Two-Dimensional Titanium and Molybdenum Carbide MXenes as Electrocatalysts for CO2 Reduction.

Electrocatalytic CO2 reduction reaction (CO2RR) is an attractive way to produce renewable fuel and chemical feedstock, especially when coupled with efficient CO2 capture and clean energy sources. On the fundamental side, research on improving CO2RR activity still revolves around late transition metal-based catalysts, which are limited by unfavorable scaling relations despite intense investigation. Here, we report a combined experimental and theoretical investigation into electrocatalytic CO2RR on Ti- and Mo-based MXene catalysts. Formic acid is found as the main product on Ti2CTx and Mo2CTx MXenes, with peak Faradaic efficiency of over 56% on Ti2CTx and partial current density of up to -2.5 mA cm-2 on Mo2CTx. Furthermore, simulations reveal the critical role of the Tx group: a smaller overpotential is found to occur at lower amounts of -F termination. This work represents an important step toward experimental demonstration of MXenes for more complex electrocatalytic reactions in the future.

Introduction

Efforts to tackle rising CO2 concentration in the atmosphere have been primarily focused on carbon capture and sequestration, as well as decarbonization of the energy and land use sectors (Walsh et al., 2017). However, an equally pressing issue of depleting fossil energy and chemical raw materials is looming. Electrocatalytic CO2 reduction reaction (CO2RR) presents an attractive pathway to achieve both decarbonization of energy economy and production of renewable fuel/chemical feedstock (De Luna et al., 2019, Seh et al., 2017), especially when coupled with increasingly affordable clean electricity (Obama, 2017).

To realize industrial-scale CO2RR, substantial challenges on both the fundamental (i.e., catalyst activity and selectivity) and system levels (i.e., mass transport, conversion rate, and energy efficiency) need to be addressed (Jouny et al., 2018, De Luna et al., 2019, Handoko et al., 2018c, Higgins et al., 2019). On the fundamental side, there has been some progress in understanding CO2RR in liquid electrolytes, with emphasis on late transition metals, particularly copper, owing to its unusual ability to convert CO2 to multi-carbon products (Hori, 2008, Huang et al., 2017). However, despite intense optimization of these transition metal catalysts (Saberi Safaei et al., 2016, Handoko et al., 2016, Mistry et al., 2016, Li et al., 2017b, Ren et al., 2016), their activity and selectivity seem to be limited. One of the most significant barriers limiting fundamental CO2RR on transition metal catalysts appears to be the linear scaling relations between the binding energies of reaction intermediates (Liu et al., 2017). These unfavorable scaling relations due to similarly bound reaction intermediates (e.g., ∗COOH, ∗CO, ∗CHO, “∗” refers to a site on the catalyst surface) limit the CO2RR overpotential that can be achieved on pure transition metal surfaces.

One of the most promising ways to improve CO2RR activity is to explore new catalyst material systems that allow stabilization of intermediates with different scaling relations. MXenes, a family of two-dimensional transition metal carbide/nitride materials with metallic-like conductivity (Anasori et al., 2017), present a viable solution. MXenes have a general formula of MXT, where M represents an early transition metal, X is carbon and/or nitrogen, with n in the range of 1–4 (Anasori and Gogotsi, 2019, Deysher et al., 2020). T represents surface termination groups, which can include –O, –F, etc. (Hope et al., 2016). The tunable surface and internal configuration of MXenes, including the possibility of mixed surface terminations, as well as mixed metal atoms in solid solution (Yang et al., 2016) or ordered structure (Anasori et al., 2015, Anasori and Gogotsi, 2019), allow for tailoring of intermediates' binding configuration and strength (Hart et al., 2019, Anasori et al., 2016, Handoko et al., 2018a, Chen et al., 2019), making it an ideal platform to search for active and selective CO2RR catalysts (Handoko et al., 2019). To date, investigations of MXenes for electrocatalytic CO2RR have been primarily based on theoretical calculations (Li et al., 2017a, Handoko et al., 2018b, Chen et al., 2019, Zhang et al., 2017b), although some works on photocatalytic CO2RR have indicated that MXenes may enhance charge separation or act as co-catalyst (Cao et al., 2018, Ye et al., 2018).

