Engineering vacancy and hydrophobicity of two-dimensional TaTe2 for efficient and stable electrocatalytic N2 reduction.

Demand for ammonia continues to increase to sustain the growing global population. The direct electrochemical N2 reduction reaction (NRR) powered by renewable electricity offers a promising carbon-neutral and sustainable strategy for manufacturing NH3, yet achieving this remains a grand challenge. Here, we report a synergistic strategy to promote ambient NRR for ammonia production by tuning the Te vacancies (VTe) and surface hydrophobicity of two-dimensional TaTe2 nanosheets. Remarkable NH3 faradic efficiency of up to 32.2% is attained at a mild overpotential, which is largely maintained even after 100 h of consecutive electrolysis. Isotopic labeling validates that the N atoms of formed NH4 + originate from N2. In situ X-ray diffraction indicates preservation of the crystalline structure of TaTe2 during NRR. Further density functional theory calculations reveal that the potential-determining step (PDS) is ∗NH2 + (H+ + e-) → NH3 on VTe-TaTe2 compared with that of ∗ + N2 + (H+ + e-) → ∗N-NH on TaTe2. We identify that the edge plane of TaTe2 and VTe serve as the main active sites for NRR. The free energy change at PDS on VTe-TaTe2 is comparable with the values at the top of the NRR volcano plots on various transition metal surfaces.

Introduction

Nitrogen fixation is a key chemical transformation for sustainable development as ammonia (NH3) is essential across modern industry and agriculture.1, 2, 3, 4 The traditional fossil-fuel-powered Haber-Bosch process remains widely employed for artificial NH3 synthesis. However, intense energy consumption (extreme reaction conditions of 300 °C–500 °C and 15–25 MPa), and massive emission of CO2 (from the reformation of fossil fuels to produce the hydrogen gas feedstock) pose severe technological, environmental, and ecological challenges.6, 7, 8, 9, 10 The electrocatalytic N2 reduction reaction (NRR), using intermittent electricity generated from renewable sources and water as the hydrogen source, is an attractive strategy for sustainable NH3 production, and has recently sparked tremendous research interest. However, the cleavage of the inert N2 molecule is difficult due to its strong dissociation energy (9.756 eV, i.e., ∼941 kJ mol−1), high first ionization energy (1,503 kJ mol−1), and short N≡N triple bond (1.098 Å).12, 13, 14, 15, 16 Another issue is that the major competitive reaction, the hydrogen evolution reaction (HER), has faster reaction kinetics and occurs under similar or even lower overpotentials during the NRR with aqueous electrolytes, causing severe energy efficiency losses. Therefore, the design and development of efficient electrocatalysts to break the N≡N bond to drive N2 conversion while simultaneously suppressing HER is extremely desirable., Despite recent efforts, most catalytic systems reported suffer from low selectivity for NH3 formation (typically less than 15% owing to the concomitant HER), large overpotential (or low energetic efficiency), and insufficient stability (usually <30 h), limiting practical use and technological commercialization.

Transition metal dichalcogenides (TMDs) are an emerging class of two-dimensional (2D) materials possessing direct and tunable band gaps.20, 21, 22 The interatomic binding in TMDs is strong due to covalent in-plane bonding. Nevertheless, the successive layers in TMD materials are bound through weaker van der Waals interlayer forces, which renders these layered materials to be easily exfoliated mechanically. Especially in transition metal tellurides, the relatively low electronegativity of tellurium frequently leads to complex scenarios of competition between metals and non-metals for the bonding electrons. Compared with O, S, and Se chalcogenides, Te has more metallic character, which is a highly desired property for electrocatalysts.24, 25, 26 2D TaTe2 has been produced by chemical vapor deposition (CVD); however, the route has drawbacks, such as low yield and complicated operation steps, among others. Large-scale synthesis of 2D TaTe2 remains a challenge. Furthermore, the catalytic properties of 2D TaTe2 for NRR are unexplored to date.

