On‐Surface Synthesis of Bismuth Monolayers through Ice‐Confined Redox Reactions
Zexiang He, Depeng Wang, Wentao Fan, Songlin Liu, Antonio Gaetano Ricciardulli, Haixia Zhong, Yiyong Mai, Paolo Samorì, Sheng Yang
TL;DR
A new ice-based method creates ultra-thin bismuth sheets that could improve electronic devices and CO2 conversion.
Contribution
A novel ice-confined synthesis method enables scalable production of atomically thin 2D bismuth with controlled thickness and high catalytic performance.
Findings
2D bismuth sheets with 1–3 atomic layers and micrometer-scale dimensions are synthesized using ice-confined redox reactions.
The produced 2D bismuth achieves a 95.6% formate Faraday efficiency in electrochemical CO2 reduction.
The method is extendable to other metals like silver, copper, and tellurium for advanced applications.
Abstract
2D bismuth possesses a unique combination of properties, such as cryogenic‐free quantum spin Hall effects and intrinsic single‐element ferroelectricity, making it highly promising for next‐generation electronic devices. However, the synthesis of 2D bismuth via exfoliation or direct growth is hindered by the low structural anisotropy of bulk bismuth crystals. To address this challenge, we demonstrate an unprecedented ice‐confined bottom‐up strategy for growing 2D bismuth. This approach involves kinetically controlled nucleation in liquid nitrogen (‐196 °C), followed by redox‐driven anisotropic growth within the confined space between ice and aluminum surfaces at −20 °C. The surfactant‐free process yields solution‐processable crystalline 2D bismuth with micrometer‐scale lateral dimensions and atomic‐level thickness, where 72% of the sheets are 1–3 layers thick. Thanks to its oxophilic…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4- —National Natural Science Foundation of China10.13039/501100001809
- —Science and Technology Commission of Shanghai Municipality10.13039/501100003399
- —Shanghai Municipal Science and Technology Major Project
- —National Foreign Expert Project of China
- —European Commission10.13039/501100000780
- —Fundamental Research Funds for the Central Universities10.13039/501100012226
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topics2D Materials and Applications · Advanced Thermoelectric Materials and Devices · Advanced Photocatalysis Techniques
Introduction
1
Bismuth is a brittle post‐transition metal featuring a puckered layered structure. Reducing its crystal dimensionality from 3D bulk to 2D gives rise to a range of unexpected properties, arising from the interplay between atomic thickness and the strong spin–orbit coupling (SOC) inherent to the relatively heavy bismuth atoms [1, 2]. For instance, bismuth monolayers supported on SiC(0001) substrates behave as quasi‐2D topological insulators with a large bandgap of 0.8 eV [1]. Many other intriguing properties, such as single‐element ferroelectric state [3], photo‐switchable electronic and spin topological phase transitions [4], and a high‐temperature quantum spin Hall (QSH) effect [1, 2, 5] have been observed in bismuth monolayers, making them promising platforms for energy‐efficient electronic devices utilizing dissipationless spin currents [1, 6]. Beyond monolayers, few‐layer bismuth structures with tunable interlayer spacing and twist angles provide tailored spin textures and band structures, holding significant potential for 2D topological spintronics [7, 8]. In addition to its electronic properties, bismuth serves as an attractive heterogeneous catalyst in organic synthesis [9], thanks to the accessibility of multiple redox couples, such as low‐valent Bi(II)/Bi(III), Bi(I)/Bi(III) as well as high‐valent Bi(III)/Bi(V). Its favorable surface chemistry, characterized by a positive standard reduction potential (Bi^3+^/Bi, 0.308 V vs standard hydrogen electrode) [10, 11] and a relatively high hydrogen overpotential, also makes bismuth particularly suitable for CO_2_ reduction reactions [12, 13, 14].
