Chemical Science EDGE ARTICLE Correlations between experiments and simulations Cite this: Chem. Sci., 2022, 13, 13409 for formic acid oxidation† All publication charges for this article have been paid for by the Royal Society Alexander Bagger, ‡*a Kim D. Jensen, ‡a Maryam Rashedi, ab Rui Luo, ac of Chemistry Jia Du, d Damin Zhang, d Inês J. Pereira, a Marı́a Escudero-Escribano, aef Matthias Arenz ad and Jan Rossmeisl a Electrocatalytic conversion of formic acid oxidation to CO2 and the related CO2 reduction to formic acid represent a potential closed carbon-loop based on renewable energy. However, formic acid fuel cells are inhibited by the formation of site-blocking species during the formic acid oxidation reaction. Recent studies have elucidated how the binding of carbon and hydrogen on catalyst surfaces promote CO2 reduction towards CO and formic acid. This has also given fundamental insights into the reverse reaction, i.e. the oxidation of formic acid. In this work, simulations on multiple materials have been combined with formic acid oxidation experiments on electrocatalysts to shed light on the reaction and the accompanying catalytic limitations. We correlate data on different catalysts to show that (i) formate, which is the proposed formic acid oxidation intermediate, has similar binding energetics on Pt, Pd and Received 15th September 2022 Accepted 25th October 2022 Ag, while Ag does not work as a catalyst, and (ii) *H adsorbed on the surface results in *CO formation and poisoning through a chemical disproportionation step. Using these results, the fundamental DOI: 10.1039/d2sc05160e limitations can be revealed and progress our understanding of the mechanism of the formic acid rsc.li/chemical-science oxidation reaction. attracted a lot of attention due to their viable energy density per Introduction mass- and volume, attractive handling/storage properties and Tremendous efforts are currently going into out-phasing fossil potential uses in other non-fuel applications, e.g. as high value fuels in favor of sustainable fuels.1 This is motivated by our chemical building blocks for industry.3 need to close the carbon cycle2 and pave the way for new fuel Some liquid fuels, such as methanol, are notoriously limited production routes.3 Electrocatalytic technologies can in the in the oxidation toward CO2 since the process goes through future possibly allow direct electrication of chemical and fuel a CO intermediate.5 CO oxidation then becomes the limiting production. Examples include reduction of CO2 towards CO, factor determining the performance of direct methanol FCs HCOOH, C2H4, C2H5OH and H2O towards H2.1 Efficient fuel (DMFCs). Formic acid as liquid fuel behaves differently; it has consumption through fuel cells (FCs) also holds great poten- a CO2-like structure with two hydrogens attached. This molec- tial.4 Liquid fuels such as formic acid and methanol have ular structure predicates that the oxidation process only requires the removal of two hydrogen atoms. Consequently, formic acid oxidation should ideally circumvent the problem of a University of Copenhagen, Department of Chemistry, Universitetsparken 5, 2100 CO-poisoning. Kbh-Ø, Denmark. E-mail: alexander@chem.ku.dk To gauge formic acid's efficiency as a fuel we compare the b College of Science, University of Tehran, Enghelab Square, Tehran, Iran single round trip efficiency of relevant closed-loop chemical c School of Environmental and Biological Engineering, Nanjing University of Science & compounds, i.e. hydrogen, formic acid, methanol and lithium Technology, Nanjing 210094, China d batteries as seen in Table 1. Here we observe that the Li-battery University of Bern, Department of Chemistry, Biochemistry and Pharmaceutical Sciences, CH-3012 Bern, Switzerland storing and release of energy exhibits the highest efficiency e Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC, Barcelona followed by hydrogen. However, both Li-batteries and H2 suffer Institute of Science and Technology, UAB Campus, 08193 Bellaterra, Barcelona, Spain from low energy density. Storing energy as methanol and formic f ICREA, Pg. Lluı́s Companys 23, 08010 Barcelona, Spain acid is very similar in terms of the cost in electrolyzer energy. † Electronic supplementary information (ESI) available: Relevant aspects of the The major difference between formic acid and methanol arises computational model and methods and the experimental methodology are when using the chemical in a fuel cell, where methanol is included herein. (PDF). Additionally, atomicstructures and analysis scripts are available on the webpage: https://chem.ku.dk/research_sections/nanochem/theoretical- limited by CO oxidation.6 electrocatalysis/. See DOI: https://doi.org/10.1039/d2sc05160e Depending on the UFAOR, direct formic acid fuel cells ‡ These authors contributed equally. (DFAFCs) can be considered an attractive alternative to © 2022 The Author(s). Published by the Royal Society of Chemistry Chem. Sci., 2022, 13, 13409–13417 | 13409 Chemical Science Edge Article   UORR  Ufuel cell reaction Table 1 Estimated round-trip efficiencies h¼  100% calculated using the difference in energy potentials. UOER þ Uelectrolyzer reaction Common for hydrogen, formic acid, and methanol we use UOER = 1.6 VRHE and UORR = 0.8 VRHE Round trip efficiency, Energy stored (electrolyzer) Energy released (fuel cell) h Hydrogen UHER = −0.1 VRHE UHOR = 0.1 VRHE z41% Formic acid UCO2/HCOOH = −0.8 VRHE UFAOR = 0.2 VRHE z25% Methanola UCO2/CO+H2 = −0.6 VRHE UCO/CO2 = 0.65 VRHE z7% Li-batteryb z90% a Here only the cost of