Electrocatalysts derived from copper complexes transform CO into C2+ products effectively in a ow cell

Electrochemical reactors that electrolytically convert CO 2 into higher-value chemicals and fuels often pass a concentrated hydroxide electrolyte across the cathode. The challenge is that this strongly alkaline medium converts the majority of CO 2 into unreactive HCO 3– and CO 32– byproducts rather than into CO 2 reduction reaction (CO2RR) products. The electrolysis of CO (instead of CO 2 ) does not suffer from this undesirable reaction chemistry because CO does not react with OH – . Moreover, CO can be more readily reduced into products containing two or more carbon atoms (i.e., C 2+ products) than CO 2 . We demonstrate here that an electrocatalyst layer derived from copper phthalocyanine ( CuPc ) mediates this conversion effectively in a flow cell. This catalyst achieved a 25% higher selectivity for acetate formation at 200 mA/cm2 than known state-of-art solid-state catalysts. A gas diffusion electrode coated with CuPc electrolyzed CO into C 2+ products at high rates of product formation (i.e., current densities ≥200 mA/ cm 2 ), and at high Faradaic efficiencies for C 2+ production ( FE C2+ ; >70% at 200 mA/cm 2 ). While operando Raman spectroscopy did not reveal evidence of structural changes of the copper molecular complex, X-ray photoelectron spectroscopy points to the catalyst undergoing a conversion to a metallic copper species during catalysis. Notwithstanding, the ligand environment about the metal still impacts catalysis, which we demonstrated through the study of a homologous CuPc bearing ethoxy substituents. These findings reveal new strategies for using metal complexes for the formation of carbon-neutral chemicals and fuels at industrially relevant conditions.


Introduction
Using renewable electricity to drive the CO 2 reduction reaction (CO2RR) is an appealing way to synthesize carbon-neutral chemicals and fuels. 1 Driving electrolytic CO2RR at a high rate of product formation requires the continuous feed of reactants to the respective electrodes in a flow cell (or "electrolyzer"). 1,2 A microfluidic flow cell is a type of flow cell proven to be particularly effective for mediating CO2RR electrolysis at high current densities (J >500 mA/cm 2 ). 3,4 The highest performance microfluidic CO2RR flow cells reported to date pass a highly alkaline aqueous solution (e.g., 7 M KOH) between an anion exchange membrane (AEM) and cathodic gas diffusion electrode (GDE). 5 This high pH serves to promote CO2RR by suppressing the competing hydrogen evolution reaction (HER); however, these strongly alkaline conditions convert up to 70% of reacted CO 2 into unreactive (bi)carbonates (Eq. 1-2). 6,7 This formation of (bi)carbonates not only impacts the energy efficiency of the reactor, it also induces solid salt formation that compromises flow cell operation. 8,9 Bicarbonate formation: CO 2 + OH -⇌ HCO 3 -Eq. 1 Carbonate formation: The use of CO as a reagent (instead of CO 2 ) can produce the same carbon-containing products as CO2RR electrolysis while bypassing the formation of (bi)carbonates that reduce both the efficiency and durability of a flow cell. [10][11][12] An additional advantage of CO reduction reaction (CORR) electrolysis is that the process saturates the electrocatalyst surface with adsorbed CO (*CO), 13,14 which increases the probability of carbon-carbon coupling to form multi-carbon (C 2+ ) products. 15,16 This ability to favor carbon-carbon bond formation reaction chemistry is appealing because C 2+ products are generally more economically valuable than the C 1 products (e.g., CO, HCOOH) that are more readily produced by CO2RR electrolysis. 17 The challenge to date is that solid-state copper is the only electrocatalyst known to mediate high rates of CORR (i.e., J >100 mA/cm 2 ). [18][19][20][21][22][23] This electrocatalyst produces a range of C 2+ products (e.g., acetate, ethylene, and alcohols), but not selectively. A CORR electrocatalyst that can favour the formation of a single C 2+ product is needed, yet the highest reported CORR selectivity is merely 48% (for the formation of acetate). 19,20,24,25 While there are a number of ways to modify solid-state electrocatalysts, our program is intrigued by the use of molecular electrocatalysts to control reaction selectivity in flow cells because of the acute synthetic control over the electronic environment of the active site as well as the primary coordination and secondary coordination spheres. 3,[26][27][28][29][30] We showed previously that a cobalt phthalocyanine complex can operate at high current densities in a CO2RR flow cell for the selective conversion of CO 2 into CO (i.e., FE CO of 95% at 150 mA/cm 2 ), 31 but the use of molecular electrocatalysts that reduce CO at high current densities in a flow cell has not yet been documented. 32 We report here that coating a gas diffusion electrode with copper phthalocyanine (CuPc, Supplementary Fig. 1) is effective at reducing CO into C 2+ products at a Faradaic efficiency (FE C2+ ) >70% at current densities up to 200 mA/cm 2 in a flow cell. Cognizant that molecular catalysts are known to degrade under such conditions, 33 we performed a series of control experiments to confirm the molecular structure of the catalysts does undergo a transformation into a metallic species during electrolysis. This molecule-derived catalyst layer yields partial current densities for C 2+ formation of 150 mA/cm 2 .

