English

• Since the Industrial Revolution, excessive carbon dioxide (CO₂) emissions have emerged as a significant concern due to their potential to exacerbate global warming, contribute to sea level rise, and harm ecosystems [1]. The conversion of CO₂ into high-value-added products through methods, such as electrocatalysis, photocatalysis, and thermal catalysis, represents an effective approach to utilize CO₂ resources and advance towards achieving carbon neutrality [2-4]. Electrocatalysis distinguishes itself among the aforementioned methods for its inherent simplicity, operability, and the ability to harness electricity generated from renewable resources [4,5]. However, the electrocatalytic reduction of CO₂ encounters significant thermodynamic barriers and slow reaction kinetics, especially during the conversion of the stable linear CO₂ molecule into the *CO₂⁻ radical anion [5]. Thus, the advancement of efficient and durable electrocatalysts with high current density and superior selectivity for desired production is essential for enhancing the kinetics of the CO₂ electroreduction.

  Currently, numerous researchers are actively exploring the applications of non-precious transition metal-based electrocatalysts in the field of the CO₂ electroreduction, due to their abundant reserves and competitive catalytic performance [2,6-8]. As illustrated in Table S1 (Supporting information), various classes of transition metal compounds have been explored for the CO₂ electroreduction, including metal nitrides, sulfides, phosphides, and antimonides. These electrocatalysts have demonstrated good catalytic performance for the CO₂ electroreduction, with certain systems achieving remarkable Faradaic efficiencies exceeding 90% for target products under optimized conditions. Despite these advances, metal borides remain relatively underexplored for the CO₂ electroreduction compared to other transition metal compounds. Recently, some researchers have discovered that incorporating trace B (< 10 wt%) into metal structures as B-doped catalysts could result in a redistribution of electron states in metal electrocatalysts [9-14]. This change could alter the adsorption of reaction intermediates, thereby impacting the catalytic performance in the electrocatalytic reduction of CO₂ [15]. For example, introducing B atoms (2.3 wt%) into bismuth (Bi) could alter its local electronic structure, thereby enhancing the catalytic activity for CO₂ electroreduction and improving the selectivity for HCOOH over a wide potential range [16]. The intercalation of B atoms (1.26 wt%) could also adjust the average oxidation state of copper (Cu), control *CO adsorption and dimerization, and then promote the generation of C₂ products [17]. Moreover, a B-doped (6.3 at%) palladium catalyst (Pd-B/C) was more effective in boosting the formation of HCOOH compared to Pd/C [18]. These results indicated that the introduction of trace B element in metal structures could affect the activity and selectivity of electrocatalytic reduction of CO₂. Therefore, we speculate that transition metal borides with a precisely defined structure, higher B content and larger CO₂ adsorption capacity [19,20] hold great potential to serve as cost-effective and high-efficiency candidates for the electrocatalytic reduction of CO₂.

  Notably, transition metal borides exhibit diverse composition, structural, and electronic characteristics, primarily including metal-rich borides (M_xB, x > 1), monoborides (MB), metal diborides (MB₂), and boron-rich borides (MB_x, x > 2) [21]. Among these borides, the well-known hexagonal AlB₂-type (P6/mmm) metal borides with high electrical conductivity and low electrical resistivity have attracted significant attention from researchers [22]. This kind of borides is characterized by close-packed transition metal sheets and graphene-like B layers, and can be synthesized with a variety of transition metals [23]. The presence of graphene-like B layers played an important role in enhancing electronic conductivity and providing more active sites [24]. It is worth noting that the choice of transition metals and the interaction between B and transition metals can have a profound effect on the physical and chemical properties of transition metal diborides, particularly in terms of their electrocatalytic performance [25]. Moreover, Li et al. theoretically demonstrated that monolayer two-dimensional transition-metal diborides exhibited a remarkable catalytic activity in selectively converting CO₂ to CH₄ [26]. Additionally, hard/superhard transition metal diborides obtained through the high pressure-high temperature (HPHT) are integral bulks, which can be directly utilized as self-supporting electrodes for electrocatalytic reduction of CO₂ [27]. In comparison to traditional powder electrocatalysts, these self-supporting electrodes can not only prevent the aggregation of electrocatalyst or the concealment of active sites caused by binder addition [28], but also eliminate the interface impedance between the conductive substrate and electrocatalyst. This facilitates efficient electron and proton transfer [29], which will ultimately enhance the catalytic performance for the electrocatalytic reduction of CO₂. The aforementioned facts motivate us to further explore the catalytic performance of various AlB₂-type metal borides self-supporting electrodes for the electrocatalytic reduction of CO₂.