Herein, we report efficient production of ultrathin metallic TaTe2 nanosheets with tuned Te vacancies simply via liquid exfoliation of bulk TaTe2. γ-Butyrolactone (γ-BL) was discovered to be an excellent organic solvent for the exfoliation. The specific surface area is maximized for a 2D structure, which affords a potentially high density of active sites and also increases surface accessibility to reactants. In addition, the in-plane electrical conductivity of TaTe2 increased after exfoliation. This provides benefits in electrochemical reactions, where the higher conductivity usually ensures more efficient utilization of the electrical energy. Importantly, the as-prepared defective 2D TaTe2 was found to be active for NRR, affording a remarkable NH3 faradic efficiency (FE) in excess of 12% and an NH3 yield rate of about 6.3 μgNH3 h−1 mgcat−1 at a low applied potential of −0.12 V versus the reversible hydrogen electrode (versus RHE). More interestingly, further surface modification of TaTe2 nanosheet electrodes by tethering with trimethoxy (1H,1H,2H,2H heptadecafluorodecyl) silane (TMHFS) could limit the proton transfer on the electrode surface without interrupting the flow of non-polar moieties, thus enhancing the availability of N2 on the electrode surface in relation to that of the protons. After such hydrophobic treatment of TaTe2 nanosheets (denoted as 2D H-TaTe2), the FE of NH3 was markedly improved, approaching 32.2%, over 15.4 times compared with bulk TaTe2. The 2D H-TaTe2 catalytic performance was maintained even after 100 h of NRR electrolysis.

Results and discussion

Synthesis and structural characterization

We successfully prepared TaTe2 nanosheets by liquid exfoliation of bulk TaTe2 (Figures 1A and 1B) under ultrasound followed by centrifugation (CF) to remove poorly exfoliated aggregates. We discovered seven organic solvents that can effectively delaminate and disperse TaTe2 (Table S1, Figure 1C), namely dimethyl sulfoxide, N-methylformamide, 1,3-dimethyl-2-imidazolidinone, γ-BL, tetrahydrofuran, cyclohexanone, and N-methyl-2-pyrrolidinone. γ-BL exhibited superior exfoliating capability. The mass of the TaTe2 material after removal of the solvent for specific volumes of dispersions allowed one to estimate the stock dispersion concentration. A sample of the stock dispersion in γ-BL was serially diluted. The absorbance per unit-cell length for each diluted sample was then measured and plotted versus TaTe2 dispersion concentration (Figure S1). The absorption coefficient (α) at 400 nm was derived to be 554.9 mL mg−1 m−1 by a straight line fit through the points. The resulting concentration (C) after CF was determined according to the Lambert-Beer law (A = αCl, where l is the cell length). The dispersion concentration increased steadily with the initial TaTe2 concentration and reached as high as 4.5 mg mL−1 (Figure S2). Analogous to many other 2D materials,, the dispersion concentration scaled inversely with CF speed (Figure S3). Of interest is that about 80% of the TaTe2 nanosheets maintained a stable dispersion against sedimentation for at least 20 days (Figure S4).

Figure 1

Structure, Surface Composition, and N2 Adsorption Property of TaTe2

(A and B) (A) Top view and (B) side view of the TaTe2 crystal structure.

(C) The absorbance (at 400 nm) per unit-cell length of TaTe2 dispersion (after CF at 3,000 rpm for 30 min) versus different organic dispersants.

(D) XRD patterns of bulk TaTe2 and 2D TaTe2.

(E and F) (E) Ta 4f and (F) Te 3d XPS spectra of 2D TaTe2.

(G–I) (G) Raman spectra, (H) EPR curves, and (I) N2-TPD profiles of bulk TaTe2 and 2D TaTe2 exfoliated in γ-BL.