Results and Discussion
All electrochemical experiments in this study were performed in a commercial three-compartment flow cell reactor (Micro Flow Cell ® ) operating in a "flow-through" configuration ( Fig. 1). This "flow-through" configuration forces convective transport of gaseous reactants through the porous electrode, differing from the traditional "flow-by" configuration whereby gas reactants reach the electrocatalyst only by diffusion (see Supplementary Fig. 2 for details). 34 The electrochemical reactor consisted of two PTFE housing pieces, a Pt/Ni anode current collector, a Ti cathode current collector, three polytetrafluoroethylene (PTFE) flow frames containing expanded PTFE meshes for turbulence, and rubber gaskets. The first flow frame delivered CO (g) at 10 standard cubic centimeters per minute (sccm) from the back side and through the carbon black immobilized CuPc coated GDE that is fixed on the cathode current collector with a titanium frame. The unreacted CO and product mixture is delivered into the second flow frame carrying 2 M KOH catholyte solution at a flow rate of 6 mL/min between the AEM (Sustainion ® ) and the electrocatalyst side of the GDE. A leak-free Ag/AgCl reference electrode was inserted into the second flow frame to measure the cathodic potential. On the opposing side of the AEM, 2 M KOH anolyte was circulated at 6 mL/min through a third flow frame between the AEM and the nickel foam fixed on the anode current collector. The outlet catholyte was recycled to a reservoir that was continuously purged through the headspace with N 2(g) at 44 sccm to carry CO and gaseous CO2RR products and hydrogen gas from the hydrogen evolution reaction (HER)s to the in-line gas chromatograph (GC) for analysis. The flow rate of the gas mixture was measured by a flow meter before the gas entered into the GC. Liquid products collected from both the catholyte and anolyte were quantified by 1 H nuclear magnetic resonance (NMR) spectroscopy using methylsulfonylmethane (MSM) and potassium hydrogen phthalate (KHP) as internal standards with corresponding calibration curves.   (Fig. 2). The chemical shifts of these products were consistent with the literature values. 35 Standard solutions were prepared and analyzed using 1 H NMR to obtain calibration curves for each liquid product ( Supplementary Fig. 3). AcOwas found to be the predominant liquid species. The FE AcO-of 36% at 200 mA/cm 2 corresponds to a turnover frequency (TOF) of 4.11 min −1 for each copper site (see Methods for details). The predominant gaseous product was C 2 H 4 with a FE C2H4 of~20% at all applied current densities that were tested. The highest partial current density for C 2+ formation (J C2+ ) was measured to be 147 ± 6 mA/cm 2 , and was achieved at a high single-pass conversion of 32%. A high single-pass conversion reduces the need to separate unreacted CO 2 from the product mixture stream. The electrolysis of CO in a batch-type "H-cell", in which CO mass transport is much lower, yields a far lower J C2+ (16 mA/cm 2 cf. 147 mA/cm 2 ) at the same potential and pH ( Supplementary Fig. 4). 21 Fig. 2 | CO electroreduction performance of a GDE coated with CuPc. a, Cumulative FE as a function of applied current density for CO electroreduction by CuPc. Error bars represent standard deviation from at least three independent measurements. b, Partial current densities of C 2+ as a function of applied potentials for CO electroreduction on CuPc. Alcohols include ethanol, propanol, and allyl alcohol.
We tested the state-of-the-art electrocatalyst OD-Cu under the same operating conditions (i.e., electrolyte composition, current density, etc.) as those used for experiments with the CuPc immobilized carbon black GDE. Our electrolysis experiments with OD-Cu yielded a product distribution of ethylene, oxygenates, and hydrogen that matched literature reports ( Supplementary Fig. 5). 21 The relative amounts of C 2+ produced matched those observed with CuPc, yet CuPc was 25% more selective towards AcOproduction than OD-Cu, while OD-Cu was 12% more selective for alcohols (e.g., ethanol and propanol).