  In this study, the AlB₂-type metal diborides (TM = Ti, Zr, Hf, V, Nb, Ta, Cr, Mn, Mo) were fabricated using the HPHT method (Fig. 1a). The detailed temperature, pressure, and duration parameters for preparing different metal diborides are provided in Table S2 (Supporting information). The X-ray diffraction (XRD) patterns of different samples in Fig. 1b confirm the successful synthesis of the hexagonal AlB₂-type metal diborides. Meanwhile, it can also be seen that these transition metal diborides synthesized by the HPHT method exhibit high crystallinity and purity. As depicted in Fig. 1c, the Rietveld refinement of XRD pattern reveals that the ZrB₂ sample corresponds to a hexagonal unit cell with the space group P6/mmm (No. 191) [30]. The lattice parameters for the ZrB₂ sample are a = b = 3.16067(3) and c = 3.52094(6) (Table S3 in Supporting information), showing excellent agreement with values previously documented in the literature. Fig. 1d shows the comparison of average distance for adjacent metal-B and B-B atoms in different metal diborides. Although the average distances of metal-B and B-B may change with the variation of metal atoms, these metal diborides adapts the typical AlB₂-type structure characterized by graphene-like B layers sandwiched between transition metal layers (Fig. S1 in Supporting information). To further investigate the electronic structure of the ZrB₂ sample, the electronic characteristic was examined using electron localization function (ELF) and density of states (DOS). The ELF analysis (Fig. 1e) reveals that there are ionic Zr-B and covalent B-B bonds in the ZrB₂ sample. Meanwhile, DOS value is not zero at Fermi energy level for the ZrB₂ sample (Fig. 1f), indicating the intrinsic metallic behavior of the ZrB₂ sample. In addition, as confirmed by the DOS result (Fig. 1f), the metallic property of the ZrB₂ sample mainly originates from d electrons of Zr layers, which can enhance its catalytic activity.

  Figure 1

  

  Figure 1.  (a) Schematic illustration of the HPHT method and procedure. (b) The XRD patterns of different metal diborides. (c) The Rietveld refinement of XRD pattern for the ZrB₂ sample. (d) The average distance of adjacent metal-B and B-B atoms in different metal diborides. (f) ELF of different planes for the ZrB₂ sample. (f) DOS of the ZrB₂ sample.

  DownLoad: Full-Size Img PowerPoint

  The morphologies of various metal diborides were also measured by scanning electron microscopy (SEM). In Fig. S2 (Supporting information), it can be observed that the surfaces of the obtained bulk metal diborides exhibit a rough and irregular texture, resulting from the abrasion of diamond griding disk. Meanwhile, high-resolution SEM images (Fig. S3 in Supporting information) reveal that different metal diborides were all composed of tightly packed, irregular nanoblocks. To further explore the morphology and structure of metal diborides, we conducted transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) tests using the ZrB₂ sample as an example. Combining high-resolution SEM (Fig. S3b) and TEM images (Fig. 2a), it can be observed that the ZrB₂ sample possesses a densely packed nano-block structure with grain size around 0.3–1.0 µm. In the HRTEM image (Fig. 2b) of the ZrB₂ sample, the interplanar spacings of 0.272 nm and 0.333 nm can be attributed to the (100) and (001) planes of ZrB₂ [31], respectively. The selected area electron diffraction (SAED) image (Fig. 2c) of the ZrB₂ sample corresponds to the (101) and (100) planes of the hexagonal ZrB₂ phase [32]. These findings are in agreement with the XRD results (Fig. 1b). The energy dispersive X-ray spectroscopy (EDS) mapping (Fig. 2d) of the ZrB₂ sample reveals a uniform distribution of the elements Zr (green) and B (red). A small amount of oxygen (O) elements is also observed on the surface of the ZrB₂ sample, which may be attributed to surface oxidation when exposed to air. To further investigate the chemical states of the elements in the ZrB₂ sample, X-ray photoelectron spectroscopy (XPS) spectra were analyzed, with the results shown in Figs. 2e and f, and Fig. S4 (Supporting information). The high-resolution XPS spectrum of Zr 3d (Fig. 2e) consists of four peaks. The binding energies at 178.7 eV (Zr 3d_5/2) and 181.2 eV (Zr 3d_3/2) correspond to Zr-B bonds in ZrB₂, while the other two peaks located at 183.0 eV (Zr 3d_5/2) and 185.5 eV (Zr 3d_3/2) are attributed to Zr-O bonds of ZrO₂ [31,33,34]. The B 1s XPS spectrum (Fig. 2f) can be resolved into two peaks. The peak at approximately 187.5 eV is assigned to B-Zr bonds in ZrB₂, while the peak at 187.9 eV belongs to B-B bonds [31,33,35]. In the O 1s spectrum (Fig. S4), the peak with a binding energy of 531.8 eV corresponds to O-Zr bonds in ZrO₂ [31,34], which may come from the surface oxidation of the ZrB₂ sample in air, consistent with the EDS mapping result (Fig. 2d). Moreover, XPS spectra of other metal diborides are presented in Figs. S5-S12 (Supporting information), respectively. The above results indicate highly quality AlB₂-type metal diborides bulks can be synthesized using the HPHT method, accompanied by slight surface oxidation when exposed to air. This distinctive structure confers special physical and chemical properties upon them, potentially enhancing their effectiveness in facilitating electrocatalytic reactions [36].

  Figure 2

  

  Figure 2.  The characterization results of the ZrB₂ sample: (a) TEM image, (b) HRTEM image, (c) SAED image, (d) EDS mapping images, (e) high-resolution XPS spectrum of Zr 3d, and (f) high-resolution XPS spectrum of B 1s.

