Organic small-molecule catalysts for electrochemical 2-methyl-3-butyn-2-ol (MBY) semihydrogenation
Oxygen-containing functional groups (OFGs), such as –OH and –COOH, can serve as Brønsted acid sites24; thus, we began this study by testing the ASH activity of simple molecules with –OH and –COOH groups. For comparison, molecules with neutral OFGs, such as C–O–C and C = O, were also tested. 2-Methyl-3-butyn-2-ol (MBY), a key stock in vitamin A production, was used as the
model substrate. The ASH activity was first studied by cyclic voltammetry (CV) (Fig. 1a). No current density variation is observed for the glassy carbon before and after adding MBY. Interestingly, a pair of redox peaks arises after introducing the molecular catalyst, indicating that electron transfer occurs between MBY and the molecular catalyst. Then, an electrolysis experiment was conducted at −1.6 V vs. Ag/AgCl (all potentials are referenced to the saturated Ag/AgCl unless otherwise stated), and the products were identified and quantified by 1H nuclear magnetic resonance (1H NMR, Supplementary Fig. 1). The four molecular catalysts exhibit ASH activity, and the molecular catalysts with –OH and –COOH show better activities (Fig. 1b). No activity was found over the pure carbon electrode (Supplementary Fig. 1). Notably, the experimental results are consistent with the DFT calculations (Figs. 1c − d). However, the reaction rate is very slow due to the inadequate contact of homogeneous molecular catalysts with the cathode. These results inspired us to construct –OH and –COOH groups in situ on conductive and stable GP.
Synthesis and characterization of EO-GP and EOR-GP
We treated the GP with an electrooxidation method followed by an electroreduction treatment to enrich it with –OH and –COOH groups (Fig. 2a). GP was first oxidized in 0.5 M H2SO4 at a current density of − 200 mA cm–2 for 0.5, 1.0, and 1.5 h to obtain electrooxidized GP (EO-GP-t, t = 0.5, 1.0, or 1.5) with different oxidation degrees. Then, a subsequent electroreduction treatment was conducted to remove reducible groups such as C–O–C and C = O groups to obtain EOR-GP-t. X-ray diffraction (XRD) patterns (Supplementary Fig. 2) of the EO-GP and EOR-GP samples exhibit diffraction peaks at 26.5°, 44.5°, 54.5°, and 77.6°, which are attributed to graphitic carbon (PDF no. 01-089-8487). Compared with those of the initial GP, the peak intensities of EO-GP and EOR-GP decrease, suggesting that the crystallinity of the GP decreases after electrochemical treatment25. Scanning electron microscopy (SEM) images (Supplementary Fig. 3) show that the GP edges become curled and thinner after electrooxidation and electroreduction. Raman spectra (Supplementary Fig. 4) show that the ID/IG ratio increases with increasing oxidation time, indicating an increase in carbon defects26. The defects remain unchanged after electroreduction.
Spectral characterization was used to characterize the OFGs on the GP before and after electrochemical treatment. X-ray photoelectron spectroscopy (XPS) survey scans (Fig. 2b and Supplementary Fig. 5a) show that a new O 1s peak emerges after electrooxidation and then becomes more obvious with increasing electrooxidation time. After electroreduction, the peak become weaker due to the removal of the reducible OFGs (Fig. 2b and Supplementary Fig. 5b). In the O 1s spectra (Fig. 2c), the initial GP shows a weak peak of –OH (533.3 eV)27 from aromatic phenols (Ar–OH) or aliphatic alcohols (C–OH), possibly due to slight oxidation during storage and cleaning. After electrooxidation, the –OH peak increases greatly, and new oxygen species of C = O (531.3 eV, ketone and carboxylic acids, etc.)24,26,27,29, O–C = O (532.2 eV, carboxylic acids, etc.)24,26,28,29, and C–O–C (534.4 eV, furan, etc.)24,26,28,29 appear. As the oxidation time increases, the three new peaks increase, while the –OH peak decreases, and O–C = O and C–O–C are the dominant groups in EO-GP-1.5 (Supplementary Fig. 6a). After electroreduction, the C–O–C peak disappears, the C = O peak becomes much weaker, while the –OH peak increases, and the O–C = O peak remains unchanged (Supplementary Fig. 6b). These results show that C–O–C and C = O are reduced to –OH, but O–C = O cannot be reduced during the electroreduction treatment. Thus, –OH and O–C = O are the dominant groups in EO-GP-1.5. Similar results are obtained for the C 1s spectra (Supplementary Fig. 7).