However, the development of efficient and reliable synthetic strategies remains a major obstacle to its widespread applications. The rhombohedral bismuth crystal exhibits limited structural anisotropy (i.e., 15 %), arising from the disparity between the in‐plane bond length (0.307 nm) and the out‐of‐plane interlayer distance (0.353 nm), which is insufficient to preserve the in‐plane architecture during exfoliation [15, 16, 17, 18]. As a result, mechanical exfoliation often yields agglomerated bismuth nanoparticles [19]. Although ion intercalation can weaken the interlayer interaction [20, 21, 22, 23], the resulting bismuth sheets are typically small and consist of multiple thick layers (e.g., 17 layers) [24]. Recently, a van der Waals squeezing technique has been developed to compress bismuth into stable monolayers encapsulated between two MoS_2_ layers [25]. Alternatively, bottom‐up approaches often rely on surfactant templates to direct growth. Nonetheless, these methods produce bismuth sheets with wide solvent spacing (10^1^–10^2^ nm) [26, 27], and thicknesses on the order of tens of nanometers [28]. Moreover, surfactant molecules or their derivatives frequently contaminate the 2D bismuth surfaces [28, 29]. Epitaxial growth can yield high‐quality 2D bismuth monolayers [1, 3], but it requires precise control over multiple critical parameters, including maintaining an ultra‐high vacuum (UHV) of 1 × 10^−10^ mbar, using a high‐purity bismuth source, mitigating substrate lattice mismatch, and finely tuning the atomic flux ratio and substrate temperature to achieve a low deposition rate [1, 3, 30]. Notably, transferring the fragile 2D bismuth onto suitable substrates without inducing structural damage remains a significant challenge.
Herein, we demonstrate a surfactant‐free on‐surface strategy for growing 2D bismuth by leveraging redox reactions within a confined space between an ice layer and an aluminum substrate. Kinetically controlled nucleation and anisotropic growth are achieved at −196°C and −20°C, respectively. This approach yields solution‐processable, atomically thin 2D bismuth sheets with lateral sizes in the micrometer range, a pristine crystal structure, and thickness down to monolayer limit (∼0.7 nm). Streamlined cyclic reactions enable technologically viable gram‐scale production. As a prototypical application, we harness the abundant active sites on the exposed surface of the as‐synthesized 2D bismuth, which exhibits high intrinsic affinity for *OCHO intermediates in the electrochemical CO_2_‐to‐formate conversion. The material achieves a high formate Faradaic efficiency of up to 95.6%. Besides, this on‐surface synthetic strategy is applicable for fabricating other functional 2D metals, including silver, copper, and tellurium, thereby paving a new avenue for developing advanced electronics and electrocatalysts.
Results and Discussion
2
The ice‐confined on‐surface synthesis of 2D bismuth is initiated by quickly injecting bismuth chloride aqueous solution (0.1 M BiCl_3_, 1 mL) onto an aluminum foil whose opposite side is exposed to liquid nitrogen (−196°C). Notably, a small amount of HCl was added to prevent hydrolysis of Bi^3^ ^+^ ions, while effectively removing any residual oxide layer on the aluminum surface. This triggers instantaneous ice formation and initiates interfacial redox reactions at the ice‐aluminum interfaces (Figure S2). Upon contact, the precursor solution rapidly freezes on the foil, leading to a transiently controlled nucleation process that forms a firmly adherent ice layer. The system is then maintained at −20°C to preserve the ice‐aluminum interface while allowing continuous growth of 2D bismuth (Figure 1a). This ice‐templated growth results in a highly ordered 2D morphology, which stands in sharp contrast to the disordered structures obtained under ice‐free conditions (Figure S3).
Ice‐confined redox synthesis of 2D bismuth. (a) Schematic comparing the ice‐confined growth strategy for 2D bismuth and the conventional exposed‐surface growth. (b) Modulating nucleation and growth kinetics through precise temperature control. (c) Collection procedure schematic, along with a digital photo of the dispersed product and its SEM image revealing the 2D morphology.
The confined redox reaction promotes 2D growth for two key reasons: (i) the ultralow temperature (−196°C) and rapid freezing jointly suppress nucleation density and slow down growth kinetics, favoring monolayer‐dominant crystallization. (ii) The diffusion and reaction of the precursors are constrained by spatial confinement at the ice‐aluminum interface, where the ion mobility is suppressed at a low temperature of −20°C, jointly promoting anisotropic 2D growth (Figure 1b). After synthesis, the 2D bismuth can be efficiently harvested by mechanically delaminating the aluminum foil, followed by ice melting at 25°C for 10 min. Through cyclic ice‐templated synthesis with optimized thermal programming, this approach achieves gram‐scale production under standard laboratory conditions (Figure S4), while retaining typical 2D characteristics (72% sheets are 1–3 layers thick). Notably, industrial scaling remains feasible through well‐established liquid nitrogen supply chains (e.g., for food and biological preservation) and conventional refrigeration systems capable of sustainable operation at −20°C. Scanning electron microscopy (SEM) confirms the micrometer‐scale lateral dimensions of the synthesized 2D bismuth (Figures 1c and 2a).
