [27, 28, 29, 30]. Han et al. prepared a Ru/C co‐decorated TiO₂ nanosheet [31], and the sample received an excellent N₂ adsorption activity and a high separation efficiency of the photoinduced carrier. On the other hand, some reports indicate that carbon‐based materials with oxygen vacancies can synergistically promote the electron‐hole separation and N₂ adsorption [32, 33].

How to efficiently develop an efficient surface strategy to fabricate a photocatalytic system with molybdenum oxides and carbon dots is promising. Herein, composite Ru/C sites have been decorated on the MoO_3‐x nanosheets successfully, which not only produce more adsorption and activation active sites of N₂ gas in pure water, but also bring a continuously enhanced LSPR effect over the visible light irradiation. Not surprisingly, therefore, the as‐prepared catalyst displays an improved photocatalytic ammonia production yield with 864 µmol·g_cal ⁻¹·h⁻¹, which is approximately 3 times to the Ru‐doped MoO_3‐x sample. The Ru‐C sites over the catalysts efficiently reduce the energy barrier for the rate‐limiting step, thus enhancing the ammonia synthesis performance.

2. Experimental Section

2.1. Materials Preparation

Ru/Carbon decorated MoO_3‐x nanosheet samples (MoRuC‐x) were prepared through a conventional process. The details of precursors prepared process are shown in the Supporting Information. 120 mg of MoC sample is put into 20 mL of high‐purity water under stirring for 60 min. Then, a certain amount of RuCl₃·3H₂O (0.3, 0.6, 0.9, and 1.2 mL with 10 mg/mL) is cautiously added dropwise and kept stirring for another 60 min to form a black‐blue suspension. Next, 5 mL of NaBH₄ solution (2 mg/mL) is dropped into the solution under continuous stirring. Finally, the received suspension is centrifuged and washed with pure water and absolute ethyl alcohol, and then heated in a vacuum oven at 333 K for 24 h. The produced samples are the MoRuC‐1, MoRuC‐2, MoRuC‐3, and MoRuC‐4, respectively. The contents of Ru element in the samples with 0.9, 1.9, 3.0, and 4.2 wt.%, respectively, for the materials are detected through the ICP‐AES measurement. For analysis, the MoRuC‐2 with 1.9 wt.% Ru has the best N₂ fixation performance, which was selected as the representative sample for characterization and designated as MoRuC. The N amounts in the as‐prepared materials are also observed and are hardly detected in the photocatalysts.

2.2. Nitrogen Photo‐Fixation Process

Photocatalytic N₂ fixation is performed in a home‐made photocatalytic reactor [34], as shown in Figure S16. A 300 W Xe lamp is then turned on. The light intensity is approximately 600 mW/cm² at the solution surface. Different filters are equipped to select the wavelength range of the light source. Namely, a 420‐nm long‐pass filter is used to create a visible light source (approximately 230 mW/cm² at the solution surface), and a variety of band‐pass filters are used to generate monochromatic light. In general, 250 mL of high‐purity water and 20 mg of as‐prepared samples are dispersed in the reactor at first, keeping it at room temperature during the experimental process. The pressure of the inlet gas remains at (0.1–0.15) MPa. At the beginning, sample‐contained solutions are continuously stirred for 0.5 h in the dark under N₂ circumstance. Then, the reactor is irradiated under lighting for 1 h with continuous stirring. Ammonia contents in the solution are detected through the colorimetric method by adding an appropriate amount of Nessler's reagent (Figure S1) and the ion chromatography (IC) method. The AQE (apparent quantum efficiency) is measured with various monochromatic light irradiations. The experimental details are shown in the Supporting Information.