Here, we investigate the electrocatalytic CO2RR activity on Ti2CT and Mo2CT MXenes using a combination of experiment and theory. These MXenes were chosen as they have Ti and Mo metal sites with opposing hydrogen binding behavior (Laursen et al., 2012). The CO2RR experiments are carried out in mixtures of acetonitrile, water, and 3-butyl-1-methyl-1H-imidazol-3-ium tetra-fluoroborate (BMIMBF4) electrolyte. Formic acid is found to be the main CO2RR product on these MXenes with Faradaic efficiency of 56.1% at −1.8 V (all potentials in this work are expressed with respect to the standard hydrogen electrode, SHE). H2 is the main side product, alongside trace amounts of CO, CH4, and other hydrocarbons. More importantly, we demonstrate that the nature of surface terminating group appears to control the CO2RR activity. Specifically, the presence of –F termination group, commonly adsorbed onto MXene surface during synthesis (Hope et al., 2016), was found to alter the binding strength of intermediates and the corresponding CO2RR limiting potential compared to fully –O terminated MXenes, as supported by density functional theory (DFT). Overall, this work provides insights on MXene electrocatalysts that circumvent traditional scaling relations in CO2RR, which can potentially be extended to other promising reactions of interest.

Results

We first investigate the effect of surface termination group T on the CO2RR activity of Ti2CT MXenes. To achieve this, we synthesized two variants of Ti2CT with different surface termination compositions using different etching procedures, namely, 18 h in 10% HF and 48 h in 4 M KF-HCl mixture (refer to the Transparent Methods in the Supplemental Information). These MXenes will be referred to as Ti2CT (HF) and Ti2CT (KF-HCl), respectively. Previous works suggest that the use of fluoride salt etching solution in place of HF can reduce the amounts of –F terminations (Hope et al., 2016, Handoko et al., 2018a). The scanning electron micrographs of both Ti2CT (HF) and Ti2CT (KF-HCl) show layered structures after the etching procedure (Figures 1A and 1B), consistent with previously reported morphologies (Anasori et al., 2017). X-ray diffraction (XRD) shows that most of the Ti2AlC precursors are successfully converted to Ti2CT, with the appearance of the characteristic (002) broad peaks of MXenes around 2θ = 11.1° and 9.2° for HF and KF-HCl etched Ti2CT, respectively (Figure 1D; see Figure S1 and Table S1 for characterization of precursors). These peaks correspond to MXene interlayer distances of around 8.0 and 9.6 Å, consistent with previous reports on smaller interlayer distance for HF-etched samples Naguib et al., 2012.

Figure 1

Morphology and Phase Characterization of Ti2CT and Mo2CT MXenes

Scanning electron micrographs of as-synthesized (A) Ti2CT (HF), (B) Ti2CT (KF-HCl), and (C) Mo2CT (HF) and (D) the corresponding X-ray diffraction data.

X-ray photoelectron spectroscopy (XPS) of the as-synthesized Ti2CT samples in the Ti 2p range detects a mix of Ti species, including Ti-T/C, Ti(II), and more oxidized Ti(IV) species (Figures 2A and 2B; see Tables S2–S5 for more details). Some degree of Ti(IV) formation is unavoidable in Ti2CT as it is less stable against oxidation compared with other MXenes with higher n such as Ti3C2T (Zhang et al., 2017a). A comparison of the F to Ti atomic ratio (Tables S3 and S4) shows that Ti2CT (HF) has significantly more –F surface terminations (F/Ti = 0.36) than Ti2CT (KF-HCl) (F/Ti = 0.21). We have previously shown that the electrocatalytic activities of MXenes are very sensitive to the surface termination (Handoko et al., 2018a), thus we posit that the variation in –F termination content on Ti2CT samples will alter their CO2RR activity.

Figure 2

Analysis of the "M" Elements of Ti2CT and Mo2CT MXenes

X-ray photoelectron spectroscopy measurements of MXenes before and after CO2RR: (A and D) Ti 2p spectra of Ti2CT (HF), (B and E) Ti 2p spectra of Ti2CT (KF-HCl), (C and F) Mo 3d spectra of Mo2CT (HF).

As an initial CO2RR activity assessment, we first compared the linear scanning voltammetry (LSV) profile of Ti2CT (HF) and Ti2CT (KF-HCl) in the presence and absence of CO2 in a mixture of 80:15:5 mol fraction of acetonitrile:water:BMIMBF4 electrolyte (Figure S2). The electrolyte mixture is used in this study owing to prevalent HER in aqueous electrolyte systems like 0.1 M KHCO3 (Figure S3). Dipolar aprotic solvents like acetonitrile have enhanced CO2 solubility compared with aqueous systems (Gennaro et al., 1990). Furthermore, imidazolium-based ionic liquid has been shown to enhance CO2RR (Asadi et al., 2016), possibly by forming complexes with CO2 molecules at moderately cathodic potentials (−0.1 V versus SHE) (Matsubara et al., 2015, Rosen et al., 2011).