Figure 1D shows powder X-ray diffraction (XRD) patterns of the starting TaTe2 and as-obtained TaTe2 nanosheets. The diffraction peaks at ∼13.2°, 30.6°, 31.2°, and 38.8° in traces A and B can be well indexed to the respective (001), (−603), (310), and (−313) reflections of TaTe2 (JCPDS no. 71-2197). This shows that the TaTe2 lattice was preserved after exfoliation. However, dramatic weakening in the relative intensity of the (001) was observed, suggesting loss of orientation in the plane as a consequence of exfoliation. The diffraction peak at 27.7° can be matched with tantalum oxide (JCPDS no. 19-1299), indicating partial oxidation under ultrasound liquid exfoliation. X-ray photoelectron spectroscopy (XPS) was employed to probe the surface chemical state of TaTe2 nanosheets (Figure 1E); the main peaks centered at 25.8 and 27.7 eV correspond to Ta 4f7/2 and Ta 4f5/2, respectively, denoting a main valence state of Ta4+., The weak peaks appearing at 22.9 and 24.8 eV arise from TaO defects. Two doublet peaks can be seen with Te 3d5/2 at 572.8 and 576.1 eV and Te 3d3/2 at 583.2 and 586.5 eV attributed to respective Te2− and Te0 (Figure 1F)., Although metallic Te and TaOx were also present in bulk TaTe2 (arising from the Te precursor remaining during CVD synthesis of TaTe2 and its surface oxidation upon exposure to air) (Figure S5), the fractions of Te0 increased after exfoliation, suggesting the formation of Te vacancies in TaTe2 nanosheets. Whereas the ratio of TaO/Ta4+ noticeably decreased after the exfoliation. An apparent O 1s XPS peak at 532.0 eV originating from adsorbed oxygen species was discernible (Figure S6). Figure 1G shows representative Raman spectra of bulk TaTe2 and TaTe2 nanosheets, with two prominent resonance peaks at around 122 and 141 cm−1 being associated with the in-plane E vibration mode and the out-of-plane A1 vibration mode, respectively. The effective restoring forces acting on the atoms increase concomitantly with layer number due to the interlayer van der Waals interaction, leading to a blue shift of the out-of-plane A1g vibration. This tendency indicates that TaTe2 nanosheets become increasingly ultrathin. Further electron paramagnetic resonance (EPR) spectroscopy was employed to detect paramagnetic signals, allowing for analysis of the unsaturated sites with unpaired electrons in catalyst (Figure 1H). Bulk TaTe2 displayed weak EPR signals, suggesting low level of defects. By contrast, exfoliated TaTe2 showed a symmetric pair of sharp peaks with the signal at g = 2.003, arising from trapped unpaired electrons by Te vacancies through adsorbed oxygen species from air (O2−), in accord with XPS results.

Bulk TaTe2 and TaTe2 nanosheets exhibited type II N2 adsorption/desorption isotherms in the Brunauer-Deming-Deming-Teller classification (Figure S7). After exfoliation, the Brunauer-Emmett-Teller surface area and average single-point total pore volume rose from <2.0 m2 g−1 to 0.0112 cm3 g−1 to 59.8 m2 g−1 and 0.160 cm3 g−1, respectively. The substantial increase in specific surface area benefits mass transport and improves surface accessibility to reactants, thus favoring NRR. We performed N2 temperature-programmed desorption (N2 TPD) to investigate the N2 adsorption ability of TaTe2. The N2 TPD profiles of TaTe2 nanosheets show two strong peaks at 338.6 °C and 411.0 °C, and a mild peak at 97.4 °C, arising from chemisorption and physisorption of N2, respectively (Figure 1I). In comparison, only weak signals of physi- and chemisorbed N2 on bulk TaTe2 were detected, implying that TaTe2 nanosheets possess more exposed active sites. The activation energy of desorption from the surface of TaTe2 nanosheets was estimated to be about 32.6 kJ mol−1, an indication of strong adsorption. The superior N2 adsorption capacity and high bonding-affinity toward N2 of 2D TaTe2 definitely provide benefits for electrocatalytic N2 reduction.

Transmission electron microscopy (TEM) images showed the formation of thin TaTe2 flakes with lateral sizes in the range of 200 nm to 1 μm, randomly stacking on top of each other (Figures 2A–2C and S8). Figure 2D reveals a set of interference fringes, with the fast Fourier transform showing d-spacings of 0.29 and 0.22 nm, and a mutual angle between the planes of 94°. These are in excellent agreement to 1T TaTe2 viewed down the [15-6] axis, and with the 0.29 nm spacing corresponding to a {111} plane. The existing oxygen means that TaTe2 flakes were weakly oxidized on the surface, with disordered edges observed for many of the nanosheets (Figures 2B, 2I, and S8A).

