Enhancing the Catalysis of Oxygen Reduction Reaction via Tuning Interfacial Hydrogen Bonds

: 20 Proton activity at the electrified interface is central to the kinetics of proton-coupled electron transfer (PCET) reactions for making chemicals and fuels. Here we employed a library of protic ionic liquids in an interfacial layer on Pt and Au to alter local proton activity, where the intrinsic ORR activity was enhanced up to 5 times, exhibiting a which is supported by ambient pressure X-ray photoelectron spectroscopy and in situ electrochemical surface-enhanced Raman spectroscopy Therefore, these results show that the kinetics of on Au and Pt could be regulated by the pK a of protic cations at the interface, X-H with decreasing Δ pK a , which is associated with enhanced ORR activity. The effect of the H-bond structure on the kinetics of PCET is examined by the Boltzmann probability ( P ν ) and vibronic coupling of proton ( S νμ ) in different quantum states. Our results reveal that stronger H-bond can increase proton tunneling kinetics ( P ν S νμ2 ) by ~10-10 3 times whereas the activation free energy ( ΔG ‡ ) of reaction remains largely 5 unchanged. This work shows compelling evidence for pK a -dependent hydrogen bond structures and their impact on the kinetics of proton tunneling and the rate-limiting PCET of ORR on Pt and Au, where altering the pK a of proton donor at the catalyst surface can change the H-bonding interaction with ORR intermediates. Our findings highlight new opportunities beyond conventional catalyst design strategies of surface 10 electronic structure tuning to control catalytic activity by tuning H-bond structures and/or solvation environments at the electrified interface.