syngas production is considered, not the full formation of methanol. b Typical charge/discharge efficiency. methanol fuel cells. Methanol provides 6 protons per reacted formation of various poisoning and site-blocking intermedi- molecule and formic acid only two and therefore methanol has ates.20 The following observations are reported in the literature, a ∼3 times higher volumetric energy density. However, the as also shown in ESI Fig. S1:† (i) FAOR onsets at low potentials potential of a formic acid fuel cell is high; even a few hundred does not necessarily correspond to high FAOR currents. (ii) millivolts reduction in overpotential can allow formic acid to Hysteresis between anodic and cathodic scans is a common output more energy than methanol per molecule. occurrence, showing a higher current in cathodic scans in than Key to understanding the limitation of formic acid oxidation anodic ones, which shows that the reaction rate depends on the is the direct link to the reverse electrochemical reaction, i.e. the prehistory. Typically, this “memory effect” is related to CO2 reduction reaction (CO2RR).7,8 CO2RR to formic acid and unwanted reactions which form species that block the surface, the formic acid oxidation reaction (FAOR) can be written in the i.e. poisoning species. (iii) Pt(111) is more active than Pd(100) in form: the low overpotential region; however, interestingly, this rela- tionship shis at higher potentials. (iv) As an observation it is ðCO2 RRÞ CO2 þ 2Hþ þ 2e /HCOOH; known that there is a difference for Pt and Pd with respect to the 0 DGCO 2 /HCOOH ¼ 0:12 eV per Hþ þ e (1) CO poisoning during FAOR.8 The ideal FAOR catalyst on the other hand should show reversible cyclic voltammetry (CV), ðFAORÞ HCOOH/CO2 þ 2Hþ þ 2e ; high activity, low onset potential and stable currents, as illus- 0 DGHCOOH/CO 2 ¼ 0:12 eV per Hþ þ e (2) trated in Fig. S2.† To understand the mechanism of this important reaction where DG0 is the thermodynamic potential per proton–electron numerous attempts have been made to map possible FAOR pair of the reaction. pathways,20–24 which we illustrate by the literature study over- CO2RR selectivity is highly dependent on the catalyst mate- view in Fig. 1. A dual-pathway mechanism for formic acid rial used and the crystal orientation.9–12 Hori et al. showed that oxidation is established by the community:25,26 The direct hydrogen is produced on Pt, Ru, Fe and Ni, carbon monoxide is pathways, which leads to the desired nal product of CO2 produced on Au, Ag, Zn, Ga and Pd (with limited amounts of through the formate adsorption,27 and another path where H2), and nally, formic acid is produced on Pb, In, Hg, Sn, Cd adsorbed CO, the poisoning species, is formed. The CO and Tl with almost 100% faradaic efficiency.11 Importantly, formation can be thought of as being formed through a so- hydrocarbons are uniquely produced on Cu.11 Using simula- called chemical disproponation reaction, where an activated tions, we were able to classify the CO2RR product distributions formic acid intermediate reacts with a hydrogen to form water towards hydrogen, hydrocarbons, CO and formic acid due to the and CO (e.g. *COOH + *H / *CO + H2O). However, the nature catalyst's affinity towards adsorbed *H and *CO.10 Interestingly, of the reactive intermediate in the direct pathway is still under we noted that the CO2RR appears selective towards formic acid strong debate and it is not given that it is the most stable when weakening *H binding (i.e. when there is no *H on the intermediate (e.g., formate) should also be the reaction inter- catalyst surface). Moreover, from this study10 we noted that the mediate. The community has focused on elucidating the reac- *OOCH vs. the *COOH intermediate cannot be used to distin- tion mechanism and attempted to circumvent poisoning issues: guish the CO or formic acid product formation. Where previous through pathway engineering,21,22 changing electrolyte compo- works used *OOCH as a descriptor for formic acid oxidation sition,23 or inclusion of sites with the ability to remove and *COOH as a descriptor for CO production, and even recent poisoning or site-blocking species.24,28–30 CO-poisoning from discussion for CO2 reduction highlights the possible formic partial HCOOH oxidation is oen considered the principal acid formation through *COOH.13 These ndings and discus- culprit28 and various works suggest CO formation can be avoi- sion are readily usable when considering the reverse reaction ded utilizing a single/dual-site catalyst.31–34 However, catalysts FAOR, which involves similar reaction intermediates to CO2RR such as Au–Pt,21 Pt–Hg/C35 and Pd–Hg/C36 exhibit limited cata- and vice versa. lytic improvement over their pure metal counterparts, for some FAOR exhibits the highest intrinsic activity on Pt and Pd.14–19 overviews see Fig. S4 in ESI.