Figure 2

Morphology and Structure Characterization of 2D TaTe2

(A) Low-magnification TEM image of 2D TaTe2.

(B) Transformed TEM image of Figure S8A.

(C and D) HRTEM images of 2D TaTe2, the inset in (D) is the corresponding fast Fourier transform pattern.

(E–H) (E) HAADF-STEM image of TaTe2 nanosheets and elemental maps of (F) O, (G) Ta, and (H) Te.

(I and J) (I) HAADF-STEM image of TaTe2 nanosheets and (J) corresponding electron energy loss spectroscopy spectrum, showing Ta and O edges.

(K) Tapping-mode atomic force microscopy image of 2D TaTe2 in γ-BL at a concentration of 0.4 mg mL−1 deposited on an SiO2/Si substrate.

(L) Corresponding line sections of image (K).

High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) observation (Figure 2E) together with in situ energy-dispersive X-ray spectroscopy mapping (Figures 2F–2H) indicated that TaTe2 nanosheets were composed of Ta, Te, and a small amount of oxygen. Ta M and O K could only be detected in the electron energy loss spectroscopy spectrum (Figure 2J), and the Ta/O atomic ratio was about 0.61, corresponding to TaO in the XRD pattern. Shown in Figure 2K are the tapping-mode atomic force microscopy images of TaTe2 dispersions deposited on an SiO2/Si substrate. Flakes with measured heights of about 5, 13, 18, and 21 nm were observed. The average thickness of TaTe2 exfoliated in γ-BL at a concentration of 0.1 mg mL−1 was estimated to be 4.5 nm (Figure S9).

Electrocatalytic property for N2 reduction

To limit water availability close to the catalyst surface, the hydrophilic electrode surface was modified by TMHFS. It is expected that diminishing proton concentrations at the electrochemical interface by such hydrophobic modification can mitigate competition with the HER and thus boost the NRR. This can be explained by two aspects. On the one hand, the rate of hydrogen production was modeled to be the first order in the electron and proton concentrations, while the rate of NH3 production was zeroth order in both. On the other hand, it is unlikely for electrochemical reduction of N2 to be kinetically limited by proton concentration in electrolyte. Figure 3A shows the contact angles of TaTe2 nanosheets before and after hydrophobic treatment. Pristine TaTe2 nanosheets is hydrophilic with a contact angle of 57.5°. In stark contrast, after hydrophobic treatment, the hydrophilic electrode showed an increased contact angle of 147.2°, indicating a hydrophobic feature of 2D H-TaTe2. The TaTe2 nanosheets before and after the hydrophobic modification were examined for NRR using a classic two-compartment H-type cell separated by a cation-exchange membrane (Nafion 117) with continuous N2 bubbling (Figure S10).

Figure 3

Contact Angle Results and Electrochemical Nitrogen Reduction Activities

(A) The water contact angles of 2D TaTe2 and H-TaTe2.

(B) The linear sweep voltammetry curves of bulk TaTe2, 2D TaTe2, and H-TaTe2 in Ar-purged (dashed line) or N2-purged (solid line) 0.1 M HCl solutions with a scan rate of 5 mV s−1.

(C and D) (C) The yield rates and (D) FEs of NH3 over 2D H-TaTe2, TaTe2 nanosheets, and bulk TaTe2. The NH3 yield rates obtained by ion selective electrode are also included, as indicated by the orange balls (ISE data). Data are represented as mean ± SE.

(E) NH3 partial geometric current densities over 2D H-TaTe2, TaTe2 nanosheets, and bulk TaTe2 at different potentials. Data are represented as mean ± SE.

(F) UV-vis absorption spectra of the electrolyte after electrolysis at −0.12 V in either Ar-saturated electrolyte (Ar gas), or without catalyst (Carbon paper), or with the binder (Binder), or at an open circuit potential (Open circuit, i.e., I = 0).