Comprehensive characterization of 2D bismuth. (a,b) Morphology examined by SEM and AFM. (c) Statistical thickness analysis from AFM images. (d–f) Crystalline structure revealed by TEM, HRTEM, and the corresponding SAED pattern. (g–i) Spectroscopic studies using XRD, Raman, and high‐resolution XPS of the Bi 4f core level, revealing the distinct structural and electronic properties of the 2D bismuth compared to its bulk counterpart.
Atomic force microscopy (AFM) was employed to analyze the thickness of the synthesized 2D bismuth flakes, which indicated the presence of monolayers with a thickness of approximately 0.7 nm (Figure 2b) [31, 32]. It should be noted, however, that AFM‐based thickness measurements often yield overestimates due to solvent adsorption, capillary forces, and substrate adhesion [31, 33, 34]. To mitigate such influences, measurements were performed on multiple substrates (Figure S5), yielding an average thickness of about 1.95 nm, corresponding to no more than 2–3 atomic layers [32]. Notably, 72% of the synthesized 2D bismuth flakes exhibited thicknesses in the range of 1 to 3 atomic layers. The structural characteristics of the 2D bismuth were further investigated using a transmission electron microscope (TEM). High‐resolution TEM (HR‐TEM) imaging (Figure 2e) revealed well‐defined lattice fringes with an interplanar spacing of 0.228 nm, aligning with the (2–10) and (−120) planes of rhombohedral (β‐phase) bismuth (Figure S1b), the thermodynamically stable polymorph under ambient conditions [35] (see also Figure S1). Moreover, the selected area electron diffraction (SAED) pattern (Figure 2f) confirmed the single‐crystalline quality of the material, with diffraction spots indexed to the [001] zone axis.
As illustrated in the X‐ray diffraction (XRD) pattern (Figure 2g), all diffraction peaks of the 2D bismuth correspond well to the rhombohedral phase (PDF 01‐085‐1331). The Raman spectrum (Figure 2h) exhibits characteristic phonon modes at 71.2 cm^−1^ (E_g_, in‐plane) and 97.4 cm^−1^ (Ag1, out‐of‐plane) with the absence of peaks between 300 and 350 cm^−1^ suggests negligible oxidation [36]. In comparison to bulk bismuth (Figure S6b; Figure 2g,h), both the XRD and Raman peaks of the 2D samples are broadened, indicating the reduced number of layers leading to decreased crystallographic periodicity and modified phonon behavior. Notably, Raman analysis confirms that 2D bismuth should be stored in a desiccated environment or as an ethanol dispersion to prevent moisture‐induced oxidation. This would otherwise lead to the formation of α‐Bi_2_O_3_ under water‐ and oxygen‐rich conditions (Figure S7). Further structural and chemical characterization was conducted using X‐ray photoelectron spectroscopy (XPS). The Bi 4f spectrum of bulk bismuth shows peaks at 159.0 and 164.3 eV (Figure 2i), indicative of slight surface oxidation, a common phenomenon for p‐block metals due to the oxophilic nature [37, 38]. In contrast, the 2D bismuth exhibits markedly reduced oxide signals, with dominant peaks appearing at 156.7 and 162.0 eV, corresponding to metallic Bi^0^ (83.14%, Table S3). This indicates a relatively pure surface with minimal chemical contamination, thanks to the surfactant‐free synthesis environment. Moreover, the use of a liquid nitrogen atmosphere and low‐temperature conditions during synthesis further suppressed oxygen incorporation, enhancing the chemical purity of 2D bismuth.
To elucidate the growth mechanism of 2D bismuth, we systematically investigated its structural evolution during both the early and post‐stages of ice formation and stabilization. In the initial stage, occurring shortly before the bismuth solution was fully frozen, bismuth(III) ions in contact with the aluminum foil underwent rapid reduction. This process involved the formation and oriented attachment of atomic clusters [39], resulting in the development of linear precursor patterns [39, 40]. These linear structures subsequently served as seeds in the second stage, facilitating the growth of quasi‐2D and 2D bismuth at the interfaces between the aluminum foil and the ice matrix (Figure 3a). Critically, the acidic Bi(III) solution eliminates native Al_2_O_3_ surface layer, exposing the underlying reactive aluminum for the subsequent reduction reaction.
