Assembly of a Metal–Organic Framework (MOF) Membrane on a Solid Electrocatalyst: Introducing Molecular‐Level Control Over Heterogeneous CO2 Reduction

Abstract Electrochemically active Metal‐Organic Frameworks (MOFs) have been progressively recognized for their use in solar fuel production schemes. Typically, they are utilized as platforms for heterogeneous tethering of exceptionally large concentration of molecular electrocatalysts onto electrodes. Yet so far, the potential influence of their extraordinary chemical modularity on electrocatalysis has been overlooked. Herein, we demonstrate that, when assembled on a solid Ag CO2 reduction electrocatalyst, a non‐catalytic UiO‐66 MOF acts as a porous membrane that systematically tunes the active site's immediate chemical environment, leading to a drastic enhancement of electrocatalytic activity and selectivity. Electrochemical analysis shows that the MOF membrane improves catalytic performance through physical and electrostatic regulation of reactants delivery towards the catalytic sites. The MOF also stabilizes catalytic intermediates via modulation of active site's secondary coordination sphere. This concept can be expanded to a wide range of proton‐coupled electrochemical reactions, providing new means for precise, molecular‐level manipulation of heterogeneous solar fuels systems.


Introduction
Metal-Organic Frameworks (MOFs) have attracted significant attention from the scientific community over the last two decades. [1] Due to their enormous surface area, high porosity,a nd the scope of reticular control over their chemical, electronic and photophysical properties,M OFs are often used in various applications,such as gas sorption, [2] gas separation, [3] catalysis, [4] sensing, [5] and artificial photosynthesis. [6] In recent years,MOFs have also shown great potential to be utilized in electrocatalytic energy-related schemes. [7] Gen-erally,e lectrocatalytic MOFs possess several key advantages over conventional, dense heterogeneous solids. [6d, 8] They can potentially be used as ap orous platform for the assembly of extremely high density of catalytically active sites (molecular catalysts) while providing mass-transport conduits accessible for diffusion of ions and catalytic-substrates towards the active sites. [9] Consequently,r esearch efforts have mainly focused on implementing electrochemically active MOFs as the catalytic component in the electrocatalytic cell, either through (a) the use of the MOF structural elements themselves (ligands or nodes) as electroactive catalysts, [10] or (b) the immobilization of active molecular catalysts within the MOF pores. [9c, 11] Yet, one can also consider the well-defined structure and chemical modularity of MOFs as another important virtue for efficient electrocatalysis,a si tc an be used to fine-tune the immediate chemical environment of the active site,a nd thus affect its overall catalytic performance. [12] In fact, in biological systems,c atalytic enzymes have evolved chemical and structural mechanisms that modulate the surroundings of the active site and hence regulate the activity and selectivity of adesired catalytic reaction. Forinstance,membrane proteins have provided precise control over the rate of proton delivery towards the active site in order to regulate redox-based proton-coupled electron transfer (PCET) reactions. [13] Moreover, an enzyme may modulate the secondary coordination sphere of the active site via the incorporation of pendant functional groups,s uch as proton relays or charge-bearing moieties. [14] These groups play ac rucial role during catalysis, as they may stabilize the intermediate species and enhance the reaction rate and selectivity.
Inspired by these concepts,w er ealized that another approach could be implemented, where anon-electrocatalytic MOF would be used as ap orous membrane layered over as olid heterogeneous electrocatalyst. Following this principle,asuitably designed MOF membrane has the potential to modify the microenvironment of the underlying heterogeneous catalyst and affect its electrocatalytic properties in awide variety of PCET reactions.