(G) 1H NMR spectra for the electrolyte after electrolysis at −0.12 V in 14N2, 15N2, and Ar-saturated 0.1 M HCl over 2D TaTe2.

Particular care was taken when carrying out NRR to avoid false positives from background NH3 in the system or external contamination. Prior to NRR measurements, the N2 feeding gas was pre-purified to eliminate possible NH3 and labile nitrogen-containing contaminants (such as nitric oxides, nitrates, or nitrites). Spectrophotometric tests verified that almost no NO3– and NO2– existed in the N2-purged electrolyte (Figures S11–S13). To evaluate the NRR activity, electrochemical tests were conducted in 0.1 M HCl solutions saturated with purified N2 or Ar (Figure S14) was used as a feed gas. Linear sweep voltammetry curves of bulk TaTe2, 2D TaTe2, and H-TaTe2 revealed an onset cathodic current related to hydrogen evolution under both Ar and N2 conditions (Figure 3B). The 2D H-TaTe2 exhibited a lower current density compared with 2D TaTe2 in 0.1 M N2-purged HCl, resulting from weak binding of hydrogen atoms on the hydrophobic surface. The NRR was found to take place with an overpotential as low as 294 mV (given the equilibrium thermodynamic potential for N2 reduction to NH4+ is 0.274 V versus normal hydrogen electrode under 298 K and 1 atm) over TaTe2 (Figure 3C). Under the reaction conditions adopted, only NH4+ ions were detected by the indophenol blue method. No N2H4 as by-product was identified by the Watt and Chrisp method within the detection limit of the method. Also, no NO2− and NO3− were detectable, excluding the occurrence of oxidation of nitrogen to NO by OH. There are marginal NH4+ adsorbed by the Nafion 117 membrane in the H-type cell (Figures S15 and S16). The NH3 formation rate over the catalysts increased within the switching potential range from −0.02 to −0.12 V, but dropped when further elevating the overpotential probably due to limitation in desorption and transport of formed NH4+ out of the catalyst surface (Figures 3C and S17). TaTe2 nanosheets attained an averaged NH3 yield rate of 6.3 μgNH3 h−1 mgcat−1 with an FE of 12.9% at −0.12 V, nearly three times higher than that of bulk TaTe2 and also significantly surpassing Te, TeO2, Ta2O5, and corresponding hybrids with 2D TaTe2 (Figure S18). The total current density increased with overpotential approaching 0.71 mA mg−1 at −0.22 V (Figure S19). The conversions of N2 fed to NH3 were estimated to be approximately 3.8 × 10−4% at −0.12 V, considerably higher than that of a recently reported liquid H2O droplet plasma system. The NH3 FE first decreased from −0.02 to −0.07 V because the HER tended to dominate over NRR at low overpotentials, but increased with applied potentials upon stepping the voltage from −0.07 to −0.17 V, beyond which the FE for NH3 formation drastically declined owing to more severe competition from the HER. This may be linked to the possibility that high reaction rates increased the local pH value, consequently favoring alkaline water reduction, which could shift the reaction selectivity toward the HER. Moreover, at high current densities the available amounts of N2 may become mass transport determined and decreased as a result of NRR. Among others, the wetted surface area within the electrode itself and occupation of active sites by ∗H (decreasing the surface coverage of ∗N2) increased as a function of the applied reductive potential, which would also intensify the HER. The NH3 FE attained was up to ∼17.3% at −0.17 V, over seven times higher than that of bulk TaTe2. Notably, H-TaTe2 nanosheets reached an NH3 FE exceeding 32.0% with an NH3 formation rate of 5.8 μgNH3 h−1 mgcat−1 at −0.12 V. The NH3 FE could be further improved to 36.2% at −0.17 V, over 15 times higher than that of bulk TaTe2 (Figure 3D). The NH3 partial geometric current density of 2D H-TaTe2 was 10.8 μA cm−2, 2.3 times higher than that of bulk TaTe2 (Figure 3E). The cathodic energy efficiency at 0.28 mA cm−2 was evaluated to be ∼24.8% (Figure S20). The N2 reduction performance was found to be most optimal at pH 1.0 (Figure S21) and a working electrode loading density of 0.5 mgTaTe2 cm−2 (Figure S22). Alternatively, both NH3 production rate and FE were observed to increase with an increase of centrifugation speed from 500 to 1,000 rpm (Figure S23), indicating that increase in vacancy concentration led to improvement of NRR activity. Whereas further increase of centrifugation speed from 1,000 to 3,000 rpm to induce more defects (i.e., VTe) resulted in decreased NRR performance. This may be due to lowering in electric conductivity and simultaneously enhanced HER. It is worth noting that the 2D H-TaTe2 nanosheets outperform most previously reported transition metal- and precious metal-based electrocatalysts in terms of NH3 FE (Table S2).