Morphological evolution of bismuth during ice‐confined synthesis. (a) Schematic of the proposed growth mechanism. (b, c) SEM images of the initial linear bismuth structures formed at b) 5°C and c) 45°C. (d, e) Subsequent evolution into (quasi‐)2D bismuth at −20°C after (d) 40 min and (e) 11 h. (f) Schematic illustrating the branch fusion mechanism. (g, h) TEM images of g) separated and h) fused linear branches. (i, j) TEM images showing i) an in‐plane grain boundary and j) a merged domain in quasi‐2D bismuth. (k, l) HRTEM images of k) quasi‐2D bismuth (inset: SAED pattern) and l) its in‐plane grain boundary (inset: FFT pattern).
Note that, Bi^3^ ^+^ exhibits reduction preference over H^+^ (supported by standard redox potentials of Bi^3^ ^+^/Bi, 2H^+^/H_2,_ and Al^3^ ^+^/Al, corresponding to +0.308, 0, and −1.662 V vs SHE, respectively) [10, 11, 41], making the following reaction ready to occur:
Controlled Diffusion and Nucleation
2.1
During the initial stage, as the solution was rapidly frozen using liquid nitrogen, the diffusion of Bi^3+^ ions toward the aluminum foil surface was suppressed. Under these constrained conditions, Bi^3+^ ions were reduced to bismuth atoms, which subsequently assembled into clusters via interfacial redox reactions:
In the confined environment, adjacent bismuth clusters (Bi_Ci,_ i = 1,2,3…) further aggregated through clustering processes, resulting in the formation of nascent faceted crystals [42] (Figure S8a):
The resulting bismuth crystals exhibited linear and fractal morphologies (Figure 3b,c). A positive correlation was observed between the initial temperature of the bismuth solution (5°C, 25°C, 45°C) and the degree of fractal feature development (Figure 3b,c; Figure S10a). Specifically, an initial temperature of 45°C yielded more pronounced fractal structures (Figure 3c). These results indicate that differences in freezing timescales and reaction rates, modulated by temperature, govern the ultimate morphological outcomes (Figure S9a).
Confined Docking and Growth
2.2
In the second stage, to activate further growth of the fractal seeds (Figure 3a), the system was transferred to a refrigerator maintained at −20°C. The relatively weak hydrogen‐bonding interactions at ice surfaces, compared to the bulk phase, reduce kinetic barriers for molecular diffusion [43]. This facilitates interfacial reaction spaces [44, 45, 46], enabling the gradual release of ice‐encapsulated bismuth precursors into the confined spaces (also termed the quasi‐liquid layer or premelting layer on ice surface) [45, 47]. This process enabled the fractal bismuth seeds to grow in a confined manner at the ice‐aluminum foil interface, evolving into quasi‐2D bismuth (Figure 3d) and ultimately 2D bismuth (Figure 3e). The sustained interfacial redox reactions continuously supplied additional bismuth clusters (Bi_C4_, Bi_C5_…), which preferentially attached to the faceted crystals (Bi_F_) generated in equation (4), filling structural vacancies within the developing 2D framework (Figure S8b):
As illustrated in figure 3f,g, the branched structures expanded toward one another and eventually docked (Figure 3h and Figure S8c), leading to the formation of a quasi‐2D architecture (Figure 3f; Figure S11d). The crystal boundaries resulting from this edge‐docking process are visible within the quasi‐2D plane (Figure 3i,j).
Notably, although the linear structure formed in the initial stage (Figure S11a–c) exhibited different exposed crystal facets from the final 2D product, the quasi‐2D intermediates generated in the second stage, including their boundary regions (Figure 3k,l), displayed the same lattice orientation along the [001] zone axis as the 2D bismuth. This suggests that the growth of 2D bismuth proceeds through two distinct regimes: a kinetically controlled initial stage favoring rapid directional growth, followed by a thermodynamically controlled stage that promotes the formation of the low‐energy (001) facet, which is the most stable surface in rhombohedral bismuth [48]. Furthermore, XRD (Figure S11e) and Raman spectroscopy analyses (Figure S11f) confirmed that both the linear and quasi‐2D bismuth intermediates adopt the rhombohedral crystal structure.