† In this work, we will take a new However, the reaction is affected by high overpotential and 13410 | Chem. Sci., 2022, 13, 13409–13417 © 2022 The Author(s). Published by the Royal Society of Chemistry Edge Article Chemical Science Fig. 1 Literature study highlighting all the conceived FAOR reaction pathways during potential cycling.38 Historically, FAOR has been split into the direct (gray)39,40 and indirect (green and yellow)16,39,41,42 path- ways. Further, partial FAOR and catalyst oxidation forming unwarranted surface blocking have been suggested, e.g. COxHy species (purple),20,25,43 CO (red)44 and hydroxide/oxides (blue).38 Even CO2RR induced CO formation by applying too cathodic potential (brown)40 have been suggested. Recently, formate (yellow)28,39,45,46 in various arrangements has gained attention as potential catalyst site-blocking agents. Fig. 2 Electrochemical data recorded at room temperature of equal loading nanoparticle catalysts: Pt/C (navy blue), Pd/C (crimson), Pt– Hg/C (blue), Pd–Hg/C (red) and Pt–Bi/C (green) systems on glassy view on all possible pathways for working catalysts. We carry out carbon in Ar-saturated 0.1 M HClO4 with 0.1 M HCOOH at 1600 rpm, electrochemical characterization on Pd/C and Pt/C catalyst as unless otherwise stated. (a) FAOR CVs at 10 mV s−1. (b) Base CVs in Ar- well as Pt–Hg/C,35 Pd–Hg/C36 and Pt–Bi/C,37 following the saturated 0.1 M HClO4 at 400 rpm and 50 mV s−1. (c) CA at 0.55 VRHE protocols reported in the literature for the preparations of these for 30 min. All measurements were repeated at least three times (shade represents standard deviation), IR-compensated and post- catalysts. Furthermore, we try to validate the possible reaction corrected, for experimental details see ESI and Fig. S5–S8.† Note, Pt– path for not/low performing catalysts (e.g., Ag) for formic acid Bi/C was unstable at higher potentials (see ESI†) hence a reduced oxidation. potential range was utilized. In this work, we address the following fundamental ques- tions in FAOR: (i) Is the FAOR activity correlated with the *COOH or the *OOCH intermediate? the catalysts' surface area arises due to the modication, (ii) How is CO formed during the FAOR? allowing for catalyst comparison. However, verifying this To probe the scientic questions, we use a combination of expectation using standard electrochemical surface area (ECSA) experimental tools based on cyclic voltammetry (CV) and evaluation (e.g. utilizing *CO-stripping or hydrogen under chronoamperometric (CA) and for simulations we use density potential deposited, H-UPD, charges) is not possible as these functional theory (DFT) calculations on the binding of carboxyl, species do not adsorb specically on Hg and Bi modied *COOH, formate bidentate, *OOCH, and hydrogen, *H. catalysts. Fig. 2a shows the formic acid oxidation CV during rotation and 10 mV s−1 scan-rate on the ve catalysts. Fig. 2b depicts the Results and discussion base CVs exhibiting suppressed hydrogen underpotential Fig. 2 summarizes experimental electrochemical data of Pt/C, deposition (HUPD) on the Pt–Bi/C, Pt–Hg/C and Pd–Hg/C Pd/C, Pt–Hg/C, Pd–Hg/C and Pt–Bi/C. Here we investigate only compared to the Pd/C and Pt/C counterparts. Fig. 2c displays known and active FAOR catalyst, and both extended surfaces as the formic acid oxidation CA at 0.55 VRHE for 30 min, illustrating Pt/C and Pd/C, but also single site catalyst Pt–Hg/C and Pd–Hg/ the loss in activity at this potential due to the formation of C and the noteworthy very active Pt–Bi/C system. These inves- poisoning or blocking species. For Pt/C and Pt–Hg/C, an tigated catalysts are synthesized from the same starting mate- apparent hysteresis is seen in the oxidation between the forward rials, i.e. premade Pd or Pt catalyst which we then modied by and backward scans of Fig. 2a indicating an irreversible change (electro)deposition of Hg or Bi at room temperature. Conse- in the catalyst going to low potentials. Interestingly, taking quently, it is reasonable to expect that no signicant impact of a combined view on Fig. 2a and c shows that forming single- © 2022 The Author(s). Published by the Royal Society of Chemistry Chem. Sci., 2022, 13, 13409–13417 | 13411 Chemical Science Edge Article sites of Pt through Hg alloying35,36 tend to improve the FAOR onset. In contrast, Pd-based catalysts generally do not exhibit any hysteresis. The Pt–Bi/C system exhibit the highest FAOR current with least hysteresis, but also with the highest over- potential. Additional relevant electrochemical studies can be found in the ESI,† represented through Fig. S10–S13.† To explain these observations in Fig. 2 we turned to DFT. Fig. 3a maps out the DFT calculated *COOH vs. *H binding energies of the most relevant model metal (111) facets and single-site catalysts, such as MNCs, PtHg4,35 PdHg4 (ref. 36) and single Pt atoms in Au, Pt1Au(111)21 (for computational details see ESI†). Via simulations we study both metals and the single- site-catalysts as they both have a linear scaling: DE*COOH = DE*H + b. However, b is about 0.29 eV as previously47 observed for metals and 0.0 eV for single-site catalysts, respectively. This allows for a fundamental destabilization of *H vs. *COOH at the single-site catalyst motifs as compared to metals, which is benecial for improved FAOR by limiting the Fig. 3 DFT calculated binding energies for metal fcc(111) (light-blue crosses) and single-site catalyst. MNC-based (black points), Pt atom in Au(111) denoted Pt1Au(111) (cyan triangle), PdHg4 (black/red square) Fig. 4 FAOR CVs at 10 mV s−1, 1600 rpm and room temperature in Ar- and PtHg4 (black/blue square). (a) *COOH vs. *H. (b) *OOCH vs. *H. (c) saturated 0.1 M HClO4 with 0.1 M HCOOH at different potential limits. *OOCH vs. *COOH, here the dashed line shows the diagonal indica- Ohmic drops were post-corrected. (a) Pt/C. (b) Pt–Hg/C. (c) Pd/C. tion the affinity towards formate bound through carbon or oxygen. Note, increasing lower potential limit minimizes the hysteresis in the Here its assumed DE0FAOR z DG0FAOR. Further, we used CO2 and H2 for CVs, lowering the upper potential limit generally decreases the activity. references when calculating *COOH, DE0FAOR is 0.12 eV per electron. Note, that after cycling in the different potential limits returning to the full range of 0.025 to 1.05 VRHE re-initializes the FAOR response i.e. no irreversible changes arise due to dissolution or sintering. 