Tuning Single-Atom Dopants on Manganese Oxide for Selective Electrocatalytic Cyclooctene Epoxidation

Selective and efficient electrocatalysts are imperative for the successful deployment of electrochemistry toward synthetic applications. In this study, we used galvanic replacement reactions to synthesize iridium-decorated manganese oxide nanoparticles, which showed a cyclooctene epoxidation partial current density of 10.5 ± 2.8 mA/cm2 and a Faradaic efficiency of 46 ± 4%. Results from operando X-ray absorption spectroscopy suggest that manganese leaching from the nanoparticles during galvanic replacement introduces lattice vacancies that make the nanoparticles more susceptible to metal oxidation and catalyst reconstruction under an applied anodic potential. This results in an increased presence of electrophilic oxygen atoms on the catalyst surface during reaction conditions, which may contribute to the enhanced electrocatalytic activity toward cyclooctene epoxidation.


A.2.1. Ir single -MnO x synthesis
BF 4 -treated MnO x nanoparticles were prepared on a carbon paper electrode (MnO x /CP) or a ITO glass electrode using a previously described method. 1 Briefly, sub-10-nm-sized MnO x nanoparticles were synthesized by using hot injection method and ligand exchange was performed with NOBF 4 to remove the organic ligand on the manganese oxide nanoparticle. Hydrophilic carbon paper electrodes were prepared by heating 14-mm diameter hydrophobic carbon papers in a muffle furnace at 600 °C for 1 hour. Next, 15 μL of ligand-exchanged MnO x nanoparticles in ethanol solution were dropped on the hydrophilic carbon paper electrode in an oven at 80 °C. The drop-casting was repeated 4 times; 2 drops on the front side and 2 drops on the back side. The electrode was then transferred to a muffle furnace. Ramp rates for a furnace were set to be 10 °C/min from room temperature to 400 °C and dwelled for 5 hours. The MnO x nanoparticle electrodes were naturally cooled to room temperature in the furnace. FeO x nanoparticles were synthesized in a similar way, using iron acetate instead of manganese acetate precursor.
To perform the Ir decoration, typically, four MnO x /CP electrodes were attached in a beaker using Kapton tape. The beaker was then placed in a 50 ℃ water bath. 40 mL of 0.1 mM K 2 IrCl 6 aqueous solution was prepared separately in a conical centrifuge tube and the temperature of the solution was adjusted to 50 ℃. We noted that the temperature of the iridium precursor-containing solution should be adjusted before adding it to the beaker. The solution was stirred by a mini magnetic bar during the decoration process ( Figure S1). After 30 min, the electrodes were removed from the beaker, gently rinsed with de-ionized water several times, and dried in the 80 ℃ oven for 30 min. A similar procedure was used to make Pt single -MnO x . 100 µM K 2 Pt(IV)Cl 6 aqueous solution was used and synthesis temperature and dwelling time were set to be 50 ℃ and 30 min, respectively. Figure S1. A water bath setup for Ir decoration on MnO x /CP.

A.2.2. Ir few -MnO x synthesis
For the synthesis of Ir few -MnO x nanoparticles, the concentration of the precursor remained same and dwelling time at 50 ℃ was set to be 2 hours. After synthesis, the electrodes were gently rinsed with de-ionized water several times and dried in the 80 ℃ oven for at least 30 min.

A.2.3. Ir few/cluster -MnO x nanoparticle synthesis
For the synthesis of Ir few/cluster -MnO x nanoparticles, increased temperature and higher precursor concentration, 0.25 mM K 2 IrCl 6 solution at 70 ℃ was used. Dwelling time at 70 ℃ was set to be 30 min.

A.2.4. Ir cluster -MnO x nanoparticle synthesis
Sub-2-nm-sized iridium clusters were formed on MnO x nanoparticles when we used 0.25 mM K 2 IrCl 6 solution at 70 ℃ with 2h of dwelling time. Compared to Ir few/cluster -MnO x nanoparticle, the portion of clustered iridium particles on the MnO x were significantly higher, which was confirmed by HR-TEM.

A.3. TEM analysis
High-resolution transmission electron microscopy (HR-TEM) images and scanning transmission electron microscopy (STEM) images were obtained using a high-resolution transmission electron microscope A.4. X-ray photoelectron spectroscopy XPS sample preparation and characterization. After electrolysis, the electrodes were rinsed with acetonitrile several times and dried under ambient air overnight. The electrodes were cut into small pieces and affixed to an XPS sample holder using conductive Cu tape. Then, the holder was transferred to the XPS sample introduction chamber. XPS spectra were collected using a Physical Electronics Versaprobe II X-ray Photoelectron Spectrometer and the data were analyzed with Casa XPS software.