Aluminum plays a critical role in the controlled synthesis of 2D bismuth by acting as an electron donor to reduce Bi^3^ ^+^ ions. The resulting Al^3^ ^+^ cations, confined at the interface, further suppress the excessive diffusion of Bi^3^ ^+^ ions toward the reaction through electrostatic repulsion. Incidentally, since redox reactions can similarly occur at ice‐metal interfaces with other oxidant ions (provided by precursor solutions such as CuCl_2_, TeCl_4_, and AgF), this method demonstrates broad applicability for synthesizing various 2D nanosheets (Cu, Te, Ag. Figure S12). The resulting nanosheets exhibit thicknesses of ∼3 nm, confirming both the versatility and technological potential of this ice‐confined redox synthesis approach.
The confined ice template method enables the preparation of 2D bismuth without surfactants, allowing full exposure of bismuth active sites. These sites exhibit high oxygen affinity and lower hydrogen affinity [37], which favors the formation of the oxygen‐bound *OCHO intermediate of formate [49] on the 2D bismuth surface during the carbon dioxide reduction reaction (CO_2_RR). As shown in Figure 4a, the electrocatalytic CO_2_RR performance of 2D bismuth was evaluated in an H‐type electrochemical cell, with commercial bulk bismuth as a reference. Given the inherent tendency of bismuth to bind surface oxygen (Figure 2i), intentionally oxidized 2D bismuth (2D bismuth‐O_x_, Figure S13) was also included as a control. In the linear sweep voltammetry (LSV) tests (Figure 4b), the current densities of bulk bismuth, 2D bismuth and 2D bismuth‐O_x_ began to increase around −0.7 V vs. the reversible hydrogen electrode (RHE) and rose with more negative cathode potentials, indicating enhanced CO_2_ reduction to formate on the bismuth surface. The formate product was determined by nuclear magnetic resonance (NMR, Figure S15), and the corresponding partial current density (j_formate_, Figure 4c) was calculated. Notably, both 2D bismuth/ 2D bismuth‐O_x_ exhibited significantly higher formate production than bulk bismuth in the potential range of −0.9–−1.3 V vs. RHE. This enhancement can be partially related to their larger electrochemically active surface area (ECSA), as estimated from the double layer capacitance (C_dl_, Figure 4d). The ECSA values of 2D bismuth/ 2D bismuth‐O_x_ (2.36, 2.19 mF cm^−2^) were higher than that of bulk bismuth (0.97 mF cm^−2^), consistent with the improved electrocatalytic activity of the 2D structures. As a supplement, the C_dl_ values were obtained from cyclic voltammograms (CV) at different scan rates (Figure S16).
*Electrocatalytic CO2 reduction to formate on 2D bismuth. (a) Schematic of the CO2RR process in H‐cell. (b) LSV curves of bulk bismuth, 2D bismuth, and 2D bismuth‐Ox in 0.5 M CO2‐saturated KHCO3. (c) Partial current density for formate (jformate) of bulk bismuth, 2D bismuth, and 2D bismuth‐Ox at different potentials. (d) Charging current density differences (Δj) vs. scan rates of bulk bismuth, 2D bismuth and 2D bismuth‐Ox. (e) Faradaic efficiencies of formate (FEformate) against potentials on 2D bismuth, 2D bismuth‐Ox and bulk bismuth. (f) The current density and formate Faradaic efficiency of 2D bismuth under constant potential of −0.9 V vs. RHE for 15 h in H‐cell. (g) Reaction energetics for CO2RR (*OCHO/*COOH) and HER (*H) on 2D bismuth (001) facet at identical adsorption sites. (h) the projected density of states (PDOS) of Bi sites with *OCHO, *COOH and H. (i) Comparison of 2D bismuth with reported electrocatalysts for formate production via CO2RR in H‐cells [52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63] (List of catalysts: (1) SW‐Cu2O, (2) BOC NFs, (3) CT/h‐BiOBr, (4) Stanene@3%BP, (5) SOR Bi@C NPs, (6) Bi/Bi2O3‐2, (7) Bi‐NS, (8) Bi‐NFs, (9) Op‐Ag1In, (10) 3 at.% Cu‐SnS2, (11) HCS/Cu‐0.12, (12) RD‐Bi).
