WorksPhysics and Astronomy
Understanding the Oxygen Reduction Reaction and Oxygen Evolution Reaction in Metal Intercalated Biphenylene Bilayers
Abstract
We conducted an ab initio study of the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) in metal-encapsulated biphenylene bilayers, B/M/B, with M = Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Nb, Ru, W, Os and Pt. In most systems, the intercalated metal sits at the square carbon sites (C) of the biphenylene lattice. Using a computational hydrogen electrode approach, we evaluated the reaction energetics at these active sites. Several B/M/B systems show competitive ORR and OER performance. Among the investigated systems, Cu, Pt, Ru, and Mn exhibit the lowest ORR overpotentials of 0.42, 0.44, 0.50, and 0.56 V, respectively, while Fe is identified as the most active catalyst for OER with an overpotential of 0.44 V. To understand the catalytic trends, we looked at the electronic structure through the metal band centers, the C band centers, and the corresponding orbital charge populations. The band centers did not give a simple polynomial dependence on the overpotentials, though they did point to favorable electronic ranges for the best catalysts. The orbital charge population of the encapsulated metal, however, correlated most clearly with activity-especially for OER-yielding volcano-type plots. From these, B/Fe/B emerges as the best OER catalyst, while B/Mn/B lies closest to the ORR optimum. The orbital population at the active carbon site also captures the main trends, albeit less strongly. Overall, these results show that straightforward electronic descriptors can predict catalytic behavior in metal-encapsulated biphenylene bilayers and guide the search for efficient catalysts where the carbon framework itself drives the reactivity.
main8_new.pdf
Understanding the Oxygen Reduction Reaction and Oxygen Evolution Reaction in Metal Intercalated Biphenylene Bilayers
Abstract
Introduction
In recent years, two-dimensional (2D) catalysts have attracted significant attention as key components for sustainable development within a green economy[1]. Their applications span fuel cells[2], zinc-air batteries[3], and water splitting[4]. In particular, single-atom catalysts (SACs) have demonstrated remarkable performance in the hydrogen evolution reaction (HER)[5, 6], oxygen reduction reaction (ORR)[7, 8, ?, ?], and oxygen evolution reaction (OER)[3, 6, ?]. This performance originates from the ability to precisely tailor the electronic structure of isolated active sites, thereby enhancing catalytic activity[?, ?, ?, ?].
Despite these advantages over commercial catalysts, SACs often face limitations in operational stability[?, ?]. Under electrocatalytic conditions, changes in oxidation state can promote dissolution of the active site into the electrolyte[?, ?, ?], resulting in progressive loss of catalytic efficiency. To mitigate these effects, several strategies have been proposed, including the deliberate introduction of vacancies to stabilize favorable oxidation states[?], engineering of the coordination environment[?, ?], and intercalation within 2D materials[?]. Among these, intercalation stands out as a particularly effective approach, as encapsulation shields the active sites from environmental degradation while preserving both catalytic activity and long-term stability[?].
The feasibility of such architectures is supported by experimental realization of layered materials with intercalated transition metals (TM), including graphene[?, ?, ?], phosphorene[?], and MoS[?, ?], as well as metal nanoparticles encapsulated within carbon nanotubes[?, ?]. These systems demonstrate that encapsulation effectively suppresses corrosion[?, ?, ?, ?, ?], thereby extending catalyst lifetime relative to conventional SACs. Beyond enhanced stability, intercalated structures, such as borophene–TM–borophene heterostructures[?] and TM dichalcogenides encapsulated within h-BN layers[?], exhibit tunable electronic properties[?, ?, ?, ?], enabling rational optimization of reduction and evolution reactions[?, ?, ?]. The electronic properties can be tuned such that catalytic activity is enhanced[?, ?, ?, ?]. In intercalated systems, electronic tuning generally occurs through charge transfer[?, ?], modifications in the work function of an intercalated material[?, ?], and hybridization between - and -orbitals[?, ?, ?]. The latter strategy offers the greatest potential for tuning catalytic activity due to the large number of elements with -orbitals in their valence shell[?, ?].
The biphenylene lattice crystallizes in the space group and remains dynamically and thermally stable up to 4000 K[?]. Its electronic structure is metallic, with pronounced electron localization at the square-planar carbon sites[?, ?, ?]. This distinctive combination of properties has spurred extensive theoretical and experimental efforts exploring biphenylene across a broad range of applications, including charge transport[?], Li and LiO battery
@bibliography electrodes[?, ?], and hydrogen storage[?]. Beyond these applications, biphenylene stands out for its exceptional catalytic performance, particularly in electrocatalytic processes such as the hydrogen evolution reaction (HER)[4, ?, ?, ?, ?], CO reduction reactions[?, ?], nitrogen reduction reaction (NRR)[?], oxygen reduction reaction (ORR)[?, ?, ?], and oxygen evolution reaction (OER)[4, ?].
In this work, we investigate the intercalation of transition metals in AA-stacked biphenylene bilayers (B/M/B, with M = Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Nb, Ru, W, Os, and Pt) as a strategy to develop ORR/OER electrocatalysts in which the metallic centers are not directly exposed to the electrolyte. We propose that intercalating metal atoms between biphenylene layers enhances resistance to corrosion while promoting catalytic activity at the surface carbon atoms. We begin by examining the energetic and structural stability of the B/M/B systems, as discussed in Sec. III A. In Sec. III B, we evaluate their catalytic activity, identify B/M/B with M = Mn, Fe, Co, Ni, Cu, and Pt are particularly promising candidates, and provide a detailed analysis of the electronic properties underlying the superior performance, relative to other B/M/B systems, introducing descriptors to guide the rational design of high-performance catalysts.
Computational Details
The calculations were performed using spin-polarized density functional theory (DFT) as implemented in the Vienna Ab initio Simulation Package (VASP)[?]. Structural optimization and electronic structure calculations were carried out using the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional[?, ?]. Dispersion interactions were accounted for using the vdW-DF2 functional[?]. The projector augmented-wave (PAW) method was employed to describe ion-electron interactions[?]. Self-interaction effects associated with the orbitals of transition metals were treated using the Hubbard correction. The parameters were determined for each transition metal via the linear response approach[?], and the corresponding values are listed in tbl:T1 of the Supplemental Materials (SM)[?]. A plane-wave energy cutoff of 500 eV was adopted based on previous studies[?, ?]. The Brillouin zone was sampled using a Monkhorst-Pack -point mesh[?], and ionic relaxations were performed with a force convergence criterion of 0.01 eV Å.
The biphenylene bilayer was modeled using a surface unit cell in the AA stacking configuration, in which the interlayer region is intercalated with transition metal atoms, as shown in Figure 1. The supercell has lattice parameters Å and Å, with a vacuum spacing of 15 Å along the out-of-plane direction. It contains 192 carbon atoms and 16 transition metal atoms. For the evaluation of electronic properties, a reduced unit cell containing 48 carbon atoms and 4 metal atoms was employed. This choice represents the smallest system that accurately reproduces the magnetic phases under investigation. Further structural stability was examined by ab initio molecular dynamics (AIMD) simulations, performed in the NVT ensemble using a Nosé-Hoover thermostat approach at 400 K, with a total simulation time of 10 ps and a time step of 0.4 fs.