13412 | Chem. Sci., 2022, 13, 13409–13417 © 2022 The Author(s). Published by the Royal Society of Chemistry Edge Article Chemical Science disproportionation reaction. Besides the scaling, a vertical- and binding vs. *COOH (below the diagonal), are basically oxidized a horizontal line indicating HUPD (DEHUPD = 0 for 1/2H2 4 *H) before they can carry out FAOR. and formic acid's thermodynamic equilibrium potential have The type of analysis illustrated in Fig. 3 is a powerful tool been included in Fig. 3a and b. Here its assumed that DE0FAOR z able to identify which catalyst suffers from disproportionation, DG0FAOR, as thermodynamic corrections and water stabilization poisoning or oxidation. By virtue of the *COOH vs. *H scaling- are expected to cancel out for the intermediates. The challenge relations,10 it gives fundamental insights into why literature to understand the implications of Fig. 3a is that one needs to historically has shown no signicant FAOR activity at the ther- compare *H as a reduction from H2O and *COOH as oxidation modynamic potential for a metal catalysts. Implicitly Fig. 3a from HCOOH. This means that *H stabilizes with negative shows disproportionation occurring at low potential arising potential, while *COOH stabilizes with positive potential during from the HUPD and the consequent creation of *CO poisoning FAOR. Essentially, from Fig. 3a it can be inferred that materials species. The creation of *CO leads to hysteresis between the to the le of approximately DEHUPD are limited by dispropor- anodic and cathodic sweeps in FAOR CVs. From this insight, tionation at low potentials. While materials to the right are one would expect that disproportionation is mitigated in the limited by weak binding *COOH. Most importantly, the CVs by simply staying above HUPD potentials. fundamental scaling between *COOH and *H matter as Fig. 4 shows FAOR CVs of Pt/C, Pt–Hg/C and Pd/C cycled with adsorption of *H on the surface leads to the possibility of varying potential limits. The potential ranges from 0.00–1.05 disproportionation towards CO. In conclusion, the single-atom VRHE reveals that Pt/C and Pt–Hg/C are poisoned in anodic catalyst scaling crosses the FAOR thermodynamic potential at sweeps. While, changing to potentials, ranging from 0.40–1.05 values signicantly above DEHUPD, which suggests that single- VRHE for the Pt/C and 0.35–1.05 VRHE for the Pt–Hg/C, signi- atom catalyst may work for FAOR near the FAOR equilibrium cantly increases the anodic activity. Hence, this allows us to potential without carrying out the disproportionation reaction. indicate that HUPD mediated disproportionation account for Fig. 3b displays the DFT calculated binding energies of poisoning through CO on Pt catalysts. For Pd/C in Fig. 5c, *OOCH vs. *H. Conversely to Fig. 3a, there is no apparent decreasing the lower potential limit has no inuence on the scaling between formate bidentate and adsorbed hydrogen. almost non-existing FAOR hysteresis. In this context, it is However the oxygen bond for *OOCH scales with *OH and (bi)- important to note that Pd is well-known to form Pd-hydride carbonate (HCO3/CO3), which have surface binding through phases51 below 0.2 VRHE, which then competes with *H adsor- oxygen. Carbonate species could be formed over time due to the bed, i.e. at potentials relevant for both FAOR and CO2RR. In equilibrium with CO2 and hence poisoning the surface, relation to CO2RR, we also note that Pd-hydride, leads to a high revealing a decaying activity as observed in Fig. 2c for all faradaic efficiency towards formate,52 whereas at higher over- working FAOR catalysts. Investigating *OOCH as a function of potentials CO and H2 will dominate.11 *H is important as *OOCH has been suggested as an important reaction intermediate. Fig. 3b shows that the binding energy of the *OOCH intermediate on Ag is slightly too strong, but very close to the thermodynamic potential and similar to Pt and Pd. Essentially, if *OOCH was the important reaction intermediate for FAOR, then Ag should be working as FAOR catalyst at very low potential, and furthermore Ag does not suffer from disproportionation (as Ag *H binding far exceeds DEHUPD). However, Ag is experimentally shown to be inert for the FAOR. The role of formate in the reaction is clearly a puzzle. For FAOR surface-enhanced IR absorption on Pt has experimentally conrmed formate on the surface27 above 0.7 VRHE and further for CO2RR in situ surface-enhanced Raman spectroscopy on Ag has experimentally conrmed formate on the surface.48 As Pt converts formic acid to CO2 and Ag converts CO2 to CO, there seems no relation between the experimental observation of formate on the surface as reaction intermediate. We note that *OOCH scales with *OH, the *OOCH binding can be considered a probe of the oxidation affinity of the catalyst, i.e. having a strong *OOCH binding results in a lower oxidation potential. Fig. 3c shows the *OOCH vs. *COOH binding for the catalyst, with a dashed line indicating the affinity towards carbon or oxygen. Depending on the catalyst we can see whether formate bidentate or carboxyl is favored. The working catalyst is Pt, which is well above the diagonal and Pd which is at the diag- Fig. 5 2 s pulsed voltammetry and CVs at 10 mV s−1 at room- onal, while poorly working FAOR catalysts e.g. Ir49 or Au50 are temperature at 1600 rpm in Ar-saturated 0.1 M HClO4 with 0.1 M slightly below the diagonal. Catalysts having stronger *OOCH HCOOH. (a) Pt/C. (b) Pd/C. Measurements repeated three times. © 2022 The Author(s). Published by the Royal Society of Chemistry Chem. Sci., 2022, 13, 13409–13417 | 13413 Chemical Science Edge Article Concerning Fig. 4, one could erroneously assume that lowering the upper potential limit would not affect the CVs, while staying above the CO oxidation potential. This is however not the case, cycling 50 times from 0.35–1.00 VRHE reveals some form of deactivation of both Pt/C, Pt–Hg/C and Pd/C. We do not know what the origin of this deactivation is. Various studies reported in the literature suggests different reasons including deactivation to irreversible metal oxidation53,54 due to insuffi- cient surface reduction or accumulation of either *OOCH,55 *COH,38 *OCOH56 or *CO species. Interesting is that in situ Fourier-transform infrared spectroscopy (FTIR) work57,58 has shown that Pt, contrary to Pd, continuously form CO above HUPD potential during FAOR. Finding that *H limits FAOR activity at low potential via disproportionation, allows one to hypothesize about the Pt–Bi and Pt–Hg systems, which perform better than Pt on two Fig. 6 FAOR data on a Cu, Ni and Ag wire at 50 mV s−1 at room- different perspectives. Pt–Bi shows no hysteresis and is active at temperature Ar-saturated 0.1 M HClO4 with 0.1 M HCOOH taken at 50 mV s−1. higher overpotential as compared to Pt. The absence of a hysteresis in the CV indicates that Pt–Bi somehow circumvent the disproportionation reaction, potentially by blocking the surface for *H at low potential as onset for Pt–Bi is higher than dissolution potentials suggesting it to be a rather poor FAOR on Pt. Pt–Hg on the other hand is more active in the negative catalyst. sweep, particular at lower potential, but does show clear Most interesting when combining the observation of similar hysteresis. Further we hypothesize that the Pt single-sites in Pt– intermediate binding energetics in Fig. 3, we identied that Ag Hg destabilizes *H shiing the onset to lower potentials and has similar binding energetics of formate as Pd and Pt, whereas increasing the activity at low potential, simply due to a limited Cu and Ni do not follow the energetic of Pd and Pt. Hence Cu disproportionation reaction. and Ni can be used as test catalysts far from known working To gauge how the FAOR is affected in the potentials regions catalysts. However, as tested here in Fig. 6 there appears no above HUPD we conducted pulsed voltammetry inspired by signicant FAOR on Ag, Cu and Ni. Clavilier et al.59 In this type of pulsed voltammetry experiments, each potential investigated is separated by a surface re- Conclusion initialization (at 1.05 VRHE) cleaning the surface for all poisons through surface oxidation. The impact from dissolu- In conclusion we have correlated the FAOR activity with simu- tion at this oxidizing potential should be minimal.60,61 lated DFT binding energies of *COOH, *OOCH and *H across Fig. 5 shows pulsed voltammograms and corresponding CVs multiple metal catalysts. We have observed that for an ideal for Pt/C and Pd/C samples. Interestingly Pt/C becomes more catalyst, the FAOR equilibrium potential should be above its active when pulsing, which is in contrast to Pd/C that does not corresponding HUPD potential in order to avoid the dispropor- show any changes from the pulsing. This experiment challenges tionation. We found that *COOH and *H binding scale on both *OOCH species as site-blocking species. Since Pd has stronger metal and single-site catalysts. This creates the fundamental relative *OOCH to *COOH binding than Pt, and it should hence limiting potential due to HUPD mediated disproportionation on be on Pd where activity was affected by blocking of *OOCH the surface. The carbon–hydrogen scaling is indeed a funda- species. One view for this observation could be given in the recent mental limitation, analog to the *OH and *OOH scaling for work by Koper and coworkers,8 who noted that formate adsorp- oxygen evolution and reduction. Experimentally, we show that tion is important for formic acid oxidation, not as an active a good performing FAOR catalyst should have the attributes of: intermediate, but more as a self-protector against CO poisoning. (i) an onset close to the fundamental derived onset, (ii) no Beyond Pd and Pt catalyst, in Fig. 6, we tested Cu, Ni and Ag hysteresis between anodic- and cathodic CV scans and (iii) high wires as FAOR catalyst. The experiments revealed that Ni indeed and stable FAOR CA currents above the derived onsets funda- seems very active to oxidize either HCOOH or Ni. However, mental limits. Interestingly, this works concludes on the direct going to potentials above 0.2 VRHE tended to bring out relation between FAOR and CO2RR; *H in combination with a yellowish tinge in the electrolyte, and by looking into Ni's *COOH forming CO in both FAOR and CO2RR. Pourbaix diagrams62 it appears to readily dissolve as Ni2+ at 0.15 VRHE at pH 1, consequently making it a poor FAOR catalyst Abbreviations (unless there is some very narrow window that Ni is stable enough to oxidize HCOOH without dissolving). Similarly, Cu shows no FAOR activity only the well-known Cu oxidation CA Chronoamperometry current at potentials larger than 0.2 VRHE is observable. Ag is CV Cyclic voltammogram/voltammetry perhaps active towards FAOR but again this occur close to Ag CO2RR CO2 reduction reaction 13414 | Chem. Sci., 2022, 13, 13409–13417 © 2022 The Author(s). Published by the Royal Society of Chemistry Edge Article Chemical Science DFAFC Direct formic acid fuel cell Metals