Direct C–C bond scission of xylitol to ethylene and propylene glycol precursors using single-atom catalysts (SACs) anchored on MgO

Currently, starch-derived ethanol and biodiesel are used as alternatives for renewable fuels. However, this idea of using edible feedstock poses a threat to food resources.¹ Lignocellulosic biomass is regarded as the most promising sustainable carbon source since it is abundantly available at a low cost worldwide. It consists of cellulose, hemicellulose, and lignin, which is ubiquitous in agricultural residues (e.g., corn wheat straw and corn stover), wood, energy crops (e.g., switchgrass, kenaf, eucalyptus, and miscanthus), and waste streams (e.g., pulp, paper mill, and bio-waste), as a potential feedstock.^2–4 Cellulose is a homopolymer composed of β-D-glucopyranose units joined via β-glycosidic bonds⁵ that can be disassembled into glucose monomers, and it accounts for 40%–50% of lignocellulosic biomass feedstock.⁶ Lignin comprises 15%–30% of biomass weight, has high aromatic functionality, and is made up of three main monomers: p-coumaryl, coniferyl, and sinapyl alcohols.⁷ Because of its high carbon content, lignin contains around 40% of the biomass’s potential energy.^6,8 Hemicellulose makes up 15%–30% of lignocellulosic biomass,⁹ and unlike cellulose, it is a branched amorphous polymer and has a low degree of polymerization. As a result, hemicellulose hydrolyzes more easily than cellulose, allowing it to be separated under moderately severe reaction conditions.⁶ Furthermore, based on the lignocellulose source, the hemicellulose structure can consist of pentoses, hexoses, and uronic acids.¹⁰ Sugars from the pentoses (xylose and arabinose) and hexose (fucose, glucose, galactose, mannose, rhamnose, etc.) can be converted into biofuels and renewable chemicals. Aside from fuel production, it is important to substitute petroleum-derived chemicals with comparable products obtained from biomass or alternative products of different properties⁶ due to the depletion of fossil fuels and environmental concerns. Studies on hemicellulose upgrading have mostly focused on the primary sugar: xylan, which accounts for 8%–25% of the biomass.¹⁰ On the other hand, xylose conversion studies mainly focus on furfural production and its derivatives owing to the value-added chemicals that can be obtained from the latter.^11–16 Another important, valued chemical that can be obtained from xylose is xylitol. Xylitol is an alcohol sugar frequently used in the pharmaceutical and food industries as a sucrose alternative due to its sweetness and low-calorie content.^17,18 It can also serve as a platform chemical for producing ethylene glycol (EG) and propylene glycol (PG),^19,20 which are frequently used as feedstock for generating hydrogen.^21,22 More so, the obtained products (EG and PG) are important chemical raw materials that are majorly used as monomers in synthesizing polyester fibers and resins, as well as functional fluids like antifreeze and coolants.^19,23 It is well known that the majority of biomass resources are made up of C–C and C–O bonds,²⁴ and the selective scission and functionalization of unstrained C–C bonds^25–28 are critical in organic chemistry, biodegradation, and the petroleum industry. Furthermore, carbohydrates, chitin, lignocellulosic, and their platform molecules, in particular, contain a large number of C(OH)–C bonds, the cleavage and functionalization of which can degrade biomass and convert it into sustainable chemicals.²⁴ EG and PG are currently produced in the industry through the multi-step conversion of non-renewable crude oil-based ethylene and propylene from their epoxide precursors.²³ Experimentally, they have been suggested to be produced from xylitol hydrogenolysis in the presence of Ca(OH)₂ by scission of the C₂–C₃ bond via the xylose intermediate, following the retro-aldol condensation mechanism proposed by Wang et al.²⁹ 