Figure 1. (a) Top view of the equilibrium geometry of pristine biphenylene bilayer (B/B); h, s, and o represent intercalation sites for the metal atoms. C indicates the carbon atom connecting the square, hexagonal, and octagonal sublattices, and C indicates the carbon atom connecting hexagonal and octagonal sublattices; (b.1) B/Fe/B with Fe at the square-planar site, while (b.2) B/V/B at the hexagonal site; (c.1) shows the spin-density of B/Fe/B at the top view; (c.2) also shows the spin-density of B/Fe/B in the respective view. The yellow and blue densities represents the spin-up and spin-down densities, respectively. The magnetic configuration presented for B/Fe/B is AFx; (d) shows the charge density of B/Fe/B with O adsorbed on C. The isosurfaces of spin-density and charge density used was 0.01 e/Å and 0.006 e/Å, respectively.
We calculated the C-1s core-level binding energies (BEs) and the corresponding core-level shifts (CLSs) for the B/M/B systems using the VASP package[?]. The calculations were performed within the self-consistent field (SCF) approach[?]. Within this framework, the C-1s core-level binding energy is defined as
where and are the total energies of the ionized system containing a C-1s core hole and the ground-state system with electrons, respectively. The binding energies were computed for all crystallographically nonequivalent carbon atoms in the B/M/B structures. As a reference, we also calculated the C-1s binding energies of pristine single-layer and AA-stacked biphenylene, denoted by . The corresponding core-level shifts were then obtained as
To facilitate comparison with experiment, the calculated CLSs can be referenced to the measured C-1s binding energy of pristine biphenylene, eV[?]. Accordingly, the calculated absolute binding energies are given by
With these values now comparable to experimental results, we also simulate the XPS spectra using the pseudo-Voigt approach[?], capturing the peaks corresponding to non-equivalent carbons in B/M/B. The parameters used in the convolution profiles are given by as Gaussian, as the Lorentzian width, and as the mixing parameter of the two distributions[?].
The thermodynamic overpotentials for ORR and OER were evaluated within the computational hydrogen electrode (CHE) model proposed by Nørskov and co-workers[?]. In this approach, catalytic activity is described in terms of the adsorption strength of key intermediates (OH*, O*, and OOH*), quantified through their adsorption free energies on the active site, expressed as . Here, and correspond to zero-point energy and entropic contributions, respectively, obtained from vibrational frequency calculations of the adsorbed species. Only the vibrational degrees of freedom of the intermediates were considered, while the catalyst surface was kept fixed. Thermodynamic and vibrational properties of gas-phase reference molecules were taken from the NIST database[?].
Solvation effects on adsorbate binding strength (OH* and OOH*), were included by accounting for hydrogen-bonding interactions. The corresponding corrections were computed for each structure using the continuum dielectric model implemented in VASPsol[?, ?, ?]. Post-processing and visualization were carried out using VASPKIT[?] and VESTA[?]. The biphenylene bilayer (B/B) has been theoretically predicted in AA[?], AB, and AX stacking configurations[?], all of which exhibit thermomechanical stability.
Results
Energetic and Structural Properties
We begin determining the ground-state configurations of transition metals embedded in biphenylene bilayers by placing the metal atoms at the square (s), hexagonal (h), and octagonal (o) interlayer sites fig:1a. Owing to the partially filled and shells of most transition metals (M), four magnetic configurations () are considered at each site: ferromagnetic (FM), antiferromagnetic (AF), and stripe antiferromagnetic states AF-x and AF-y. The latter are defined by antiferromagnetic (ferromagnetic) coupling along the () direction for AF-x, and along the () direction for AF-y.
The structural stability of the metal-intercalated biphenylene bilayers (B/M/B) is assessed via the intercalation energy , defined as
where and are the total energies of the intercalated system (for ) and the pristine bilayer, respectively. For each geometry, the lowest-energy magnetic phase is considered. Here, is the energy of an isolated metal atom and is the number of intercalates. Negative indicates exothermic intercalation. We obtain that the most of intercalated metals preferentially occupy the square sites, B/M/B fig:1b.1, whereas Nb, V, and Ti favor hexagonal sites, B/M/B fig:1b.2. The corresponding intercalation energies and ground-state magnetic phases () are summarized in Table I. Additional data for other geometries are provided in the SM in Tables T2–T5[?].
| M | |||||
| Cr | AF- | ||||
| Mn | AF | ||||
| Fe | AF- | ||||
| Co | AF | ||||
| Ni | AF | ||||
| Cu | FM | ||||
| Ru | AF- | ||||
| W | AF | ||||
| Os | FM | ||||
| Pt | FM | ||||
| M | |||||
| Ti | AF | ||||
| V | FM |
Table I. Intercalation energy (in eV/M-atom) of the metal atoms between the square (), octagonal (), and hexagonal () sites of the biphenylene bilayer, and the (respective) magnetic phase (). The binding energies (in eV) of () adsorbed on are listed in the rightmost column.
The intercalated metal atoms form chemical bonds with the carbon atoms of the biphenylene bilayer, where the coordination environment depends on the intercalation site. At the square site, the intercalated metal atom () exhibits fourfold coordination with the atoms. In contrast, at the hexagonal site (), the intercalated metal atom forms two bonds with the atoms and four bonds with the atoms. Bader charge analysis[?] further reveals that the net charge transfer from the intercalated metal to the biphenylene bilayer also depends on the intercalation site. In particular, the amount of charge transferred to the and atoms differs between the and configurations. The corresponding equilibrium structural parameters and Bader charges are summarized in Table T5 and T6 of the Supplemental Material.
Next, we examine the dependence of the C-1s core-level binding energies on the metal intercalation site. As a reference, we first calculate the C-1s binding energies of pristine biphenylene (3). The simulated XPS spectrum, shown as the shaded region in Figure 2, exhibits two distinct binding-energy () peaks. The lower- peak is associated with the fourfold-coordinated atoms (283.90 eV), whereas the higher- peak originates from the sixfold-coordinated atoms (284.42 eV), resulting in a binding-energy difference of eV.
Figure 2. Simulated XPS spectra for B/M/B with M = Mn, Fe, Co, Ni, Cu, and Pt, for metals intercalated in the s configuration (black curve) and h configuration (red curve). The region in gray indicates the experimental peak position[?] and the convolution XPS spectra simulation for the biphenylene monolayer.
Metal intercalation systematically shifts the C-1s binding energies toward lower values relative to pristine biphenylene, resulting in negative core-level shifts (CLSs). This behavior is consistent with the net charge transfer from the intercalated metal atoms () to the biphenylene bilayer, as revealed by the Bader charge
analysis. The calculated CLSs of the atoms () for the B//B and B//B structures, together with the corresponding binding-energy separations, , are summarized in Table II. Representative simulated C-1s XPS spectra for B//B systems with = Mn, Fe, Co, Ni, Cu, and Pt are presented in Figure 2(a)–(f). The spectra reveal that both the magnitude of the CLSs and the binding-energy separation between the C and C atoms depend on the chemical identity of the intercalated metal and its coordination site within the biphenylene bilayer, being systematically larger in B/M/B. These findings demonstrate that the B/M/B systems exhibit distinct XPS fingerprints and provide valuable guidance for future experimental spectroscopic characterization.
| B/M/B | B/M/B | |||
| CLS | (C − C) | CLS | (C − C) | |
| Cr | −0.63 | −0.19 | −0.77 | −0.52 |
| Mn | −0.65 | −0.24 | −0.74 | −0.58 |
| Fe | −0.28 | −0.32 | −0.31 | −0.48 |
| Co | −0.28 | −0.47 | −0.68 | −0.62 |
| Ni | −0.31 | −0.05 | −0.55 | −0.67 |
| Cu | −0.42 | −0.16 | −0.56 | −0.55 |
| Ru | −0.43 | −0.07 | −0.64 | −0.41 |
| W | −0.59 | −0.03 | −0.63 | −0.64 |
| Os | −0.91 | +0.04 | −0.68 | −0.26 |
| Pt | −0.25 | +0.36 | −0.72 | −0.37 |
| Ti | −0.80 | −0.40 | −0.63 | −0.63 |
| V | −0.64 | −0.28 | −0.63 | −0.60 |
Table II. TABLE II. Calculated core-level shifts of C with respect to the pristine biphenylene bilayer, CLS, and the C-1s binding energy difference between the C and C carbon atoms of B/M/B with the metal intercalated at the square (M) and hexagonal (M) sites, (C − C). The energies are in eV.
ORR and OER Activity of B/M/B
Single-atom catalysts (SACs) have demonstrated remarkable catalytic activities for both the ORR and OER, often surpassing those of commercially available benchmark catalysts[6, ?]. However, their practical application is frequently limited by insufficient structural stability under operating conditions, where migration, aggregation, or detachment of the metal atom from the supporting material can lead to the progressive degradation of the catalytic performance[2, ?, ?, ?, ?, ?].
In contrast, the metal atoms in B/M/B are fully encapsulated between the two biphenylene layers, providing an effective confinement that hinders metal migration and detachment and is therefore expected to enhance the long-term structural stability of the catalyst. Moreover, charge analysis reveals electron transfer from the intercalated metal atoms to the biphenylene framework, leading to a more electron-rich carbon network[?]. Such charge redistribution is expected to modify the C-2p electronic states and, consequently, tune the adsorption strength of ORR/OER intermediates, thereby enhancing the catalytic activity[?].
Reaction Free Energies of ORR/OER intermediates
We begin by examining the O adsorption on the biphenylene bilayer encapsulated by transition metals,
The carbon atoms forming the square sublattice [C in Figure 1(a)] are identified as the active sites because they exhibit stronger O adsorption than the C atoms and possess a higher local electron density (see Figs. S3 and S4 of the Supplemental Material[?]), favoring the adsorption of ORR/OER reaction intermediates on the biphenylene framework[?]. For example, in B/Fe/B, the calculated O binding energies are −0.57 and −0.06 eV for adsorption at the C and C sites, respectively. The optimized adsorption geometry of O on the C site, together with the corresponding charge transfer upon adsorption, is shown in Figure 1(d).
The calculated O2 binding energies () at the C active sites for all B/M/B systems are summarized in Table I. With the exception of B/Nb/B and B/W/B[?] all systems exhibit negative O binding energies, indicating that O adsorption on the C active sites is thermodynamically favorable.
Focusing on the ORR and OER processes, in an alkaline solution, the former can be described as the following steps,
Likewise, in an alkaline solution, the OER steps proceed in a reverse order.
We evaluate the ORR and OER activities of all B/M/B systems using the computational hydrogen electrode (CHE) model developed by Nørskov et al.[?]. Within this framework, the adsorption free energies of the key reaction intermediates, namely OOH*, O*, and OH*, are calculated following Ref. 13 as
where is the adsorption energy obtained from density functional theory, is the zero-point energy correction, is the entropic contribution derived from vibrational frequency calculations of the adsorbed species (with the catalyst surface kept fixed), and is the solvation correction. Further details of the free-energy calculations and the ORR/OER mechanisms are provided in Sec. III of the Supplemental Material[?]. Through the Gibbs free-energy profiles of the elementary ORR and OER reaction steps, we determine the onset potentials, i.e., the minimum (maximum) electrode potential at which all ORR (OER) elementary steps become thermodynamically downhill with respect to the Gibbs free energy, [?]. The onset potentials for ORR and OER, denoted by and , respectively, are used to calculate the corresponding overpotentials (),
where V is the standard electrode equilibrium potential.
The Gibbs free-energy diagrams for the ORR and OER on the B/M/B systems with M = Mn, Fe, Co, Ni, Cu, and Pt (Figure 3 and 4) show that (i) at zero electrode potential relative to the reversible hydrogen electrode (), the Gibbs free-energy profiles exhibit a cascade-like behavior, characteristic of efficient ORR/OER catalysts [?, ?], whereas (ii) at the equilibrium potential (), the rate-determining step depends on the intercalated metal. Specifically, the potential-limiting step is O* OH* for Mn and Cu, OH* HO for Fe, and O OOH* for Co, Ni, and Pt. The Gibbs free-energy diagrams for the remaining B/M/B systems are provided in fig:S5–S10 of the Supplemental Material (SM)[?].