In this work, as amodel PCET Scheme,wechose to focus on the electrochemical reduction of CO 2 [15] using an Ag electrocatalyst. Ag is well known for its ability to effectively catalyze the reduction of CO 2 into CO. [16] Nevertheless, proton reduction to H 2 constitutes ac ompetitive sidereaction, often impeding the activity and selectivity of the desired conversion of CO 2 . [17] Consequently,afavorable activation of CO 2 over H + is ar equisite for an efficient Ag-based electrocatalytic system. To this end, we prepared athin layer of Zr 6 -oxo based MOF (UiO-66), directly assembled on an Ag CO 2 reduction electrocatalyst. We found that the electrochemical CO 2 reduction properties of the Ag catalyst were precisely tuned by the MOF overlayer,l eading to ad rastic improvement in electrocatalytic activity and selectivity.D etailed electrochemical investigation and material characterizations reveal that the effect of the MOF layer is threefold. First, it serves as ap orous membrane that physically attenuates the mass transport of reactants (CO 2 ,H + ) towards the Ag electrocatalyst. By doing so,t he local concentrations of reactants at the vicinity of the catalytic sites are significantly altered compared to those in the bulk solution, thus shifting catalysis pathways.Second, the MOFs Zr 6 -oxo nodes contain pendant Brønsted acidic groups proximal to the catalytically active surface;t hese groups accelerate electrocatalysis through the stabilization of activated *COO À intermediates.T hird, post-synthetic modification of the MOF with ap ositively charged ligand, (3carboxypropyl)trimethylammonium (TMA), imposes an electrostatic ionic gating mechanism yielding finer control over the H + flux towards the catalyst surface.O verall, the combination of all 3m echanisms allowed the systematic tuning of CO 2 -to-CO selectivity from 43 %for bare Ag, up to 89 %( at apotential of À0.8 Vvs. RHE).

Results and Discussion
Thin films of UiO-66 were grown on flat Ag electrocatalysts,a si llustrated in Figure 1a.T he synthetic strategy was adapted from ap reviously reported synthetic procedure (for experimental details,s ee Supporting Information). [18] In short, aD MF solution of ZrOCl 2 ,a cetic acid, and 1,4benzene-di-carboxylic acid (BDC) was drop-casted on af lat Ag catalyst. Thereafter,t he Ag foil was placed in ac losed container and kept at 110 8 8Ci nt he presence of DMF and acetic acid vapor for 4hours.
Scanning electron microscopy (SEM) images show that the UiO-66 grown on Ag formed continuous,h omogeneous, and porous,t hin films ( Figure S1). By increasing the volume of precursor solution drop-casted on the Ag film, we were able to obtain four samples with increasing UiO-66 thickness ( Figure S2, and Table S1). Fora ll samples,p owder X-ray diffraction (PXRD) confirms the successful growth of UiO-66 thin films ( Figure 3a). Moreover,acareful analysis at low angles reveals two additional broad peaks located at low 2q (4.48 8 and 5.48 8), indicating the presence of nanoregions of MC defects in all four UiO-66 thin films ( Figure 3b). [19] UiO-66-A exhibits the most intense signals in this region of the PXRD pattern, thus suggesting ah igher density of MC defects in this sample.N 2 physisorption ( Figure S3) measurements reveal that the BET surface area of UiO-66-A is considerably lower compared to the surface area of UiO-66-(B-D) (Table S2), as aresult of the excessively high degree of defects in UiO-66-A. [19d] The surface area of UiO-66-(B-D) matches well with literature reports of defective UiO-66. [9,19] TheO1s X-ray photoelectron spectroscopy (XPS) further confirms the presence of MC defects in all samples (Figure 4, Figure S4 and S5). Ty pically,the O1s signal of UiO-66 shows 3 distinct peaks,e ach assigned to as pecific oxygen-based species at the Zr 6 -oxo node: [20] "Zr-O-Zr" (530.5 AE 0.2 eV), "Zr-O-H" (533.4 AE 0.2 eV), and "Zr-O-C" (532.0 AE 0.1 eV), i.e., Zr 6 -oxo nodes,exposed hydroxy groups,and BDC linkers present within the UiO-66 layers,r espectively.
Fora ll samples,t he percentages of each oxygen-based species are plotted in Figure 4e.U iO-66-A clearly presents al ower relative density of "Zr-O-Zr" than the rest of the samples,which all exhibit arather similar "Zr-O-Zr" relative   density.T he relative density of Zr 6 -oxo nodes is inversely correlated to the relative density of MC defects;thus,one can monitor the extent of MC defects by probing the "Zr-O-Zr" signal. In other words,among all samples,UiO-66-A contains the largest number of MC defects,i na greement with the PXRD results.F urthermore,g enerally an MC defect is characterized by am issing Zr 6 -oxo cluster alongside all of its 12 associated linkers, [19b] thus leaving behind neighboring Zr 6 -oxo clusters with high density of missing linkers and ac orresponding large concentration of exposed pendant, Brønsted acidic OH groups,a si llustrated in Figure 1b. Indeed, the relative intensity of the "Zr-O-H" O1s peak increases with increasing density of MC defects (Figure 4e), reaching amaximum for UiO-66-A. As seen in Figure S5, this notion was further supported via determination of Zr 6 nodes density for UiO-66-(A-D) via inductively coupled plasma optical emission spectrometry (ICP-OES). It shows aconsiderably lower Zr 6 node density for UiO-66-A while the density of Zr 6 node in UiO-66-(B-D) is comparable to each other. Thus,w ew ere able to tune the thickness and chemical composition of UiO-66 thin films grown on Ag electrodes, which in turn will serve to modulate their respective electrocatalytic CO 2 reduction activity,a sw ill be further discussed below.