  DownLoad: Full-Size Img PowerPoint

  To assess the catalytic performance of as-synthesized metal diborides for the electrocatalytic reduction of CO₂, a series of electrochemical tests were conducted in a H-type cell employing a 1 mol/L [Bmim]PF₆-containing acetonitrile ([Bmim]PF₆(1M)-MeCN) solution as the electrolyte. In this study, the metal diborides, including TiB₂, ZrB₂, HfB₂, VB₂, NbB₂, TaB₂, CrB₂, MnB₂, and MoB₂, formed integral bulk materials, which were directly utilized as self-supporting electrodes for the electrocatalytic reduction of CO₂. The controlled potential electrolysis was initially carried out over different metal diborides at −2.2 V vs. Ag/Ag⁺ in CO₂-saturated [Bmim]PF₆(1M)-MeCN electrolytes. The products were detected by gas chromatography (GC) and nuclear magnetic resonance (NMR) spectroscopy. Upon analysis, it was found that the main gaseous products on these metal diborides were CO and H₂, with no liquid products being generated (Fig. 3a). At −2.2 V vs. Ag/Ag⁺, the ZrB₂ electrode exhibited a CO Faradaic efficiency of 92.2%, significantly outperforming other metal diborides. Meanwhile, the current densities of different electrodes were normalized according to their geometric surface areas. In Fig. 3a, the ZrB₂ electrode achieved a total current density of 51.5 mA/cm², which was 3.4, 3.4, 2.5, 2.1, 4.4, 4.7, 1.8, and 3.9 times higher than those of TiB₂, HfB₂, VB₂, NbB₂, TaB₂, CrB₂, MnB₂, and MoB₂ electrodes, respectively. These results suggest that the type of metal elements in the AlB₂-type metal diborides has a significant impact on the electrochemical reduction of CO₂ to CO. Meanwhile, linear sweep voltammetry (LSV) curves of the ZrB₂ electrode were recorded in Ar- or CO₂-saturated electrolytes (Fig. S13 in Supporting information), which indicated that the ZrB₂ electrode exhibits the ability to catalyze CO₂ reduction. In addition, to confirm the carbon source of CO product on the ZrB₂ electrode, the controlled potential electrolysis experiment was conducted under the same conditions, except that the CO₂ atmosphere was replaced with the Ar atmosphere. From Fig. S14 (Supporting information), it can be observed that no CO was detected under the Ar atmosphere, indicating that CO primarily originated from the reduction of CO₂. In a more comprehensive investigation, the relationship between the current density (or CO Faradaic efficiency) and the applied potential was analyzed for each metal diboride over a potential range from −2.0 V to −2.5 V vs. Ag/Ag⁺ (Fig. 3b, Figs. S15 and S16 in Supporting information). When the applied potential was shifted negatively, the current densities over various metal diborides exhibited increasing trends (Fig. 3b and Fig. S15), and the Faradaic efficiencies of CO on different metal diborides followed volcano plots (Fig. 3b and Fig. S16). Impressively, the ZrB₂ electrode demonstrated an outstanding CO Faradaic efficiency of 92.2% at a potential of −2.2 V vs. Ag/Ag⁺, accompanied by a current density of 51.5 mA/cm² (Fig. 3b). As the applied potential decreased to values less than −2.2 V vs. Ag/Ag⁺, the CO Faradaic efficiencies on the ZrB₂ electrode correspondingly diminished. As shown in Figs. S15 and S16, only the MnB₂ electrode achieved the maximum CO Faradaic efficiency (74.7%) at −2.2 V vs. Ag/Ag⁺, with corresponding current density of 34.9 mA/cm². The maximum CO selectivity at the surfaces of HfB₂ (72.0%), VB₂ (61.2%), NbB₂ (51.9%), and CrB₂ (67.0%) electrodes was observed at −2.3 V vs. Ag/Ag⁺, with corresponding current densities of 49.5, 27.9, 43.5, and 27.7 mA/cm², respectively. Meanwhile, the highest CO Faradaic efficiencies on TiB₂, TaB₂, and MoB₂ electrodes were 72.6%, 84.8%, and 82.4% at −2.4 V vs. Ag/Ag⁺, respectively. Although all of these metal diborides had the ability to electrocatalytically convert CO₂ into CO, the ZrB₂ electrode can efficiently facilitate this conversion and achieve the highest CO Faradaic efficiency (92.2%) at a relatively lower applied potential (−2.2 V vs. Ag/Ag⁺) compared to other metal diborides, indicating that the ZrB₂ electrode was more favorable for the electrocatalytic reduction of CO₂ to CO.

  Figure 3

  

  Figure 3.  (a) The current density and CO Faradaic efficiency on various metal diborides at −2.2 V vs. Ag/Ag⁺. (b) The current density and CO Faradaic efficiency on the ZrB₂ electrode at different applied potentials. (c) Long-term stability of the ZrB₂ electrode in CO₂-saturated [Bmim]PF₆(1M)-MeCN electrolytes at −2.2 V vs. Ag/Ag⁺. (d) The duration and current density of different B-containing electrocatalysts for the electrocatalytic reduction of CO₂ to CO [9-13].