A significant challenge for effective biomass utilization and upgrading is catalysis. Several catalysts have been tested for biomass processing,^{11,18,30–34} but only a few are efficient, time-stable, recyclable, and selective to the target product. Catalyst minimization and efficiency can be accomplished by combining active sites (metals), supports, and promoters.⁶ Metals are the most effective catalytic systems for chemical reaction processes, energy transformation, and environmental cleanup. They are used as catalysts in various applications, including crude oil refining, fuel cells, emission reduction, and the production of chemical precursors, pharmaceuticals, and agrochemicals. It is a well-known fact that catalysis is a surface phenomenon, where the reaction occurs on metal surfaces. Meanwhile, any metal atoms inaccessible to the reactants are largely wasted. As a result, the fraction of these metal atoms used as catalysts should be minimized.³⁵ Supported noble metal nanoparticles are essential catalysts that facilitate many critical technologies, such as energy production and environmental cleanup.^35–37 Noble metals (e.g., Pt, Pd, Ru, Au, and Ag), on the other hand, are expensive and scarce. Therefore, any advancement that reduces these metal usage, enhances their activity and selectivity, and/or improves their long-term stability is extremely crucial.^35,38–40 Liu³⁵ reported that metal nanoparticle size reduction alters their catalytic properties in various ways such as (i) surface effects, in which the fraction of unsaturated coordination bonds increases; (ii) quantum size effects, in which electron confinement results in increased energy levels and the widening of the HOMO‒LUMO gap; (iii) metal support interactions, in which chemical bonding between metal and support improves and charge transfer may ensue; and (iv) cluster configurations, in which the atom positions and their number within a cluster can significantly alter their physicochemical properties. The use of metal clusters anchored on supports is a promising alternative for biomass processing, as the interaction between metals and supports modifies the catalyst’s surface properties. These modifications can greatly enhance catalytic activity, alter selectivity to the targeted product, and boost catalyst stability in the presence of biomass-derived impurities and harsher reaction conditions.⁶ Small metal clusters can be perceived as accumulating from single atoms, dimers, trimers, etc., and should act very much like atoms or molecules than surface atoms on large metal NPs, which function like single-crystal surfaces. A supported single-atom catalyst (SAC) is made up of only secluded individual atoms as the primary active centers that are highly dispersed on and coordinated with the surface atoms of suitable support.^41–43 SACs not only boost the atom efficiency of costly metals but also offer a new strategy for tuning the activity and selectivity of a catalytic reaction.^38,43 New supported metal cluster catalysts must be specifically optimized for particular reactions to ensure high reaction rates, selectivity, and yields.⁴² SACs have been employed simultaneously with their parent single-crystal surfaces for various chemical reactions and show excellent performance in terms of activity and yield.^41,44–48 Anchored onto high-surface-area supports, SACs have the potential to profoundly reshape the field of heterogeneous catalysis, which has been crucial in facilitating several important innovations. Although there is no clear description of xylitol hydrogenolysis into EG and PG, it has been proposed that polyols (such as sorbitol) hydrogenolysis generally involves two major steps, namely, dehydrogenation of polyols to the respective carbonyl intermediates on metal catalysts and subsequent C–C bond cleavage with bases, most likely via the retro-aldol condensation.^23,29 In this study, we employed density functional theory (DFT) techniques to investigate the activities of magnesium oxide (MgO) supported SACs on the direct C–C bond cleavage of xylitol into EG and PG precursors. Our SACs are made up of a mix of noble and transition metal (TM) catalysts (Co, Ni, Ru, Rh, Pd, Ir, and Pt). In accordance with this, we investigated the aforementioned reaction on the thermodynamic stable single-crystal slab (SCS) surfaces of the above metals, namely, Co (0001), Ni (111), Ru (0001), Rh (111), Pd (111), Ir (111), and Pt (111), respectively. MgO with a melting point of ∼3250 K⁴⁹ is used as catalyst support because of its abundant availability, high surface area, stoichiometry and composition, cation valence, redox properties, acid-base features, and crystal and electronic configuration.³⁶ By transferring electrons between the parent catalyst and MgO as support, the presence of MgO as a catalyst support results in a synergetic effect in fine-tuning the electronic state of the overall catalyst.³⁶ MgO usually shows extraordinary stability and activity when used as support for various chemical reactions^50,51 and has shown superior performance to other catalyst supports used for a similar process.³⁶ Several MgO-supported TMs have been synthesized and tested for various chemical processes.^52–58 Under realistic experimental conditions, SACs are prepared from the metal precursors while reducing the amount of metal loading.³⁵ They are frequently seen to be dispersed on various sites of the desired supports when observed via a scanning transition electron microscope (STEM).³⁵ As a result, in this study, our designed SACs constitute an isolated TM(s) anchored on the most stable site of MgO (100).

