Introduction

Brominated chemicals are versatile intermediates for synthesizing value-added fine chemicals and pharmaceuticals1,2,3,4,5,6. For example, p-bromo-anisole, with a consumption of more than 20,000 tons per year in China, is widely served as the basic molecule to combine with other functional groups via the well-known Suzuki-coupling reaction. 4-Bromo-3-methylanisole (10,000 T/a in China) is the intermediate to make colorant. Bromoxynil octanoate (150,000 T/a in the world) is the main herbicide. Owing to the unique function of bromine, demands for late-stage bromination of bioactive molecules and pharmaceuticals to regulate their physicochemical properties, such as pKa and metabolic rate, are rapidly going up6,7,8,9. The current strategies of bromination rely heavily on liquid Br2; more than 0.5 million tons of liquid Br2 are consumed each year globally that are produced via traditional high energy consumption and high discharge processes4. Liquid bromine can ensure fast bromination reactions, but suffers from poor bromination selectivity and the unavoidable generation and release of equimolar HBr, causing poor atom-efficiency and bringing serious environmental pollution. The pungent odor as well as the corrosive nature largely hinder its transportation, storage, and handling processes for practical applications3. NBS (N-bromosuccimimide) can fulfill bromination with moderate reaction rates, which is, unfortunately, derived from liquid Br210. Besides, the stability of NBS is problematic as it is sensitive to light, water, and reductive reagents10,11.

The electrochemical method offers an interesting approach to synthesize bromine for bromination reaction6,8,9. The electrochemical strategy, driven by electricity, is commonly operated under mild conditions without the requirement of strong or toxic redox reagents12,13,14, which is considered as a promising sustainable technology to enable a green industry for useful chemicals and/or pharmaceuticals synthesis in the coming carbon-neutral century. Owing to a spatially-decoupled redox system at anode and cathode, accurate electrochemical windows and real-time electrochemical signals can ensure good reaction selectivity and high energy efficiency15. More importantly, the ease of scaling up electrochemical cells facilitates the large-scale production of chemicals or pharmaceuticals via electrochemical synthesis13. Over the past decades, various electrochemical syntheses have been developed to produce brominated organic molecules using bromide salt as the green bromine source16,17,18,19,20,21,22,23. Commercial electrodes such as carbon, Pt was roughly applied as the anode in an electrochemical cell to drive an electrochemical Br- oxidation reaction to in-situ produce Br2 to brominate organic molecules, which suffers from very high overpotential, poor-Faradaic efficiency and substrate over-oxidation (Supplementary Table 1). It is thus urgent to deeply understand the mechanism of electrochemical bromine evolution reaction (BER) and then guide the rational design of suitable catalytic materials for high-performance BER.

In this work, we fabricated an oxygen-vacancy-rich Co3O4 catalyst that exhibits high bromine selectivity and reaction rate at room temperature in air to synthesize Br substituted molecules, natural products, and drugs using NaBr as the cheap bromine source. It delivers ultralow operating potential within 50 mV to catalyze BER at 10 mA cm-2. These enable the substitution of oxygen evolution reaction (OER) by BER to couple with hydrogen evolution reaction (HER) for energy-saving electrochemical hydrogen production (Supplementary Fig. 1, 2). Based on a collection of electrochemical and in-situ spectroscopic measurements together with density functional theory (DFT) calculations, it is found that the oxygen vacancy on Co3O4 can effectively regulate the adsorption and desorption of Br* to promote BER at low overpotentials. The in-situ produced Br2 works very effectively in the bromination of arenes, achieving overall bromination Faradaic efficiencies greater than 90%. The as-developed electrochemical bromine evolution approach can be easily scaled up to synthesize a kilogram scale of bioactive molecules.