  DownLoad: Full-Size Img PowerPoint

  To further investigate the practicality of the ZrB₂ electrode, we conducted long-term stability testing in [Bmim]PF₆(1M)-MeCN electrolytes at a potential of −2.2 V vs. Ag/Ag⁺. Considering the volatilization of MeCN and the accumulation of carbonaceous compounds on the ZrB₂ electrode surface during the electrolysis experiment, the electrolyte was changed and the electrode surface was polished by diamond griding disk every 5–7 h. It is evident from Fig. 3c that the current density and CO Faradaic efficiency remained relatively stable over the 61-h electrolysis period. Combining the SEM image and XPS spectra of the used ZrB₂ electrode (Fig. S17 in Supporting information), we discovered the good stability of the ZrB₂ electrode in effectively catalyzing the electroreduction of CO₂ to CO over prolonged operational periods. Despite the relatively stable current density maintained over the 61-h electrolysis period, we observed minor fluctuations in the current density for the electrocatalytic reduction of CO₂ with the electrolysis time. These fluctuations might be linked to factors such as the replacement of the electrolyte and the polishing of the electrode surface. Furthermore, slight variations in the charge-transfer resistance (R_ct) of the ZrB₂ electrode before and after the CO₂ electrolysis, as revealed by electrochemical impedance spectroscopy (EIS) measurements, could also contribute to changes in the current density [37]. As shown in Fig. S18 (Supporting information), the ZrB₂ electrode demonstrated relatively low R_ct values both before and after CO₂ electrolysis. This suggested that the ZrB₂ electrode possessed favorable electronic transfer kinetics, thereby facilitating the electrocatalytic reduction of CO₂. Moreover, a slight decrease in the R_ct values of the ZrB₂ electrode was observed from 3.2 Ω pre-electrolysis to 2.2 Ω post-electrolysis conditions, indicating a minor reduction in R_ct values that could influence the electrocatalytic performance.

  To better demonstrate the advantages of the self-supporting ZrB₂ electrode, we prepared a comparative ZrB₂ powder as a control sample. The ZrB₂ powder was drop-casted onto carbon paper to fabricate a conventional electrode, designated as the p-ZrB₂ electrode. In Fig. S19a (Supporting information), the current density and CO Faradaic efficiency of the p-ZrB₂ electrode were lower than those of the self-supporting ZrB₂ electrode across the entire potential range, primarily due to the adverse effects of the Nafion binder required for its preparation. The binder might not only induce aggregation of ZrB₂ particles but also block active sites, thereby impairing the electrocatalytic performance of the CO₂ reduction. Moreover, the p-ZrB₂ electrode demonstrated limited stability (Fig. S19b in Supporting information), maintaining the CO Faradaic efficiency and current density for only 3 h, as the crushed ZrB₂ sample with relatively large particles tended to detach from the electrode surface during the electrolysis. Further EIS measurements (Fig. S19c in Supporting information) revealed that the superior performance of the self-supporting ZrB₂ electrode might correlate with its lower R_ct value, which facilitated faster electron transfer rates and consequently enhanced the CO₂ reduction process. To the best of our knowledge, the stability and current density for the efficient electrocatalytic reduction CO₂ to CO observed on the ZrB₂ electrode surpasses most recently reported state-of-the-art B-containing electrocatalysts (Fig. 3d and Table S4 in Supporting information) [9-13], which is beneficial for the industrial application and development of CO₂ electroreduction. Meanwhile, the performance of Zr-containing electrocatalysts for the electrocatalytic reduction of CO₂ is also shown in Table S5 (Supporting information). It can be observed that the primary product of these Zr-containing electrocatalysts for this reaction was mainly CO, with the exception of the bimetallic Cu-Zr electrocatalyst (ZrO₂/Cu-Cu₂O) [38], whose main product was ethylene (C₂H₄). Compared to Zr-containing electrocatalysts that produced CO, the ZrB₂ electrode exhibits superior electrocatalytic performance. In addition, the ZrB₂ electrode displayed better stability compared to the transition metal compounds reported in the literature (Table S1).