Additionally, soft X-ray absorption near edge structure (XANES) was used to further confirm the OFGs. In the O K-edge XANES spectra (Fig. 2d), a weak peak at 534.5 eV for the initial GP is attributed to the aromatic phenol (σ* Ar–OH)24,28 After electrooxidation, a broad peak attributed to C = O (542.1 eV)21, 25, O–C = O (539.7 eV)24,26,28,29, and C–O–C/C–OH (537.4 eV)24,26,28,29 appear simultaneously. Another new peak at 533.5 eV is attributed to the π* C = O in quinone. After electroreduction24,26,28, the π* C = O and σ* C–O–C peaks decreased significantly, while the O–C = O and C–OH peaks are preserved. In the C K-edge XANES spectra (Fig. 2e), four new peaks at 286.9, 287.7, 286.5, and 288.5 eV arise in the spectrum of EO-GP-1.5, which are attributed to π* C–O–C, π* C = O, C–OH and σ* O–C = O, respectively24,28,29. After electroreduction, only σ*, O–C = O and C–OH are retained in EOR-GP-1.5. These results are consistent with the XPS data. A similar variation is observed in attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy (Supplementary Fig. 8). These results demonstrate the successful construction of –COOH- and –OH-enriched EOR-GP.
EOR-GP for electrochemical 2-methyl-3-butyn-2-ol (MBY) semihydrogenation
The electrochemical ASH performance of GP and EOR-GP was tested. CV curves (Fig. 3a and Supplementary Fig. 9) show that the hydrogen evolution reaction (HER) performance of EOR-GP is better than that of GP, and EOR-GP-1.5 shows the best HER performance. After the addition of MBY, GP exhibits a higher onset potential and a lower current density compared with that of no MBY, possibly due to the larger electrolyte resistance caused by MBY. In contrast, EOR-GP shows a much lower onset potential and a noticeable increase in current density after the introduction of MBY, indicating the rapid conversion of MBY on EOR-GP. Constant potential electrolysis experiments show that the electrochemical ASH performance increases with increasing oxidation degree, and EOR-GP-1.5 exhibits the best performance of 72% FE, 96% yield, 99% selectivity and 0.48 mmol cm− 2 h− 1 yield rate of MBE at the optimum potential of − 1.5 V (Figs. 3b−d and Supplementary Fig. 10). Notably, the EOR-GP-1.5 performance is comparable to that of Pd-based catalysts30. In comparison, GP shows negligible performance for electrochemical ASH (Fig. 3d). The ASH performance and catalyst composition can be maintained for 40 h, verifying the good durability of EOR-GP-1.5 (Fig. 3e and Supplementary Fig. 11).
Control experiments were performed to gain insight into the origin of the activity of EOR-GP. To eliminate the activity difference caused by the active area, the normalized yield rates of the electrochemical surface areas (ECSAs) were calculated (Supplementary Fig. 12). EOR-GP-1.5 shows much better activity, ruling out the contribution of a high ECSA to the performance enhancement (Supplementary Fig. 13). Furthermore, to observe whether the defects contributed to the performance enhancement, EOR-GP-1.5 was further treated in a mixed hydrogen/argon atmosphere (denoted as EOR-GP-1.5-H2) to remove the OFGs but retain the defects31. XPS and Raman data (Supplementary Fig. 14) reveal that this hydrogen reduction process removes most of the OFGs while the defects are still retained. As a result, EOR-GP-1.5-H2 shows very negligible electrochemical ASH activity (Fig. 3f), excluding the contribution of defects. Additionally, we treated GP via chemical oxidation through concentrated nitric acid at 100 ℃ for 25 h, and the obtained COR-GP was enriched with –COOH and –OH and exhibited activity comparable to that of EOR-GP-1.5 (Fig. 3f and Supplementary Figs. 15−17). These results suggest that the activity of EOR-GP-1.5 originates from the OFGs on the carbon surface.