Results

Catalyst preparation and characterization

Electrodeposition combined with calcination (Fig. 1a) was applied to prepare oxygen-vacancy-rich cobalt oxide on carbon paper (Co-OV@CP). The carbon paper (CP) is made of carbon fibers with diameters of several micrometers. Co(OH)2 was electrodeposited on CP (Supplementary Fig. 3a) followed by calcination in a tube furnace to form black pristine Co3O4 nanosheets (Co3O4@CP). Further calcination in a 5% H2/Ar atmosphere at 300 °C for 15 min created oxygen vacancies on the surface of Co3O4 nanosheets (Co-OV@CP) (Fig. 1b–g)24,25 and the mass loading of Co on carbon paper is determined to be about 4.65 mg cm-2 (by ICP-MS, Supplementary Table 2). Both Co3O4@CP and Co-OV@CP display ultrathin honeycomb-like nanosheet arrays on CP (Supplementary Fig. 3b, c)26. The elemental mapping results reveal uniform distribution of Co and O elements across the nanosheet framework (Fig. 1d, e and Supplementary Fig. 3). High-resolution transmission electron microscopy (TEM) images show clear lattice spacings that can be indexed to the (220) plane of cubic Co3O4. The atomic discontinuity near the Co3O4 crystal lattice may be due to the formation of oxygen vacancies27,28.

Fig. 1: Catalyst preparation and characterization.
Fig. 1: Catalyst preparation and characterization.
Full size image

a Schematic illustration showing preparation of Co-OV@CP. Typical SEM (b) and TEM (c) images of Co-OV and the corresponding elemental distribution of Co (d) and O (e). HRTEM image (f) and schematic crystal structure (g) of Co-OV. (h) EPR spectra, (i) O 1 s XPS spectra and (j) the estimated surface oxygen vacancy content.

The crystal structure of Co3O4@CP and Co-OV@CP were examined by X-ray diffraction (XRD). As shown in Supplementary Fig. 4, the diffraction peaks at 31.4°, 36.9°, 44.9°, 59.3° and 65.3° can be assigned to the (220), (311), (400), (511) and (440) planes of cubic Co3O4 (JCPDS 42-1467) and the diffraction peaks at 18.2°, 26.6° and 54.9° are resulted from CP26,29. It is worth noting that the diffraction peaks of Co-OV@CP are almost identical to those of Co3O4@CP, suggesting minimum bulk crystal structure change before and after H2 treatment (the short-time reduction treatment only induces formation of oxygen vacancies on surface of Co3O4 (Fig. 1f, g)). Electron paramagnetic resonance (EPR) measurements (Fig. 1h) show an enhanced EPR signal at g = 2.0045 for Co-OV@CP, which can be attributed to more unpaired electrons resulting from Co2+, suggesting the presence of abundant oxygen vacancies on the surface of Co-OV@CP30,31. The oxygen vacancies on the surface of Co-OV@CP were further studied by X-ray photoelectron spectroscopy (XPS). The survey spectra (Supplementary Fig. 5a) show both Co and O on the surface of Co3O4@CP and Co-OV@CP. The high-resolution Co 2p XPS spectra (Supplementary Fig. 5b) exhibit two peaks at around 780 and 796 eV, which can be attributed to Co 2p3/2 and Co 2p1/2, respectively. Compared to Co3O4@CP, the Co-OV@CP displays two new satellite peaks centered at 786 and 803 eV, which may be due to the reduction of Co3+ to Co2+ along with the generation of oxygen vacancies. The high-resolution O 1 s XPS spectra, as shown in Fig. 1i further verify the existence of oxygen vacancies on Co-OV@CP. The deconvoluted peaks at 529.6, 531.2, and 532.8 eV are attributed to the lattice oxygen, the oxygen vacancy, and the adsorbed hydroxyl species, respectively26,27,28,32,33,34,35. The obviously decreased lattice oxygen peak (529.6 eV) and enhanced oxygen vacancy peak (531.2 eV) indicate a higher concentration of oxygen vacancies on the surface of Co-OV@CP. The concentration of surficial oxygen vacancies was estimated according to the deconvoluted O 1 s XPS peaks. By considering the coordination number of Co (binding with six O atoms; oxygen vacancy may increase fivefold formation of Co(Ov)-O)) with O (binding with four Co atoms; oxygen vacancy may induce triple decrease of Co-O bond) atoms as well as the increased or decreased percentage of Co-O or Co(Ov)-O compared with the total oxygen bonding with Co or Co(Ov) before H2 treatment, the content of surficial oxygen vacancies was determined to be in the range of 6%-8% (Fig. 1j).

