Fig. 1. (a) Photograph of the used setup with the homemade lid which enables tilt correction of the BHE over the IRE. (b) Scheme of the BHE placed above the Ge-IRE. By pumping electrolyte through the middle of the BHE, a radial flow profile is obtained between BHE and IRE. The catalyst is immobilized on a glassy carbon ring (black). Formed products and depleted reactants are detected in the evanescent wave above the IRE. In this configuration, a cylindric volume element is formed under the BHE. The relative distance between BHE and IRE is exaggerated for clarity.

Fig. 2. Simulation of the flow conditions in the 50 µm thin electrolyte layer underneath the BHE at a flow rate of 5 µL min-1 calculated with finite elements using COMSOL Multiphysics 5.5. (a) 2D simulation through one half of the BHE. (b) Magnification of the flow profile close to the borehole. The yellow curve indicates the velocity profile in radial direction at a height of 25 µm.

Fig. 3. LSM images of the different loadings on glassy carbon disks with a diameter of 8 mm. The magnified images correspond to areas of 1063 µm times 1418 µm.

Fig. 4. CVs as a function of the different NixB loadings (210, 52, 13 µg cm-2) in 1 M KOH (a) and b)) as well as 0.1 M glycerol/1 M KOH (c) and (d)). (a) 10 cycles recorded as conditioning with 100 mV s-1 between 1.04 to 1.64 V vs. RHE. The following CVs were obtained at lower scan rates of 10 mV s-1 ranging to a higher upper potential of 1.84 V vs. RHE. One scan was recorded in (b), five in (c) and one in (d). The CVs (a) to (c) were recorded at 1600 rpm; (d) was recorded under static conditions.

Fig. 5. Current density response as a function of time to the potential profile where 100 mV potential steps (10 min each) ranging from 1.3 V to 1.8 V vs. RHE were applied intermitted by a 4 min rinsing step at 1.0 V where no GOR is expected. The distance between BHE and IRE was set to 50 µm and a flow rate of 5 µL min-1 was applied. Corresponding spectra are shown in Fig. 6.

Fig. 6. Spectra recorded during the potential step profile shown in Fig. 5. (a) Recorded spectra from 1.3 to 1.8 V vs. RHE for a catalyst loading of 210 µg cm-2, (b) of 52 µg cm-2 and (c) of 13 µg cm-2. Solid lines represent the last spectra during each step (recorded from 8 to 9.5 min during each step) and dashed lines the last spectrum during the rinsing step after the respective elevated potential step (recorded from 2 to 3.5 min during each rinsing step). Positive (upwards pointing) bands are caused by formed species and consumed species (glycerol) are represented by negative bands due to the log(R-1) scale. F is formate, GCol is glycolate, C is carbonate, Ox is oxalate and G is glycerol.

Fig. 7. Chronoamperometric data obtained for HPLC analysis of electrolyzed 0.1 M glycerol in 1 M KOH electrolyte over NixB at 1.8 V for 2 h. The insert shows a zoom of the first 10 min.

Fig. 8. Conversion of glycerol determined via HPLC analysis of the pumped electrolyte during the chronoamperometric measurements for catalyst loadings of 210, 52 and 13 µg cm-2 shown in Fig. 7 of two different samples obtained for 30 min each. Acquisition started after 1 h pumping time.

Fig. 9. Selectivities for different catalyst loadings in the first and second sample. The inset is a magnification excluding formate.

Fig. 10. Simplified reaction scheme of the GOR representing the complex and interwoven pathway from C3 to C1 species.