Figure 3. Reaction free energy diagrams for the elementary steps of the ORR on the B/M/B, for M = Mn, Fe, Co, Ni, Cu and Pt, respectively in panels (a)-(f), at different applied potentials: V (black), V (blue), and the onset potential (red). All energies are referenced to standard conditions.
The calculated thermodynamic onset (limiting) potentials and corresponding overpotentials for the ORR () and OER (), summarized in Table III, reveal that the B/M/B systems with M = Mn, Fe, Co, Ni, Cu, Ru, and Pt exhibit ORR catalytic activities comparable to that of the theoretical Pt(111) benchmark. Notably, B/Cu/B and B/Pt/B outperform Pt(111), exhibiting values of 0.42 and 0.44 V, respectively.
| M | ORR | ORR | OER | OER |
| Ti | ||||
| V | ||||
| Cr | ||||
| Mn | ||||
| Fe | ||||
| Co | ||||
| Ni | ||||
| Cu | ||||
| Ru | ||||
| Os | ||||
| Pt | ||||
| Pt(111) | - | - | ||
| IrO(110) | - | - |
Table III. Calculated ORR and OER onset potentials (, in V) and overpotentials (, in V) in B/M/B for all M with . The benchmark Pt(111)[?] and IrO[?] are included for comparison.
values of 0.42 and 0.44 V, respectively. Likewise, the OER activities of the B/M/B systems with M = Fe, Co, Ni, and Pt are comparable to that of the theoretical IrO benchmark. Among them, B/Fe/B stands out with an overpotential of V, which is substantially lower than that of IrO. The Gibbs free-energy profiles at the equilibrium potential ( V, blue curves) and at the corresponding onset potential (, red curves) are also shown for both reactions in Figs. 3 and 4. These results indicate that, beyond protecting against structural degradation through the shielding effect of the bilayer architecture[?], M-intercalated biphenylene bilayers (B/M/B) constitute a promising class of electrocatalysts for both the ORR and OER[3, ?].
To further assess the catalytic performance of the B/M/B systems, we compare our calculated overpotentials with predictions based on established adsorption free-energy descriptors. Figure 5(a) presents the ORR overpotential as a function of , whereas Figure 5(b) shows the OER overpotential as a function of , the universal activity descriptor for the OER[?, ?]. The resulting volcano plots indicate that the minimum overpotentials are obtained for eV in the ORR and eV in the OER. These optimal descriptor values are consistent with previous theoretical studies[?, ?].

Figure 5. Volcano plots of the overpotentials for ORR and OER as a function of their respective universal descriptors, where panel (a) shows against , while panel (b) shows against .
Regarding structural stability, the calculated intercalation energies, , indicate that the insertion of metal atoms between the square lattices of biphenylene is an exothermic process. The intercalation energies range from to eV per metal atom for B/Cr/B and

Figure 4. Reaction free energy diagrams for the elementary steps of the OER on the B/M/B, for M = Mn, Fe, Co, Ni, Cu and Pt, respectively in panels (a)-(f), at different applied potentials: V (black), V (blue), and the onset potential (red). All energies are referenced to standard conditions.
B/Pt/B, respectively. Focusing on the B/M/B systems identified as promising ORR/OER electrocatalysts, the calculated intercalation energies range from eV per metal atom for B/Cu/B to eV per metal atom for B/Ni/B.
The thermal stability of the B/M/B systems was further assessed by ab initio molecular dynamics (AIMD) simulations performed at 400 K for 10 ps. This temperature was chosen to approximate the operating conditions of proton exchange membrane fuel cells, which typically operate in the 350–400 K temperature range
and involve both ORR and OER processes[?]. As shown in Figure 6, the intercalated metal atoms (M = Cu, Fe, and Pt) remain stably confined within the biphenylene bilayer throughout the AIMD simulations, with no evidence of metal diffusion or clustering, confirming the structural stability of the B/M/B systems. The corresponding AIMD results for the other B/M/B structures, presented in fig:S12 of the Supplemental Material[?], exhibit similar behavior, demonstrating that the metal atoms remain mostly at the squared site between the biphenylene layers during the entire simulation.