Electrocatalytic performance

The electrochemical BER on Co3O4@CP and Co-OV@CP were probed by linear sweep voltammetry (LSV) in a three-electrode configuration. As displayed in Fig. 2a, the onset potential of the anodic oxidation reaction is clearly decreased from 1.12 V vs. SCE in Na2SO4 electrolyte to 0.82 V vs. SCE in NaBr electrolyte due to the more sluggish four-electron-transfer OER as compared to the two-electron-transfer BER. The Co-OV@CP electrode only requires a potential as low as 0.89 V vs. SCE to drive BER at a current density of 10 mA/cm2 in 0.5 M NaBr aqueous solution, which is 320 mV lower than that of OER in Na2SO4 electrolyte. Considering that a large amount of the electricity to overcome the high overpotential is consumed by OER during water electrolysis36,37, the BER coupled hydrogen production can decrease 30% of the total electricity consumption. Besides, the on-site generated Br2 from BER can be directly used for the synthesis of valuable Br-substituted chemicals and pharmaceuticals at ambient temperature and pressure. To create a compatible media for organic substrate and increase organic substrate solubility for the bromination reaction, acetonitrile was added into the NaBr aqueous solution. As shown in Fig. 2b, it can be seen that the addition of organic substrate into the NaBr electrolyte does not shift the LSV curve of bare glassy carbon (GC) electrode. Thus, optimizing the surface property of BER electrocatalyst is expected to improve the effectiveness of the bromination reaction. Both Co3O4@CP and Co-OV@CP show much enhanced BER activity as compared to CP, among which, Co-OV@CP displays the lowest BER onset potential at 0.82 V vs. SCE, which is about 50 mV and 150 mV lower as compared to Co3O4@CP and pristine CP, respectively (Fig. 2c and Supplementary Fig. 6). We made the corresponding Tafel plots using the iR corrected LSV curves. The Tafel slope for Co-OV@CP is estimated to be ∼48 mV dec-1, suggesting favorable kinetics for the two-electron transfer Br- oxidation over Co-OV@CP. While the Tafel slope (Fig. 2d) for Co3O4@CP is calculated to be ∼114 mV dec-1, indicating relatively sluggish Br- oxidation on Co3O4@CP38,39. Next, an electrolysis system coupling the Co-OV@CP-catalyzed BER with cathodic HER was established to verify the feasibility of substituting OER by BER to couple with HER for energy-saving hydrogen production. We evaluated the Faradaic efficiency (FE) for hydrogen production, and the measured volume of produced H2 gas matched well with the theoretical value (Supplementary Fig. 7), reaching a FEH2 as high as more than 99%. The possible products in the liquid phase were examined by NMR. As depicted in Supplementary Fig. 8, the NMR peaks at around 2.0 and 4.7 ppm of chemical shifts could be assigned to CH3CN and D2O, with no detectable peaks originating from CH3CN decomposition products, indicating the stability of acetonitrile in electrolysis.

Fig. 2: Electrochemical measurements.
Fig. 2: Electrochemical measurements.
Full size image

a LSV curves of Co-OV@CP recorded in 0.5 M Na2SO4 and 0.5 M NaBr aqueous solution, respectively. b CV curves recorded on glassy carbon (GC) electrode in 5 ml acetonitrile plus 5 ml 0.5 M NaBr aqueous solution with or without anisole as substrate. Scan rate: 100 mV/s. iR-corrected LSV curves (c) and the corresponding Tafel plots (d) in 0.5 M NaBr aqueous solution. Yields (e) and FE values (f) for bromination of anisole on Co-OV@CP and Co3O4@CP at different bias potentials. Potentiometric curve for electrochemical BER (g) (inset compares the applied potential20,21,22 required to drive BER at a current density of 10 mA cm-2) and cyclic bromination experiments (h) over Co-OV@CP. i Digital photograph showing a sunlight-driven electrochemical bromination cell using Co-OV@CP as the anode. The error bars correspond to data from three independent experiments.