Also, we elucidate the role of oxygen vacancy in the MgO support and its effects on the SACs conformation and xylitol C–C bond scission activity. It is also well-known that oxygen vacancies can be created on the surface of MgO under certain conditions. This is because, even though MgO is thermodynamically stable, it can still undergo a reduction in the presence of reactive gases, such as hydrogen or carbon monoxide. The reduction process leads to the formation of oxygen vacancies, which are sites where oxygen atoms are missing from the surface.⁵⁹ 

Unlike supported sub-nanoparticles, the structure and chemical reactivity of the support surfaces tends to be more important for supported metal clusters, particularly for single metal atoms. Furthermore, it is worth noting that reactions on some catalyst systems, including supported SACs, can occur on the interphase between the catalyst and the supports,⁶⁰ resulting in complexity in the metal’s activity trend. This observation will be investigated during the adsorption of the reacting species in this study. Herein, we investigate the direct C–C bond cleavage reaction of xylitol into EG and PG precursors on the thermodynamically stable surfaces of selected TMs [refer to single-crystal catalysts (SCCs)] and MgO-supported SACs (see Fig. 1). Furthermore, we examined and compared the electronic structures of the SACs with their parent single-crystal surfaces. Our findings revealed distinct differences between these surfaces, laying the groundwork for future research into SACs for various biomass refining processes.

It is necessary to ascertain the stability of SACs to avoid agglomeration during the reaction, which may adversely affect their activities. Earlier studies revealed that single atoms are generally stabilized by the lattice oxygen of the oxide supports or alkali metals.⁷⁴ In this study, all the investigated SACs are stabilized on the lattice oxygen of the pristine MgO. Subsequently, we investigated the stability of the SACs on a defective MgO, i.e., when a surface O is removed from the MgO (100) support, designated as MgO_Ovac. Prior to this, we obtained the surface oxygen vacancy formation energy to be 5.99 eV, which agrees with the previously calculated value.⁷⁵ On MgO_Ovac support, all the SACs were stabilized at the O vacant site. Table I shows the binding energies of the SACs on pristine (E_{b_p}) and the defective (E_{b_d}) MgO (100) support. On both surfaces, the values show high exothermic binding; however, the existence of O vacancy on the MgO surface further enhances the stability of the SACs.

Following this, we calculated the formation energies of dimer for the various SACs, i.e., going from M₁/MgO to M₂/MgO. The results from Table II demonstrated that pristine MgO support has a higher dimer formation capability than defective MgO support. Although the formation energy of the dimer is important in understanding the overall behavior of the catalyst systems, it does not necessarily indicate the reactivity of the catalyst in the reaction of interest. Hence, we will only conduct the present study on the pristine and defective MgO-supported SACs.

For both the SCCs and SACs, various sites on the surfaces and several adsorbate conformations were investigated to obtain the most stable binding conformations for each species involved in the xylitol C–C bond scission. Table III gives the adsorption energies of the most stable C₅, C₃, and C₂ conformations on the investigated surfaces based on the PBE functional, whereas Table S1 gives the comparison of xylitol (C₅) adsorption energies on SACs supported pristine MgO based on the PBE and PBE + D3 functionals.