Under N2 purging, both Ti2CT samples display sharp onset potential at around −1.8 V (Figures S2A and S2B), which can be attributed to the hydrogen evolution reaction (HER). The introduction of CO2 into the reaction environment changes the LSV profile for both Ti2CT samples significantly. Most notably, some cathodic features are observed around −1.4 to −1.8 V. These features were especially clear on Ti2CT (KF-HCl), indicating more significant ∗CO2 (or related intermediates) adsorption and interaction (Salehi-Khojin et al., 2013) compared with Ti2CT (HF). Additionally, the HER onset potentials on both Ti2CT samples are delayed by about −0.1 to −0.2 V with reduced current density. As the electrolyte pH (∼1.24) and reference electrode are unaffected by CO2 purging (Figure S4), this HER “poisoning” effect can be attributed to ∗CO2 (or related intermediates) adsorption that competes with ∗H. Such an effect is expected on transition metal surfaces that show weak ∗H binding strength like Ti (Huang et al., 2017, Zhang et al., 2014).

To quantify their CO2RR activity, both Ti2CT samples are subjected to chronoamperometric (constant voltage) measurement for 100 min under continuous CO2 gas purging in a two-compartment cell (Figure S5). Both Ti2CT samples retain most of their characteristic XPS peaks after CO2RR, with slight increase in the more oxidized Ti(IV) signals due to ambient exposure during testing (Figures 2D and 2E). Online gas chromatography (GC) and nuclear magnetic resonance (NMR) were used to quantify gaseous and liquid products from CO2RR, revealing formic acid and H2 to be the main products on both Ti2CT samples (Figures 3 and S6). It is clear from the Faradaic efficiency (FE) plot (Figure 3A) and partial current density (j) plot (Figure 3B) that Ti2CT (KF-HCl) shows enhanced selectivity and turnover for CO2RR to formic acid compared with Ti2CT (HF), although the CO2RR onset potentials are similar for both samples at −1.5 V. At a potential of −1.8 V, Ti2CT (KF-HCl) displays 56.1% FEHCOOH compared with 20.7% on Ti2CT (HF), with a corresponding 2.5 times higher jHCOOH normalized to geometric surface area. Plots of jHCOOH normalized to electrochemical surface area are also shown in Figure S7 and Table S6.

Figure 3

CO2RR Selectivity and Activity on Ti2CT and Mo2CT MXenes

Comparison of (A) Faradaic efficiency and (B) partial current density normalized by geometric surface area for CO2RR to formic acid on Ti2CT and Mo2CT MXenes. Error bars represent one standard deviation of three independent measurements.

To gain insight on the experimental results, we turn to DFT calculations to systematically investigate the effect of varying amounts of –F termination group on the theoretical CO2RR overpotential on Ti2CT samples. In particular, three different Ti2CT structures were modeled with T groups comprising (1) 0.0% –F, 100.0% –O, (2) 33.3% –F, 66.7% –O, and (3) 66.7% –F, 33.3% –O (Figure S8). The T compositions were selected based on the range of –F termination previously studied on Ti-based MXenes (Handoko et al., 2018a). On fully –O terminated Ti2CT surface, CO2RR to formic acid is completed in four steps, including a CO2 adsorption step, two consecutive proton-coupled electron transfer (PCET) steps, and finally an HCOOH (l) desorption step (Figure 4A). In this case, only the ∗COOH pathway was considered, as the alternative route through HCOO intermediate was found to be unfavorable (Figure S9), consistent with the literature (Li et al., 2017a). The ∗CO2 + H+ + e− → ∗COOH step (PCET-1) is predicted to be potential limiting with a free energy change (ΔGelem, at 0 V applied potential) of 0.85 eV. Hence, the theoretical CO2RR limiting potential (ULCO2 = -ΔGelem/e) can be calculated to be −0.85 V.

Figure 4

Density Functional Theory Calculations for CO2RR to Formic Acid on Ti2CT and Mo2CT MXenes

Calculated free energy diagram at 0 V applied potential for CO2RR to formic acid on (A) Ti2CT and (B) Mo2CT MXenes with varying fractions of –F and –O surface terminating groups. Blue, purple, red, brown, and white spheres represent Ti, Mo, O, C, and H atoms, respectively.