We then optimized the solvent and the concentration of NaBr for electrochemical bromination of anisole at a constant applied potential of 1.1 V vs. SCE on Co-OV@CP. The mixture of equal volume of 0.5 M NaBr aqueous solution and acetonitrile was found to give the best performance in electrochemical bromination of anisole. Besides, the yield of Br substituted anisole varied with respect to the applied potential and pH of the reaction environment (Fig. 2e). Neutral reaction media favored the electrochemical bromination reaction. Less than 40% yield was obtained at 0.9 V vs. SCE after 4 h in neutral electrolyte and the yield increased to 94% if the applied potential was increased to 1.1 V vs. SCE. These results directly indicate that the Br2 generation rate determines the total electrochemical bromination rate, which can help us conclude that the bromination reaction is rate-limited at the oxidation of Br-. Faradaic efficiency (FE) offers a key parameter to assess the selectivity of electrochemical reaction. The FE for the formation of Br substituted anisole increased from 69.1% to ~91.6% with increasing applied anodic potential from 0.9 to 1.1 V vs. SCE within 2 h, while decreased gradually to 83.1% if the applied potential further increased to 1.3 V vs. SCE due to promoted OER (Fig. 2f). The FEs on Co3O4@CP show the same potential dependence, but are lower than those on Co-OV@CP across the entire potential window. Besides activity and selectivity, the Co-OV@CP electrode also displays good stability (Fig. 2g). Compared to the reported commercial electrodes such as carbon20, Pt19,21 and RuO222 electrodes, the Co-OV@CP electrode requires a much less overpotential to drive BER (Fig. 2g inset, Supplementary Table 1). The stability of Co-OV@CP can be further verified by cyclic electrochemical bromination experiments. Within four cycles of Br substituted anisole electrosynthesis over Co-OV@CP at 1.1 V vs. SCE, stable yield of more than 90% could be achieved (Fig. 2h). After electrochemical reaction, the Co-OV@CP electrode was examined by XPS and XRD (Supplementary Fig. 9–11), which displayed well-preserved crystal structure and valence states, and the surface oxygen vacancy signals of Co(OV)-O on the Co3O4-Ov catalyst remained almost unchanged, indicating good stability of this catalyst. Furthermore, we can make the electrochemical bromination reaction portable for on-site production of Br substituted chemicals through powering the electrosynthesis cell by commercial solar cells as demonstrated in Fig. 2i. This strategy is attractive that can transform solar energy into chemical bonds for green organic synthesis through the photovoltaic conversion.