From Table III, we noted that, overall, the binding strength of the reacting species increased from C₅ via C₃ to C₂. For the investigated surfaces, in addition to C₂ and C₃ binding energies on defected MgO-supported SACs (Pd₁ and Pt₁), the SCCs have the lowest affinity for the biomass species C₅, C₃, and C₂. Table S1 showed that incorporating the van der Waals forces (D3 model) into the PBE functional helped to further stabilize the xylitol on the surfaces. However, recent studies^64,65 have shown that this has minimal effect on the kinetic barriers. To that end, we used only the PBE functional for the rest of the studies. Refer to Table S1 for the comparison between the xylitol C–C bond scission barrier based on PBE and PBE + D3 functionals.

Unlike long-chain alcohol, shorter ones are beneficial as feedstock for hydrogen production. More so, the product selectivity of a metal catalyst for higher oxygenate reforming is determined by carbon–carbon bond-cleaving activity as opposed to carbon–oxygen bond-cleaving activity.⁷⁶ Furthermore, it has been demonstrated that C–C bond-cleaving reactions are rate limiting in the decomposition of various hydrocarbon^77–79 because they are thermodynamically unfavorable due to the relatively high stability of the C–C bond (∼4.0 eV).⁸⁰ In addition, via first-principle studies of ethanol dissociation on Pt (111), it has been shown that the rate constants for C–C bond scission are lower than that for ethane because of destabilization caused by the presence of oxygen in ethanol.⁸¹ In another development, experimental studies showed that even at temperatures near 500 K on Pt, C–C bond-scission rates are slower than dehydrogenation reactions for ethanol.⁸² On the other hand, Pt (331) surface has been reported to cleave the C–C bond in ethanol at low temperatures (200–300 K).⁸³ These findings suggest that alcohol reactions on transition metal catalysts, particularly Pt, are structure sensitive. From these earlier reports, we can conclude that alcohol's C–C bond scission is an energetically demanding process. As a result, designing a catalyst that can easily break the C–C bond of large molecular alcohol or biomass oxygenates would be advantageous to the biomass upgrading industries. Our proposed mechanism for the C–C bond scission of xylitol to EG and PG precursors begins with the adsorption of vapor phase xylitol and requires a one-step process (see Scheme ). In this section, we investigated and compared the activities of the parent SCS in conjunction with MgO-supported SACs. The energetics of this process will be discussed later.

First, we investigated the thermodynamic feasibility of xylitol direct C–C bond scission on various surfaces. As previously mentioned, our proposed mechanism for xylitol conversion to EG and PG precursors begins with the adsorption of vapor-phase xylitol followed by a one-step C₂–C₃ bond scission. Table IV shows the reaction energies for the C–C bond scission of xylitol on the investigated surfaces.

The results in Table IV showed that on Co (0001), Ni (111), and M₁/MgO_Ovac (M = Ni, Pd, and Pt), the direct C–C bond scission of xylitol is highly thermodynamically unfeasible. As a result, these surfaces would be eliminated for further studies. However, on the rest of the surfaces, the direct C–C bond scission of xylitol is either slightly unfeasible or highly feasible. To that end, the kinetic of xylitol’s direct C–C bond scission is investigated on these surfaces. Notwithstanding, because our objective is to compare the activities of SACs with those of their SCCs counterparts, Co (0001) and Ni (111) surfaces, which showed thermodynamic unfeasibility, were also included for the kinetic study. Figure 2 shows xylitol, TS intermediate’s, and C₂ and C₃ species on rhodium surfaces as the representation of the entire surfaces, while Table V shows the kinetic barrier for the direct C–C bond scission of xylitol on the investigated surfaces. In addition, Fig. 3 shows the relationship between the xylitol C–C bond scission barrier and xylitol adsorption energy on the investigated surfaces. We could see a similar trend on both the SCCs and M₁/MgO surfaces. Overall, a higher xylitol binding strength resulted in a lower kinetic barrier on both surfaces. On the contrary, oxygen vacancy on the MgO surface modified the stability and C–C bond scission of xylitol on M₁/MgO_Ovac, as revealed in Fig. 3(c).