We then set to examine the effect of the MOF coating on the electrocatalytic CO 2 reduction properties of the underlying Ag catalyst. Electrochemical analysis was carried out using ag as-tight H-cell three-electrode setup,i nC O 2saturated 0.1 MN aHCO 3 electrolyte (pH 6.8). Platinum foil, Ag/AgCl (saturated KCl), and UiO-66-coated Ag foils were used as counter, reference,a nd working electrodes,r espectively.T he working and reference electrodes were separated from the counter electrode by an ion-exchange membrane, Nafion 117 (see Supporting Information Experimental Section for details).
Thee lectrocatalytic activity of bare Ag and all UiO-66coated electrodes were measured through bulk electrolysis experiments within ap otential window of À0.6 Vt oÀ1.0 V vs.RHE ( Figure S6). As can be seen in Figure 5a,b,adrastic improvement in electrocatalytic selectivity towards CO production was achieved for all UiO-66-modified electrodes compared to ab are Ag foil. Notably,a tt he potential of À0.8 Vv s. RHE, the CO production selectivity was systematically enhanced from 43 %for bare Ag up to 79 %for UiO-66-B (we note that for all experiments,the combined (CO + H 2 )f aradaic efficiencyw as essentially 100 %, as seen in Figure S7). It is important to mention that the selectivity towards CO production by bare Ag can vary widely depending upon several important parameters as the type of electrolyte solution, [15e,f] surface exposed crystal facets, [15g] porosity of the catalytic surface, [15h] crystallinity and particle sizes, [15i,j] and surface roughness. [16b] Moreover,t he partial catalytic current densities corresponding to CO (j CO )a nd H 2 (j H 2 )p roduction were plotted to understand the manner in which the MOF membrane affects the catalytic operation (Figure 5c,d, and Figure S8). All UiO-66-coated samples exhibited notably higher j CO and rather similar j H 2 in comparison to the bare Ag foil, thus hinting that the kinetics of electrocatalytic CO 2 reduction to CO is accelerated by the presence of the MOF membrane.
Furthermore,t he electrochemically active surface area (EASA) of all samples,m easured by cyclic voltametric analysis in the non-faradic region at different scan rates,w as essentially unchanged ( Figure S9, and Table S3). To monitor the chemical and electronic properties of the Ag electrocatalyst during the growth of the MOF,X PS measurements were performed for the UiO-66-(A-D) coated Ag films by removing the MOF layer with an inert Ar-ion gun. TheA g 3d 3/2 and 3d 5/2 peaks were not affected by the presence of the MOF membrane ( Figure S10). This suggests that the synthetic condition of the UiO-66 does not generate AgO on the UiO-66-Ag interface and thus not affect the chemical/electronic nature of the catalyst. Furthermore,t oe liminate any ambiguity in the XPS measurement arising from the use of the Arion gun during etching of the UiO-66-(A-D), ac ontrol experiment was performed with bare Ag electrode treated in an identical condition as that of the MOF synthesis (i.e., kept for 4hours at 110 8 8Ct emperature in ac losed container in presence of DMF + acetic acid vapor without using the precursor of the UiO-66 MOF,t ermed Ag-110 8 8C-DMF). Since Ag-110 8 8C-DMF did not contain any MOF coating,the  XPS measurement could be performed without any etching by Ar-ion gun. Acomparison of the XPS of Ag-110 8 8C-DMF and bare (unreacted) Ag, suggests no change of the oxidationstate of the Ag ( Figure S11). Furthermore,t he XRD pattern of Ag-110 8 8C-DMF showed no additional peak of crystalline silver oxide ( Figure S12). TheR aman spectrum ( Figure S13) of the Ag-110 8 8C-DMF was also similar to the one of bare Ag. All these results support the claim of no conversion of Ag to AgO during the synthesis of the UiO-66 thin film. Hence,itis evident that the synthesis of the UiO-66 did not change the chemical and electronic nature of the silver catalyst.