B.1. Electrochemical methods
Electrochemical experiments were conducted with a sandwich-type one-compartment electrochemical cell. Platinum foil and M 1 -M 2 O x loaded carbon paper were used as the counter and working electrode, respectively. An Ag/AgCl electrode (3.4 M KCl leak-free 2.0 mm diameter, Innovative Instruments) was used as the pseudo-reference electrode, and aluminum foil was used as the current collector. Hydrophobic carbon paper was placed behind the nanoparticle-loaded carbon paper to prevent leakage of the solvent. Acetonitrile S5 with 0.11 M TBABF 4 was used as the solvent with varying concentrations of cyclooctene and water. The total volume of electrolyte was set to be 4 mL. The resistance value at open circuit potential (OCP) was measured by electrochemical impedance spectroscopy techniques. In Figure 4A & B, the potentials were 100% iR compensated (85% automatic compensation with the last 15% manually compensated for potential-dependence experiments), and in other experiments, the potentials were 85% IR-compensated automatically. The electrochemical measurements were conducted with VMP3 Multi-channel potentiostat from BioLogic. For the potential dependent and water-dependent experiments, 10 coulombs were passed. For the substrate-dependent analysis, 5 C was passed for 10, 25, and 50 mM cyclooctene, and 10 C were passed for 100 and 200 mM cyclooctene.
The Ag/AgCl pseudo-reference electrode was calibrated against the ferrocene/ferrocenium (Fc/Fc + ) redox potential, which was obtained by averaging the potential of the oxidative and reductive peaks from the cyclic voltammetry with a scan rate of 50 mV/s. The calibration solutions were prepared by adding ~5 mM of All cell parts including the counter electrode were washed with acetone and MilliQ water after every experiment and sonicated for at least 15 minutes in 10% nitric acid aqueous solution to remove any contaminants. The cell parts were washed with MilliQ water again after the sonication and oven-dried at 80 °C for at least 20 minutes before cell assembly to remove any residual moisture.

B.2. Long-term chronoamperometry
Long-term electrolysis with a small amount of stagnant electrolyte can cause the accumulation of products, which can be detrimental to the catalyst performance. Therefore we refreshed the electrolyte (4.0 mL) every ~15 min during the chronoamperometry experiment for over two hours. When switching the electrolyte, the electrolysis was paused and restarted within 1 min. The potential-time plot in Figure S3 includes those interval times as well as the electrolysis times.

C. Product analysis
After the electrolysis, post-electrolysis electrolytes were collected using glass pipette. The collected electrolyte was often less than 4 mL, so ~ 0.2 mL of acetonitrile was added to the electrolyte chamber, rinsed the chamber, and then added the solution to the post-electrolysis electrolyte until the total volume of collected electrolyte becomes 4 mL. We sampled 2 mL of the collected electrolyte, and ~ 13 mL of additional water, ~ 5 mL of hexane, and the internal standard 1,3,5-trimethoxybenzene were added to the electrolyte. Cyclooctene, its oxidation products, and the internal standard moved to the hexane layer after extraction with the help of centrifugation at 3000 rpm for 30 seconds. Extraction with hexane was required to exclude TBABF 4 that can overlap with the products on 1 H-NMR spectra. Hexane extraction was performed three more times (~ 20 mL of hexane was used in total), and the combined hexane solution was quantified using 1 H-NMR. The 1 H-NMR spectra were measured on a Bruker Avance Neo spectrometer operating at 500.34 MHz with solvent suppression of the three largest peaks from hexane. Samples were prepared by combining 50 μL of the extracted hexane layer with a 600 μL CDCl 3 . The 1 H probe was automatically tuned, the lock was achieved on the CDCl 3 in the sample, and gradient shimming and auto-gain were used. 32 scans were collected per sample, and relaxation delay was set to 6 seconds. Products were quantified by comparing the areas of the peaks of interest to those of S7 the internal standard, 1,3,5-trimethoxybenzene. Auto phase correction and baseline correction with Whittaker Smoother method from the MestReNova software was performed on the spectra before the integration of each peak.

E. Discussion on the cyclooctene epoxidation mechanism
The proposed mechanism in our previous work includes chemical reactions as below.  is the resting state and the population of it during the reaction cycle outcompetes that of other surface species, other terms in the dominator except for the second term can be neglected, giving a theoretical Tafel slope of 118 mV/dec and first-order dependencies on cyclooctene and water activity. However, when other surface species are present in comparable amounts to the metal-oxo species, it is difficult to expect a cardinal Tafel slope, and non-linearity would be observed in the Tafel plot. Therefore, the observed Tafel slope of 93±8 mV/dec ( Figure S4) for Ir single -MnO x catalysts toward cyclooctene epoxidation may reflect a more distributed surface species population. The order dependencies can be complicated as well when there are no dominant species. However, in our system, the given rate equation indicates that the first order in the water activity would be conserved in any case, and first-order in cyclooctene would be conserved as well if the sum of occupied sites is much bigger than the number of vacant sites ( Another complication in this system is that both iridium and manganese can be the active site for cyclooctene epoxidation. Iridium sites could have larger reaction constants ( 1 2 3 4 , , , k k k k ) than manganese which would result in a high epoxidation rate. Manganese leaching could create more active sites ( all  ) on the surface, which would contribute to the increase in cyclooctene epoxidation rate. Moreover, we should consider competing reactions to account for trends in selectivity. Competition between olefin epoxidation and water oxidation can be governed by the following two reactions.

S16
If the iridium active sites and/or modified manganese active sites in Ir single -MnO x have larger 15 / kk ratio compared to the manganese active sites in MnO x , the selectivity toward epoxidation will increase. On iridium clusters, oxygen evolution could exhibit a binuclear oxygen-coupling mechanism, which could decrease selectivity toward epoxidation if the new route for oxygen evolution is favored.