2 volcano-shaped dependence on the pKa of the ionic liquid. The enhanced ORR activity was attributed to favorable proton transfer kinetics for strengthened hydrogen bonds between the ionic liquid to the ORR product with comparable pKa. This proposed mechanism was supported by in situ surface-enhanced Fourier-Transform Infrared Spectroscopy and our simulation of PCET kinetics based on computed proton 5 vibrational wavefunction at the H-bond interface. These findings highlight opportunities in using non-covalent interactions of hydrogen bond structures and solvation environments at the electrified interface to tune the kinetics of ORR and beyond. 10 Understanding proton coupled electron transfer (PCET) processes are critical to control the reaction kinetics in bio-1, 2, 3 , organic 2, 4 , inorganic 5,6,7 , environmental chemistry 8 and electrochemistry 4, 9 . In particular, PCET steps have been used extensively in recent research on catalysts for making energy carriers or solar fuels 4, 10, 11, 12 including water splitting 4, 12, 13 , and for converting energy carriers in fuel cells to generate work 7, 9 15 including oxygen reduction reaction (ORR) 9 . Current thinking of the ORR mechanism 9, 14 involves PCET steps on surface metal sites, which can involve one key intermediate on metals such as Au 15 for two-electron ORR (O2 + 2H + + 2e -→ H2O2, 0.68 VRHE), -4 can serve as intermolecular proton relay between proton in the bulk electrolyte and that near the metal surface (Scheme 1), where altering of protic cation with different proton activity (or pKa) provides opportunities to examine how local proton activity near the active sites can influence the ORR kinetics. A library of protic cations with different pKa values were used (Scheme 1) [C4C1im] + pKa=23. 3 51  pKa-dependent ORR activity measurements of protic-cation-modified catalysts 5 The ORR activity of Au/C (6.0 nm Au) and Pt/C (TKK, 19wt% 1.7 nm Pt) with and without protic cations was measured using rotating disk electrode measurements, as shown in Fig. 1a and 1b, respectively. The ionic liquid layer thickness with protic cations on the surface of Au/C and Pt/C was estimated to ~1 nm (details in Fig. S1).
The catalysts with protic cations showed enhanced activity compared to those without, 5 as indicated by the positive shifts in the half-wave potential in the voltammetry data.
ORR kinetic currents were extracted from the data in Fig. 1a, 1b and Fig. S2 using the Koutecky-Levich analysis, from which the specific and mass activity was obtained by normalizing the kinetic ORR current (Fig. S3) by the electrochemically surface area from CV measurements ( Fig. S4-5) and metal mass (Fig. S6), respectively. The specific 10 ORR activity for Au/C at 0.4 VRHE can be enhanced up to ~5 times with protic cations relative to pristine Au/C shown in this work while that of Pt/C at 0.9 VRHE can be increased by ~3 times compared to pristine Pt/C (TKK 19%, ~190 A/cm 2 Pt) shown in this work and previous studies 57, 58, 59, 60, 61 . Similar activity trends and enhancement were found for polycrystalline Au (~4 times at 0.3VRHE, Fig. S7a) and Pt (~2 times at 15 0.9 VRHE, Fig. S7b). The exchange current density of ORR was extracted for Au/C (O2 + 2H + + 2e -=> H2O2, 0.68 VRHE) and for Pt/C (O2 + 4H + + 4e -=> 2H2O, 1.23 VRHE) using Butler-Volmer in Fig. S8 and Fig. S9 respectively. These observations suggest that the local proton activity at the metal/electrolyte interface can considerably influence the ORR kinetics. 20 The ORR activity of both Au/C and Pt/C was found to first increase and then decrease, exhibiting a volcano trend, as a function of the pKa value of protic cations, as shown in Fig. 1c and Fig. 1d, respectively. The maximum ORR activity enhancement for Au/C was found for [DEMA][NTf2] with pKa of 10.3, which is similar to the ORR product on Au , H2O2 of pKa of 11.6 62 . In acid, the first PCET step of ORR on Au (O2 + H3O + 25 6 + e -=> OOHAu + H2O) has similar kinetics to the second step (OOHAu + H3O + + e -=> H2O2 + H2O), as indicated by the 0.09 eV difference in activation barrier from DFT 21, 63 . As increasing pH is shown to promote the rate of the first step by several orders of magnitude 46 , the second step becomes rate-limiting to the overall kinetics with increasing pH. Therefore, we propose that the ORR kinetics on Au with interfacial ionic 5 liquids are limited by the PCET kinetics of OOHAu to form H2O2 ( OOHAu + N-H + + e -=> H2O2 + N) 15, 21 (the detailed kinetic analysis depicted in Fig. S10). This hypothesis is in agreement with previous kinetic measurements of ORR in organic solvents, where the first electron transfer step (O2 + e -=> O2 -) 64 is more than 10 3 times faster than the second PCET step (O2 -+ e -+ H2O => OOH -+ OH -) 65 . Further support comes from the 10 observation that OOHAu has been detected as the stable ORR intermediate accumulating on Au by in situ ATR-SEIRAS 16,17 and in situ surface enhanced Raman 66 . Therefore, it is proposed that different pKa altered the kinetics of the second PCET step for ORR on Au in ionic liquid (OOHAu + N-H + + e -=> H2O2 + N) and tuned the overall ORR kinetics consequently. This hypothesis departs from the decoupled proton-electron 15 transfer mechanism outlined by Koper,9,41 where increasing kinetics with increasing pH for one-electron reduction can be attributed to enhanced electron-transfer kinetics on the RHE scale. form H2O is rate-limiting (OHPt + H + + e -=> Pt + H2O), 14 which is supported by ambient pressure X-ray photoelectron spectroscopy 18 and in situ electrochemical surface-enhanced Raman spectroscopy 19 . Therefore, these results show that the kinetics of ORR on Au and Pt could be regulated by the pKa of protic cations at the interface, 25 which can work as the proton donor near the active sites catalyzing the rate-limiting PCET, having the maximum ORR activity obtained with minimum pKa difference between the protic ionic liquid (proton donor) and rate-limiting ORR product (proton acceptor).
As the thermodynamic driving force of proton coupled electron transfer (PCET) 5 reactions diminishes with minimized pKa difference between the proton donor and acceptor 67 , we propose that the enhancement in the ORR activity can be attributed to difference in the interfacial hydrogen bond structure as predicted previously by PCET theory for homogeneous reactions 5, 6, 7 , which will be examined by in situ ATR-SEIRAS experiments and computation below.  (Fig. 2a), resulting from the stretching mode of C=N-H + and C=N in the cation 68 , respectively, grew with decreasing potential. The increased peak intensities can be attributed to having more protic cations on the Au surface with decreasing potential, which is in agreement with increasing peak intensities of C-N 25 8 stretching 69,70 from protic cations at 1231 cm -1 (Fig. 2b). Similar peaks to those found on Au were found on [MTBD][NTf2]-modified Pt thin-film surface (Fig. S11), which also grew with decreasing potential.
Two new peaks, at 1263 cm -1 and 3238 cm -1 emerged ( Fig. 2b and 2c) (Table S1), which also show the two fine peak features at ~3000 cm -1 for