As the properties of the active Ag electrocatalyst were not altered by the growth of the UiO-66 membrane,other factors must dictate the remarkable increase in CO selectivity for UiO-66-(A-D). As illustrated in Figure 6a,b,t wo additional mechanisms could take place.( i) Thep orous UiO-66 membrane could attenuate the mass transport of reactants towards the Ag electrocatalyst surface (Figure 6a). In 0.1 MNaHCO 3 electrolyte,the bulk concentration of dissolved CO 2 (33 mM) is more than four orders of magnitude larger than the H + concentration (1.6 10 À7 M);i no ther words,a lthough local concentration gradients will be generated for both species, compared to CO 2 ,p rotons are more prone to be depleted at the catalyst-electrolyte interface during catalytic operation, shifting the catalytic reaction towards CO 2 -to-CO conversion. [16b,17b,21] (ii)T he porous UiO-66 membrane could modulate the secondary coordination sphere of the catalyst (secondarysphere effect, Figure 6b)a tt he catalyst-MOF interface.M C defects in UiO-66 generate pendant OH groups at the Zr 6 -oxo nodes,acting as Brønsted acidic sites at the vicinity of the Ag catalyst. In turn, these sites can act as proton relays to stabilize *COO À intermediates and hence increase the rate of CO production. As such, decoupling the contribution of each of these mechanisms is difficult because both are expected to occur simultaneously during electrocatalysis.Y et, we have realized that by capping the pendant proton-relaying OH groups at the Zr 6 -oxo nodes,o ne would eliminate the secondary-sphere effect of the UiO-66 membrane,w hile still attenuating mass transport. Hence,t he UiO-66-coated Ag electrodes were reacted with benzoic acid (BA) to cap the exposed OH groups at the MOF nodes (hereafter referred as UiO-66_BA), as illustrated in Figure 7a (for details,s ee Supporting Information Experimental Section). [22] PXRD ( Figure S14) and SEM ( Figure S15) analysis confirm that the post-synthetic modification with BA did not change the crystal structure and morphology of the resulting MOF. Raman spectroscopy was employed to verify the tethering of BA to the Zr 6 -oxo nodes ( Figure S16). Ther elative intensity of the two peaks appearing at 1448 cm À1 and 1431 cm À1 (OCO symmetric stretching of carboxylate linkers) gets modified due to comprehensive contribution of OCO stretching of bound BA. [9a,19c] TheB Al oading in UiO-66-(A-D)_BAwere determined by 1 HNMR analysis ( Figure S17,18). As expected, the BA loading for all samples follows as imilar trend as for the density of Zr 6 nodes.M eaning,U iO-66-(B-D)_BA exhibit comparable BA loading while UiO-66-A_BAsB A loading was lower. Additionally,t he prominent decrease in the percentage of "Zr-O-H" species in the O1s XPS (Figure S19) indicates the capping of the exposed node-bound OH groups by BA.Inthat manner,BA-capped Zr 6 -oxo nodes would lack the ability to stabilize the Ag-bound *COOintermediate by secondary-sphere interaction. Consequently, comparing the CO selectivity of UiO-66, UiO-66_BA, and bare Ag should disclose the respective contribution of each of the two mechanisms governing the enhancement in electrocatalytic performance.Asexpected, for all MOF membranes thicknesses,t he electro-catalytic CO selectivity of UiO-66_BA( measured at À0.8 Vv s. RHE) was found to be considerably lower than that of UiO-66, yet higher than bare Agss electivity (Figure 7b). Fore ach UiO-66 thickness,t he difference in CO selectivity between UiO-66 and UiO-66_BA, i.e., DCO% (UiO-66-UiO-66_BA), should represent the contribution of the pendant OH groups to the overall electrocatalytic performance of the system (Figure 7c).  Assuming this is correct, DCO% (UiO-66-UiO-66_BA) should strongly correlate with the relative coverage of pendant OH groups at the catalyst surface.A sm entioned earlier, an increased concentration of MC defects induces an increase of exposed OH groups (missing linkers) at the remaining Zr 6 -oxo clusters,a lbeit with al ower surface coverage of clusters,i .e., al ow amount of Zr 6 -oxo clusters and pendant groups in contact with the Ag. As such, the ratio of Zr 6 -oxo clusters to pendant OH groups (i.e., the XPS signal ratio of Zr-O-Zr to Zr-OH) should correlate with the relative surface coverage of pendant OH groups.I ndeed, as seen in Figure 7c, DCO% (UiO-66-UiO-66_BA) for the different samples is closely correlated with the relative surface coverage of pendant OH groups (calculated as the ratio of Zr-O-Zr to Zr-OH). Hence,the effect of the attenuated mass-transport mechanism was disclosed by the CO selectivity difference between UiO-66_BAand bare Ag, i.e., DCO% (UiO-66_BA-Ag), as seen in Figure 7d.C learly,a ll UiO-66 thicknesses exhibit noticeable DCO% (UiO-66_BA-Ag), reaching am aximum value of 21 %f or UiO-66-B.T hus,i tc an be concluded that the MOF membrane-induced attenuation of mass transport plays an important role in modulating the electrocatalytic selectivity of the system. Overall, due to the interplay of both mechanisms (stabilization of the *COOintermediate by secondary-sphere interactions,a nd attenuated mass transport), the maximum selectivity for CO was achieved using UiO-66-B,w hich possesses an optimum combination of pendant Brønsted acidic sites and membrane thickness.