To probe possible interference, if any, from the environment and assess the origin of the detected NH3, a set of rigorous control and verification experiments (for multiple repeats at each condition) were carried out. Almost no NH3 was obtained either in Ar-saturated solution (under the same conditions as the NRR experiments), or in the absence of catalyst, or with just the background Nafion solution binder, or at an open circuit potential (Figures 3F and 3G). Isotopic labeling using 15N2 in combination with isotope-sensitive proton nuclear magnetic resonance (1H NMR) was further performed. The 1H NMR spectrum of the NRR product exhibited a doublet coupling (∼73 Hz) typical for 15NH4+ compared with a triplet coupling (∼52 Hz) for 14NH4+ (Figure 3G). The absolute dominant 15NH4+ doublets confirmed that the N in NH3 stemmed from the gaseous N2 supplied. In addition, we would like to emphasize that the TaTe2 catalysts neither contain nitrogen in their structures nor are prepared from nitrates, nitrides, or ammonium precursors, excluding extraneous nitrogen sources. These strongly suggest that the NH3 was generated from the reduction of dissolved N2 accelerated by the 2D TaTe2 electrocatalyst.

To evaluate the stability of 2D H-TaTe2, we conducted alternating electrolysis between Ar and N2-saturated electrolytes, which showed that the NH3 evolved remained essentially unchanged for three cycles (Figure 4A). TaTe2 nanosheets also exhibited good stability with nearly constant NH3 yield rates, FEs, and partial geometric current densities (JNH3) for 20 times by use of 20 batches of 2D H-TaTe2 at −0.12 V (Figure 4B). Strikingly, negligible decay in JNH3 occurred even after continuous electrolysis for 100 h, indicating considerable catalytic stability of 2D H-TaTe2 (Figure 4C). After 100 h of NRR, the catalytic performance of post-NRR 2D H-TaTe2 and TaTe2 in fresh N2-saturated electrolyte was also determined. As displayed in Figure 4D, there are no obvious decreases in both NH3 yield and FE at −0.12 V, suggesting that H-TaTe2 nanosheets still maintained high activity after long-term use. No obvious leaching of TaTe2 into the electrolyte (<0.06 wt %) was observed even after 100 h of electrolysis by inductively coupled plasma atomic emission spectroscopy analysis. The stability of TaTe2 nanosheets was also investigated by in situ XRD analysis (Figures 5A and S24). It can be clearly seen that the peaks corresponding to TaTe2 remained consistent throughout the entire electrolysis, accounting for the good durability of such VTe-rich TaTe2. Post characterization by EPR (Figure S25) showed preservation of defects in 2D TaTe2 even after electrolysis, accounting for maintenance of the NRR performance.

Figure 4

Stability Test Results

(A) The NH3 yield rate (bar) and FE (ball) with alternated Ar and N2 cycles over 2D H-TaTe2 at −0.12 V.

(B) The NH3 yield rates (bar) and FEs (ball) for 20 times by use of 20 batches of 2D H-TaTe2 at −0.12 V.

(C and D) (C) Long time stability and (D) NH3 yield rates and FEs at −0.12 V over initial 2D H-TaTe2 and TaTe2 nanosheets after subjected to 100 h of operation. Data in (D) are represented as mean ± SE.