3. Results and Discussion

XRD patterns of the materials have been illustrated in Figure 1a. The peaks in the curves are all consistent with the MoO_3‐x characteristic peaks (JCPDS #76‐1003). And the two main diffraction peaks of MoO_3‐x in the MoRu sample show no significant shift compared to the pristine material, indicating that the Ru element has been decorated on the MoO_3‐x surface in the metallic state. For the MoRuC sample, these two peaks shift toward the lower angles by approximately 0.04°and 0.08°, accompanied by an increase in the lattice constant of molybdenum oxide. The pure Ru⁰ (JCPDS #06‐0663) can be successfully generated when the reaction solution only contains the RuCl₃·3H₂O precursor and NaBH₄ reductant (Figure 1b), and the diffraction peaks of (1 0 0) at approximately 38.4°and (1 1 0) at approximately 69.4° can also be found in the MoRuC sample in Figure 1c. This proves the existence of Ru⁰ in the catalytic system. The Raman spectroscopy for the sample is also collected in Figure 1d. Two peaks at approximately 1355 and 1585 cm⁻¹ of D and G bands in carbon are detected from the carbon decorated samples, and carbon dots are also confirmed in the as‐prepared materials [35, 36].

FIGURE 1.

(a, b) XRD patterns of the photocatalysts and pure metallic Ru⁰, (c) The enlarged XRD images of M and MoRuC samples, (d) Raman spectroscopy of the MoC and MoRuC samples, (e) XPS survey curves of the various samples, High‐resolution XPS spectra of Mo 3d (f), Ru 3p (g) and O 1s (h) for the various samples, (i) EPR spectra under dark and 10 min light irradiation for the M and MoRuC samples.

The XPS results are shown in Figure 1e–h and Figure S2. From Figure 1e, all the elements of Mo, O, and Ru are easily detected in the various samples. The Mo 3d XPS spectra are shown in Figure 1f. The signals located at 236.2 and 232.8 eV give credit to Mo 3d_3/2 and Mo 3d_5/2 for the Mo⁶⁺, respectively. The others correspond to the Mo 3d_3/2 and 3d_5/2 states of Mo⁵⁺ with the energies of 235.3 and 232.1 eV [37, 38]. The Mo⁵⁺ fractions of 4.6%, 6.2%, 7.3%, and 11.7% are measured for the M, MoC, MoRu, and MoRuC, respectively. Figure 2g presents the spectra of Ru 3p. The peaks at approximately 464 eV of Ru 3p_3/2 and 486 eV of Ru 3p_1/2 come from the Ru‐O oxide states, and the others at around 462.5 and 485.5 eV are the Ru⁰ with zero valence, respectively. The results further verify that the RuCl₃ precursor can be successfully reduced to Ru⁰. No significant differences are observed for the Ru element states between the MoRu and MoRuC, and the introduction of the carbon phase does not affect the surface states of Ru⁰. But for the XPS spectra of O 1s in Figure 1h, the surface oxygen vacancy (O_v) fraction of 11.7% according to the peak area for the MoRuC sample is higher than that of 5.2% for the pristine material [39, 40]. These findings offer preliminary evidence that the increased O_v fraction in the system is attributed to the carbon‐derived adjustment of Mo‐O conditions. The EPR measurements are shown in Figure 1i. A signal with a higher intensity at a g value of about 2.003 is observed in MoRuC, showing that more O_v are provided in this sample. Furthermore, the photo‐generated signal in MoRuC is also greater than that of pristine MoO_3‐x, and more in situ O_v are also produced under light irradiation. The in situ extra vacancies and Ru⁰ sites offer more possibilities for the improved photo‐catalytic performances [41, 42].

FIGURE 2.

HRTEM analyses of M (a–c), pure Ru⁰ particles (d–f), and MoRuC (g–j) samples; SEM‐EDS mapping images of the MoRuC sample.