Depending on the catalyst and reactions involved, increasing the temperature of a reaction alters the reactant’s kinetic energy, which ultimately affects the reaction rate. In the current study, to account for the temperature effect on the kinetics of xylitol C–C bond scission, the DFT energy barriers and reaction energies earlier generated at 0 K were converted to the operating temperature of 473 K and displayed on Gibb’s free energy profile diagrams, as shown in Fig. 4. In addition, Table VI gives the kinetic enthalpy, kinetic Gibb’s free energy, and thermodynamic reaction energies at the operating temperature of 473 K on all the investigated surfaces.

Recall that the C–C bond energy is ∼4.0 eV. In this study, values of xylitol’s C–C bond scission Gibb’s reaction barriers obtained for the various SCCs such as 2.70, 3.09, 2.70, 2.01, 2.24, 2.09, and 2.31 eV on Co (0001), Ir (111), Ni (111), Pd (111), Pt (111), Rh (111), and Ru (0001) are still too high, implying that xylitol’s C–C bond scission appears to be a near-impossible task on SCCs. Meanwhile, various first-principle studies on biomass oxygenate (particularly, ethanol and EG) reforming have demonstrated that our obtained results for xylitol C–C bond scission are not peculiar. For example, Faheem et al.⁸⁴ obtained a Gibbs’ free reaction barrier of 2.10 eV for the C–C bond scission of EG (CH₂OHCH₂OH) into its monomer (CH₂OH) on Pt (111) in the vapor phase. In the current study, we obtained the xylitol C–C bond-cleaving Gibbs’ free reaction barrier on Pt (111) surface to be 2.24 eV. Both results are quite close, even though our biomass oxygenate has more carbon and oxygen atoms than the EG used in their study. Moreover, they discovered that the C–C bond-cleaving barrier decreases as EG dehydrogenation progresses. In another study using Cu (100), Wu et al.⁸⁵ found the C–C bond-cleavage barrier for ethanol-dehydrogenated intermediates (CH₃CO, CH₂CO, and CHCO) to be 1.03, 1.02, and 1.17 eV, respectively. They also obtained improved results in an alkaline medium, although their values depend on the final product formed during the cleaving process. Xu et al.,⁸⁶ on the other hand, studied the C–C and C–H bond cleavage of ethanol oxidized intermediates on Cu₂O (111) surface using the first-principle technique with and without the inclusion of the Hubbard U parameters. They obtained values of 1.24, 3.06, and 3.43 eV for CH₃CO, CH₂CO, and CHCO when a U value was included, and 1.22, 1.97, and 2.73 eV when the U value was not included, respectively. However, contrary to the case of Faheem et al.,⁸⁴ for EG, the C–C bond-cleavage barrier for ethanol dehydrogenated intermediates obtained by Wu et al.⁸⁵ and Xu et al.⁸⁶ tends to increase as the dehydrogenation progresses. In the current study, even though the xylitol C–C bond scission on the investigated SCCs gave higher barriers (>2.0 eV), the values obtained from our designed catalysts demonstrated that SACs could significantly improve the C–C bond-cleaving reaction of biomass oxygenates. Here, Iridium SACs (Ir₁/MgO and Ir₁/MgO_Ovac) showed the highest decrease in Gibb’s free reaction barrier from 3.09 eV on its parent SCC [Ir(111)] to 0.73 and 1.41 eV on Ir₁/MgO and Ir₁/MgO_Ovac (see Table VI), representing ∼76% and 54% reduction. More so, a trade-off exists between the sintering tendency of our designed SACs (M₁/MgO and M₁/MgO_Ovac) and their activities. In this case, all the M₁/MgO SACs have higher activities than their M₁/MgO_Ovac counterparts, although they show a higher tendency to undergo sintering than their M₁/MgO_Ovac counterparts (see Table II). Notwithstanding, it is not unexpected to see prominent transition metal catalysts, such as the Ni-SCC and its SACs, underperform in the xylitol C–C bond scission reaction. Previous experimental reports have demonstrated that the Ni catalysts have low selectivity for C–C bond scission.²³ 