Figure 5

In-situ Measurements and Kinetics Analysis

(A) In situ XRD profiles of 2D TaTe2 at −0.12 V for 2 h of operation.

(B and C) (B) Tafel plots and (C) Nyquist profiles of bulk TaTe2, 2D TaTe2, and H-TaTe2.

To gain insight into the outstanding activity of TaTe2 nanosheets, the Tafel plot and electrochemical impedance were evaluated. The Tafel slope was ∼146.3 mV dec−1 for 2D H-TaTe2, ∼163.5 mV dec−1 for TaTe2 nanosheets, much lower than that of bulk TaTe2 (∼269.5 mV dec−1) (Figure 5B). This implies that exfoliated TaTe2 nanosheets possess more rapid reaction kinetics, and the first electron transfer process to yield ∗N–NH (∗ represents the surface adsorbed species) is the rate-determining step. Nyquist plots (Figure 5C) revealed a significantly lower charge transfer resistance for H-TaTe2 and TaTe2 nanosheets than that for bulk TaTe2, in accordance with its observed superior NRR activity.

To help understand the reaction mechanism for NRR and origin for enhanced NRR activity on TaTe2 nanosheets, we conducted DFT calculations. We used monoclinic-TaTe2, which agrees with the as-obtained crystalline structure of TaTe2 from the XRD results (Figure 1D). The (−603) facet, which is the basal plane of the exfoliated TaTe2, was considered. EPR characterization (Figure 1H) identified the emerging Te vacancy (VTe) after exfoliation, and thus we also considered the NRR activity at VTe sites. The VTe site is modeled by eliminating one Te atom in TaTe2(–603) (Figure S26), the most stable vacancy site is considered, denoted as VTe-TaTe2(–603). The TEM and STEM images (Figures 2B and 2I) showed the disordered edges. Making a representative DFT model for amorphous solid is limited due to the need for reasonable calculation times. Alternatively, we used the crystalline edge of TaTe2(010) instead, assuming that the local moiety of the amorphous edge to be similar to the crystalline edge. To investigate the effect of the TaO in the edge site, we modeled the oxygen atom containing TaTe2(010) edge sites. Also, we constructed Te(10-10) facet to consider the remaining Te which were found in XPS (Figure 1F). The optimized geometries of calculation models are shown in Figures 6A and S27. We consider surface Ta atom as an active site in TaTe2(010) and VTe-TaTe2(–603). For TaTe2(–603), Te atom is regarded as an active site since Ta atom is not exposed.

Figure 6

Calculation Models and Free Energy Diagrams for NRR

(A) Top view (upper panel) and side view (lower panel) of the optimized structures of TaTe2(–603) (left) and TaTe2(010) (right).

(B) The free energy change for ∗N–NH formation on TaTe2(–603), VTe-TaTe2(–603), and TaTe2(010).

(C) The free energy diagram for NRR at 0 V (versus RHE) on VTe-TaTe2(–603).

(D) The free energy diagram for NRR at 0 V (versus RHE) on TaTe2(010). Free energies for less stable intermediate are represented by the blue line. Sky blue, yellow, blue, and white balls represent Ta, Te, N, and H atoms, respectively.

Among the many intermediate steps in NRR, we first focused on ∗N–NH formation. This has been identified as the largest free energy demanding step in various catalysts and, hence, the ∗N–NH formation energy is a good descriptor for estimating the NRR activity. We compared the ∗N–NH formation free energy on (−603), VTe-TaTe2(–603), and TaTe2(010), obtained by G(∗N–NH)–G(∗)–G(N2(g))–G(H+ + e−) (Figure 6B). The ∗N–NH formation free energy is noticeably decreased on TaTe2(010) (1.47 eV) and VTe-TaTe2(–603) (0.82 eV) compared with that on TaTe2(−603) (2.53 eV). We also found that N2 does not chemically bind at Te atom in TaTe2(–603) (Figure S28). This result indicates that the Ta atom site rather than Te atom facilitates ∗N–NH formation and plays an important role in enhancing N2 activation. Meanwhile, the ∗N–NH formation energies of Te(10-10) (3.41 eV) and TaO (1.62–3.44 eV) are higher than that of VTe-TaTe2(–603) (0.82 eV) or TaTe2(010) (1.47 eV) (Figure S27C). Thus, the Te and TaO are less reactive for NRR than VTe-TaTe2 and TaTe2(010), which qualitatively agrees with the experimental result (Figure S18). We further conducted DFT calculations with the oxygen-doped TaTe2(010), which showed the reliable ∗N–NH formation energy (1.66 eV, 1.63 eV) to consider the effect of the oxygen.