during Oxygen Evolution in Acidic Media, DMFC Direct methanol fuel cell ChemCatChem, 2014, 6(8), 2219–2223, DOI: 10.1002/ DFT Density functional theory cctc.201402194. fcc Face centered cubic 4 B. C. Ong, S. K. Kamarudin and S. Basri, Direct Liquid Fuel FAOR Formic acid oxidation reaction Cells: A Review, Int. J. Hydrogen Energy, 2017, 42(15), FC Fuel cell 10142–10157, DOI: 10.1016/j.ijhydene.2017.01.117. HER Hydrogen evolution reaction 5 T. Yajima, H. Uchida and M. Watanabe, In-Situ ATR-FTIR HUPD Hydrogen underpotential deposition Spectroscopic Study of Electro-Oxidation of Methanol and MNC Metal–nitrogen–carbon Adsorbed CO at Pt-Ru Alloy, J. Phys. Chem. B, 2004, 108(8), OER Oxygen evolution reaction 2654–2659, DOI: 10.1021/jp037215q. ORR Oxygen reduction reaction 6 D. S. Mekazni, R. M. Arán-Ais, A. Ferre-Vilaplana and RHE Reversible hydrogen electrode E. Herrero, Why Methanol Electro-Oxidation on Platinum ESI Electronic supporting information in Water Takes Place Only in the Presence of Adsorbed OH, ACS Catal., 2022, 1965–1970, DOI: 10.1021/ acscatal.1c05122. Data availability 7 S. Nitopi, E. Bertheussen, S. B. Scott, X. Liu, A. K. Engstfeld, Data is available here: https://nano.ku.dk/english/research/ S. Horch, B. Seger, I. E. L. Stephens, K. Chan, C. Hahn, theoretical-electrocatalysis/katladb/formic_acid_oxidation_ J. K. Nørskov, T. F. Jaramillo and I. Chorkendorff, Progress reaction/. and Perspectives of Electrochemical CO2 Reduction on Copper in Aqueous Electrolyte, Chem. Rev., 2019, 119(12), 7610–7672, DOI: 10.1021/acs.chemrev.8b00705. Author contributions 8 X. Chen, L. P. Granda-Marulanda, I. T. McCrum and The manuscript was written through the contributions of all M. T. M. Koper, How Palladium Inhibits CO Poisoning authors. All authors have given approval to the nal version of during Electrocatalytic Formic Acid Oxidation and Carbon the manuscript. Dioxide Reduction, Nat. Commun., 2022, 13(1), 1–11, DOI: 10.1038/s41467-021-27793-5. 9 Y. Hori, I. Takahashi, O. Koga and N. Hoshi, Selective Conflicts of interest Formation of C2 Compounds from Electrochemical There are no conicts to declare. Reduction of CO2 at a Series of Copper Single Crystal Electrodes, J. Phys. Chem. B, 2002, 106(1), 15–17, DOI: 10.1021/jp013478d. Acknowledgements 10 A. Bagger, W. Ju, A. S. Varela, P. Strasser and J. Rossmeisl, We would like to thank the Center for High Entropy Alloy Electrochemical CO2 Reduction: Classifying Cu Facets, ACS Catalysis (CHEAC) funded by the Danish National Research Catal., 2019, 9(9), 7894–7899, DOI: 10.1021/ Foundation (DNRF 149) and the Villum Foundation through the acscatal.9b01899. Villum Center for the Science of Sustainable Fuels and Chem- 11 Y. Hori, H. Wakebe, T. Tsukamoto and O. Koga, icals (#9455) for funding this work. M. E.-E. and K. D. J. would Electrocatalytic Process of CO Selectivity in like to thank the Independent Research Fund Denmark for the Electrochemical Reduction of CO2 at Metal Electrodes in award of a DFF-Research Project 1 grant (9041-00224B). M. E.-E. Aqueous Media, Electrochim. Acta, 1994, 39(11–12), 1833– acknowledges the Villum Foundation for nancial support 1839, DOI: 10.1016/0013-4686(94)85172-7. through a Villum Young Investigator Grant (project number: 12 P. Sebastián-Pascual, S. Mezzavilla, I. E. L. Stephens and 19142). J. D. D. Z. would like to thank the funding from the M. Escudero-Escribano, Structure-Sensitivity and China Scholarship Council (CSC). A. B would like to thank the Electrolyte Effects in CO2 Electroreduction: From Model Carlsberg foundation (CF21-0144). Studies to Applications, ChemCatChem, 2019, 11(16), 3626– 3645, DOI: 10.1002/cctc.201900552. References 13 T. Zheng, C. Liu, C. Guo, M. Zhang, X. Li, Q. Jiang, W. Xue, H. Li, A. Li, C.-W. Pao, J. Xiao, C. Xia and J. Zeng, Copper- 1 Z. W. Seh, J. Kibsgaard, C. F. Dickens, I. Chorkendorff, Catalysed Exclusive CO2 to Pure Formic Acid Conversion J. K. Nørskov and T. F. Jaramillo, Combining Theory and via Single-Atom Alloying, Nat. Nanotechnol., 2021, 16(12), Experiment in Electrocatalysis: Insights into Materials 1386–1393, DOI: 10.1038/s41565-021-00974-5. Design, Science, 2017, 355(6321), eaad4998, DOI: 10.1126/ 14 M. Baldauf and D. M. Kolb, Formic Acid Oxidation on science.aad4998. Ultrathin Pd Films on Au(Hkl) and Pt(Hkl) Electrodes, J. 2 S. C. Peter, Reduction of CO2 to Chemicals and Fuels: A Solution Phys. Chem., 1996, 100(27), 11375–11381, DOI: 10.1021/ to Global Warming and Energy Crisis, ACS Energy Lett., 2018, jp952859m. 3(7), 1557–1561, DOI: 10.1021/acsenergylett.8b00878. 15 L. A. Kibler, A. M. El-Aziz, R. Hoyer and D. M. Kolb, Tuning 3 S. Cherevko, A. R. Zeradjanin, A. A. Topalov, N. Kulyk, Reaction Rates by Lateral Strain in a Palladium Monolayer, I. Katsounaros and K. J. J. Mayrhofer, Dissolution of Noble © 2022 The Author(s). Published by the Royal Society of Chemistry Chem. Sci., 2022, 13, 13409–13417 | 13415 Chemical Science Edge Article Angew. Chem., Int. Ed., 2005, 44(14), 2080–2084, DOI: 28 A. Ferre-Vilaplana, J. V. Perales-Rondón, J. M. Feliu and 10.1002/anie.200462127. E. Herrero, Understanding the Effect of the Adatoms in the 16 A. Ferre-Vilaplana, J. V. Perales-Rondón, C. Buso-Rogero, Formic Acid Oxidation Mechanism on Pt(111) Electrodes, J. M. Feliu and E. Herrero, Formic Acid Oxidation on ACS Catal., 2015, 5(2), 645–654, DOI: 10.1021/cs501729j. Platinum Electrodes: A Detailed Mechanism Supported by 29 J. V. Perales-Rondón, A. Ferre-Vilaplana, J. M. Feliu and Experiments and Calculations on Well-Dened Surfaces, J. E. Herrero, Oxidation Mechanism of Formic Acid on the