  It is widely recognized that the composition of the electrolyte plays a significant role in regulating the catalytic activity and product distribution of the electrocatalytic reduction of CO₂. In this study, we first evaluated the catalytic performance of the ZrB₂ electrode in an aqueous solution with potassium bicarbonate (KHCO₃, 1 mol/L) as supporting electrolyte and a non-aqueous electrolyte, specifically an MeCN solution with a supporting electrolyte of tetrabutylammonium hexafluorophosphate (TBAPF₆(1M)-MeCN). In 1 mol/L KHCO₃ aqueous electrolyte, the ZrB₂ electrode exhibited almost no catalytic activity towards CO₂ reduction, with the main product being H₂. It can be seen in Fig. 4a that the Faradaic efficiency of H₂ over the ZrB₂ electrode was approximately 100.0% at all applied potential. In contrast, the highest CO Faradaic efficiency for the ZrB₂ electrode was 64.1% at −2.5 V vs. Ag/Ag⁺ in TBAPF₆(1M)-MeCN electrolytes, with a current density of 11.7 mA/cm² (Fig. 4b). The high catalytic activity and selectivity of CO observed in the TBAPF₆(1M)-MeCN electrolyte compared to 1 mol/L KHCO₃ electrolyte suggest that the non-aqueous solution containing TBAPF₆ contributes to enhancing the catalytic performance of the ZrB₂ electrode. Based on this, we endeavored to utilize the ZrB₂ electrode in non-aqueous electrolytes to investigate its electrocatalytic performance for CO₂ reduction. In non-aqueous electrolytes, ionic liquids are commonly used as a supporting electrolyte for the electrocatalytic reduction of CO₂, especially those including imidazole cations. These ionic liquids can regulate the adsorption of CO₂ and its intermediates, thereby changing the catalytic activity of the electrode and achieving a co-catalytic effect [39,40]. Thus, the catalytic activity of the ZrB₂ electrode was further tested in MeCN solutions consisting of imidazole ionic liquids, including [Bmim]PF₆, 1‑butyl‑3-methylimidazolium perchlorate ([Bmim]ClO₄), 1‑butyl‑3-methylimidazolium acetate ([Bmim]OAc), 1‑butyl‑3-methylimidazolium nitrate ([Bmim]NO₃), and 1-ethyl-3-methylimidazolium hexafluorophosphate ([Emim]PF₆) (Fig. 4c). The LSV curves in Fig. S20 (Supporting information) show that the ZrB₂ electrode exhibited a higher catalytic activity for CO₂ electroreduction in the [Bmim]PF₆(1M)-MeCN electrolyte compared to electrolytes containing other ionic liquids. In Fig. 4c, the current densities and CO Faradaic efficiencies in [Bmim]PF₆(1M)-MeCN and [Emim]PF₆(1M)-MeCN electrolytes at −2.2 V vs. Ag/Ag⁺ were higher than those in TBAPF₆(1M)-MeCN electrolyte, indicating that ionic liquids with imidazole cations in the electrolytes could promote the generation of CO. In the [Emim]PF₆(1M)-MeCN electrolyte, the CO Faradaic efficiency (88.7%) and total current density (39.2 mA/cm²) were notably lower than those in the [Bmim]PF₆(1M)-MeCN electrolyte. This discrepancy may be attributed to the cation variation in the ionic liquids, influencing the solubility of CO₂ and thus modulating the catalytic performance of the electrocatalytic reduction of CO₂ [41-43]. Furthermore, substituting the anions in the [Bmim]PF₆ with ClO₄⁻, OAc⁻, and NO₃⁻ resulted in decreased reaction kinetics and CO Faradaic efficiency on the ZrB₂ electrode. This result could be a consequence of these anions altering the properties of the electric double layer near the surface of the ZrB₂ electrode, thereby decreasing the rate of CO generation [44]. Meanwhile, the strong association between the PF₆⁻ anion and the CO₂ molecule can improve the solubility of CO₂. Based on the above results, it can be concluded that the supporting electrolyte [Bmim]PF₆ was beneficial for the electrocatalytic reduction of CO₂. The perfect combination of an MeCN solution with a supporting electrolyte of [Bmim]PF₆ and the ZrB₂ electrode may promote CO₂ adsorption, which is crucial for effective CO₂ electroreduction. Fig. 4d also displays the impact of varying concentrations of [Bmim]PF₆ in the electrolyte on the electrocatalytic activity for the CO₂ electroreduction reaction using the ZrB₂ electrode at −2.2 V vs. Ag/Ag⁺. In the pure acetonitrile electrolyte, only trace amounts of CO were detected with negligible current density. This was due to the inherently low ionic conductivity of acetonitrile as a non-aqueous solvent, which severely limited charge transport and impeded the kinetics of the electrocatalytic reduction of CO₂. As [Bmim]PF₆ concentration increased (< 1 mol/L), both current density and CO Faradaic efficiency showed progressive enhancement, attributable to improved solution conductivity through ionic dissociation ([Bmim]⁺ and [PF₆]⁻) and facilitated charge transfer processes. However, exceeding 1 mol/L of [Bmim]PF₆ in the electrolyte led to declining performance metrics, primarily due to increased solution viscosity and intensified ion-pair ([Bmim]⁺ and [PF₆]⁻) interactions that collectively impeded mass transport and charge transfer kinetics. These observations established the optimal concentration of [Bmim]PF₆ was 1 mol/L, effectively enhancing the current density and CO Faradaic efficiency of the electrocatalytic CO₂ reduction.

  Figure 4

  

  Figure 4.  The current density and CO Faradaic efficiency on the ZrB₂ electrode in different electrolytes: (a) 1 mol/L KHCO₃ aqueous solution, (b) TBAPF₆(1M)-MeCN electrolyte, (c) different ionic liquids(1M)-MeCN electrolyte, and (d) [Bmim]PF₆−MeCN electrolyte with varying concentrations of [Bmim]PF₆. (e) In-situ ATR-SEIRS spectra on the ZrB₂ electrode collected within the potential range of −1.8 V to −2.2 V vs. Ag/Ag⁺. (f) The energy barriers for *CO desorption on different ZrB₂ surfaces.

  DownLoad: Full-Size Img PowerPoint

  In order to delve deeper into the potential reaction pathway for the electrocatalytic reduction of CO₂ to CO, in-situ attenuated total reflection surface-enhanced infrared reflection spectroscopy (ATR-SEIRS) was utilized to identify the reaction intermediates on the ZrB₂ electrode (Fig. 4e). The spectra were obtained within the potential range of −1.8 V to −2.2 V vs. Ag/Ag⁺, using the measurement at open circuit potential (OCP) as the reference. It is noticeable that no vibration peak is observable in the spectra of the ZrB₂ electrode at OCP. In the potential range of −1.8 V to −2.2 V vs. Ag/Ag⁺, newly emerged peaks at ~1710 cm⁻¹ could be assigned to the C=O stretching of the *COOH intermediate [45-49] during the electrocatalytic reduction of CO₂. Meanwhile, the peaks observed at ~2090 cm⁻¹ corresponded to *CO species, with their weak intensities possibly resulting from the rapid consumption of *CO for the formation of CO [50]. Moreover, the prominent characteristic peaks at ~2345 cm⁻¹ are attributed to the adsorbed CO₂ on the surface of the ZrB₂ electrode. As the applied potential decreased, the intensity of these peaks gradually strengthened, indicating an increase in the consumption of the adsorbed CO₂ [51]. Based on the above results, potential reaction pathway for the electrocatalytic reduction of CO₂ to CO on the ZrB₂ electrode can be speculated as follows: (1) ^* + CO₂ → ^*CO₂; (2) ^*CO₂ + H⁺ + e⁻ → ^*COOH; (3) ^*COOH + H⁺ + e⁻ → ^*CO; (4) ^*CO → ^* + CO, where ^* represents the adsorption site on the surface of the ZrB₂ electrode.