We then tested how robust the GDE was in the flow cell by measuring the FE of the different products over 20 h of electrolysis at 100 mA/cm 2 periodically replenishing the KOH catholyte to ensure the concentrations of liquid products remained within the calibration ranges for 1 H NMR analyses ( Supplementary Fig. 6). A high FE C2+ of >60% was maintained throughout the 20 h of electrolysis. We did observe a 10% decrease in FE C2+ over this period, but this decline in performance compares favorably to solid-state Cu electrocatalysts operated at the same current densities in microfluidic cells. 20 Our system is likely more stable than these previous reports because of the cell design: We operated the microfluidic flow cell in a "flow-through" configuration rather than a "flow-by" configuration (i.e., we do not claim that CuPc is more stable than a solid-state catalyst). The flow-through configuration pushes gaseous CO through the GDE to mitigate "flooding", a common flow cell failure mode due to excessive electrolyte percolation into the GDE. 36 The durability of the GDE with the electrocatalyst layer was tested by analyzing the gas diffusion electrodes before, during and after electrolysis using a variety of techniques. SEM imaging of the GDEs and powder XRD diffractograms before and after electrolysis did not detect any structural changes for the copper molecular complex ( Supplementary Fig. 7-9). Operando Raman spectroscopy also did not reveal any changes in the molecular complexes ( Supplementary Fig. 10). However, X-ray photoelectron spectroscopy (XPS) did indicate chemical changes of CuPc at the GDE surface after 4 h of continuous electrolysis at 200 mA/cm 2 . The deconvolved Cu 2p spectra indicate the reduction of Cu 2+ (935.4 eV) to a combination of Cu + (934.1 eV) and Cu 0 (933.1 eV) species (Fig. 3a, 3c). The high-resolution N 1s spectrum of CuPc catalyst also showed the emergence of a new peak at 400.3 eV after electrolysis (Fig. 3b, 3d). This peak has been previously ascribed to pyrrolic nitrogen (N-H) of non-metallated H 2 Pc (Supplementary Fig. 11). These results preclude our ability to claim that the molecular species remains intact during catalysis. Previous studies have claimed that CuPc can undergo reversible ligation during electrolysis experiments tested in an H-cell. 33,37 This reversible ligation is unlikely to be occurring in our experiments recorded in a flow cell: Quantitative XPS element analyses showed a 4-fold reduction in the relative atomic amount of N to Cu present in the GDE following electrolysis (Fig. 4). This experiment is consistent with the ligands being liberated from the metal and removed from the flow cell, and the formation of a metallic copper species during electrolysis.
We note that in-situ generated metallic copper clusters are air-sensitive and can be easily reoxidized, which can complicate the XPS analysis of the copper oxidation states. For example, air exposure of the post-electrolysis sample for three days resulted in a significant increase of Cu 2+ component, yet no change in N:Cu ratio was observed (Fig. 4, Supplementary Fig. 12). From this experience, the quantitative elemental analysis of N:Cu ratio should be performed to monitor the molecular integrity of CuPc.
While the reduced form of the electrocatalyst is difficult to define at this stage, the nature of the ligand still clearly plays a role during catalysis. We prepared an analogous catalyst bearing ethoxy substituents about the ligand that we refer to here as "CuPc-OEt". Compared with the unmodified CuPc, these ethoxy group-functionalized CuPc-OEt catalysts exhibit a 38% higher FE C2+ at 300 mA/cm2 ( Supplementary Fig. 13). This finding points to the ligand environment affecting the nature of the catalytically active species, a subject that will be the focus of future investigations. 33