Figure 6. Energy variation as a function of time of the AIMD simulation for the B/Cu/B (panel (a)), B/Fe/B (panel (b)) and B/Pt/B (panel (c)) systems at 400 K. Insets show ball-and-stick representations of the atomic configurations at the initial stage (left) and after 10 ps (right), and time steps of 0.4 fs.
Additionally, we performed an AIMD simulation for the B/Fe/B system in the presence of water at its experimental density (). After water stabilization, the simulation was carried out at 400 K for 5 ps using a time step of 1 fs. As shown in fig:S13 of the Supplemental Material[?], no reactions between the water molecules and the biphenylene surface were observed throughout the simulation. These results indicate that the biphenylene bilayer effectively acts as a protective barrier, shielding the intercalated metal centers from direct interaction with the aqueous environment where ORR/OER occur.
Electronic Properties and catalytic-activity descriptors
Next, we analyze the electronic structure of the B/M/B systems to establish correlations between their catalytic activity and their electronic properties. We show in detail the band structures, density of states (DOS), and projected density of states (PDOS) for all B/M/B systems in Section IV in SM[?].
We first consider the relative positions of the orbitals of M and the orbitals of C. These positions are quantified by the − and −band centers, respectively, as defined in Eqs. (13) and (14). The −band center[?, ?] is given by
where is the spin-dependent PDOS projected onto the M- orbitals. The Kohn-Sham eigenvalues () are chosen in relation to the Fermi level. This relation can be used as a descriptor for ORR[?] and OER[?], where the range of should be , in relation to the Fermi level, to minimize the overpotentials in Metal-Organic Framework systems as verified in Ref.[?].
For metal-free catalysts[?, ?, ?], the energy position of the −orbitals has also been used as a descriptor for ORR and OER overpotentials[?, ?, ?, ?]. In this context, the energy position of the band center (), defined as
Here, is the PDOS projected onto the C-2 orbitals for each spin channel. Therefore, we calculated and for all B/M/B systems and summarized these results in tbl:IV.
Neither nor exhibits a clear linear dependence on the − or −band center. The B/M/B systems span a narrow range of eV, indicating that the C −orbitals remain largely localized despite changes in the intercalated metal. In contrast, the −band center exhibits a more discernible relation with catalytic activity. The M = Mn, Fe, Co, Ni, Ru, and Pt systems fall within the optimal range reported for oxygen electrocatalysis, eV[?]. B/Cu/B is a notable exception, with eV while exhibiting ORR and OER overpotentials comparable to those of Pt(111)[?] and IrO[?], respectively.
Although the −band center is a useful descriptor for transition-metal active sites[3, ?], its applicability to B/M/B systems is less direct because the active site is C. We therefore examine the relative positions of the − and −band centers, as variations in their relative energy have been shown to correlate with enhanced catalytic activity[?, ?, ?]. Figure 7 shows and as functions of in panels (a) and (b), respectively.

Figure 7. ORR and OER overpotentials, respectively in panels (a) and (b), as a function of the difference between the −band center and the −band center for each B/M/B with eV and eV are indicated by blue squares. The −band center and the −band center are calculated relative to the Fermi level.
For both reactions, the overpotentials exhibit a piecewise-linear dependence on , resulting in a volcano-like relationship consistent with the Sabatier principle[?]. For ORR, shown in panel (a), B/Cu/B and B/Pt/B were excluded from the linear fits because they deviate markedly from the trend followed by the remaining B/M/B systems. For OER [panel (b)], the M = Cr, Mn, and Ru systems were similarly excluded. After these exclusions, the corresponding values[?] for the left and right branches are 0.95 and 0.45 for ORR and 0.95 and 0.83 for OER, respectively. The resulting volcano peaks occur at B/Ru/B for ORR, with eV, and at B/Fe/B for OER, with eV. These results suggest that the relative position of the − and −band centers may
serve as a descriptor of catalytic activity in B/M/B systems.
Considering the volcano-like dependence on shown in Figure 7, we next examine how this electronic descriptor is related to the orbital charge populations. As shown in Figs. 8(a) and 8(b), the charge populations of the metal −orbitals and the C −orbitals exhibit a nearly linear dependence on . The charge populations were obtained by projecting the Kohn-Sham wave functions onto atom-centered orbitals[?]. This linear relationship establishes a direct connection between the relative alignment of the metal states and the C states and their corresponding orbital occupations, thereby motivating the use of orbital charge populations as alternative electronic descriptors for catalytic activity.
Notably, in Figure 8, both the ORR and OER overpotentials exhibit volcano-like dependences on the charge populations of the orbitals [panels (c) and (d)] and orbitals [panels (e) and (f)]. Table IV summarizes the charge populations of the −orbitals of the intercalated metals and the −orbitals of C. For the −orbital charge population, the ORR volcano peak occurs at , with and 0.10 for the left and right branches, respectively. For OER, the peak is located at , with and 0.82 for the left and right branches, respectively.
| M | Band center | Charge population | |||
| −orbital | −orbital | ||||
| Ti | −2.89 | +0.32 | −3.21 | 1.760 | 2.581 |
| V | −2.85 | −0.45 | −2.40 | 1.758 | 3.534 |
| Cr | −2.96 | −1.66 | −1.30 | 1.752 | 4.498 |
| Mn | −2.63 | −2.35 | −0.28 | 1.752 | 5.062 |
| Fe | −2.95 | −1.76 | −1.18 | 1.753 | 6.571 |
| Co | −2.73 | −2.75 | −0.29 | 1.748 | 7.457 |
| Ni | −2.68 | −2.78 | +0.09 | 1.744 | 8.835 |
| Cu | −2.75 | −4.41 | +1.66 | 1.740 | 9.415 |
| Nb | −2.90 | +0.09 | −2.99 | 1.752 | 5.062 |
| Ru | −2.70 | −2.03 | −0.67 | 1.747 | 6.579 |
| W | −3.32 | −1.12 | −2.19 | 1.758 | 4.174 |
| Os | −3.45 | −2.08 | −1.37 | 1.753 | 6.300 |
| Pt | −2.53 | −2.91 | +0.39 | 1.734 | 8.080 |
Table IV. Calculated − and −band centers for each B/M/B systems, in relation to the Fermi level, the difference between − and −band centers (both in eV), and charge population of −orbital from intercalated metal and −orbital from C (in ) for all B/M/B.