The microstructures of the pristine MoO_3‐x nanosheet are observed via the HRTEM measurement in Figure 2a–c. The sample shows the nanosheet structure with the lattice of 0.19 nm along with the (2 2 0) plane. After the co‐decorated process, the lattice constant (0.19 nm) remains basically unchanged compared to the original sample in Figure 2g–j. The thickness of the nanosheets is also detected through the AFM measurement in Figure S3, and it is approximately 1.3 nm for the as‐prepared MoRuC sample. The attaching lower crystallinity nanoparticles with the lattice constant of 0.32 nm for the (0 0 2) plane (Figure 2k) can be considered as the carbon particles, which is consistent with the results of broad G and D peaks in the Raman spectra [43]. The pure Ru⁰ particle with a lattice of 0.21 nm for the (1 0 1) plane is also detected in Figure 2j. Both microscopic morphology and lattice spacing of this particle are highly similar to the pure Ru⁰ sample in Figure 2d–f, indicating that the metallic Ru⁰ particles are generated in the system. The elemental mapping analyses of the MoRuC sample clearly present the existence of Mo, O, and Ru (Figure 2l; Figure S4), and they are all evenly distributed. The Ru element content is about 1.6 wt.% in this catalyst. Furthermore, metallic Ru⁰ and carbon particles are decorated on the molybdenum oxides, and the Ru‐C composite structure is observed on the catalyst surface.

The NH₄ ⁺ production rates are illustrated in Figure 3a. No ammonia is produced for the pristine MoO_3‐x under light illumination, and trace amounts of NH₄ ⁺ are detected over the MoC sample. The rate catalyzed by the MoRu sample is 178 µmol·g_cal ⁻¹·h⁻¹ under lighting. Furthermore, all of the modified materials have excellent photocatalytic performances. The absorption features and N₂ photo‐fixation performances of the Ru‐C/MoO_3‐x samples can be optimized by adjusting Ru‐C loading and their ratios. Among them, the highest ammonia production rate is 864 µmol·g_cal ⁻¹·h⁻¹ under full‐spectrum light illumination, and increasing by more than 3 times compared to the MoRu. The production rate over the MoRuC sample under the visible light irradiation is approximately 645 µmol·g_cal ⁻¹·h⁻¹. To confirm the ammonia capacity in solution, an ion chromatograph (IC) measurement is used to detect the NH₄ ⁺ ions (Figure S5). The produced rate is in good agreement with the data from Nessler's reagent method. As shown in Figure 3b, the photocatalytic rates at the fifth cycle are still more than 690 or 503 µmol·g_cal ⁻¹·h⁻¹ over the full spectrum or visible light illumination, respectively. There are no obvious changes in the photocatalyst characteristics after the photoreaction process (Figure S6). The microstructures of the MoRuC sample after the reaction are also collected in Figure S7. The nanosheets remained almost unchanged after the experiment, and the lattice contents have also barely budged. The production amounts catalyzed by a 1 g sample (MoRuC) are displayed in Figure 3c. No NH₄ ⁺ ions are generated when the inlet gas is Ar, showing that the N element in the products originates from the original N₂ gas. Under the N₂ gas feeding system, the NH₄ ⁺ production continues to increase. Moreover, the ammonia produced yield under alternative lighting with a 0.5‐h interval is also detected in Figure 3d. The results show that the catalyst has no catalytic properties when the lights are turned off. The photoreaction is driven by the light directly. Furthermore, there is no hydrogen in the outlet gas (Figure S8). The sample exhibits a higher selectivity toward ammonia in the N₂ fixation process. The by‐products of oxygen and nitrate during the oxidation half‐reaction process under lighting for the MoRuC sample are also proven (Figures S8 and S9).

FIGURE 3.

(a) Ammonia generation performance for the catalysts; (b) the cycle stabilities of the MoRuC sample; (c, d) The ammonia yields catalyzed by 1 g of MoRuC under lighting; (e) 1H NMR detections of the solution products after the photoreaction for the MoRuC sample under ¹⁴N₂ and ¹⁵N₂ gas; (f) AQEs of the MoRuC sample under different monochromatic light illumination; (g) The selected photocatalytic N₂ fixation results for the similar catalytic systems [5, 10, 23, 32, 44, 45, 46, 47, 48, 49, 50].