Mechanistic studies

To dig out the excellent BER performance over Co-OV@CP, we estimated ECSA of Co-OV@CP and Co3O4@CP samples as shown in Supplementary Fig. 12. The ECSA of Co-OV@CP is only slightly increased as compared to that of Co3O4@CP. We then performed in-situ Raman spectroscopy measurements. As shown in Fig. 3a, the in-situ Raman spectra well illustrate the potential-dependent Br- adsorption in 0.5 M NaBr aqueous solution. Recorded on Co3O4@CP, the peak at 692 cm-1 with A1g symmetry originates from the octahedral CoO6 sites in crystalline Co3O4. The intensity of this A1g peak as well as Eg (487 cm-1) and F2g (529 cm-1) peaks over Co3O4@CP keeps nearly unchanged with increasing applied anodic potential (Fig. 3a and Supplementary Fig. 13a)40,41,42. Over Co-OV@CP, A1g, Eg and F2g peaks can also be clearly observed in the Raman spectra (Fig. 3b and Supplementary Fig. 13b), indicating that mild H2 reduction treatment would not damage the bulk Co3O4 spinel structure41. The new vibrational peak at 378 cm-1 compared with that in 0.5 M Na2SO4 aqueous solution (Supplementary Fig. 14) can be identified in the in-situ Raman spectra as the Ag phonon mode of Co-Br bond43,44, whose intensity varies insignificantly on Co3O4@CP while gradually decreases on Co-OV@CP with increasing applied anodic potential, indicating possibly easier desorption of *Br species on Co-OV@CP than on Co3O4@CP. Considering that the Co-O vibration in the as-developed electrocatalysts is quite stable during BER, the Co-Br to Co-O peak intensity ratio is able to tell the relative *Br coverage under the reaction conditions (Fig. 3c). Specifically, the peak intensity of the Co-Br vibration decreased by 33% when the applied bias potential increased from 0.6 to 0.8 V vs. SCE (Fig. 3c), which could be attributed to the activation of BER that accelerated desorption of adsorbed *Br on Co-OV. Note that both Co3O4@CP and Co-OV@CP display the Ag phonon mode of Co-Br even at the open circuit potential (OCP) in NaBr solution, suggesting spontaneous adsorption of Br-. Density functional theory (DFT) calculation was then performed to study the Br- adsorption and BER process. As shown in Fig. 3d-f and Supplementary Fig. 15, the adsorption energy of Br- on different sites of Co3O4 and Co-OV were carefully calculated. Br- can be easily adsorbed on spinel Co site of Co3O4 with Gibbs free energy of -0.74 eV, which explains the Ag phonon mode (378 cm-1) of Co-Br bond observed in the in-situ Raman spectra. After introduction of oxygen vacancies, the Br- adsorption on Co site next to the oxygen vacancy (denoted as Co(OV), Fig. 3g) was enhanced and showed a favorable adsorption energy of -0.79 eV. The adsorption of Br- over Co3O4@CP and Co-OV@CP in NaBr solution can be clearly observed in the XPS spectra (Supplementary Fig. 16)45, matching well with the DFT calculation results. While Br- adsorption energy on the Co site around the oxygen vacancy (denoted as Co(O), Fig. 3g) is 0.163 eV, considering that BER involves two up-hill electron transfer processes, this moderate unfavorable adsorption of Br- may facilitate the desorption of *Br to form Br2 over Co-OV@CP. The desorption energy barrier for *Br over Co-OV (also the rate-determining step of BER) is calculated to be 0.377 eV, which is much smaller than that over Co3O4 (1.28 eV) and matches well with the experimentally determined Tafel slope. All these results reveal that the Co site around the oxygen vacancy is likely the active site for BER over Co-OV@CP. Considering the coordinating structure of spinel Co3O4: typically, four Co atoms coordinate with one oxygen atom, and one oxygen vacancy can induce four Co(OV). The introduction of oxygen vacancies will create both strong Br adsorption sites (Co(OV)) and weak Br adsorption sites (Co(O), Supplementary Fig. 17). This contributes to explain the peak intensity decrease of Co-Br vibration in the in-situ Raman spectra over Co-OV@CP as shown in Fig. 3c. Removing oxygen vacancies via calcining Co-OV@CP in air at 300 °C for 1 h greatly decreased the BER performance (Supplementary Fig. 18), highlighting the significant role of oxygen vacancy on Co3O4 played for the BER. Additionally, projected density of states (pDOS) of Br on different Co sites of Co-OV were further calculated (Fig. 3f). The overlapping between Co 3 d and Br 4p orbitals was observed near but above the Fermi level in the pDOS of Br-Co(Ov), revealing the possible Br-Co(Ov) covalency, while overlapping between Co 3d and Br 4p orbitals was observed below the Fermi level in the pDOS of Br-Co(O), indicating unfavorable covalency of Br-Co(O) within the oxygen vacancy-rich Co3O4 framework, further explaining the calculated adsorption energies and confirming the Co atom (Co(O) site) next to Co(OV) as the active site over Co-Ov for electrochemical BER.

Fig. 3: Reaction mechanism.
Fig. 3: Reaction mechanism.
Full size image

In-situ Raman spectra for BER recorded on Co3O4@CP (a) and Co-OV@CP (b) at different bias potentials (potential vs. SCE). c The in-situ Raman peak intensity for Co-Br vibration at 380 cm-1 recorded on Co3O4@CP and Co-OV@CP at different bias potentials. The calculated adsorption energy of *Br on Co3O4 (d) and Co-OV (e). Projected density of states (pDOS) of Co-OV after adsorbing *Br on Co(Ov) site or Co (O) site (f). g Schematic illustration showing possible *Br adsorption sites on Co3O4 and Co-OV samples. The error bars correspond to data from three independent measurement.

Under optimized reaction conditions (Supplementary Table 3), the reaction scope was carefully investigated (Fig. 4). Representative substrates bearing electron-donating or electron-withdrawing substituents such as methoxy-, methyl- or cyan- on arenes (2a-2m) could be efficiently converted to the corresponding brominated products (with 76-95% yields) at the most electrophilic position6,9,10. Para-position of strong electron-donating functional group is more preferrable for bromination (2 d, 2e), while ortho-position is preferred only if the para-position is blocked (2b, 2c). Sensitive functionalities including alkyne, amine, amide, ester, and ketone, were found to be well-compatible, giving the desired brominated products (e.g., 2c, 2e, 2i, 2 m etc.) in good yields. Substrates composed of halogen substituents, such as Br- (2j) and Cl- (2b), could be smoothly brominated in high yields (90% and 88%, respectively), indicating the possibility of stepwise bromination. Dual bromination of certain substrates (e.g., 2 h) could be achieved by increasing the applied voltage (e.g., 1.3 V vs. SCE). It is delightful to see the highly selective bromination of naphthalene derivatives, which furnished the value-added products 2n and 2o in good to excellent yields. Importantly, bromination of heterocyclic compounds such as thiophene (2p), parazole (2q), indoline (2r), quinoline (2 s) and 1,2,3,4-tetrahydroquinoline (2t) were also successfully achieved, confirming the wide compatibility of the electrochemical bromination strategy. The electrochemical bromination strategy could be further extended to chlorination (2u-2v) by simply replacing NaBr with NaCl. The Faradaic efficiency for some above products in Fig. 4 can be calculated in the range of 70–86% (Supplementary Table 4).