Figure 8. Charge populations of the and orbitals as a function of are shown in panels (a) and (b), respectively. The B/M/B systems are the same as those considered in Figs. 6 and 7. Each system is represented by a blue square, while the red lines denote linear fits to the corresponding charge populations. The ORR and OER overpotentials as a function of the -orbital charge population of the intercalated metal are presented in panels (c) and (d), respectively. B/M/B systems with and are highlighted by blue squares. The corresponding ORR and OER overpotentials as a function of the -orbital charge population are shown in panels (e) and (f), respectively, with the same systems highlighted. The red lines denote linear fits of and as functions of the orbital charge population, revealing volcano-like relationships for both reactions.
In panel (c), B/Mn/B lies closest to the volcano summit. Although B/Cu/B exhibits a lower overpotential than B/Mn/B, it deviates from the volcano-like trend and was therefore excluded from the corresponding linear fits. In panel (d), B/Fe/B lies at the volcano summit, corresponding to the most favorable catalytic performance for OER. The B/Ru/B system was excluded from the linear fits because its −orbital charge population is close to that of B/Fe/B but exhibits a significant deviation from the corresponding trend.
Similarly, in panel (e), the ORR volcano summit occurs at a −orbital charge population of , with B/Mn/B located at the optimum. For OER [panel (f)], the summit occurs at , with B/Fe/B at the optimum. To maximize the values of the linear fits, three systems were excluded in each case. For ORR, B/Cr/B, B/Cu/B, and B/Pt/B were excluded, yielding and 0.15 for the left and right branches, respectively. For OER, B/Cr/B, B/Mn/B, and B/Ru/B were excluded, resulting in and 0.75 for the left and right branches, respectively. The origin of these deviations from the trends observed for the remaining B/M/B systems warrants further investigation.
These results establish a direct relationship between the relative positions of the − and −band centers and the charge populations of the corresponding orbitals. They further suggest that the − and −orbital charge populations may serve as qualitative descriptors of catalytic activity in B/M/B systems, as both exhibit volcano-like relationships with the ORR and OER overpotentials, consistent with the Sabatier principle.
Conclusions
We performed an ab initio investigation of the catalytic activity (ORR and OER) in metal-encapsulated biphenylene bilayers, B/M/B, with M = Os, W, Pt, Fe, Ni, Co, Ru, Cu, Cr, Mn, Nb, V, and Ti. Except for the last three elements, we find that metal intercalation occurs between the carbon atoms forming the square sites in biphenylene, C.
The evaluation of the ORR and OER processes at the surface C sites, based on the theoretical framework developed by Nørskov et al.[?], reveals that the M-intercalated biphenylene bilayer (B/M/B) exhibits catalytic performance competitive for ORR and OER. The catalytic activity was quantified through overpotentials (), and for B/M/B systems, the follows the usual linear relationships with for ORR and for OER.
The analysis of the electronic structure of the B/M/B systems showed that the relative position of the -band center with respect to the -band center exhibits a tendency to describe the Sabatier principle, particularly for OER. The occurrence of a volcano-plot-like behavior of the overpotentials and indicates an intrinsic correlation between the catalytic activity and the electronic structure, modulated by the chemical environment, of the B/M/B systems. Additionally, we show that is directly related to the charge populations of the and -orbitals.
Furthermore, the metal -orbital charge population exhibits a more robust correlation with the overpotentials, particularly for OER, enabling the construction of volcano-like relationships. These trends identify B/Fe/B as the optimal system for OER and B/Mn/B as the system closest to the ORR optimum, further supporting the use of simple electronic descriptors to rationalize catalytic activity in B/M/B systems.
The -orbital charge population of the active C site also exhibits volcano-like relationships with the overpotentials, yielding the same optimal systems for ORR and OER. However, compared with the metal -orbital charge population, the -orbital charge exhibits weaker correlations, requiring the exclusion of more B/M/B systems and yielding lower values. Thus, while the active-site -orbital charge captures general trends in catalytic activity, the metal -orbital charge population provides a more robust electronic descriptor for rationalizing and predicting the catalytic performance of B/M/B systems.
These findings suggest that such electronic features constitute key design principles for nonmetallic catalysts mediated by the encapsulation of metallic elements.
Acknowledgments
The authors acknowledge the Laboratório Nacional de Computação Científica (LNCC-SCAFMat2), Centro Nacional de Processamento de Alto Desempenho (CENAPAD-SP) for computer time. HGM and RHM acknowledge financial support from CNPq, INCT-Nanocarbono, INCT-Materials Informatics, Coordenação de Aperfeiçoamento de Pessoal de Ensino Superior (CAPES). WO and PHS acknowledge financial support from the National Agency for Research and Development of Chile (ANID) through project FONDECYT 1230138 and FONDECYT POSTDOC 3240440.
References
within the PAW formalism implemented in VASP[?], the partial charge associated with the angular momentum channel of atom , denoted by , is given by
Where the is the number of -points used to sample the Brillouin zone, is the occupation number of the Kohn–Sham state corresponding to band and wave vector , and denotes the atom-centered orbital of atom characterized by angular momentum quantum number and magnetic quantum number , i.e. spherical harmonics. The summation over accounts for all orbitals associated with a given angular momentum channel. Therefore, provides a measure of the electronic population associated with the (), (), and () orbitals of atom .
Supplementary Material: Understanding the Oxygen Reduction Reaction and Oxygen Evolution Reaction in Metal Intercalated Biphenylene Bilayers
Henri G. Mendonça; Pedro H. Souza; Walter Orellana; Roberto H. Miwa
suplementar.pdf
Supplementary Material: Understanding the Oxygen Reduction Reaction and Oxygen Evolution Reaction in Metal Intercalated
Biphenylene Bilayers
Structural Details
In the table below, we show the values in eV for all B/M/B, calculated by linear response according to [1]. For this calculation, we use the supercell, with 48 carbon atoms and 4 M atoms and the most stable magnetic phase, where the geometry used is with M on C biphenylene.
| TM | Cr | V | Mn | Os | W | Nb | Ti | Fe | Pt | Co | Ni | Cu | Ru |
| 3.59 | 3.17 | 4.85 | 1.01 | 0.82 | 1.15 | 2.97 | 3.29 | 1.65 | 3.53 | 5.50 | 4.20 | 3.74 |
Table T.1. The values calculated by linear response [1] for transitions metals.
We calculated the most stable magnetic phase for each of the B/M/B systems by considering four possible magnetic configurations: FM, AF-x, AF, and AF-y, as represented in panels (a)-(d) of Figure S.1 for the B/Fe/B system.