The isotopic labeling detections of ¹⁴N₂ and ¹⁵N₂ are also performed to further confirm the source of the N element in the products. In Figure 3e, three strong peaks with a 52 Hz distance and two weak peaks with 72 Hz are clearly observed in the 1H NMR spectra under the ¹⁴N₂ and ¹⁵N₂ gas atmosphere, respectively. The AQEs during the N₂ photoreduction process catalyzed by the MoRuC sample are also detected in Figure 3f. The considerate AQEs are obtained in the visible light region, and the data at 420 nm can reach 1.9%. The photocatalytic performances of the as‐prepared sample and the reported similar catalysts under comparable experimental conditions are illustrated in Figure 3g and Table S1. The catalytic system proposed in the research displays a competitive N₂ fixation rate and AQE. In order to determine the SCCE of the as‐prepared sample, N₂ photo‐fixation reactions are also carried out in distilled water under a simulated AM 1.5 G solar irradiation (1sun) in Figure S11. The SCCE for the N₂ fixation on MoRuC is determined to be approximately 0.052% and remains almost unchanged after prolonged illumination.

The UV–vis DRS spectra of the as‐prepared catalysts are displayed in Figure 4a and Figure S12. The pristine MoO_3‐x exhibits an absorption edge in the UV region of 400 nm. The MoC sample has a similar edge of near 400 nm, but the more effective visible light responses are obtained. This edge is moving to the infrared region with the formation of co‐decorated photocatalysts, and they also have a broad absorption range to the near‐infrared region. We can receive the band gap values via linear fitting results according to the Kubelka‐Munk function (Figure 4b). The bandgap is 3.08 eV, 3.12 eV, 2.95 eV and 2.98 for the M, MoC, MoRu, and MoRuC, respectively. The positions of the flat band (E_fb) can be obtained from the measured Mott‐Schottky curves (Figure 4c), and the E_fb of M, MoC, MoRu and MoRuC are 0.34 V, 0.21 V, 0.03 and −0.14 (vs. NHE), respectively. The E_cb of the n‐type semiconductors is usually more negative than the E_fb by approximately 0.2 V [51, 52, 53]. The E_cb of 0.14, 0.01, −0.17, and −0.34 V for the M, MoC, MoRu, and MoRuC samples are obtained. The positions of the valence band (E_vb) are calculated from the E_cb and band gap with 3.22, 3.11, 2.78, and 2.64 V for the M, MoC, MoRu, and MoRuC samples, respectively. The potential difference between the E_vb and E_f (Fermi level) can be concluded via the linear fitting results in the VB‐XPS curves. According to Figure 4d, the E_f positions of the M, MoC, MoRu, and MoRuC samples with 0.11, 0.19, −0.12, and −0.21 V are calculated. The band structures (vs. NHE) are illustrated in Figure S13. The upshifting Fermi level can decrease the work function in theory. Simultaneously, this is also favorable to the charge transfer and separation under the visible light excitation.

FIGURE 4.

(a) UV–vis diffuse reflection spectra, (b) The linear fitting results according to the Kubelka‐Munk theory, (c) Mott‐Schottky curves, (d) VB‐XPS curves, and (e–h) Simulated work function curves of the catalysts; (i) The diagram of photoinduced electron flow over MoRuC during the N₂ fixation.

As the results in previous reports [54, 55], inherent oxygen vacancies in MoO_3‐x over the Mo⁶⁺/Mo⁵⁺ pairs can improve the injection of free charges into the Mo 4d level and accompany the defect level generations, inducing the appearance of LSPR effects. The UV–vis DRS spectra (Figure 4a) have increased LSPR peaks at around 575 nm. Remarkably, enhanced defect levels have been detected in VB‐XPS patterns for the two Ru‐contained materials. The defects serve as the active sites to promote the adsorption of N₂ molecules on the material surface [56]. The same results have also been obtained in the previous report [31]. Besides, an additional intense peak is also detected at approximately 490 nm in the MoRu and MoRuC, illustrating that a strong LSPR effect is triggered with the introduction of the Ru element [57]. This may increase the local temperature around the Ru sites, thereby emitting hot electrons to enhance the reduction. As a result, the two samples exhibit promising NH₄ ⁺ production rates in photocatalytic N₂ fixation under visible light irradiation.