Additionally,t odemonstrate the generality of utilizing secondary-sphere interactions to facilitate PCET processes, we used OH À as an electrochemical probe. [23] To do so,for all samples (UiO-66-(A-D) and bare Ag), we compared the reversible electrochemical adsorption and desorption of OH À (see detailed discussion in Supporting Information). The overpotential required for OH À adsorption and desorption on all UiO-66-coated electrodes was found to be lower than for bare Ag ( Figure S21, and Table S4). This result is an outcome of the stabilizing interactions between the MOFsp endant Brønsted acidic sites and Ag-bound hydroxides,y ielding an increased OH À binding energy at the MOF-modified Ag surface ( Figure S22). Hence,t his experiment provides yet another strong support to the argument of asecondary-sphere stabilization of the *COO-intermediate during electrocatalytic CO 2 reduction performed with the UiO-66-coated Ag electrodes.
To validate the generality of the MOF-membrane concept, 2s ets of experiments were conducted. First, to understand if the effect UiO-66 is not limited only to flat surfaces and can also influence the electrocatalytic performance of high surface area electrodes,A gn anoparticles were grown within UiO-66-B (termed Ag@UiO-66-B;s ee Experimental Section in Supporting Information). As shown in Figure S23, higher catalytic current density,and comparable CO product selectivity as that of UiO-66-B could be achieved, thus pointing that the presented concept could indeed be applied to high surface area electrodes.S econd, to verify the suitability of the MOF-membrane approach for different types of electrocatalysts,U iO-66 thin film was grown on Au (termed UiO-66-Au), awell-known electrocatalyst for CO 2 to CO conversion. [21,24] As clearly seen in Figure S24, higher catalytic currents and better selectivity towards CO formation could be achieved for UiO-66-Aucompared to bare Au over aw ide potential range.I no ther words,t he electrocatalytic CO 2 reduction improvement by Ag@UiO-66-B and UiO-66-Au demonstrate the generality and flexibility of the MOFmembrane notion, to facilitate PCET processes.
Next, we postulated that the concept of mass-transport attenuation could be further extended by endowing the MOF membrane with permselectivity; [25] tethering aligand bearing fixed cationic charge within the MOFsp ores should allow electrostatic tuning of cations (e.g., H + )i ngress throughout the membrane towards the catalytically active sites,w ithout affecting the permeability of neutral reactants,e .g.,CO 2 (see illustration in Figure 8a).