Figure S.1. We present in panels (a)-(d) the respective magnetic phases FM, AF-y, AF, and AF-x, after convergence, for the B/Fe/B system. For B/Fe/B, the most stable magnetic phase at the C site is AF-x.
TABLES T.2-T.4 compile, for the four considered magnetic orders FM, AF-y, AF, and AF-x, the energetic and magnetic results for each of the three structural geometries (s, h, and o). For each case, we list the energy difference relative to the lowest-energy configuration, defined as , in eV per cell; positive values indicate configurations that are less stable relative to the minimum. Additionally, we report the local magnetic moments, in , of the four transition metals atoms in the cell. Reference configurations () have been highlighted in red in the tables.
| s | h | o | ||||||||||||||
| Co | ||||||||||||||||
| FM | 0.666 | 0.196 | 0.000 | 1.7 | 1.7 | 1.7 | 1.7 | 2.0 | 2.0 | 2.0 | 2.0 | 1.9 | 1.9 | 1.9 | 1.9 | |
| AF-y | 0.254 | 0.601 | 0.060 | 1.6 | 1.6 | -1.6 | -1.6 | 2.0 | 2.0 | -2.0 | -2.0 | 2.0 | 2.0 | -2.0 | -2.0 | |
| AF | 0.000 | 0.542 | 0.081 | 1.7 | -1.7 | -1.7 | 1.7 | 2.0 | -2.0 | -2.0 | 2.0 | 2.0 | -2.0 | -2.0 | 2.0 | |
| AF-x | 0.154 | 0.000 | 0.059 | 1.7 | -1.7 | 1.7 | -1.7 | 1.9 | -1.9 | 1.9 | -1.9 | 1.9 | -1.9 | 1.9 | -1.9 | |
| Cr | ||||||||||||||||
| FM | 0.093 | 0.024 | 0.107 | 4.1 | 4.1 | 4.1 | 4.1 | 4.2 | 4.2 | 4.2 | 4.2 | 4.4 | 4.4 | 4.4 | 4.4 | |
| AF-y | 0.259 | 0.000 | 0.089 | 3.8 | 3.8 | -3.8 | -3.8 | 4.1 | 4.1 | -4.1 | -4.1 | 4.4 | 4.4 | -4.4 | -4.4 | |
| AF | 0.687 | 0.529 | 0.083 | 3.7 | -3.7 | -3.7 | 3.7 | 4.1 | -4.1 | -4.1 | 4.1 | 4.4 | -4.4 | -4.4 | 4.4 | |
| AF-x | 0.000 | 0.015 | 0.000 | 3.9 | -3.9 | 3.9 | -3.9 | 4.2 | -4.2 | 4.2 | -4.2 | 4.4 | -4.4 | 4.4 | -4.4 | |
| Cu | ||||||||||||||||
| FM | 0.000 | 0.000 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||
| AF-y | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||||
| AF | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||||
| AF-x | 0.000 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | |||
| Fe | ||||||||||||||||
| FM | 0.588 | 0.190 | 0.000 | 1.7 | 1.7 | 1.7 | 1.7 | 3.0 | 3.0 | 3.0 | 3.0 | 3.3 | 3.3 | 3.3 | 3.3 | |
| AF-y | 0.555 | 0.911 | 0.241 | 1.7 | 1.7 | -1.7 | -1.7 | 2.7 | 2.7 | -2.7 | -2.7 | 3.1 | 3.1 | -3.1 | -3.1 | |
| AF | 0.683 | 0.895 | 0.278 | 1.5 | -1.5 | -1.5 | 1.5 | 2.7 | -2.7 | -2.7 | 2.7 | 3.2 | -3.2 | -3.2 | 3.2 | |
| AF-x | 0.000 | 0.000 | 0.236 | 2.1 | -2.1 | 2.1 | -2.1 | 2.9 | -2.9 | 2.9 | -2.9 | 3.3 | -3.3 | 3.3 | -3.3 | |
| Mn | ||||||||||||||||
| FM | 0.041 | 0.000 | 0.018 | 4.2 | 4.2 | 4.2 | 4.2 | 4.5 | 4.5 | 4.5 | 4.5 | 4.5 | 4.5 | 4.5 | 4.5 | |
| AF-y | 0.123 | 0.015 | 0.106 | 4.3 | 4.3 | -4.3 | -4.3 | 4.5 | 4.5 | -4.5 | -4.5 | 4.6 | 4.6 | -4.6 | -4.6 | |
| AF | 0.000 | 0.356 | 0.047 | 4.2 | -4.2 | -4.2 | 4.2 | 4.4 | -4.4 | -4.4 | 4.4 | 4.6 | -4.6 | -4.6 | 4.6 | |
| AF-x | 0.196 | 0.359 | 0.000 | 4.2 | -4.2 | 4.2 | -4.2 | 4.4 | -4.4 | 4.4 | -4.4 | 4.6 | -4.6 | 4.6 | -4.6 |
Table T.2. Relative energy difference (, eV per cell) between the FM, AF-y, AF, and AF-x magnetic configurations for the s, h, and o geometries of , , , and , along with the local magnetic moments of the four atoms (, , and , in ). The indicates the most stable configuration; values in red highlight the most stable geometry for each magnetic phase.
| s | h | o | |||||||||||||
| Nb | |||||||||||||||
| FM | 0.209 | 0.361 | 0.117 | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 1.1 | 0.7 | 0.7 | 0.7 | 0.7 |
| AF-y | 0.080 | 0.189 | 0.074 | 1.3 | 1.3 | -1.3 | -1.3 | 1.1 | 1.1 | -1.1 | -1.1 | 1.0 | 1.0 | -1.0 | -1.0 |
| AF | 0.855 | 0.000 | 0.000 | 0.02 | -0.02 | -0.02 | 0.02 | 1.1 | -1.1 | -1.1 | 1.1 | 1.1 | -1.1 | -1.1 | 1.1 |
| AF-x | 0.000 | 0.329 | 0.085 | 1.2 | -1.2 | 1.2 | -1.2 | 1.1 | -1.1 | 1.1 | -1.1 | 0.7 | -0.7 | 0.7 | -0.7 |
| Ni | |||||||||||||||
| FM | 0.009 | 0.019 | 0.268 | 0.0 | 0.0 | 0.0 | 0.0 | 0.9 | 0.9 | 0.9 | 0.9 | 0.9 | 0.9 | 0.9 | 0.9 |
| AF-y | 0.008 | 0.247 | 0.000 | 0.1 | 0.1 | -0.1 | -0.1 | 0.9 | 0.9 | -0.9 | -0.9 | 1.0 | 1.0 | -1.0 | -1.0 |
| AF | 0.000 | 0.234 | 0.002 | 0.2 | -0.2 | -0.2 | 0.2 | 0.9 | -0.9 | -0.9 | 0.9 | 1.0 | -1.0 | -1.0 | 1.0 |
| AF-x | 0.009 | 0.000 | 0.015 | 0.0 | 0.0 | 0.0 | 0.0 | 0.9 | -0.9 | 0.9 | -0.9 | 1.0 | -1.0 | 1.0 | -1.0 |
| Os | |||||||||||||||
| FM | 0.000 | 0.000 | 0.000 | 0.8 | 0.8 | 0.8 | 0.8 | 0.3 | 0.3 | 0.3 | 0.3 | 1.1 | 1.1 | 1.1 | 1.1 |
| AF-y | 0.075 | 0.001 | 4.653 | 0.7 | 0.7 | -0.7 | -0.7 | 0.3 | 0.3 | -0.3 | -0.3 | 0.0 | 0.0 | 0.0 | 0.0 |
| AF | 0.169 | 0.098 | 4.654 | 0.8 | -0.8 | -0.8 | 0.8 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| AF-x | 0.035 | 0.098 | 0.001 | 0.7 | -0.7 | 0.7 | -0.7 | 0.0 | 0.0 | 0.0 | 0.0 | 1.1 | -1.1 | 1.1 | -1.1 |
| Pt | |||||||||||||||
| FM | 0.000 | 0.000 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | |
| AF-y | 0.001 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | ||
| AF | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | |||
| AF-x | 0.035 | 0.000 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | |
| Ru | |||||||||||||||
| FM | 0.142 | 0.049 | 0.223 | 1.2 | 1.2 | 1.2 | 1.2 | 0.8 | 0.8 | 0.8 | 0.8 | 2.8 | 2.8 | 2.8 | 2.8 |
| AF-y | 0.126 | 0.000 | 0.129 | 1.2 | 1.2 | -1.2 | -1.2 | 0.9 | 0.9 | -0.9 | -0.9 | 2.6 | 2.6 | -2.6 | -2.6 |
| AF | 0.138 | 0.076 | 0.201 | 1.1 | -1.1 | -1.1 | 1.1 | 0.9 | -0.9 | -0.9 | 0.9 | 2.6 | -2.6 | -2.6 | 2.6 |
| AF-x | 0.000 | 0.135 | 0.000 | 1.4 | -1.4 | 1.4 | -1.4 | 0.8 | -0.8 | 0.8 | -0.8 | 2.6 | -2.6 | 2.6 | -2.6 |
Table T.3. Relative energy difference (, eV per cell) between the FM, AF-y, AF, and AF-x magnetic configurations for the s, h, and o geometries of M = Nb, Ni, Os, Pt and Ru, along with the local magnetic moments of the four M atoms (, , and , in B). The indicates the most stable configuration; values in red highlight the most stable geometry for each magnetic phase.
| Ti | s | h | o | ||||||||||||
| FM | 0.000 | 0.085 | 0.000 | 1.3 | 1.3 | 1.3 | 1.3 | 1.5 | 1.5 | 1.5 | 1.5 | 1.3 | 1.3 | 1.3 | 1.3 |
| AF-y | 0.127 | 0.607 | 0.975 | 1.2 | 1.2 | -1.2 | -1.2 | 1.5 | 1.5 | -1.5 | -1.5 | 1.6 | 1.6 | -1.6 | -1.6 |
| AF | 0.164 | 0.000 | 0.975 | 1.1 | -1.1 | -1.1 | 1.1 | 1.7 | -1.7 | -1.7 | 1.7 | 1.6 | -1.6 | -1.6 | 1.6 |
| AF-x | 0.077 | 0.030 | 1.288 | 1.2 | -1.2 | 1.2 | -1.2 | 1.5 | -1.5 | 1.5 | -1.5 | 1.4 | -1.4 | 1.4 | -1.4 |
| V | s | h | o | ||||||||||||
| FM | 0.366 | 0.000 | 0.238 | 2.1 | 2.1 | 2.1 | 2.1 | 2.4 | 2.4 | 2.4 | 2.4 | 2.9 | 2.9 | 2.9 | 2.9 |
| AF-y | 0.000 | 0.248 | 0.096 | 2.3 | 2.3 | -2.3 | -2.3 | 2.6 | 2.6 | -2.6 | -2.6 | 3.0 | 3.0 | -3.0 | -3.0 |
| AF | 0.127 | 0.234 | 0.096 | 2.1 | -2.1 | -2.1 | 2.1 | 2.5 | -2.5 | -2.5 | 2.5 | 3.0 | -3.0 | -3.0 | 3.0 |
| AF-x | 0.134 | 0.045 | 0.000 | 2.3 | -2.3 | 2.3 | -2.3 | 2.5 | -2.5 | 2.5 | -2.5 | 2.9 | -2.9 | 2.9 | -2.9 |
| W | s | h | o | ||||||||||||
| FM | 0.721 | 0.000 | 1.7 | 1.7 | 1.7 | 1.7 | 1.0 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | |
| AF-y | 0.261 | 0.025 | 0.000 | 2.2 | 2.2 | -2.2 | -2.2 | 0.8 | 0.8 | -0.8 | -0.8 | 0.0 | 0.0 | 0.0 | 0.0 |
| AF | 0.000 | 0.005 | 2.0 | -2.0 | -2.0 | 2.0 | 0.7 | -0.7 | -0.7 | 0.7 | 0.0 | 0.0 | 0.0 | 0.0 | |
| AF-x | 0.658 | 0.108 | 1.8 | -1.8 | 1.8 | -1.8 | 0.6 | -0.6 | 0.6 | -0.6 | 0.0 | 0.0 | 0.0 | 0.0 |
Table T.4. Relative energy difference (, eV per cell) between the FM, AF-y, AF, and AF-x magnetic configurations for the s, h, and o geometries of M = Ti, V and W, along with the local magnetic moments of the four M atoms (, , and , in ). The indicates the most stable configuration; values in red highlight the most stable geometry for each magnetic phase.
In the table below, we show the bond length between the metal transition intercalated in the B/B and the closest C in the biphenylene bilayer, for the M intercalated in s and h configuration. Also, we show the distance () between the B/B in the same configurations. The average bilayer distance is 3.93 Å and 3.78 Å for the s and h configurations, respectively. For the B/B without intercalated metals the Å
| M | s configuration | h configuration | ||
| Bond length M-C (Å) | (Å) | Bond length M-C (Å) | (Å) | |
| Ti | 2.31 | 4.13 | 2.38 | 3.88 |
| V | 2.25 | 3.98 | 2.32 | 3.71 |
| Cr | 2.28 | 4.06 | 2.45 | 4.05 |
| Mn | 2.31 | 4.12 | 2.42 | 3.95 |
| Fe | 2.11 | 3.68 | 2.30 | 3.68 |
| Co | 2.11 | 3.67 | 2.25 | 3.55 |
| Ni | 2.06 | 3.59 | 2.30 | 3.67 |
| Cu | 2.16 | 3.78 | 2.35 | 3.79 |
| Nb | 2.38 | 4.27 | 2.42 | 3.96 |
| Ru | 2.20 | 3.92 | 2.28 | 3.64 |
| W | 2.29 | 4.07 | 2.30 | 3.68 |
| Os | 2.20 | 3.86 | 2.25 | 3.55 |
| Pt | 2.25 | 3.99 | 2.46 | 4.06 |
Table T.5. the bond length between the metal transition intercalated in the B/B and the closest C, and the distance () between the B/B.
For the analysis of binding energy of O, we can define the binding energy as
where is the total energy of the systems B/M/B when the O is adsorbed on C, is the total energy of B/M/B and is the total energy of O molecule in the triplet configuration.
At the point of view of O adsorption, only B/Nb/B and B/W/B present the , indicating the unfavorable adsorption in this systems. This results are due to the fact that when O is adsorbed onto B/M/B, the M atoms throughout the structure are displaced from the square sites of the B/B, causing the system energy to increase considerably, as show in the Figure S.2.