  DFT calculations were also carried out to clarify the intrinsic activity for the electrocatalytic performance in CO₂ reduction on the ZrB₂ surface. The B-terminated ZrB₂ (001) and Zr-terminated ZrB₂ (001) were utilized as model electrocatalysts to explore the active sites of the ZrB₂ electrode for the electroreduction of CO₂ (Figs. S21 and S22 in Supporting information). In Fig. S23 (Supporting information), we found that the desorption of *CO is an endergonic process on both the B-terminated ZrB₂ (001) and Zr-terminated ZrB₂ (001) surfaces. Compared to the B-terminated ZrB₂ (001) model (1.84 eV), the Zr-terminated ZrB₂ (001) model exhibits a lower energy barrier for *CO desorption (0.76 eV) suggesting that metal edge sites of the ZrB₂ (001) surface can enhance *CO desorption, thereby promoting the formation of CO. To further validate the role of metal edge sites in the ZrB₂ electrode, the energy barriers for *CO desorption on the ZrB₂ (101) and ZrB₂ (100) surfaces, which can be observed in the XRD and TEM results (Figs. 1b, 2b and 2c), were also investigated. The results in Fig. 4f further demonstrate that the Zr-terminated ZrB₂ surface can facilitate *CO desorption compared to the B-terminated ZrB₂ surface, indicating that Zr sites in the ZrB₂ electrode are the primary active sites for the electrocatalytic reduction of CO₂ to CO. Moreover, the thermodynamic limiting potentials for the electrocatalytic reduction of CO₂ (U_L(CO₂)) and HER (U_L(H₂)) on the Zr-terminated ZrB₂ (001) are shown in Fig. S24a (Supporting information). The difference (ΔU_L) between U_L(CO₂) and U_L(H₂) is a positive value, which means a higher selectivity of the electrocatalytic reduction of CO₂ [52]. This is also consistent with experimental findings (Fig. 3b). These results indicate that the metal sites on the ZrB₂ (001) surface can lower the desorption energy of *CO intermediates, suppress the HER, and ultimately enhance the activity and selectivity for the electrocatalytic reduction of CO₂ to CO.

  In conclusion, we have successfully developed a range of bulk metal diborides through the HPHT method, utilizing them as self-supporting electrodes for the electrocatalytic reduction of CO₂ in ionic liquid-based electrolytes. The as-obtained ZrB₂ electrode demonstrated outstanding electrocatalytic performance in the conversion of CO₂ to CO, achieving an impressive Faradaic efficiency for CO of up to 92.2% at −2.2 V vs. Ag/Ag⁺. More importantly, the ZrB₂ electrode also exhibited extraordinary stability, lasting up to ~60 h. Mechanistic investigations corroborated that metal sites on the ZrB₂ (001) surface played multiple roles of facilitating the desorption of *CO intermediates, inhibiting the HER, thereby accelerating the formation of the CO product. The superior electrocatalytic performance of AlB₂-type metal borides positions them as highly promising candidates for CO₂ conversion applications, thereby likely to spur the progression and exploration of other metal borides in the field of CO₂ electroreduction.

  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.

  CRediT authorship contribution statement

  Yidan Mao: Writing – original draft, Data curation, Investigation. Bingyu Li: Data curation, Investigation, Validation. Shuailing Ma: Methodology, Funding acquisition. Siwen Cui: Methodology, Investigation. Zihan Zhang: Data curation, Validation. Pinwen Zhu: Supervision, Funding acquisition. Kongsheng Qi: Formal analysis, Validation. Xiaodong Li: Supervision, Writing – review & editing. Weiwei Dong: Supervision, Writing – review & editing, Formal analysis, Funding acquisition. Wei Luo: Writing – review & editing, Funding acquisition. Rajeev Ahuja: Funding acquisition, Writing – review & editing. Dexin Yang: Writing – review & editing, Funding acquisition, Conceptualization, Supervision. Tian Cui: Supervision, Writing – review & editing, Funding acquisition.

  Acknowledgments

  The authors thank the National Natural Science Foundation of China (Nos. 22003058, 12204254), National Key Research and Development Program of China (No. 2023YFA1608902), the Program for Science and Technology Innovation Team in Zhejiang (No. 2021R01004) and the National Major Science Facility Synergetic Extreme Condition User Facility Achievement Transformation Platform Construction (No. 2021FGWCXNLJSKJ01). Wei Luo and Rajeev Ahuja thank Swedish Research Council (VR) (No. 2020–04410) for financial support.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111675.

