Stable Operation of Paired CO2 Reduction/Glycerol Oxidation at High Current Density

Despite the considerable efforts made by the community, the high operation cell voltage of CO2 electrolyzers is still to be decreased to move toward commercialization. This is mostly due to the high energy need of the oxygen evolution reaction (OER), which is the most often used anodic pair for CO2 reduction. In this study, OER was replaced by the electrocatalytic oxidation of glycerol using carbon-supported Pt nanoparticles as an anode catalyst. In parallel, the reduction of CO2 to CO was performed at the Ag cathode catalyst using a membraneless microfluidic flow electrolyzer cell. Several parameters were optimized, starting from the catalyst layer composition (ionomer quality and quantity), through operating conditions (glycerol concentration, applied electrolyte flow rate, etc.), to the applied electrochemical protocol. By identifying the optimal conditions, a 75–85% Faradaic efficiency (FE) toward glycerol oxidation products (oxalate, glycerate, tartronate, lactate, glycolate, and formate) was achieved at 200 mA cm–2 total current density while the cathodic CO formation proceeded with close to 100% FE. With static protocols (potentio- or galvanostatic), a rapid loss of glycerol oxidation activity was observed during the long-term measurements. The anode catalyst was reactivated by applying a dynamic potential step protocol. This allowed the periodic reduction, hence, refreshing of Pt, ensuring stable, continuous operation for 5 h.


Morphology and Structure of the As-Prepared Catalyst Layers
SEM images captured from the Ag (d<100 nm)+10 wt% CST ionomer-coated GDE (B) and from the Ir black+15 wt% Nafion-coated GDE (D).The diffractogram recorded for the Ir sample was smoothed using an FFT filter (points of window=5), but the original dataset is presented behind the smoothed curve with 50% transparency.

Ag NPs:
Four diffractions can be identified in the XRD recorded for the Ag nanoparticles corresponding to the diffraction of the Ag (111), ( 200), ( 220) and (311) lattice planes, respectively. 1According to the SEM image presented in Figure S3B, Ag NPs are evenly distributed along the Freudenberg H23C6 carbon paper gas diffusion layer substrate.The presence of 10 wt% CST can't be unequivocally identified.

Ir black:
Three diffractions can be observed in the XRD recorded for the Ir black NPs, which correspond to the (111), ( 200), (220) lattice planes of Ir, characteristic to an fcc crystal structure. 2Similar to the Ag GDE, the Ir black NPs formed a homogeneous coating on the surface of the GDL.The presence of the 15 wt% Nafion ionomer can be spotted on the SEM image (Figure S3D) manifested as an amorphous matrix around the NPs.After the potentiodynamic experiments, the GOR activity of the Pt/C samples was studied by performing potentiostatic electrolysis between +0.88 V RHE and +1.13 V RHE with 50 mV increments (Figure S5B).The goal was to narrow this relatively wide potential window by finding the potentials at which GOR can be performed at the highest reaction rate with good stability.We note here that galvanostatic protocols are usually applied in the scientific community or the cell voltage is controlled rather than the anode potential.However, by controlling the anode potential, we can better monitor the performance fading of the catalyst that is most likely related to PtO x formation.In terms of the trends, similar conclusions can be drawn as in the case of the CVs: highest current densities were measured if the ionomer content was equal or less than 5 wt%.The measured current density monotonously increased with the increasing potential reaching a maximum of 250 mA cm -2 , in the case of the Pt/C+2.5wt%CST sample at +1.08 V RHE .High current densities were accompanied with considerable stability in the case of the 2.5 wt% and 5 wt% samples.When the amount of Nafion was higher than 5 wt%, considerably higher fraction of the charge was consumed by lactate and oxalate formation.The measured current densities decreased to 25 mA cm -2 and 19.5 mA cm -2 in the case of the 10 wt% and 15 wt% Nafion-containing catalysts, respectively, which might limit the comparability of our results.The electrolyte flow rate was varied in between 1 cm 3 cm -2 min -1 and 6 cm 3 cm -2 min -1 .The CO 2 flow rate was maintained at 12 cm 3 cm -2 min -1 .Measurements were performed applying E = +1.03V RHE anode potential in 1 M KOH+0.5 M glycerol solution.Pt/C+5 wt% CST anode catalyst.The voltammogram was recorded applying 5 mV s -1 scan rate in between 0 V RHE and 1.25 V RHE .CO 2 was fed to the cathode side, while Ar was fed to the anode side of the cell with 12 cm 3 cm -2 min -1 flow rate.The purpose of this was to allow the transfer of the formed gas phase products towards the gas analyzer MS.The electrolyte flow rate was 5 cm 3 cm -2 min -1 .(B) The corresponding current density vs. t plot.(C) Partial pressure of the CO 2 formed at the anode during the electrochemical protocol presented in "A" and carried by the Ar flow to a residual gas analyzer MS, allowing the monitoring of the amount of formed products in situ.The delay between the electrolyzer cell and the MS has been corrected accordingly.