Figure S.2. Perspective views of the optimized geometries for the B/Nb/B (a) and B/W/B (b) systems following the adsorption of the at the square-planar sites ().
Bader Charge
In this section, we present the results of charge transfer, via Bader charge analysis [2] and net charge density, from the M atoms to the sites, as summarized in Table T.5. For each B/M/B system, we show the average charge lost by M and gained by the atoms.
| Cr | V | Mn | Os | W | Nb | Ti | |
| Fe | Pt | Co | Ni | Cu | Ru | ||
Table T.6. Mean value of Bader charge of transition metals atoms () and mean value of Bader charge of carbon atoms in Square (). While other carbon atoms also accept electron from the M, the electron transfer to the carbon atoms at the square-planar sites is substantially more pronounced. The mean values of Bader charge are in [] unit.
We performed the net charge density difference of B/M/B for M=Fe, given by
where , and are the charge density of B/M/B, M e B, respectively. Considering, B/Fe/B the are represented in Figure S.3. The carbons in the square gain electrons, as indicated by T.6, whereas the Fe atoms lose electrons.

Figure S.3. The for B/Fe/B. Panel (a) is the b-view of B/Fe/B, and panel (b) is the c-view of B/Fe/B. Yellow color indicated the region to gain electrons and blue color indicated the regions to lose electrons. Isosurface: 0.01 .
In the Fig. S.4, we present the project bands (left panel) and PDOS (right panel) of the -orbital of the square-site carbon atoms (C) and the non-square carbon atoms (C) for B/B. It is notable that near , the majority contribution to the -orbital comes from the C sites. This suggests that the greater density of states at C compared to C renders C more reactive [3], and indicates that C should be the preferential adsorption sites for O, OH, and OOH.