Moreover, the Faradaic efficiency of formate (FE_formate_) was evaluated under various conditions (Figure 4e). Remarkably, 2D bismuth maintained FE_formate_ above 90% across a broad potential window from−0.9 to −1.2 V vs. RHE, reaching a maximum of 95.6% at −1.1 V. The durability test presented in figure 4f demonstrates excellent stability, with negligible decay in current density and a sustained FE_formate_ exceeding 93.5%. Even after extended operation for 80 h at −0.8 V vs RHE, the electrode material maintains a stable current density while retaining a FE_formate_ above 94.0% (Figure S17), which confirms its long‐term operational durability. It is also worth noting that 2D bismuth‐O_x_ and pristine 2D bismuth showed comparable catalytic activity and selectivity (Figure 4c,e), suggesting that minor surface oxidation has a negligible impact on their performance. This similarity can likely be attributed to the reductive conditions in the cathode region during activation and testing, which effectively remove most surface oxygen from 2D bismuth‐O_x_ (Figure S13b,c; Table S3).
The excellent catalytic performance of 2D bismuth for CO_2_‐to‐formate conversion was further proved by theoretical calculations. In the CO_2_RR process, multiple competing reaction pathways exist, each associated with distinct intermediates (Figure S18) and corresponding mechanisms. Density functional theory (DFT) results revealed that on the 2D bismuth (001) facet, the formate pathway exhibits the lowest energy barrier (0.71 eV) for the rate‐determining step, compared to the carbon monoxide pathway (1.40 eV) and the hydrogen evolution reaction (HER) pathway (1.01 eV), as shown in figure 4g. Moreover, the projected density of states (PDOS) of key intermediates adsorbed at Bi sites were analyzed (Figure 4h). The highest peak for *OCHO was located closer to the Fermi level than those for *COOH and *H, indicating the lowest occupancy of anti‐bonding states and thus the strongest interaction with the bismuth surface [50, 51]. These electronic structure insights confirm that the formate pathway proceeds more readily and preferentially on the 2D bismuth (001) facet. As a result, 2D bismuth achieves high activity and selectivity in the electrocatalytic conversion of CO_2_ to formate, reaching a FE_formate_ of 95.6% and a large formate partial current density of 47.8 mA cm^−2^ at −1.1 V vs. RHE in an H‐Cell. These performances metrics surpass those of mainstream catalysts reported in previous studies [52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63] (Figure 4i).
Conclusion
3
In conclusion, we have developed a confined redox reaction system at the ice/metal interface to regulate the in‐plane anisotropic growth of bismuth monolayers. The combination of kinetic and thermodynamic control was crucial for fine‐tuning ion diffusion and reaction rates, thereby enabling the synthesis of free‐standing, atomically thin 2D bismuth sheets with micrometer‐scale lateral dimensions, a pristine crystal structure, and thickness down to 0.7 nm. Besides, the design of streamlined cyclic reactions allow for gram‐scale production. The superior material quality is directly translated into outstanding catalytic performance, as demonstrated by a formate Faradaic efficiency of up to 95.6% in CO_2_RR. This simple and efficient synthesis strategy offers a scalable route to producing low‐dimensional metals and other functional materials, thereby providing a versatile platform for a wide range of future applications in catalysis, electronics, and beyond.
Experimental Section
4
Chemicals
4.1
Hydrochloric acid (36%‐38%) and aluminum foil (≥99.0%, 0.1 mm thick) were purchased from Sinopharm Chemical Reagent Co., Ltd. Liquid nitrogen was supplied by Shanghai Lvmin Gas Co., Ltd. Bismuth(III) chloride (99%) was sourced from Shanghai Adamas Reagent Co., Ltd. Bulk bismuth (99.99%, ≥200 mesh), deuterium oxide (≥99.9 atom% D), dimethyl sulfoxide (≥99.9%), and potassium bicarbonate (≥99.5%) were acquired from Shanghai Aladdin Biochemical Technology Co., Ltd. Isopropyl alcohol (≥99.5%) was purchased from Shanghai Macklin Biochemical Technology Co., Ltd. Deionized (DI) water (18.2 MΩ cm) was produced by a Sartorius Arium water purification system and deoxygenated via freeze‐pump‐thaw cycles prior to use.