pKa-dependent PCET kinetics in ORR facilitated by protic cations
Here we compute the exchange current densities of PCET kinetics (  Table S10 and S11, where one or two states dominate the overall kinetics, showing much higher rate constant than protons across the contributing states ( Fig. 4e and 4f). Sνμ 2 was quantified by the integral overlap of proton vibrational wavefunctions between reactant (oxidized state, before 12 ET) and product (reduced state, after ET) (Table S8) indicating the vibronic coupling of proton through PCET process and Pν as the Boltzmann probability of different quantum states is shown in Table S9. As shown in Fig. 4c, the energy of reactant higher states for N-H + •••OOHAu than the ground state (state 0) is much higher (0.35-0.4 eV in Table S8), so the (0, 0) transition is the contributing state for overall kinetics, the PCET  Table S10, which is in agreement with experimental results (Fig. 1). In contrast to Au, as the energy of state 10 1 for reactant of N-H + •••OHPt is only 0.1-0.2 eV higher than the ground state (Table S9) [MTBD]N-H + and H2O, respectively. The predicted J0 is in good agreement with those measured experimentally (Table S10 and S11). As the Pν and Sνμ reflect the properties of H-bonds vibronic states between selected protic cations and ORR intermediates, we reason that the kinetics of PCET relevant step on Au and Pt can be tuned by altering Hbond structures at the interface. This concept reflects the "Polanyi rules" discovered in 5 fundamental reaction dynamic studies, which describes that vibrational energies play the most important role in the late-barrier atom-diatom reactions (H + X2 => HX + X, X=Cl, Br) 77, 78 . This mechanism stands apart from previous studies on homogeneous reaction rates such as nitrogen reduction 79 , hydrogen evolution 80, 81 and CO2 reduction 82 , where pKa-dependent kinetics has been attributed largely to the driving force of proton 10 (chemical potential) and electron (electro-potential tuned by redox potential of metal center ) and pKa-dependent activation energy. Therefore, we propose that the stretching frequency and PνSνμ 2 of a H-bond, describing the vibrational features of intermediates, can be considered as descriptors for PCET kinetics. This study highlights the important role of vibrational features on reaction kinetics and new opportunities into electrolyte 15 and interface tuning to control reaction kinetics.

Conclusions
This study shows that ORR activity forms a volcano relationship with pKa of ion liquids (serving as proton donor) on the surface of Au and Pt in acid. The optimized 20 pKa for proton donor is around 15 and close to the pKa value of water for Pt while the optimized pKa for proton donor on Au is around 11 and close to the pKa value of H2O2.
In situ ATR-SEIRAS provides direct evidence for red-shifted stretching frequency of

Material and Methods
Pt/C and Au/C and Ag/C catalysts: 15 Pt/C catalysts was supplied by Tanaka Kikinzoku (TKK TEC10E20A), with weight fraction 19%. The Au nanoparticles were synthesized following a reported approach. 83 Briefly, Tetralin (10 mL), Oleylamine (OAm) (10 mL), and HAuCl4·3H2O (0.1 g) were mixed at room temperature and magnetically stirred for 10 min under N2 atmosphere to make precursor solution. Then, the reducing solution of 0.5 mmol of TBAB, tetralin (1 20 mL), and OAm (1 mL) was injected into the precursor solution, the reaction was carried out with water-ice bath to maintain the temperature at 2 ⁰C for 1 h. After reaction, the nanoparticles were wash with acetone and collected by centrifugation. As prepared Au nanoparticles and carbon black were dispersed in solvent 1:1 isopropanol and hexane by 20-minute ultrasonicate treatment to synthesize the Au/C catalysts, resulting in 30wt.% 25 Au loading. The products were collected by centrifugation.  25

Synthesis of ionic-liquid-modified catalysts
The ionic-liquid-modified catalysts were synthesized by a reported protocol 47 .