Fig. 4: Substrate dependent electrochemical bromination/chlorination reaction.
Fig. 4: Substrate dependent electrochemical bromination/chlorination reaction.
Full size image

Reaction conditions: Co-OV@CP (size: 10 mm × 10 mm × 0.3 mm) anode and platinum plate (size: 10 mm × 10 mm × 0.2 mm) cathode in an H-type cell divided by Nafion membrane, SCE as reference electrode; NaBr (0.5 M in deionized water, 7.5 mL), CH3CN (7.5 mL) in each chamber, 1 (0.5 mmol), constant voltage (0.9–1.3 V vs. SCE) at room temperature for 4–8 h. Isolated yields are shown. *GC-MS yield. #The reaction was carried out under standard conditions for 2 h followed by adding p-chloroaniline (0.5 mmol).

Late-stage bromination of natural products, drugs and their derivatives are of great significance for biological applications and drug developments6,7,8,9,10 (Fig. 5). Solvents were further optimized for increasing the solubility of natural products with large molecular weight. A variety of biomolecules such as guaiacol derivative (2w), psoralen ether derivative (2x), gemfibrozil (2 y) and naproxen-derived methyl esters (2z) could be facilely brominated with excellent isolated yields. The dual bromination of a flavoring agent was demonstrated, affording 2,5-dibromide product 2aa with 78% yield. As one type of spice, phenoxyethyl isobutyrate (2ab) could be brominated at the para position of the alkoxy functional group. Adapalene methyl ester, a more complex drug derivative, reacted smoothly to produce ortho-bromination product 2ac in 85% yield. For δ-tocopherol derived substrate, bromination tended to take place on the less hindered position (2ad). This tool was also applied to deuterated biomolecules, to furnish the corresponding brominated products (2ae-ag), which could be served as important building blocks for the synthesis of deuterated drugs. The electrochemical bromination reaction can be easily scaled up in a flow cell system (photograph in Fig. 5, Supplementary Fig. 19). As displayed in Fig. 5, Naproxen (2x’), psoralen(2z’) and δ-tocopherol (2ad’) derivatives could be brominated to produce gram-scale products with good isolated yields, pinpointing the great potentials for flow-synthesis of valuable brominated chemicals and pharmaceuticals. Impressively, an unprecedented 0.61-kilogram-scale synthesis of brominated naproxen derivative was achieved using 4 tandem cells with an electrode area of 3 × 3.5 cm2, illustrating the great potential for industrial-scale bromination of this method with sustainable and cheap NaBr. Compared to traditional bromination routes via Br2, NBS or other non-natural bromination reagents, this strategy will benefit from unnecessary complex processes with high energy-consumption, tough transportation/storage and/or environmental unfriendliness (Supplementary Fig. 20). As can be seen from the economic analysis (inset figure in Fig. 5 and economic analysis in Supplementary Information including Supplementary Fig. 21–23 and Supplementary Tables 5, 6), the electrochemical BER route to produce 2z can save near 50% of the production cost, considering the compatibility of this method with solar power plant (Fig. 2i). In addition, we also conducted the experiments of in-situ electrochemical bromination of anisole using simulated brine (50 g L-1 Cl- + 3 g L-1 Br-) as the electrolyte. LSV curves (Supplementary Fig. 24) show that the BER is more favorable than ClER on the Co3O4-Ov/CP catalyst with a ~ 0.2-V negatively shifted onset potential, even though the concentration of Cl- is remarkably higher than that of Br-. This is a consequence of the lower theoretical potential of Br--to-Br2 conversion (1.08 V vs. SHE) than that of Cl--to-Cl2 conversion (1.35 V vs. SHE). Thus, a potentiostatic electrolysis can be conducted at around 1.1 V vs. SCE that only ignited BER but not ClER for in-situ electrochemical bromination reaction even in low-cost brine electrolytes.