Mechanistic studies
Further study was conducted to elucidate the reaction mechanism. Generally, two mechanisms are involved in electrochemical reduction: electrocatalytic hydrogenation (ECH) and direct electroreduction32. ECH involves strong interactions with the electrode surface and reactants, requiring Hads produced via H2O splitting as a hydrogen source. In direct electroreduction, protonation occurs in solution and does not require direct chemical interactions with the electrode surface. A previous study on metal electrodes for electrochemical ASH proposes an ECH mechanism2. However, carbon-based electrodes are reported to favour the direct electroreduction mechanism in the electrochemical reduction reaction32, 33. This finding inspires us to explore the mechanism of ASH on EOR-GP-1.5. First, a kinetic isotopic effect (KIE) for the HER is observed for EOR-GP-1.5 (Fig. 4a). The CVs in Fig. 4b show that MBY reduction also experiences a KIE with a similar magnitude to that of the HER, indicating that the ASH reaction may share a common step with the HER and involve reactions with Hads (ref.32). As expected, a decreased production rate of MBE is observed in the D2O electrolyte at − 1.5 V (Fig. 4c). Moreover, a long-chain thiol, which can block surface sites but allow electron tunneling32, 34, are employed to modify the EOR-GP-1.5 electrode. Almost no MBE is detected after thiol modification, demonstrating that the specific adsorption of MBY or H is
indispensable for MBY reduction (Fig. 4d). Similar results are obtained by electrochemical in situ Raman spectroscopy (Fig. 4e). In the spectrum of EOR-GP-1.5, peaks from the stretching vibrations of C ≡ C (2107 cm− 2) and C − H (2943 and 2992 cm− 2)35 in MBY increase as the applied potential increases. Accordingly, the peak attributed to the stretching vibration of C = C (1448 cm− 2)35 arises at − 1.3 V and increases with increasing potential, which is consistent with the experimental results. In comparison, no Raman signals of MBY or MBE are observed for the glassy carbon electrode without OFGs. These results indicate that the adsorption of MBY is the precondition for realizing MBY reduction and that the OFGs probably play a key role in promoting MBY adsorption, as also suggested by DFT calculations (Supplementary Fig. 18). Moreover, the reaction intermediates of the enol and active H species are confirmed by high resolution mass spectrometry (HRMS, Fig. 4f). Based on the above analysis, a possible ECH mechanism is proposed (Fig. 4g). First, MBY adsorbed on the EOR-GP-1.5 surface and activated by receiving an electron from the cathode. Then, the activated substrate processes the addition of two active H from H2O splitting to produce MBE, which can desorb quickly to avoid overhydrognation.
Computational analysis
DFT calculations were conducted to further understand the activities of the OFGs towards ASH (see the supporting information for calculation details, Supplementary Tables 1–3 and Supplementary Fig. 19). The adsorption energy of C5H9O* (Gad (M-C5H9O*)) is used as the descriptor of MBY reduction activity based on the scaling relationships between the adsorption energies of the adsorption intermediates involved in the reaction (Supplementary Fig. 20−21). The limiting potential UL is defined as the lowest potential36 at which the free energy of all reaction steps decreases and is considered a metric of activity. Figure 5 and Supplementary Fig. 22 show the calculated UL as a function of Gad (M-C5H9O*) based on the OFGs grafted onto graphene (graphene-OFGs, blue dots) or the OFGs containing molecules (molecular-OFGs, red dots). Among the different graphene-OFGs, the –COOH and –OH groups on the basal plane and at the edge of the graphene are located near the summit of the volcano. Thus, we reasonably speculate that –COOH and –OH on EOR-GP-1.5 are both active sites for electrochemical ASH. For molecular-OFGs, we found that –COOH is located closest to the summit of the volcano, with a theoretical overpotential of 0.53 V vs. RHE. –OH and C = O are the second closest to the volcanic summit, with overpotentials of 0.92 and 1.00 V vs. RHE, respectively. C–O–C is slightly farther from the summit of the volcano, with an overpotential of 1.14 V vs. RHE. Pure C exhibited the least activity on the right side of the volcano plot due to the weak bond energy of M-C5H9O*, with an overpotential of 1.54 V vs. RHE. The calculation results are not exactly the same for graphene-OFGs or molecular-OFGs, probably due to the difference in the adsorption scaling relationship between the two catalysts (Supplementary Fig. 21). For example, although graphene-OH and molecular-OH have similar binding energies for MBY (C5H8O*), graphene-OH has a much stronger binding energy for the further hydrogenated product of MBY (M-C5H9O*), resulting in better molecular-OH activity (Fig. 5). These DFT results confirm that the activity probably originates from the –COOH and –OH groups on the GP.
Utility and universality study
The utility and universality of the catalyst were studied. We successfully implemented a flow electrolytic cell for the electrosynthesis of gram-scale MBE from MBY electroreduction with 81% FE and 10.4 g yield (Fig. 6a and Supplementary Fig. 23). Moreover, EOR-GP-1.5 is also applicable in gas-phase acetylene electrocatalytic semihydrogenation, where a 92% FE is achieved at a potential of − 1.7 V (Fig. 6b). Additionally, a series of alkynes with electron-donating and electron-withdrawing groups on the phenyl ring and aliphatic terminal alkynes are converted to alkenes with high selectivity and yield (Fig. 6c and Supplementary Fig. 24). Furthermore, EOR-GP-1.5 is suitable for other electrocatalytic hydrogenation reactions, such as the hydrogenation of nitrobenzene to aniline and the dehalogenation of halides (Fig. 6d and Supplementary Fig. 25), showing the promising potential of this metal-free catalyst in other electrocatalytic hydrogenation reaction systems.