Conclusion
We show in this study how a copper complex, CuPc, coated on a carbon-based GDE can mediate the electrolytic conversion of CO into C 2+ products in a microfluidic flow cell at current densities of 150 mA/cm 2 . The molecular species does not remain intact during electrolysis, but the electrocatalyst does behave as an effective electrocatalyst layer. These findings present the opportunity to use the rich library of metal-complexes catalysts available in the literature, as well as new derivatives, for effective CO electrolysis at industrially-relevant rates of product formation. We contend that further investigations into metal complexes electrocatalysts in flow reactors will broaden the scope of energy-dense products that can be produced efficiently from CO.

Eq. 3 =
Where q k is the number of electrons exchanged, F is Faraday's constant (F = 96,485 C mol -1 ), x k is the mole fraction of the gas k in the gaseous mixture analyzed, F m is the molar flow rate in mol/s, and I is the total current in A. The molar flow rate is derived from the volume flow rate F v by the relation , with p being the atmospheric pressure in Pa, R the ideal gas constant of 8.314 J mol -1 = / K -1 , and T the temperature in K. The FE of a liquid product j was determined by equation 4.

Eq. 4 =
Where q j is the number of electrons exchanged, F is Faraday's constant (F = 96,485 C mol -1 ), n j is the moles of product j, I is the applied current, t is the electrolysis time. Turnover frequency (TOF) was calculated using equation 5 for each experiment 38 : Where q is the number of electrons exchanged, F is Faraday's constant (F = 96,485 C mol -1 ), n cat is the moles of catalyst deposited on the gas diffusion electrode, I is the applied current, FE is Faraday efficiency. The consumption of CO (in mL/min) was obtained by equation 6. Then, the single pass conversion was calculated using equation 7 (example for AcO -). The total single pass conversion was calculated by dividing the sum of CO consumed (in mL/min) for each product by the inlet CO flow rate of 10 mL/min.  Figure 1 Exploded view of the micro uidic ow cell architecture used in this study. The graphic shows the positioning of the components, the feeds going into the inlet, and the products (C2+ includes C2H4, AcO-, C2H6O, C3H8O, C3H6O) exiting the outlet, as well as the electrodes that are physically connected to the external power supply. inset, photograph of the micro uidic ow cell.  X-ray Photoelectron Spectra measurements of CuPc before and after electrolysis. a, Cu 2p and b, N 1s of CuPc before the electrolysis. c, Cu 2p and d, N 1s after the electrolysis for 4 h at 200 mA/cm2. Post electrolysis samples were stored under N2 before the analysis.

Figure 4
Quantitative X-ray Photoelectron elemental measurements of CuPc coated electrodes before and after electrolysis. Green, element analysis of the GDE before the electrolysis; red, element analysis of the GDE after the electrolysis at 200 mA/cm2 for 4 h; grey, analysis of GDE after the electrolysis and air exposure for three days.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SRCuPcCORRSINC3rdforupload1.pdf