Fig. 5: Direct bromination of natural products and drug derivatives.
Fig. 5: Direct bromination of natural products and drug derivatives.
Full size image

Reaction conditions: Co-OV@CP (size: 10 mm × 10 mm × 0.3 mm) anode and carbon paper (size: 10 mm × 10 mm × 0.2 mm) cathode in an H-type cell divided by Nafion membrane, SCE as reference electrode; NaBr (0.5 M in deionized water, 12.5 mL), CH3CN (12.5 mL) in each chamber, 1 (0.5 mmol), constant voltage (1.1-1.3 V vs. SCE) at room temperature for 4-8 h. Isolated yields are shown. bReaction conditions: Co-OV@ carbon felt (size: 35 mm × 30 mm × 5 mm) anode and carbon felt (size: 35 mm × 30 mm × 5 mm) cathode. Substrate (4 mmol), solvent (0.5 M NaBr in deionized water (10 mL): CH3CN: CH3OH: CH2Cl2 = 10:10:10:3), flow cell, two-electrode system and constant voltage (4-5 V), at room temperature, flow rate: 5 mL/min. Isolated yields are shown. cReaction conditions: the same as condition b except for 4 tandem cells, constant voltage (6 V), 18 h, flow rate: 50 mL/min, 518 g substrate.

With fast development of solar, wind or other green power technology, the further reduced electricity cost will enable a further 20% cost cut for the BER method in the near future (inset figure in Fig. 5 and Supplementary Fig. 23) and will achieve a sustainable and efficient bromine industry compared to traditional energy-intensive and environmentally unfriendly bromination method10,11,46,47.

Discussion

In conclusion, we rationally designed and prepared oxygen-vacancy rich Co3O4 nanosheets on carbon paper to facilitate electrochemical BER in aqueous NaBr solution for on-site synthesis of value-added natural products and drug derivatives. The as-developed electrochemical synthesis process is efficient and sustainable with broad substrate scopes that is able to achieve kilogram-scale synthesis of drug derivatives. Our findings shall contribute to the design of thermodynamically favorable oxidation reactions coupled with hydrogen evolution to make value-added products in the carbon neutrality century.

Methods

Chemicals

Cobalt(II) nitrate hexahydrate (Adamas-beta Reagent Co., Ltd, RG, 99.9%), acetonitrile (Adamas-beta Reagent Co., Ltd, AR, 99%), sodium bromide (Aladdin, AR, 99%), anisole (Aladdin, GC, 99.9%), ultra-pure water (18.2 MΩ cm−1, Millipore Corporation, Billerica), chloroform-D (Adamas, RG, 99.8 % D), and other chemicals were used as received without further purification.

Electrode preparation

Carbon paper (Shanghai Hesen Electric Co., Ltd., 0.2 mm thick) was cut into 10 × 20 mm2 pieces and washed twice with deionized water and ethanol, respectively. After being dried in a vacuum oven, the tailored carbon paper was immersed in a 0.1 mol/L cobalt nitrate aqueous solution. Cyclic voltammetry (CV) from 0 V to -1.2 V vs. saturated calomel electrode (SCE) at a scan rate of 100 mV/s was then performed in a three-electrode configuration, in which carbon paper, platinum foil, SCE were used as the working electrode, counter electrode and reference electrode, respectively. After the reaction, the as-obtained carbon paper was rinsed with deionized water and calcined at 300 oC for 2 h in air (named as Co3O4@CP). To introduce oxygen vacancies onto Co3O4, the Co3O4@CP was calcined again at 300 oC for 15 min in 5% H2/Ar atmosphere (named as Co-OV@CP).