The work functions (WF) have been calculated via the density functional theory (DFT) in Figure 4e–h, and four computational models, the original MoO_3‐x (M), MoO_3‐x/C (MoC), MoO_3‐x/Ru (MoRu), and MoO_3‐x/Ru/C (MoRuC), are built in Figure S14. The WF values of the samples are 6.80, 6.09, 5.06, 4.86 eV, respectively. Compared to the other three samples, a decrease in the WF for MoO_3‐x@Ru/C is observed. Theoretically, the band bending due to a lower WF effectively enhances the transfer and separation of the photo‐generated charges [58, 59]. Combining the above results, the process of electron flow during photocatalytic N₂ fixation over MoRuC is depicted in Figure 4i. Under lighting, the photoinduced electrons on the light‐harvesting unit of MoO_3‐x/Ru migrate to the adsorbed N₂ molecules via the charge transfer mediator of the MoO_3‐x/C unit, thus promoting the separation of photogenerated carriers.

The separation efficiencies of the photo‐generated charges are studied through the steady‐state fluorescence (PL) spectroscopy. In Figure 5a, all the as‐prepared catalysts have an emission signal at approximately 450 nm due to the charge recombination. Clearly, the signal intensity for the MoC and MoRuC samples is much lower than the other two, illustrating that the improved separation efficiencies of electron‐hole pairs occur over the C sites. TRPL spectra are collected in Figure 5b. The fitting results indicate that the average decay time of the MoRuC sample is much longer than that of the MoC, demonstrating that Ru‐C composite species decoration is more conducive to promoting the separation efficiency and extending the photo‐excited carrier lifetime. Furthermore, the MoRuC also has the strongest electron transport capability due to the defects enriching sites in Figure 5c. The steady‐state photovoltage (SPV) curves also confirm that the MoRuC sample has a higher SPV intensity than M and MoRu in Figure 5d, and the photovoltage induced by MoRuC is 15.2 times and 2.1 times that of M and MoRu, which further proves that the modification process effectively enhances the electron transfer.

FIGURE 5.

(a, b) PL and TRPL spectra, (c) photocurrent response curves and (d) surface photovoltage spectra under ambient conditions of the various samples.

The gas chemisorption process on the photocatalysts is a key step during the N₂ photo‐fixation [60]. N₂‐TPD measurements (Figure 6a) are used to evaluate the gas adsorption abilities of the materials. The weak peaks at (100°C–200°C) are usually regarded as the physisorption of N₂. A pristine MoO_3‐x sample has a poor chemisorption ability, and three weak chemisorption peaks at around 510°C, 600°C, and 709°C are detected. A slight increase in adsorption capacity is obtained in the MoRu, but the positions of the chemisorption peaks are similar with the M sample. However, the Ru/C anchoring MoO_3‐x (MoRuC) sample significantly increases the chemical adsorption capacity at approximately 434°C. Three chemisorption peaks located at 434°C, 639°C, and 760°C, respectively, are observed in the curve for the MoRuC sample. And the enhanced desorption temperature shows the improved adsorption energy.

FIGURE 6.

N₂‐TPD measurements (a), simulated N₂ adsorption energies (b), comparison of PDOS (c) for the catalysts; Simulated differential charge density images for the various samples after N₂ adsorption, where the sky blue and light yellow iso‐surfaces are related to the electron depletion and accumulation zone, respectively.

In Figure 6b, the MoRuC sample has the most negative energy of −1.0 eV compared to those of M (−0.03 eV) and MoRu (−0.78 eV) through the DFT simulations, which are in agreement with the N₂ adsorption results. The simulated pDOS curves are displayed in Figure 6c. Much more overlapping peaks between Ru and N₂ pDOS curves are observed in the MoRuC sample, and the reaction sites are more easily able to absorb the N₂ molecules after introducing carbon dots in the photocatalyst. Moreover, the lengths of N≡N bonds after the adsorption are detected. The length of the N≡N bond for the M sample is 1.1 Å, which is very close to that of the pristine N₂ (1.1 Å) in Figure S15. The lengths increase to 1.112 and 1.121 Å for the MoRu and MoRuC samples, indicating that the Ru‐C composite sites play a key role in the N₂ activation on the materials. Simulated charge density difference images are used to estimate the electron distributions in Figure 6d. For the adsorbed model of the M sample, the electrons are mainly localized at the Mo sites, which are far from the adsorbed N₂ molecule. After the generation of a Ru‐decorated surface, the gathered electrons on the Ru site can be transferred to the end‐on adsorbed N₂. When the carbon is anchored on the catalyst, a large number of electrons are accumulated around it to further promote the charge transfer efficiency. It is the synergistic effects that greatly improve the photocatalytic nitrogen fixation performance of the materials.