Synthesis of 2D Bismuth
4.2
The experimental procedure was initiated by diluting concentrated hydrochloric acid to 2 mol/L with deionized water. A 0.1 mol/L BiCl_3_ aqueous solution was then prepared by dissolving bismuth(III) chloride in the dilute hydrochloric acid under continuous stirring. An aluminum foil (0.1 mm thick) was cleaned with dilute hydrochloric acid and placed on a 4 × 4 cm^2^ quartz sheet substrate. This assembled quartz‐aluminum substrate was positioned inside a Dewar flask, and liquid nitrogen was carefully added until the aluminum foil was submerged to a depth of approximately 11 mm. Next, 1 mL of the BiCl_3_ solution was rapidly dispensed onto the aluminum surface, resulting in immediate ice formation. After 5 min of immersion in liquid nitrogen, the frozen quartz‐aluminum‐ice composite was transferred to a −20°C refrigerator, and held for specific durations (t = 0 for linear bismuth, 40 min for quasi‐2D bismuth, and 11 h for 2D bismuth). The aluminum foil was mechanically delaminated and removed, leaving the ice layer. After the ice melted, the samples were sequentially rinsed with dilute hydrochloric acid and deionized water, then vacuum‐dried at room temperature. For the preparation of 2D bismuth‐O_x_, the as‐prepared 2D bismuth was dispersed in deionized water (200 mg in 20 mL) and stirring vigorously at 1500 rpm for 3 h under ambient air. The resulting product was washed several times with deionized water and then dried for further use.
Structural Characterization
4.3
Field‐emission scanning electron microscopy (FE‐SEM, Gemini 300, Zeiss) was used to characterize the morphology of the as‐prepared bismuth samples. Transmission electron microscopy (TEM) images and selected area electron diffraction (SAED) patterns were acquired using a Talos F200X TEM (Thermo Scientific) at 200 kV accelerating voltage. X‐ray diffraction (XRD) patterns were recorded on an ARL EQUINOX 3500 diffractometer (Thermo Scientific) with Cu‐Kα radiation. X‐ray photoelectron spectroscopy (XPS) analysis was performed on an ESCALAB Xi+ spectrometer (Thermo Scientific). Raman spectra were collected using a LabRAM Soleil spectrometer (Horiba) with 638 nm laser excitation. Atomic force microscopy (AFM) measurements were conducted on a Dimension XR system (Bruker). ^1^H nuclear magnetic resonance (^1^H NMR) spectra were obtained using an Avance III HD 500 MHz spectrometer (Bruker).
Electrochemical Measurements
4.4
All electrochemical measurements were performed in a three‐electrode H‐cell separated by a Nafion 212 proton exchange membrane, using a Biologic VMP3 electrochemical workstation. The working electrode was prepared by uniformly mixing 5 mg of the sample, with 500 µL of isopropyl alcohol and 20 µL of Nafion D520, followed by sonication for 30 min. Then, 50 µL aliquots of the resulting ink were drop‐cast onto carbon paper (CeTech, GDL340, 1 × 1 cm^2^), yielding a catalyst loading of ∼0.5 mg cm^−^ ^2^. An Ag/AgCl (saturated KCl) and a platinum mesh (2 × 2 cm^2^) were used as the reference and counter electrodes, respectively. The electrolyte consisted of 0.5 M KHCO_3_ pre‐saturated with CO_2_, with continuous CO_2_ bubbling maintained at 20 sccm using a digital flowmeter (Sevenstar CS‐200A). Prior to testing, each electrode was activated by applying a constant potential of −0.7 V vs RHE for 30 min. All reported potentials were converted to the reversible hydrogen electrode (RHE) using the equation E RHE = E Ag/AgCl+0.197 + 0.059 × pH without iR compensation. Linear sweep voltammetry (LSV) was conducted at a scan rate of 20 mV s^−1^. Cyclic voltammetry (CV) was performed within the potential range of 0.2–0.3 V vs. RHE at scan rates from 20 to 120 mV s^−1^ to determine the double layer capacitance (C_dl_). The liquid products were quantified by ^1^H nuclear magnetic resonance (^1^H NMR) spectroscopy using DMSO as an internal standard.
Electrochemically Active Surface Area (ECSA)
4.5
The ECSA was estimated from the double layer capacitance (C_dl_) of catalyst [64]. The C_dl_, charging current (i_c_) and scan rate (v) follow the equation:
Based on the relationship, the C_dl_ was obtained by plotting the linear curve of (J_a_ – J_c_)/2 vs. scan rate and calculating its slope. J_a_ and J_c_ denote the anodic and cathodic current densities, respectively. Then, the ECSA was determined by C_dl_ using the following equation:
where C_s_ is the specific capacitance.