Characterization

The morphological information was obtained by scanning electron microscopy (SEM) (MIRA3, TESCAN). The TEM measurement was taken with a HITACH HT7700 microscope operated at an acceleration voltage of 80 kV. The crystal structure was characterized by X-ray diffraction (XRD) (Ultima IV, Rigaku) operated at 40 kV and 40 mA (Cu Kα X-ray radiation source) at a scanning speed and step interval of 6°/min and 0.01°, respectively. X-ray photoelectron spectroscopy (XPS) measurements were conducted on a Thermo Scientific K-Alpha XPS system using a monochromated Al Kα radiation. The inductively coupled plasma (ICP) emission experiments were conducted on an ICP-OES Optima 7000DV spectrometer (Perkin Elmer) or on an ICP-MS spectrometer (PerkinElmer NexION 300X). Electron Paramagnetic Resonance (EPR) tests were conducted on a BRUKER A300-10/12 instrument. Nuclear magnetic resonance (NMR) tests were performed on a Bruker AVANCE III NMR spectrometer (600 MHz). The organic products were analyzed by gas chromatography mass spectrometry (GC-MS, Agilent 7890 A) using 0.3 mmol toluene as the external standard. The high-resolution mass spectrometry (HRMS) was conducted by a Q Exactive GC Orbitrap GC-MS/MS (Thermo Scientific). In-situ surface-enhanced Raman spectra were recorded at the following applied potentials: 0.6 V, 0.8 V, 1.0 V, 1.2 V, and 1.4 V (vs. SCE) in 0.5 M NaBr aqueous solution at room temperature.

Electrochemical measurements

Linear sweep voltammetry (LSV) and cyclic voltammetry (CV) were carried out in an undivided cell filled with 0.5 mol/L NaBr aqueous solution mixed with an equal volume of acetonitrile. All electrochemical measurements were performed on a CHI 660E electrochemical workstation in a three-electrode configuration with glass carbon electrode, platinum foil, SCE as the working, counter, and reference electrode, respectively. The electrochemical synthesis of brominated chemicals was performed on Co-OV@CP in a H-cell. Typical reaction conditions: each cell was filled with 7.5 mL acetonitrile and 7.5 mL 0.5 mol/L NaBr aqueous solution. 0.5 mmol substrate was dissolved in the anodic cel,l and constant voltage was applied. For large-scale production, DC power supply and carbon felt cathode was used.

DFT calculations

Spin-polarized density functional theory (DFT) calculations were performed using the Vienna ab initio Simulation Program (VASP)48,49. The electron exchange-correlation potential was treated by the Perdew-Burke-Ernzerhof (PBE) function50 within the generalized gradient approximation. To simulate the adsorption of Br on Co3O4 surface, the preferred oriented crystallographic (110) plane of Co3O4 cubical phase as suggested from the TEM and XRD characterization, was constructed by a six-layer slab with a vacuum region of 20 Å along the z-axis to minimize the periodic image interactions. An energy cutoff was set to 520 eV to expand the wave functions. Brillouin zone integration was sampled with 2 × 3 × 1 Monkhorst-Pack50 mesh k-points grids for geometry optimizations and 20 × 30 × 1 k-points grids for electronic property calculations of Co3O4 (110) slab, respectively. The outer valence electron arrangement of each element was set 2s22p4 for O, 3d74s2 for Co, 4s24p5 for Br, respectively. DFT + U method was adopted with a typical U value of 3.4 eV for the consideration of the on-site Coulomb and exchange interactions of the strongly localized Co 3 d electrons.

All geometries were fully optimized with the convergence tolerances set to be 0.015 eV/ Å for forces on all atoms. The energy corrections for the slab-to-slab dipole interaction along the surface perpendicular direction was employed51,52. The adsorption energy (\({E}_{*{{{\rm{Br}}}}}\)) of Br on the Co3O4 surface was calculated as follows:

$${E}_{*{Br}}={E}_{{Surf}+{Br}}-{E}_{{Surf}}-{1/2E}_{{{{{\rm{Br}}}}}_{2}}$$

where \({E}_{{Surf}+{Br}}\) is the total energy of Co3O4 (110) slab adsorbed with Br, \({E}_{{Surf}}\) is the energy of the Co3O4 (110) surface and \({E}_{{{Br}}_{2}}\) is the energy of the adsorbed bromine. Then, the free energy of the adsorbed state is calculated as:

$$\Delta {G}_{*{Br}}=\Delta {E}_{*{Br}}+{\Delta E}_{{ZPE}}-T\Delta S$$

where \({\Delta E}_{{ZPE}}\) is the difference corresponding to the zero point energy and \(\Delta S\) is the entropy.