Mater. Chem. A, 2017, 5(41), 21773–21784, DOI: 10.1039/ Bismuth Adatom-Modied Pt(111) Surface, J. Am. Chem. c7ta07116g. Soc., 2014, 136(38), 13110–13113, DOI: 10.1021/ja505943h. 17 T. Shen, S. Chen, R. Zeng, M. Gong, T. Zhao, Y. Lu, X. Liu, 30 N. M. Marković and P. N. Ross Jr, Surface Science Studies of D. Xiao, Y. Yang, J. Hu, D. Wang, H. L. Xin and Model Fuel Cell Electrocatalysts, Surf. Sci. Rep., 2002, 45(4), H. D. Abrunã, Tailoring the Antipoisoning Performance of 117–229, DOI: 10.1016/S0167-5729(01)00022-X. Pd for Formic Acid Electrooxidation via an Ordered PdBi 31 Á. Cuesta, M. Escudero-Escribano, B. Lanova and Intermetallic, ACS Catal., 2020, 10(17), 9977–9985, DOI: H. Baltruschat, Cyclic Voltammetry, FTIRS, and DEMS 10.1021/acscatal.0c01537. Study of the Electrooxidation of Carbon Monoxide, Formic 18 R. Rizo and B. Roldan Cuenya, Shape-Controlled Acid, and Methanol on Cyanide-Modied Pt(111) Nanoparticles as Anodic Catalysts in Low-Temperature Electrodes, Langmuir, 2009, 25(11), 6500–6507, DOI: Fuel Cells, ACS Energy Lett., 2019, 4(6), 1484–1495, DOI: 10.1021/la8041154. 10.1021/acsenergylett.9b00565. 32 K. Jiang, H. X. Zhang, S. Zou and W. B. Cai, Electrocatalysis 19 E. Plaza-Mayoral, I. J. Pereira, K. Nicole Dalby, K. D. Jensen, of Formic Acid on Palladium and Platinum Surfaces: From I. Chorkendorff, H. Falsig, P. Sebastián-Pascual and Fundamental Mechanisms to Fuel Cell Applications, Phys. M. Escudero-Escribano, Pd–Au Nanostructured Chem. Chem. Phys., 2014, 16(38), 20360–20376, DOI: Electrocatalysts with Tunable Compositions for Formic 10.1039/c4cp03151b. Acid Oxidation, ACS Appl. Energy Mater., 2022, 5(9), 10632– 33 A. Cuesta, At Least Three Contiguous Atoms Are Necessary 10644, DOI: 10.1021/acsaem.2c01361. for CO Formation during Methanol Electrooxidation on 20 A. Capon and R. Parsons, The Oxidation of Formic Acid at Platinum, J. Am. Chem. Soc., 2006, 128(41), 13332–13333, Noble Metal Electrodes Part 4. Platinum + Palladium DOI: 10.1021/ja0644172. Alloys, J. Electroanal. Chem., 1975, 65(1), 285–305, DOI: 34 W. Zhong, Y. Qi and M. Deng, The Ensemble Effect of Formic 10.1016/0368-1874(75)85124-0. Acid Oxidation on Platinum-Gold Electrode Studied by First- 21 P. N. Duchesne, Z. Y. Li, C. P. Deming, V. Fung, X. Zhao, Principles Calculations, J. Power Sources, 2015, 278, 203–212, J. Yuan, T. Regier, A. Aldalbahi, Z. Almarhoon, S. Chen, DOI: 10.1016/j.jpowsour.2014.12.071. D. Jiang, N. Zheng and P. Zhang, Golden Single-Atomic- 35 S. Siahrostami, A. Verdaguer-Casadevall, M. Karamad, Site Platinum Electrocatalysts, Nat. Mater., 2018, 17(11), D. Deiana, P. Malacrida, B. Wickman, M. Escudero- 1033–1039, DOI: 10.1038/s41563-018-0167-5. Escribano, E. Paoli, R. Frydendal, T. W. Hansen, 22 J. M. Feliu and E. Herrero, Formic Acid Oxidation, Handb. I. Chorkendorff, I. E. L. Stephens and J. Rossmeisl, Fuel Cells, 2010, DOI: 10.1002/9780470974001.f206048. Enabling Direct H2O2 Production through Rational 23 J. Joo, T. Uchida, A. Cuesta, M. T. M. M. Koper and M. Osawa, Electrocatalyst Design, Nat. Mater., 2013, 12(12), 1137– Importance of Acid-Base Equilibrium in Electrocatalytic 1143, DOI: 10.1038/nmat3795. Oxidation of Formic Acid on Platinum, J. Am. Chem. Soc., 36 A. Verdaguer-Casadevall, D. Deiana, M. Karamad, 2013, 135(27), 9991–9994, DOI: 10.1021/ja403578s. S. Siahrostami, P. Malacrida, T. W. Hansen, J. Rossmeisl, 24 N. M. Marković, H. A. Gasteiger, P. N. Ross, X. Jiang, I. Chorkendorff and I. E. L. Stephens, Trends in the I. Villegas and M. J. Weaver, Electro-Oxidation Electrochemical Synthesis of H2O2: Enhancing Activity and Mechanisms of Methanol and Formic Acid on Pt-Ru Alloy Selectivity by Electrocatalytic Site Engineering, Nano Lett., Surfaces, Electrochim. Acta, 1995, 40(1), 91–98, DOI: 2014, 14(3), 1603–1608, DOI: 10.1021/nl500037x. 10.1016/0013-4686(94)00241-R. 37 A. Sáez, E. Expósito, J. Solla-Gullón, V. Montiel and A. Aldaz, 25 A. Capon and R. Parsons, The Oxidation of Formic Acid at Bismuth-Modied Carbon Supported Pt Nanoparticles as Noble Metal Electrodes Part III. Intermediates and Electrocatalysts for Direct Formic Acid Fuel Cells, Mechanism on Platinum Electrodes, J. Electroanal. Chem., Electrochim. Acta, 2012, 63, 105–111, DOI: 10.1016/ 1973, 45(2), 205–231, DOI: 10.1016/S0022-0728(73)80158-5. j.electacta.2011.12.076. 26 A. Betts, V. Briega-Martos, A. Cuesta and E. Herrero, 38 S. G. Sun, J. Clavilier and A. Bewick, The Mechanism of Adsorbed Formate Is the Last Common Intermediate in Electrocatalytic Oxidation of Formic Acid on Pt (100) and the Dual-Path Mechanism of the Electrooxidation of Pt (111) in Sulphuric Acid Solution: An Emirs Study, J. Formic Acid, ACS Catal., 2020, 10(15), 8120–8130, DOI: Electroanal. Chem., 1988, 240(1–2), 147–159, DOI: 10.1016/ 10.1021/acscatal.0c00791. 0022-0728(88)80319-X. 27 A. Miki, S. Ye and M. Osawa, Surface-Enhanced IR 39 W. Gao, J. A. Keith, J. Anton and T. Jacob, Theoretical Absorption on Platinum Nanoparticles: An Application to Elucidation of the Competitive Electro-Oxidation Real-Time Monitoring of Electrocatalytic Reactions, Chem. Mechanisms of Formic Acid on Pt(111), J. Am. Chem. Soc., Commun., 2002, 2(14), 1500–1501, DOI: 10.1039/b203392e. 2010, 132(51), 18377–18385, DOI: 10.1021/ja1083317. 13416 | Chem. Sci., 2022, 13, 13409–13417 © 2022 The Author(s). Published by the Royal Society of Chemistry Edge Article Chemical Science 40 C. Rice, S. Ha, R. I. Masel and A. Wieckowski, Catalysts for Using in Situ EXAFS and XRD, Phys. Chem. Chem. Phys., Direct Formic Acid Fuel Cells, J. Power Sources, 2003, 2003, 5(15), 3220–3225, DOI: 10.1039/b302956e. 115(2), 229–235, DOI: 10.1016/S0378-7753(03)00026-0. 