To test this hypothesis,w em odified the best performing sample (UiO-66-B) with 3(carboxypropyl)-trimethyl-ammonium (TMA;s ee Supporting Information Experimental Section for details). Ther esulting TMA-modified MOF membrane (hereafter UiO-66-B_TMA) was characterized by PXRD,SEM and XPS ( Figure S25-S28) and was found to retain the crystal structure and morphology characteristics of the parent UiO-66-B MOF.Confirmation of successful binding of TMA ligands to the MOF nodes was obtained by the recorded alteration of the relative intensity of UiO-66-B_TMA Raman peaks located at 1448 cm À1 and 1431 cm À1 as aresult of contribution of symmetric OCO stretching of the TMA ligand bound to the Zr 6 -oxo nodes ( Figure S29). [9a, 19c] Theloading level of the TMA was found to be 0.4 per Zr 6 -oxo node,a sd etermined by 1 HNMR and ICP-OES analysis ( Figure S30, S31) (see detailed discussion in Supporting Information). We then compared the electrochemical response of UiO-66-B and UiO-66-B_TMA in the presence of ac ationic redox probe,R u(NH 3 ) 6 3+ ( Figure S32). Cyclic voltammograms recorded for UiO-66-B_TMA ( Figure S32a) show ab roadened, cathodically shifted reduction wave for Ru(NH 3 ) 6 3+ compared to UiO-66-B,asexpected for asluggish electrochemical reaction due to the slower diffusion of the cationic species towards the Ag electrode surface.
In addition, reductive potential step chronoamperometry measurements were conducted to extract the Ru(NH 3 ) 6 3+ diffusion coefficients for both electrodes (see detailed discussion in Supporting Information). [9c] As seen in Table S5, the diffusion of Ru(NH 3 ) 6 3+ within UiO-66-B_TMA is four times slower than in UiO-66-B,which confirms the ability to modulate the cation permeability in UiO-66-B_TMA.
With this knowledge in mind, we set to assess the electrocatalytic CO 2 reduction performance of UiO-66-B_TMA (Figure 8b-d). Remarkably,UiO-66-B_TMA exhibits enhanced CO selectivity over the tested potential range. At À0.8 Vv s. RHE, CO selectivity reaches am aximum of 89 %, improving the catalytic selectivity of bare Ag and UiO-66-B by 46 %a nd 10 %r espectively (Figure 8b). Noticeably, the CO partial current densities for UiO-66-B_TMA and UiO-66-B are similar (Figure 8c). Nevertheless,the H 2 partial currents of UiO-66-B_TMA are significantly lower than those of UiO-66-B (Figure 8d). Hence,itisapparent that while the rate of CO 2 reduction to CO is not affected by the cation exclusion (ion-gating) mechanism, [26] the kinetics of H 2 evolution is greatly suppressed due to the slower delivery of protons towards the active sites.
Finally,t he stability of the best-performing electrocatalytic system, UiO-66-B_TMA, was examined through a5hour bulk-electrolysis analysis conducted at À0.8 Vvs. RHE. To counteract the low solubility of CO 2 in aqueous NaHCO 3 solution, we purged the solution with CO 2 every hour for 30 mins (see Supporting Information Experimental Section). As seen in Figure S33, catalytic currents remained essentially constant over the entire period of measurement, while maintaining over 90 %ofthe initial CO selectivity.Additionally,a ss een in Figure S34, XPS analysis of UiO-66-B_TMA before and after bulk-electrolysis reveals at least ap artial preservation of the MOF-installed TMA ligand (as evident by the retained N1s peak). In addition, SEM images ( Figure S35) and PXRD characterization ( Figure S36) show ap reserved MOF morphology and crystallinity upon electrolysis.

Conclusion
In this work, we show that an on-catalytic MOF membrane layered over as olid electrocatalyst, can systematically tune its catalytic performance at the molecular level, without alteration of the active sitesproperties.Specifically,asaproof of concept, athin layer of UiO-66 MOF was assembled on an Ag CO 2 reduction electrocatalyst. It was found that the presence of the MOF overlayer leads to as ubstantial improvement in the systemse lectrocatalytic activity and selectivity towards CO production. Thee nhanced electrocatalysis was achieved as aresult of three key mechanisms:i ) the MOF acts as ap orous membrane that attenuates mass transport of reactants towards the catalytic surface,g enerating local concentration gradients near the active sites,a nd hence modulates the catalytic reaction;ii) the stabilization of the reactive *COO À intermediate by as econdary-sphere interaction introduced by the MOF nodespendant Brønsted acidic groups positioned at the catalyst-MOF interface;iii)the electrostatic ionic gating of protons:f unctionalization of MOF pores with fixed positively charged group enables control over delivery of H + to the catalytic sites,t hus improving the electrocatalytic selectivity.I mportantly,t he presented concept could in principle be extended for any given proton-coupled electron transfer (PCET) reaction and hence provide an ew,g eneral tool kit for the design and manipulation of efficient heterogeneous electrocatalytic systems.