In this section, we discussed the differences in electronic structures between the SACs (M₁/MgO and M₁/MgO_Ovac) and their parent SCCs. Figure 5 shows the partial density of states (DOS) of the SACs and their parent SCCs for cobalt- and iridium-type catalysts. Refer to Fig. S2 in the supplementary for the DOS of nickel-, palladium-, platinum-, rhodium-, and ruthenium-type catalysts. Here, we can see that the DOS of all investigated M₁ @ M₁/MgO and M₁ @ M₁/MgO_Ovac SACs is different from their corresponding SCSs. Similarly, there is a slight difference between the DOS of M₁/MgO and M₁/MgO_Ovac, indicating that oxygen vacancy on the MgO support modifies the electronic structure of the SACs. The calculated d-band center (ε_d) is inserted in the plots. Meanwhile, it has been established from the d-band theory that the up-shift of the d-band center toward the Fermi level tends to increase binding strength between adsorbate and transition metal surface and vice versa.⁸⁷ In the current study, there is an up-shift of the DOS for the surface M₁ atoms in M₁/MgO and M₁/MgO_Ovac compared to the surface atoms of their respective SCCs, leading to M₁ higher d-band center, which confirms stronger binding of adsorbate(s) in contrast to the SCCs. However, the d-band model was not able to explain the reduction in activities of M₁/MgO_Ovac SACs compared to M₁/MgO in terms of adsorbate(s) binding strength. To explain these discrepancies, we correlated the binding energies of xylitol on M₁/MgO_Ovac and M₁/MgO SACs with the density of states near the Fermi level (DOS_FL), as shown in Fig. 6. Generally, a higher value of DOS_FL leads to stronger binding and vice versa. In this section, we compared the values of DOS_FL and xylitol's binding energy for each identical atom in M₁/MgO and M₁/MgO_Ovac SACs. Our findings indicated that, apart from Ir₁/MgO, all M₁/MgO SACs that had stronger binding with xylitol than their counterpart M₁/MgO_Ovac SACs had higher values of DOS_FL. Also, note that the presence of O vacancy on the MgO surface modifies the stability and activities of the SACs. Hence, the activity trends of the SACs on both supports (MgO and MgO_Ovac) differ.

In another development, we conducted the surface charge analysis by Bader approach⁸⁸ on the vacuum MgO and MgO_Ovac supported SACs (see Table VII). On both groups of supported SACs, charge accumulation was observed on the surfaces. However, there was higher charge accumulation on all MgO_Ovac-supported SACs than on the MgO-supported SACs. The extremely higher surface charges on M₁/MgO_Ovac compared to M₁/MgO SACs were attributed to the oxygen vacancy created on the MgO surface, which induced modifications in the stability and activities of the SACs.

In this study, we performed periodic DFT calculations to investigate the direct decomposition of xylitol (C₅) biomass into ethylene glycol (C₂) and propylene glycol (C₃) precursors on various catalyst surfaces. Our model catalysts include single-crystal (SC) slabs and single-atoms (SA) supported on MgO (without and with surface oxygen vacancy). First, we investigated the stability of the SACs by testing their sinter-ability on the pristine and O vacant MgO support surfaces. Our DFT calculated reaction energies for dimer formation revealed that pristine MgO support has a higher dimer formation capability than the defective MgO support.