Next, we obtained a free energy diagram for NRR and identified the lowest free energy pathway. The NRR activity is estimated by comparing the free energy change at the potential-determining step (PDS), the largest free energy requiring electrochemical step. The PDS is ∗N2 + (H+ + e–) → ∗N–NH on TaTe2(–603) and TaTe2(010) (Figures S28 and 6D), while that on VTe-TaTe2(–603) is ∗NH2 + (H+ + e–) → ∗NH3 (Figure 6C). We noted that a stronger N binding on VTe-TaTe2(–603) than TaTe2(010) and TaTe2(–603) leads to the different PDS. The theoretical volcano plot for NRR on transition metal surfaces showed a trend that the PDS of strong N binding metal and weak N binding metal is usually ∗NH2 + (H+ + e–) → NH3 and ∗ + N2 + (H+ + e–) → ∗N–NH, respectively. Also, we found that ∗N–NH, the first electron transfer step, shows the highest apparent energy in the free energy diagram of VTe-TaTe2(–603) and TaTe2(010). This result matches the Tafel slope analysis, which reveals that the rate-determining step is the first electron transfer step.

As summarized in Figure S30, the free energy change at PDS increases in the order of TaTe2(010) < VTe-TaTe2(–603) < TaTe2(–603). Interestingly, the free energy change at PDS on TaTe2(010) and VTe-TaTe2(–603) is decreased by ∼1.8 eV compared with that of TaTe2(–603), indicating that the exposed Ta atoms in TaTe2, rather than surface Te atom, act as the major active sites for NRR. We note that the free energy change at PDS on TaTe2(010) and VTe-TaTe2(–603) (0.95 eV) is comparable with the values (∼1.0 eV) at the top of the NRR volcanoes on various transition metal surfaces. In the case of oxygen-doped TaTe2(010) planes, the PDS is ∗N2 + (H+ + e–) → ∗N–NH and the free energy changes at the PDS are 0.96 and 0.93 eV (Figure S29), which are similar to that of TaTe2(010) (0.95 eV). These results indicate that the presence of an amorphous TaO may not have a critical effect on catalytic activity.

Conclusions

In summary, we have demonstrated high yield of stably dispersed few-layer metallic TaTe2 nanosheets rich in Te vacancies by ultrasonication of bulk TaTe2 in γ-butyrolactone. The defective 2D TaTe2 efficiently facilitated electrochemical N2 fixation, delivering a large NH3 FE (∼12.9%) and high NH3 formation rate (∼6.3 μgNH3 h−1 mgcat−1) at an applied potential of −0.12 V. Facile modification of TaTe2 electrodes by using trimethoxy (1H,1H,2H,2H heptadecafluorodecyl) silane substantially inhibited the parasitic HER and greatly increased the FE for NH3 formation, reaching ∼32.2% at −0.12 V with an NH3 yield rate of 5.8 μgNH3 h−1 mgcat−1. The NH3 FE was further improved to ∼36.2% at −0.17 V, over 15.4 times higher than on bulk TaTe2. Of particular interest is that the H-TaTe2 nanosheets retained NRR performance even after 100 h of operation. DFT calculations showed that the exposed Ta atom sites dramatically enhanced N binding and decreased the free energy change at PDS, thereby boosting N2 reduction. We envision that the integral strategy by engineering anion vacancies and surface hydrophobicity of 2D TMDs will offer profound implications for design and preparation of efficient NRR electrocatalysts.