  
  
• 1. [1]

     B. Ekwurzel, J. Boneham, M.W. Dalton, et al., Clim. Change 144 (2017) 579–590. doi: 10.1007/s10584-017-1978-0

  2. [2]

     D. Johnson, Z. Qiao, A. Djire, ACS Appl. Energy Mater. 4 (2021) 8661–8684. doi: 10.1021/acsaem.1c01624

  3. [3]

     G. Xie, W. Guo, Z. Fang, et al., Angew. Chem. Int. Ed. 63 (2024) e202412568. doi: 10.1002/anie.202412568

  4. [4]

     M. Melchionna, P. Fornasiero, M. Prato, M. Bonchio, Energy Environ. Sci. 14 (2021) 5816–5833. doi: 10.1039/D1EE00228G

  5. [5]

     S. Das, J. Pérez-Ramírez, J. Gong, et al., Chem. Soc. Rev. 49 (2020) 2937–3004. doi: 10.1039/C9CS00713J

  6. [6]

     J. Zhu, Q. Zhang, C. Wang, et al., Nanoscale 17 (2025) 2709–2717. doi: 10.1039/D4NR04843A

  7. [7]

     K. Lian, J. Ding, J. Zhang, et al., J. Mater. Chem. A 13 (2025) 2010–2018. doi: 10.1039/D4TA06503D

  8. [8]

     S. Gao, H. Li, Z. Lu, et al., Carbon Energy 7 (2025) e637. doi: 10.1002/cey2.637

  9. [9]

     Y. Liu, Y. Fang, Q. Yuan, J. Lu, H. Wang, Green Chem. 25 (2023) 1339–1344. doi: 10.1039/D2GC04659H

  10. [10]

      X. Ma, J. Du, H. Sun, et al., Appl. Catal. B: Environ. 298 (2021) 120543. doi: 10.1016/j.apcatb.2021.120543

  11. [11]

      X. Song, W. Guo, X. Ma, et al., Green Chem. 24 (2022) 1488–1493. doi: 10.1039/D1GC04146K

  12. [12]

      T. Shao, D. Duan, S. Liu, et al., Nanoscale 14 (2022) 833–841. doi: 10.1039/D1NR06545A

  13. [13]

      C. Jia, W. Ren, X. Chen, W. Yang, C. Zhao, ACS Sustain. Chem. Eng. 8 (2020) 6003–6010. doi: 10.1021/acssuschemeng.0c00739

  14. [14]

      J. Song, X. Lei, J. Mu, et al., Small 19 (2023) 2305666. doi: 10.1002/smll.202305666

  15. [15]

      X. Liu, Y. Jiao, Y. Zheng, S.Z. Qiao, ACS Catal. 10 (2020) 1847–1854. doi: 10.1021/acscatal.9b04103

  16. [16]

      D. Wu, R. Feng, C. Xu, et al., Nano-Micro Lett. 14 (2021) 38.

  17. [17]

      Y. Zhou, F. Che, M. Liu, et al., Nat. Chem. 10 (2018) 974–980. doi: 10.1038/s41557-018-0092-x

  18. [18]

      B. Jiang, X.G. Zhang, K. Jiang, D.Y. Wu, W.B. Cai, J. Am. Chem. Soc. 140 (2018) 2880–2889. doi: 10.1021/jacs.7b12506

  19. [19]

      X. Lu, Y. Hu, S. Cao, et al., Phys. Chem. Chem. Phys. 25 (2023) 18952–18959. doi: 10.1039/D2CP05449C

  20. [20]

      S.H. Mir, V.K. Yadav, J.K. Singh, ACS Appl. Mater. Interfaces 14 (2022) 29703–29710. doi: 10.1021/acsami.2c02469

  21. [21]

      H. Chen, X. Zou, Inorg. Chem. Front. 7 (2020) 2248–2264. doi: 10.1039/D0QI00146E

  22. [22]

      F. Ling, K. Luo, L. Hao, et al., ACS Omega 5 (2020) 4620–4625. doi: 10.1021/acsomega.9b04260

  23. [23]

      Z. Pu, T. Liu, G. Zhang, et al., Small Methods 5 (2021) 2100699. doi: 10.1002/smtd.202100699

  24. [24]

      P.R. Jothi, Y. Zhang, K. Yubuta, et al., ACS Appl. Energy Mater. 2 (2018) 176–181.

  25. [25]

      V. Mazánek, H. Nahdi, J. Luxa, Z. Sofer, M. Pumera, Nanoscale 10 (2018) 11544–11552. doi: 10.1039/C8NR02142B

  26. [26]

      H. Yuan, Z. Li, J. Yang, J. Phys. Chem. C 123 (2019) 16294–16299. doi: 10.1021/acs.jpcc.9b04221

  27. [27]

      M. Wallace, High-pressure high-temperature (HPHT) synthesis of functional materials, in: S. Igor (Ed. ), Sintering of Functional Materials, IntechOpen, Rijeka, 2017, pp. 166–182.

  28. [28]

      B. Ren, H. Cui, C. Wang, Electrochem. Energy Rev. 5 (2022) 32.

  29. [29]

      Y. Chen, Y. Yao, Y. Xia, et al., Nano Res. 13 (2020) 2777–2783. doi: 10.1007/s12274-020-2928-0

  30. [30]

      H. Li, L. Zhang, Q. Zeng, et al., Comp. Mater. Sci. 49 (2010) 814–819. doi: 10.1016/j.commatsci.2010.06.027

  31. [31]

      C. Zhou, P. Wang, C. Wei, et al., J. Alloys Compd. 698 (2017) 27–32. doi: 10.1016/j.jallcom.2016.12.244

  32. [32]

      P. Wang, Y. Qi, S. Zhou, et al., Phys. Status Solidi B 253 (2016) 1590. doi: 10.1002/pssb.201552752

  33. [33]