Gradual –F substitution of the –O T group in Ti2CT results in significant variation of reaction free energy (Figure 4A). Specifically, the PCET-1 potential limiting step becomes more endergonic (larger ΔGelem) at a higher fraction of –F substitution. As a result, the ULCO2 becomes more negative at −0.89 and −1.26 V for 33.3% and 66.7% –F, respectively, possibly due to increasingly unstable ∗COOH conformations (Figure S8). As ∗HCOOH adsorption energy does not change substantially with –F substitution, we observe that the PCET-1 reaction step remains potential limiting throughout. This means that the least negative ULCO2 can only be achieved on fully –O terminated Ti2CT. These simulation results explain the higher FEHCOOH and jHCOOH observed on Ti2CT (KF-HCl) as it has significantly lower –F surface termination compared with Ti2CT (HF).

To study if the effect of –F substitution is common to other MXenes, we expand the investigation to Mo2CT. In this case, only HF etching is used, as milder KF-HCl etchant is not able to etch Mo2Ga2C precursor. Nonetheless, Mo2CT could potentially be a better CO2RR catalyst as it tends to have fewer –F terminations than Ti2CT even when a harsher etching condition using 48% HF is used (Halim et al., 2016). The morphology of Mo2CT is slightly different than Ti2CT with less obvious delamination between the layers (Figure 1C), consistent with previous studies (Halim et al., 2016). XRD analyses also indicate successful etching of most Mo2Ga2C precursor, with a broad peak at 2θ = 9.0° indicative of a (002) interlayer distance of around 9.8 Å (Figure 1D; see Figure S1 and Table S1 for characterization of precursors). XPS shows a relatively large Mo-T/C component with no signs of Mo(VI) (Figure 2C). The F/Mo atomic ratio is estimated to be 0.03 (Table S5).

Initial LSV assessments under N2 purging show earlier HER onset potential on Mo2CT around −1.3 V (Figure S2C), which is consistent with its superior HER activity (Handoko et al., 2018a, Seh et al., 2016). However, it is unusual that Mo2CT also displayed a similar HER poisoning effect under CO2 purging since, unlike Ti, Mo should bind to ∗H strongly (Laursen et al., 2012). The observation of HER poisoning on both MXenes suggests that the interaction of ∗CO2 or other intermediates with the T groups is possibly more critical than that with the base metal sites.

GC and NMR analysis during chronoamperometric measurements revealed that Mo2CT starts forming formic acid at much less negative potential (−0.9 V) than both types of Ti2CT (−1.5 V). This is consistent with the cathodic features observed in LSV under CO2 purging (Figure S2C). Although the maximum FEHCOOH of Mo2CT (32.6% at −1.3 V, Figure 3A) is lower compared with Ti2CT (KF-HCl, 56.1% at −1.8 V), the former shows significantly higher jHCOOH up to −2.5 mA cm−2geo at less negative potentials (Figure 3B), suggesting that it is quite active in reducing CO2 to formic acid. The stability of CO2RR on Mo2CT is assessed by conducting continuous electrolysis for 500 min. Some fluctuation in FEHCOOH from 39.9% in the 100th minute to 25.7% in the 500th minute was observed (Figure S10A). We attribute the apparent FEHCOOH instability to the loss of volatile acetonitrile in the electrolyte during continuous purging rather than catalyst deactivation, as no degradation of jHCOOH is observed (Figure S10B). Gradual evaporation of acetonitrile would alter the electrolyte composition and increase the water proportion, leading to higher HER turnover that dominates the total current density (Figure S10B). Loss of volatile electrolyte component is a common issue for analytical electrochemical systems that require continuous sampling (Lazouski et al., 2020).

Mo2CT was found to retain most of its characteristic XPS peaks after CO2RRR as well (Figure 2F). The earlier onset and higher turnover for CO2RR to formic acid on Mo2CT is unexpected owing to its predisposition for catalyzing HER (Seh et al., 2016, Pan, 2016, Handoko et al., 2018a). As such, Mo-based MXenes have not been considered active for CO2RR in previous computational studies (Handoko et al., 2018b, Morales-García et al., 2018, Chen et al., 2019), with the exception of a hypothetical Mo3C2(OH)2 structure (Li et al., 2017a).