For the xylitol C–C bond scission study, first, the most stable conformations of xylitol, EG, and PG precursors were obtained on the individual surfaces. Following this, we employed the climbing image nudged elastic band (CI-NEB) technique to search for the transition state, leading to the minimum energy path from which we estimated the kinetic barriers on the various surfaces. For the frequency calculations, it is important to note that TSs are critical intermediates in catalytic reactions, and the accuracy of their determination is crucial for understanding the reaction mechanism. We conducted a vibrational frequency analysis to verify all TSs in this paper. Thereafter, we converted the obtained DFT kinetic barriers (at 0 K) to enthalpies and Gibb’s free energies (at 473 K) to account for the experimental condition of the xylitol C–C bond scission reaction. Our results revealed that the overall binding strength of the reaction species increased from C₅ via C₃ to C₂ on all the investigated surfaces, with the single-crystal catalysts (SCCs) having the lowest affinity for these species. The enthalpies and Gibb’s free energies of activation obtained for all the investigated SCCs demonstrate that xylitol's C–C bond scission appears to be a near-impossible task on these surfaces. On the contrary, values obtained from our designed catalysts demonstrated that SACs could significantly improve xylitol's C–C bond-cleaving reaction. For example, at a practical operating temperature of 473 K, we obtained Gibb’s free energies of activation for xylitol direct C–C bond cleavage as 3.09, 0.73, and 1.41 eV on a conventional iridium SC surface [Ir (111)] and our designed SACs (Ir₁/MgO and Ir₁/MgO_Ovac), respectively, which represent ∼76% and 54% reduction. More so, we noted that a trade-off existed between the sintering tendency of our designed SACs (M₁/MgO and M₁/MgO_Ovac) and their activities. Here, while all the M₁/MgO SACs have higher activity than their M₁/MgO_Ovac counterparts, all M₁/MgO_Ovac SACs have a lower tendency to undergo sintering than M₁/MgO SACs. In addition, electronic structure calculations revealed an up-shift in the DOS for the surface M₁ atoms in all M₁/MgO and M₁/MgO_Ovac supported SACs compared to the surface atoms of their respective SCCs, resulting in M₁ higher d-band center and stronger adsorbate(s) binding. Furthermore, as expected, there was extremely higher charge accumulation on all SACs on the MgO_Ovac than on the MgO support, which was attributed to the oxygen vacancy created on the MgO surface. These extremely higher charges induced modifications in the stability and activities of the SACs. This study highlights the importance of SACs for boosting the atom efficiency of costly metals while also offering a new strategy for tuning the activity of catalytic reactions.

See the supplementary material for xylitol adsorption conformations on M1/MgO surfaces, representation of the projected density of states (DOS) of the d-band surface atom(s) of the investigated SCSs and SACs, and comparison of the adsorption energies of C5 and its C2–C3 scission barriers on the studied M1/MgO based on the PBE and PBE-D3 functionals, respectively.

This work was supported by the Hydrogen Energy Innovation Technology Development Program of the National Research Foundation of Korea (NRF) funded by the Korean government [Ministry of Science and ICT(MSIT)] (Grant No. NRF-2020M3E6A1043955); the New & Renewable Energy Core Technology Program of the Korea Institute of Energy Technology Evaluation and Planning (KETEP), granted financial resource from the Ministry of Trade, Industry & Energy, Republic of Korea (Grant No. 20213030040080); the INHA UNIVERSITY Research Grant; and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (Grant No. NRF-2020R1A2C1099711).

The authors have no conflicts to disclose.

Shedrack G. Akpe: Conceptualization (lead); Data curation (lead); Formal analysis (lead); Investigation (lead); Methodology (lead); Visualization (lead); Writing – original draft (lead). Sun Hee Choi: Investigation (supporting); Resources (supporting); Supervision (supporting). Hyung Chul Ham: Conceptualization (lead); Funding acquisition (lead); Investigation (supporting); Resources (lead); Supervision (equal); Writing – review & editing (lead).

The data that support the findings of this study are available from the corresponding author upon reasonable request.