      L. Scatteia, R. Borrelli, G. Cosentino, et al., J. Spacecr. Rockets 43 (2006) 1004–1012. doi: 10.2514/1.21156

  34. [34]

      S.H. Lee, Y. Sakka, Y. Kagawa, J. Am. Ceram. Soc. 91 (2008) 1715–1717. doi: 10.1111/j.1551-2916.2008.02343.x

  35. [35]

      Y. Panayiotatos, Diam. Relat. Mater. 12 (2003) 1151–1156. doi: 10.1016/S0925-9635(02)00318-7

  36. [36]

      F. Guo, Y. Wu, H. Chen, et al., Energy Environ. Sci. 12 (2019) 684–692. doi: 10.1039/C8EE03405B

  37. [37]

      B. Qin, Y. Li, H. Wang, et al., Nano Energy 60 (2019) 43–51. doi: 10.1016/j.nanoen.2019.03.024

  38. [38]

      P.P. Guo, Z.H. He, S.Y. Yang, et al., Green Chem. 24 (2022) 1527–1533. doi: 10.1039/D1GC04284J

  39. [39]

      D. Yang, Q. Zhu, B. Han, The Innovation 1 (2020) 100016.

  40. [40]

      X. Tan, X. Sun, B. Han, Natl. Sci. Rev. 9 (2022) nwab022. doi: 10.1093/nsr/nwab022

  41. [41]

      B.A. Rosen, A. Salehi-Khojin, M.R. Thorson, et al., Science 334 (2011) 643–644. doi: 10.1126/science.1209786

  42. [42]

      C. Cadena, J.L. Anthony, J.K. Shah, et al., J. Am. Chem. Soc 126 (2004) 5300–5308. doi: 10.1021/ja039615x

  43. [43]

      G. Hong, J. Jacquemin, P. Husson, et al., Ind. Eng. Chem. Res. 45 (2006) 8180–8188. doi: 10.1021/ie0605377

  44. [44]

      D. Faggion, W.D.G. Gonçalves, J. Dupont, Front. Chem. 7 (2019) 1–8. doi: 10.3389/fchem.2019.00001

  45. [45]

      J. Pei, T. Wang, R. Sui, et al., Energy Environ. Sci. 14 (2021) 3019–3028. doi: 10.1039/D0EE03947K

  46. [46]

      B. Ratschmeier, B. Braunschweig, J. Phys. Chem. C 125 (2021) 16498–16507. doi: 10.1021/acs.jpcc.1c02898

  47. [47]

      N.J. Firet, W.A. Smith, ACS Catal. 7 (2016) 606–612.

  48. [48]

      Y. Katayama, F. Nattino, L. Giordano, et al., J. Phys. Chem. C 123 (2018) 5951–5963.

  49. [49]

      H. Chang, F. Gao, A. Luo, et al., Chem. Commun. 59 (2023) 7647–7650. doi: 10.1039/D3CC01834B

  50. [50]

      K. Yao, J. Li, H. Wang, et al., J. Am. Chem. Soc. 144 (2022) 14005–14011. doi: 10.1021/jacs.2c01044

  51. [51]

      M. Dai, W. Ni, Z. Zeng, et al., Energy Environ. Sci. 17 (2024) 1531–1539. doi: 10.1039/D3EE03226D

  52. [52]

      X. Li, W. Bi, M. Chen, et al., J. Am. Chem. Soc. 139 (2017) 14889–14892. doi: 10.1021/jacs.7b09074

• 

  Figure 1  (a) Schematic illustration of the HPHT method and procedure. (b) The XRD patterns of different metal diborides. (c) The Rietveld refinement of XRD pattern for the ZrB₂ sample. (d) The average distance of adjacent metal-B and B-B atoms in different metal diborides. (f) ELF of different planes for the ZrB₂ sample. (f) DOS of the ZrB₂ sample.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  The characterization results of the ZrB₂ sample: (a) TEM image, (b) HRTEM image, (c) SAED image, (d) EDS mapping images, (e) high-resolution XPS spectrum of Zr 3d, and (f) high-resolution XPS spectrum of B 1s.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) The current density and CO Faradaic efficiency on various metal diborides at −2.2 V vs. Ag/Ag⁺. (b) The current density and CO Faradaic efficiency on the ZrB₂ electrode at different applied potentials. (c) Long-term stability of the ZrB₂ electrode in CO₂-saturated [Bmim]PF₆(1M)-MeCN electrolytes at −2.2 V vs. Ag/Ag⁺. (d) The duration and current density of different B-containing electrocatalysts for the electrocatalytic reduction of CO₂ to CO [9-13].

  下载: 全尺寸图片 幻灯片
  

  Figure 4  The current density and CO Faradaic efficiency on the ZrB₂ electrode in different electrolytes: (a) 1 mol/L KHCO₃ aqueous solution, (b) TBAPF₆(1M)-MeCN electrolyte, (c) different ionic liquids(1M)-MeCN electrolyte, and (d) [Bmim]PF₆−MeCN electrolyte with varying concentrations of [Bmim]PF₆. (e) In-situ ATR-SEIRS spectra on the ZrB₂ electrode collected within the potential range of −1.8 V to −2.2 V vs. Ag/Ag⁺. (f) The energy barriers for *CO desorption on different ZrB₂ surfaces.

  下载: 全尺寸图片 幻灯片
•