To explain the experimental finding on Mo2CT, we turn back to DFT calculations to examine CO2RR steps on this surface. Similar to Ti2CT, we construct fully –O terminated Mo2CT and gradually replace the T with –F, but this time at lower fractions of 11.1%, 22.2%, and 33.3% (Figure S11) owing to fewer –F terminations detected experimentally. On fully –O terminated Mo2CT, the reaction step ∗COOH + H+ + e−→ ∗HCOOH (PCET-2) with ΔGelem of 0.47 eV (and ULCO2 of −0.47 V) is found to be potential limiting. Here, –F substitution on Mo2CT appears to affect CO2RR intermediates more significantly than on Ti2CT (Figure 4B). We observe on Mo2CT that the ΔGelem of PCET-1 increases significantly by 0.35 eV after substitution of only 11.1% of the –O termination to –F. This is in contrast to the 0.05 eV change observed on Ti2CTx when 33.3% of the –O T group is replaced by –F (Figure 4A).

More interestingly, we note that the presence of –F T group on Mo2CT changes the potential limiting step from PCET-2 on fully –O terminated surface to PCET-1 after 11.1% –F substitution. This means that the lowest ULCO2 can be found at low fraction of –F substitution, between 0% and 11.1%, that equalize the ΔGelem of PCET-1 and PCET-2. Quadratic functions fitting of the ∗COOH free energy and the respective limiting potential at different fractions of –F substitution suggest that such a minimum can be found at around 4.2% –F (Figure S12A).

One possible reason for the unique CO2RR behavior on Ti2CT and Mo2CT MXenes could lie in the preference toward the ∗HCOOH pathway. Unlike late transition metal catalysts, where a majority of CO2RR goes through the ubiquitous ∗CO intermediate (Peterson and Nørskov, 2012), the reaction path through ∗HCOOH intermediate is favorable on MXenes owing to the hydrogen-bond interaction between ∗HCOOH and the T groups, particularly –O (Figure S13). The preference toward the ∗HCOOH pathway results in non-linear scaling with ∗COOH in terms of their binding energies, as these two intermediates are coordinated differently on the MXene surfaces (Figure 5A). In general, ∗COOH binds to the –O T groups on MXene surfaces through C atom, whereas ∗HCOOH binds through the H atom (Figures S8, S11, and S13). A volcano-like plot of limiting potentials can then be constructed, with boundaries drawn to mark the neutral potential of PCET-1 (orange line) and PCET-2 (brown line) reaction steps with respect to ∗COOH binding energy (Figure 5B). A majority of the catalysts are governed by the PCET-1 step, where less negative ULCO2 is achieved at stronger ∗COOH binding energy, up to a point where protonation of ∗COOH to ∗HCOOH becomes difficult. It can be seen that fully –O terminated Mo2CT sits atop the volcano, near to the ideal case where both PCET-1 and PCET-2 can proceed at the same limiting potential.

Figure 5

Relations between Binding Energies of Different CO2RR Reaction Intermediates on Ti2CT and Mo2CT MXenes

(A) ∗HCOOH binding energy plot against ∗COOH binding energy showing deviation from linear scaling relations.

(B) Limiting CO2RR potentials for elementary steps. The lines represent the calculated potential where the most negative reaction steps are neutral as a function of ∗COOH binding energy (PCET-1: ∗CO2 + H+ + e− → ∗COOH; PCET-2: ∗COOH + H+ + e− → ∗HCOOH).

(C) Calculated ULCO2-ULH2 plot with respect to ULCO2 on all variants of Ti2CT and Mo2CT theoretical models with different –F T termination fractions in this study.

We also examined the HER reaction steps on both Ti2CT and Mo2CT surfaces at various –F substitution using DFT calculation (Figure S14). This is important as HER is the main competition to CO2RR on these surfaces. Compared with CO2RR, HER is a simpler reaction, which can be represented by: ∗ + H+ + e− → ∗H → ½ H2 (g) + ∗. It is found that –F substitution generally destabilizes ∗H adsorption due to much weaker H-F interaction. The weaker ∗H results in a more negative HER limiting potential (ULH2) nearly on all cases, except on Mo2CT with 11.1% –F substitution where the limiting potential is close to the ideal value (−0.02 V, Figure S12B).

The difference between the limiting potentials of CO2RR and HER (ULCO2-ULH2) could then be used to gauge the selectivity of the catalysts toward CO2RR (Hong et al., 2016, Shi et al., 2014). In addition to having the least negative ULCO2 value, we found that –O terminated Mo2CT also possesses the least negative ULCO2-ULH2 difference of −0.1 V (Figure 5C). Quadratic functions fitting of ULCO2-ULH2 on Mo2CTx identifies a minimum at 3.8% –F substitution (Figure S12B), indicating that low amounts of –F T presence may be beneficial to CO2RR.