Glycerol concentration dependent GOR activity and selectivity
A constant flow of Ar was fed to the anode while the gas outlet was connected to the MS to monitor the composition of the gas stream in real-time, while no CO 2 was fed to the cathode (HER proceeded here).A CV in between 0 V RHE and +1.25 V RHE was recorded applying 5 mV s -1 sweep rate.A very small amount of CO 2 was detected from approximately +0.65 V RHE , and its amount increased gradually until reaching the upper potential limit.This is an upper estimate for the onset potential of CO 2 formation, as the majority of CO 2 is immediately dissolved in the form of CO 3 2- and only a small amount can escape from the electrolyte solution (after saturating the locally formed carbonate buffer).Still, this experiment proves that CO 2 formation is unavoidable even at much lower potential than the GOR peak maximum.Scheme S2 summarizes GOR pathways based on all detected (and quantified products).

FigureFigure S2 .
Figure S1.(A) XRD pattern measured for the as-prepared Pt/C catalyst layer.The obtained data was smoothed using an FFT filter (Points of window = 5).The original dataset is presented behind the smoothed curve with 50% transparency.(B) TEM image recorded for the Pt/C catalyst and (C) Pt nanoparticle size distribution calculated from the TEM data measuring the diameter of 150 nanoparticles from multiple images, recorded at different sites on the sample.

Figure S3 .Figure S4 .
Figure S3.SEM image (A) captured for the as-prepared Pt%C GDE+5wt% CST and Pt distribution (B) in the nanocarbon support scrutinized with SEM-EDX.

Figure S6 .
Figure S6.(A) Cyclic voltammograms recorded for the Pt/C samples prepared with different ionomer content applying 100 mV s -1 scan rate.N = Nafion.10 cycles were recorded in between 0 V RHE to +1.25 V RHE .(B) Potentiostatic holds in between +0.88 V RHE and +1.13V RHE (ΔE = 50 mV) recorded for the Pt/C samples prepared with varying the ionomer content.The duration of each hold was 2 minutes.All presented data was recorded in 1M KOH (+0.5 M glycerol) electrolyte solution.(C) Potentiostatic hold at +1.03 V RHE for 10 minutes recorded for the Pt/C catalyst layers prepared by varying the ionomer content.(D) Glycerol oxidation product distribution as the function of samples prepared with varying the ionomer content.Samples were taken during the potentiostatic holds presented in C. All measurements were performed in a membraneless microfluidic flow electrolyzer cell.

Figure S7 .
Figure S7.Potentiostatic measurements, depicting the influence of the applied electrolyte flow rate on the GOR activity (A) and on the GOR selectivity (B).The electrolyte flow rate was varied in between 1 cm 3 cm -2 min -1 and 6 cm 3 cm -2 min -1 .The CO 2 flow rate was maintained at 12 cm 3 cm -2 min -1 .Measurements were performed applying E = +1.03V RHE anode potential in 1 M KOH+0.5 M glycerol solution.

Figure S8 .Figure S9 .Figure S10 .
Figure S8.(A) Cyclic voltammograms recorded for the Pt/C samples prepared with the addition of 5 wt% CST ionomer, applying 100 mV s -1 scan rate.CVs were collected in the membraneseparated microfluidic flow electrolyzer cell.10 cycles were recorded in between 0 V RHE to +1.25 V RHE .(B) Peak glycerol oxidation current densities derived from A. (C) Product distribution at the cathode measured by GC during a 10 min-long potentiostatic hold at E = +0.97VRHE .

Figure S12 .
Scheme S2.Schematic summary of the various glycerol oxidation reaction pathways.Redox reactions (purple arrows) and non-redox reactions (teal arrows) are marked with different colors.All products that have been identified and quantified are framed by pink and the ones, which were only identified are marked with a purple frame.

Figure S14 .
Figure S14.(A)-(C) TEM images captured for the pristine Pt/C+5 wt% CST catalyst layer and for the catalyst layers after performing either the "static" or "dynamic" long-term electrochemical protocols.(D) Pt nanoparticle size distribution calculated from the TEM data in A-C measuring the diameter of 150 nanoparticles.