52 D. Gao, H. Zhou, F. Cai, D. Wang, Y. Hu, B. Jiang, W. B. Cai, 41 N. V. Rees and R. G. Compton, Sustainable Energy: A Review X. Chen, R. Si, F. Yang, S. Miao, J. Wang, G. Wang and X. Bao, of Formic Acid Electrochemical Fuel Cells, J. Solid State Switchable CO2 Electroreduction via Engineering Active Electrochem., 2011, 15(10), 2095–2100, DOI: 10.1007/ Phases of Pd Nanoparticles, Nano Res., 2017, 10(6), 2181– s10008-011-1398-4. 2191, DOI: 10.1007/s12274-017-1514-6. 42 M. Neurock, M. Janik and A. Wieckowski, A First Principles 53 J. M. Tura, P. Regull, L. Victori and M. D. de Castellar, XPS Comparison of the Mechanism and Site Requirements for and IR (ATR) Analysis of Pd Oxide Films Obtained by the Electrocatalytic Oxidation of Methanol and Formic Acid Electrochemical Methods, Surf. Interface Anal., 1988, 11(8), over Pt, Faraday Discuss., 2009, 140, 363–378, DOI: 10.1039/ 447–449, DOI: 10.1002/sia.740110807. b804591g. 54 V. A. Saveleva, V. Papaehimiou, M. K. Daletou, W. H. Doh, 43 X. H. Xia and T. Iwasita, Inuence of Underpotential C. Ulhaq-Bouillet, M. Diebold, S. Zafeiratos and Deposited Lead upon the Oxidation of HCOOH in HClO4 E. R. Savinova, Operando Near Ambient Pressure XPS at Platinum Electrodes, J. Electrochem. Soc., 1993, 140(9), (NAP-XPS) Study of the Pt Electrochemical Oxidation in 2559–2565, DOI: 10.1149/1.2220862. H2O and H2O/O2 Ambients, J. Phys. Chem. C, 2016, 120(29), 44 A. Tripković, K. Popović and R. Adžić, Structural Effects in 15930–15940, DOI: 10.1021/acs.jpcc.5b12410. Electrocatalysis: Oxidation of Formic Acid on Single Crystal 55 E. Herrero and J. M. Feliu, Understanding Formic Acid Platinum Electrodes and Evidence for Oscillatory Oxidation Mechanism on Platinum Single Crystal Behaviour, J. Chim. Phys., 1991, 88, 1635–1647, DOI: Electrodes, Curr. Opin. Electrochem., 2018, 9, 145–150, DOI: 10.1051/jcp/1991881635. 10.1016/j.coelec.2018.03.010. 45 J. Joo, T. Uchida, A. Cuesta, M. T. M. Koper and M. Osawa, 56 T. Al Najjar, N. Ahmed and E. N. El Sawy, Mechanistic Effects The Effect of PH on the Electrocatalytic Oxidation of of Blending Formic Acid with Ethanol on Pd Activity towards Formic Acid/Formate on Platinum: A Mechanistic Study by Formic Acid Oxidation in Acidic Media, RSC Adv., 2021, Surface-Enhanced Infrared Spectroscopy Coupled with 11(37), 22842–22848, DOI: 10.1039/d1ra01209f. Cyclic Voltammetry, Electrochim. Acta, 2014, 129, 127–136, 57 H. Miyake, T. Okada, G. Samjeské and M. Osawa, Formic DOI: 10.1016/j.electacta.2014.02.040. Acid Electrooxidation on Pd in Acidic Solutions Studied by 46 M. R. Gennero de Chialvo, G. C. Luque and A. C. Chialvo, Surface-Enhanced Infrared Absorption Spectroscopy, Phys. Formic Acid Electrooxidation on Platinum, Resolution of Chem. Chem. Phys., 2008, 10(25), 3662, DOI: 10.1039/ the Kinetic Mechanism in Steady State and Evaluation of b805955a. the Kinetic Constants, ChemistrySelect, 2018, 3(34), 9768– 58 G. Samjeské and M. Osawa, Current Oscillations during Formic 9772, DOI: 10.1002/slct.201801725. Acid Oxidation on a Pt Electrode: Insight into the Mechanism 47 A. Bagger, W. Ju, A. S. Varela, P. Strasser and J. Rossmeisl, by Time-Resolved IR Spectroscopy, Angew. Chem., 2005, Single Site Porphyrin-like Structures Advantages over 117(35), 5840–5844, DOI: 10.1002/ange.200501009. Metals for Selective Electrochemical CO2 Reduction, Catal. 59 J. Clavilier, Pulsed Linear Sweep Voltammetry with Pulses of Today, 2017, 288, 74–78, DOI: 10.1016/j.cattod.2017.02.028. Constant Level in a Potential Scale, a Polarization 48 D. Bohra, I. Ledezma-Yanez, G. Li, W. de Jong, E. A. Pidko Demanding Condition in the Study of Platinum Single and W. A. Smith, Lateral Adsorbate Interactions Inhibit Crystal Electrodes, J. Electroanal. Chem., 1987, 236(1–2), HCOO − While Promoting CO Selectivity for CO2 87–94, DOI: 10.1016/0022-0728(87)88020-8. Electrocatalysis on Silver, Angew. Chem., Int. Ed., 2019, 60 E. Pizzutilo, S. Geiger, S. J. Freakley, A. Mingers, S. Cherevko, 58(5), 1345–1349, DOI: 10.1002/anie.201811667. G. J. Hutchings and K. J. J. Mayrhofer, Palladium 49 R. Gómez and M. J. Weaver, Electrochemical Infrared Electrodissolution from Model Surfaces and Nanoparticles, Studies of Monocrystalline Iridium Surfaces. 3. Adsorbed Electrochim. Acta, 2017, 229, 467–477, DOI: 10.1016/ Nitric Oxide and Carbon Monoxide as Probes of Ir(100) j.electacta.2017.01.127. Interfacial Structure, J. Phys. Chem. B, 1998, 102(19), 3754– 61 S. Cherevko, N. Kulyk and K. J. J. Mayrhofer, Durability of 3764, DOI: 10.1021/jp980098s. Platinum-Based Fuel Cell Electrocatalysts: Dissolution of 50 A. Cuesta, G. Cabello, F. W. Hartl, M. Escudero-Escribano, Bulk and Nanoscale Platinum, Nano Energy, 2016, 29, 275– C. Vaz-Domı́nguez, L. A. Kibler, M. Osawa and C. Gutiérrez, 298, DOI: 10.1016/j.nanoen.2016.03.005. Electrooxidation of Formic Acid on Gold: An ATR-SEIRAS 62 A. Jain, S. P. Ong, G. Hautier, W. Chen, W. D. Richards, Study of the Role of Adsorbed Formate, Catal. Today, 2013, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder and 202(1), 79–86, DOI: 10.1016/j.cattod.2012.04.022. K. A. Persson, Commentary: The Materials Project: A 51 A. Rose, S. Maniguet, R. J. Mathew, C. Slater, J. Yao and Materials Genome Approach to Accelerating Materials A. E. Russell, Hydride Phase Formation in Carbon Innovation, APL Mater., 2013, 1(1), 011002, DOI: 10.1063/ Supported Palladium Nanoparticle Electrodes Investigated 1.4812323. © 2022 The Author(s). Published by the Royal Society of Chemistry Chem. Sci., 2022, 13, 13409–13417 | 13417