Although the ULCO2-ULH2 values on all variants of Ti2CT and Mo2CT are still generally negative across all fractions of –F substitutions, we recognize that the kinetics of both HER and CO2RR can be significantly altered by the reaction environment (König et al., 2019). In addition to the enhanced CO2 solubility in acetonitrile, the formation of BMIM-CO2 complex has been shown to enhance CO2 mass transport and significantly boost CO2RR activity (Rosen et al., 2011). Furthermore, the hydrophobic BMIM cation has been proposed to populate near the catalyst surface and suppress HER upon application of CO2RR-relevant cathodic potentials (Rosen et al., 2012).

Apart from formic acid, we also detected up to 1.1% FE of CO and trace amounts of CH4 and multi-carbon products (Figures S15–S17). Our finding partially validates the CO2RR route via a ∗HCOOH intermediate (Li et al., 2017a, Handoko et al., 2018b, Chen et al., 2019), although further optimization of MXene surface terminations (Table S7 and S8 and Figures S18–S22) is necessary to enhance the ∗HCOOH intermediate stability for the production of more reduced moieties like CH4 and multi-carbon products.

Discussion

In this work, we report a combined experimental and theoretical CO2RR investigation on Ti2CT and Mo2CT MXenes. Formic acid is found to be the main CO2RR product with maximum FE exceeding 56% at −1.8 V versus SHE on Ti2CT (KF-HCl). In addition, CO, trace amounts of CH4, and other multi-carbon products are also detected. More importantly, we found that the CO2RR activity appears to be correlated with the fraction of –F and –O surface termination groups (T). Ti2CT (HF) with large amounts of –F (less –O) shows poorer CO2RR activity and selectivity than Ti2CT (KF-HCl) with lower –F (more –O). Even higher CO2RR activity, up to −2.5 mA cm−2geo is observed on Mo2CT catalysts with minimal –F fraction.

DFT simulations indicate that the presence of –F destabilizes ∗COOH and ∗H, thus causing the limiting potential of both CO2RR and HER to become more negative. An exception is found on Mo2CT where small amounts of –F substitution is predicted to balance the individual limiting potentials for PCET-1 and PCET-2 steps and yield the smallest overpotential for CO2RR to formic acid. The CO2RR activity on MXene surfaces could be attributed to unique intermediate-T interaction, which results in stabilization of differently coordinated ∗COOH and ∗HCOOH intermediates. The dissimilar binding coordination leads to deviation in linear scaling relations typically observed on late transition metal-based CO2RR catalysts like copper.

Although the activity and selectivity toward formic acid can be further improved, this work represents an important step toward experimental demonstration of MXenes for electrocatalytic CO2RR. The detection of trace amounts of CH4 and other hydrocarbon products indicate that CO2RR to more complex moieties on MXenes is possible as well. Work is ongoing to further enhance the activity and selectivity of MXene catalysts through material design strategies including surface engineering (Hart et al., 2019), doping (Yu et al., 2019, Li et al., 2018, Gao et al., 2019), and formation of composite/hybrid structures (Handoko et al., 2019, Chi et al., 2018). In addition, we believe that MXenes could be extended to other technologically important electrocatalytic reactions such as nitrogen reduction or methane oxidation, via surface/composition tailoring and catalytic reaction engineering in the future.

Limitations of the Study

This study investigates Ti- and Mo-based MXenes for CO2RR to HCOOH and demonstrates different reaction pathways that can break the scaling relations seen in pure transition metal catalysts. However, the prevalent HER activity on MXenes requires the use of non-aqueous electrolyte, which can be challenging for reaction scale up. Gradual loss of volatile component of the non-aqueous electrolyte also causes fluctuation in the HCOOH selectivity. Further optimization of MXene surface terminations is necessary to enhance the intermediate stability for the production of more reduced moieties like CH4 and multi-carbon products. The DFT computational methods employed here only lend insight from the thermodynamic point of view. The actual experimental result also depends on reaction kinetics that is challenging to probe in complicated systems with multiple intermediates like CO2RR.

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Zhi Wei Seh (sehzw@imre.a-star.edu.sg).

Materials Availability

This study did not generate new unique reagents.

Data and Code Availability

This study did not generate/analyze datasets/code.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.