Product Analysis
4.6
The Faradaic efficiencies of formate (FE_formate_) were calculated as below:
where n_formate_ (mol) is the molar production of formate, N (= 2) is the number of electrons transferred per formate molecule, F (= 96485 C mol^−1^) is the Faraday constant, i (A) is the recorded current, t is the reaction time (s).
The formate partial current density (j_formate_) was calculated as follows:
in which j (mA cm^−2^) denotes the total current density.
Theoretical Calculations
4.7
All DFT calculations were carried out using the Vienna ab initio simulation package (VASP) [65] with the Perdew‐Burke‐Ernzerhof (PBE) exchange‐correlation functional [66]. The projector augmented wave (PAW) [67] pseudopotentials were used to describe electron‐ion interactions. A plane wave basis set with a cutoff energy of 420 eV and the 3 × 3 × 1 Monkhorst‐Pack k‐point grid was adopted for geometry optimization of the Bi (001) slab [68]. The energy and force criterion for convergence of the electron density were set at 10^−5^ eV and 0.02 eV/Å, respectively. The Bi (001) slab consisted of six atomic layers and a vacuum space of 15 Å along the z‐direction, of which the top two layers were allowed to relax during optimization. Note that, the solvent corrections were included in surface calculations. The solvent effect on adsorbates was achieved using the Poisson–Boltzmann implicit solvation model with a dielectric constant of 80 [69].
On the uniformly defined Bi(001) facet adsorption sites, the free energy was computed for all species as:
where E DFT, E ZPE, S, and T are electronic energy, zero‐point energy, entropy, and system temperature (298.15 K), respectively. For absorbates, E ZPE and S were determined by vibrational frequencies calculations, where all 3N degrees of freedom were treated as harmonic vibrations, without including contributions from the slab.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting File: adma72157‐sup‐0001‐SuppMat.pdf
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1F. Reis , G. Li , L. Dudy , et al., “Bismuthene on a Si C Substrate: A Candidate for a High‐Temperature Quantum Spin Hall Material,” Science 357 (2017): 287–290.28663438 10.1126/science.aai 8142 · doi ↗ · pubmed ↗
- 2S. Sun , J.‐Y. You , S. Duan , et al., “Epitaxial Growth of Ultraflat Bismuthene with Large Topological Band Inversion Enabled by Substrate‐Orbital‐Filtering Effect,” ACS Nano 16 (2022): 1436–1443.34918901 10.1021/acsnano.1c 09592 · doi ↗ · pubmed ↗
- 3J. Gou , H. Bai , X. Zhang , et al., “Two‐Dimensional Ferroelectricity in a Single‐Element Bismuth Monolayer,” Nature 617 (2023): 67–72.37020017 10.1038/s 41586-023-05848-5PMC 10156600 · doi ↗ · pubmed ↗
- 4B. Peng , G. F. Lange , D. Bennett , K. Wang , R.‐J. Slager , and B. Monserrat , “Photoinduced Electronic and Spin Topological Phase Transitions in Monolayer Bismuth,” Physical Review Letters 132 (2024): 116601.38563950 10.1103/Phys Rev Lett.132.116601 · doi ↗ · pubmed ↗
- 5M. S. Lodge , S. A. Yang , S. Mukherjee , and B. Weber , “Atomically Thin Quantum Spin Hall Insulators,” Advanced Materials 33 (2021): 2008029.10.1002/adma.20200802933893669 · doi ↗ · pubmed ↗
- 6H. Du , X. Sun , X. Liu , et al., “Surface Landau Levels and Spin States in Bismuth (111) Ultrathin Films,” Nature Communications 7 (2016): 10814.10.1038/ncomms 10814 PMC 479296126964494 · doi ↗ · pubmed ↗
- 7J. Gou , L. Kong , X. He , et al., “The Effect of Moiré Superstructures on Topological Edge States in Twisted Bismuthene Homojunctions,” Science Advances 6 (2020): aba 2773.10.1126/sciadv.aba 2773 PMC 726965432537502 · doi ↗ · pubmed ↗
- 8L. Zullo , D. Ninno , and G. Cantele , “Twist‐Tunable Spin Control in Twisted Bilayer Bismuthene,” Physical Review B 110 (2024): 165411.