The hydrogenation free energy on the adsorbed N₂ of the distal or alternate pathway for the MoRuC sample is calculated in Figure 7a. The rate‐limiting step (RDS) energy barrier in the alternating pathway (^*NHNH₂‐ ^*NH₂NH₂, 2.03 eV) is higher than that for the distal pathway (^*NH₂‐^*NH₃, 1.71 eV). Therefore, the distal pathway is thermodynamically more favorable than MoRuC. The hydrogenation process (^*N‐ NH₂→^*N) is the RDS over MoRu, and the uphill energy is up to 5.67 eV. The results show that the RDS has been changed over the co‐decorated sample and led to a substantial reduction in the free energy barrier. Furthermore, to confirm the ammonia formation during the N₂ fixation, FTIR spectra with various light irradiation durations over the MoRuC sample are illustrated in Figure 7b. In general, the peak locked at around 1140 cm⁻¹ is contributed by the N─N stretching mode, and the signals at 1415 and 2900 cm⁻¹are attributed to the absorbed NH₄ ⁺species [61, 62]. The peak for the ─N─H bending mode at approximately1620 cm⁻¹ is also obviously present in the curves [63, 64]. The intensity of the peaks increases gradually with the reaction time, showing that the NH₄ ⁺ ions are successfully generated through the hydrogenation reaction over the adsorbed N₂ molecule. The hydrogenation process from N₂ to NH₃ on the catalysts can be described in Figure 7c. At first, an absorbed N₂ molecule is activated on the sample, and a ─N─NH species is formed by receiving a charge (e⁻)/proton (H⁺) pair. Then, another H⁺/e⁻ is attached on the distal site of the ‐N‐NH to generate a ^*N‐NH₂. Next step, the first NH₃ is generated in the reaction. Finally, ^*N, ^*NH₂, and NH₃ can also be produced sequentially in turn to react with one H⁺/e⁻ pair on the previous intermediate product.

FIGURE 7.

(a) Calculated the free energy for N2 photoreduction on MoRuC and MoRu, (b) FTIR spectra for the various lighting durations over the MoRuC sample, (c) Possible process for the photocatalytic N₂ fixation.

4. Conclusion

In this study, we have developed a Ru‐C/MoO_3‐x photocatalyst with an enhanced LSPR effect for N₂ photo‐fixation. The catalysts not only provide abundant adsorption active sites but also promote photocatalytic performance by reducing the N≡N bond dissociation activation energy. An additional LSPR peak at around 490 nm is observed in the UV–vis DRS curves with the introduction of the Ru element over the catalysts. On the other hand, increased N₂ adsorption activity can be obtained through the carbon decoration. The as‐prepared catalyst displays the improved photocatalytic ammonia produced rates with 864 or 645 µmol·g_cal ⁻¹·h⁻¹ under full spectrum or visible light irradiation, respectively. Especially, a remarked AQEs with more than 1.9% is achieved under the visible light excitation at 420 nm, which is one of the highest values for the photocatalytic N₂ fixation ever reported over the metal oxide catalysts. The DFT calculations illustrate that the Ru‐C composite active sites in the catalytic system can significantly reduce the rate‐determining step energy barrier in the alternating pathway, thereby promoting the N₂ photoreduction. This study paves a promising strategy to adjust the interfacial electronic structures in the catalysts for improved photocatalytic activities.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll74504‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.