Figure S.4. The left panel shows the band structure projected onto the orbitals of the square-site carbon atoms (black dots - C) and the non-square carbon atoms (red dots - C). The right panel displays the PDOS of the orbital for C (black curve) and C (red curve). The dashed line in both panels indicates the Fermi level. It is notable that near , the majority contribution to the -orbital comes from the C sites. This suggests that the greater density of states at C compared to C renders C more reactive [3], and indicates that C should be the preferential adsorption sites for O, OH, and OOH.
Gibbs Free Energy
We estimated the reaction energies for the ORR and OER on TB/M/B by calculating the adsorption free energies of three key intermediates (OOH*, O*, and OH*) [4] at the square site of the biphenylene framework [3]. For standardization, these energies were referenced to the free energies of equivalent stoichiometric amounts of gaseous HO and H, utilizing the formalisms described in the following equations. The asterisks denote species adsorbed on the square B site.
The reaction of the ORR can be written as [4],
where the Gibbs free energy at zero potential, relative to the reversible hydrogen electrode (RHE), for each chemical reaction steps are given by,
The reaction of the OER can be written as [4],
The Gibbs free energy at zero potential, relative to RHE, for OER are given by:
The onset potential () for ORR/OER is determined by,
and the overpotential () for each reaction is defined by
where .
Numerical Results for Gibbs free energies
In this section, we present the numerical results obtained from the catalysis calculations using DFT. In the TABLE T.6 contains the calculated values for and at K. TABLE T.7 presents the values for , , and , as per Eqs. (18), for each B/M/B system, including solvation corrections. TABLE T.8 provides the values for the ORR steps , , , and , given by Eqs. (23), respectively, also including solvation corrections. Also in TABLE T.8 shows the values for the OER steps , , , and , given by Eqs. (28), with solvation corrections. Finally, TABLE T.9 presents the solvation correction values calculated via VASPsol [5, 6, 7] for each B/M/B system upon adsorption of and .
TABLE T.7: Zero-point energies () and entropic contributions () at K for each B/M/B system with adsorbed , , and species on the catalyst. For these calculations, the B/M/B substrate was kept fixed while the , , and adsorbates were allowed to relax freely.
| Metals | ||||||
| Cr | 0.081 | 0.048 | 0.386 | 0.071 | 0.454 | 0.096 |
| V | 0.082 | 0.047 | 0.382 | 0.077 | 0.464 | 0.131 |
| Mn | 0.081 | 0.050 | 0.387 | 0.068 | 0.447 | 0.104 |
| Os | 0.084 | 0.046 | 0.384 | 0.075 | 0.467 | 0.130 |
| W | 0.103 | 0.039 | 0.383 | 0.075 | 0.465 | 0.131 |
| Nb | 0.092 | 0.041 | 0.390 | 0.065 | 0.444 | 0.112 |
| Ti | 0.081 | 0.053 | 0.389 | 0.067 | 0.460 | 0.132 |
| Fe | 0.082 | 0.045 | 0.387 | 0.068 | 0.456 | 0.142 |
| Pt | 0.079 | 0.050 | 0.382 | 0.076 | 0.461 | 0.136 |
| Co | 0.082 | 0.046 | 0.386 | 0.070 | 0.457 | 0.143 |
| Ni | 0.087 | 0.033 | 0.388 | 0.069 | 0.460 | 0.138 |
| Cu | 0.081 | 0.048 | 0.390 | 0.066 | 0.463 | 0.130 |
| Ru | 0.082 | 0.047 | 0.390 | 0.067 | 0.451 | 0.165 |
Table T.7.
| Metals | (eV) | (eV) | (eV) |
| Ti | 1.98 | ||
| V | 2.23 | 0.75 | 1.75 |
| Cr | 5.24 | 5.23 | 2.42 |
| Mn | 4.20 | 1.95 | 1.27 |
| Fe | 3.25 | 1.74 | 0.47 |
| Co | 4.46 | 2.54 | 1.51 |
| Ni | 4.40 | 2.39 | 1.47 |
| Cu | 4.07 | 1.91 | 1.09 |
| Nb | |||
| Ru | 4.14 | 1.96 | 1.23 |
| W | |||
| Os | |||
| Pt | 4.13 | 2.05 | 1.17 |
Table T.8. Gibbs free energy adsorption for each M considered the species , and , and with the solvation effect includes.
| Metals | (eV) | (eV) | (eV) | (eV) | (eV) | (eV) | ||
| Ti | ||||||||
| V | ||||||||
| Cr | ||||||||
| Mn | ||||||||
| Fe | ||||||||
| Co | ||||||||
| Ni | ||||||||
| Cu | ||||||||
| Nb | ||||||||
| Ru | ||||||||
| W | ||||||||
| Os | ||||||||
| Pt |
Table T.9. Gibbs free energy steps for each M in the ORR with - and to the OER with -.
| Metals | (eV) | (eV) |
| Ti | ||
| V | ||
| Cr | ||
| Mn | ||
| Fe | ||
| Co | ||
| Ni | ||
| Cu | ||
| Nb | ||
| Ru | ||
| W | ||
| Os | ||
| Pt |
Table T.10. Binding energy energy for solvation correction for each metals.
We present the onset potential () and overpotentials () for B/Nb/B and B/W/B, where for ORR we obtain V and V for B/Nb/B, and V and V for B/W/B. Additionally, for OER we obtain V and V for B/Nb/B, and V and V for B/W/B.
Here, we present the Gibbs free energies diagrams for each B/M/B, for the ORR, Figs. S.5, S.6 and S.7 and OER, Fig. S.8, S.9 and S.10, reactions pathways. All energies are referenced to standard conditions.

Figure S.5. ORR Gibbs free energy for B/M/B (M = Co, Cr, Cu and Mn) at different applied potentials: V (black), V (blue), and the onset potential (red).

Figure S.6. ORR Gibbs free energy for B/M/B (M = Nb, Ni, Os and Pt) at different applied potentials: V (black), V (blue), and the onset potential (red).

Figure S.7. ORR Gibbs free energy for B/M/B (M = Ru, Ti, V and W) at different applied potentials: V (black), V (blue), and the onset potential (red).

Figure S.8. OER Gibbs free energy for B/M/B (M = Co, Cr, Cu and Mn) at different applied potentials: V (black), V (blue), and the onset potential (red).

Figure S.9. OER Gibbs free energy for B/M/B (M = Nb, Ni, Os and Pt) at different applied potentials: V (black), V (blue), and the onset potential (red).

Figure S.10. OER Gibbs free energy for B/M/B (M = Ru, Ti, V and W) at different applied potentials: V (black), V (blue), and the onset potential (red).
Notably, the B/Os/B, B/Ti/B, B/W/B, and B/Nb/B systems exhibit the poorest catalytic performance among the B/M/B systems. InFig. S.11, we illustrate a contributing factor to this low catalytic activity. Adsorption of any species-O*, OH*, or OOH*-onto these specific systems induces a displacement of the M=Os, Ti, W and Nb atoms from the C sites. This general shift of the metal atoms leads to a considerable increase in the energy of the +O (or +OH or +OOH) system, thereby raising the energy barriers and consequently increasing the overpotential for both the ORR and OER.

Figure S.11. Top views of the optimized geometries for the B/Os/B (a.x), B/Ti/B (b.x), B/W/B (c.x) and B/Nb/B (d.x) systems following the adsorption of the (z.1), (z.2), and (z.3) intermediates at the square-planar sites.
We also show the AIMD for the B/M/B systems with , , , and in Figure S.12, i.e., other B/M/B high-catalytic-performance systems for ORR and OER. With these results, it is possible to consider that the B/M/B systems with relatively high catalytic performance are stable at . Additionally, we can infer that the intercalated metals do not cluster and remain in the square-planar site of the B/B framework.

Figure S.12. AIMD simulation of B/M/B with M = Mn, Co, Ni and Ru for 10 ps.
Additionally, we performed the AIMD of B/Fe/B in the presence of water, at the experimental concentration (). These results indicate that the biphenylene bilayer effectively acts as a protective barrier, shielding the intercalated metal centers from direct interaction with the aqueous environment where ORR/OER occur.

Figure S.13. AIMD simulation of in the presence of for 5 ps with experimental density of water 1 g cm.
Electronic Structure
In this section, we present the electronic properties of the systems. Figure S.14 Figure S.15 Figure S.16 Figure S.17 and Figure S.18 shows the spin-polarized band structures, black lines are the spin-up component, while red lines are the spin-down components, for all systems in their most stable magnetic phase with M positioned at the site. The magnetic phase is essentially localized on the M atoms, whereas for the carbon atoms, the magnetization is essentially zero, with no spin component distinction for the carbon atoms. The choice of M at the site is justified because if M were positioned at the hexagonal or octagonal site, the electronic density loses localization at the square sites, hindering surface adsorption [8].

Figure S.14. The spin-polarized band structures of B/M/B with M = Ti, V and Cr, relative to the vacuum level. The dashed line indicates the Fermi energy level.

Figure S.15. The spin-polarized band structures with , and , relative to the vacuum level. The dashed line indicates the Fermi energy level.

Figure S.16. The spin-polarized band structures B/M/B with M = Ni, Cu and Nb, relative to the vacuum level. The dashed line indicates the Fermi energy level.

Figure S.17. The spin-polarized band structures B/M/B with , W and Os, relative to the vacuum level. The dashed line indicates the Fermi energy level.

Figure S.18. The spin-polarized band structures B/M/B with M = Pt, relative to the vacuum level. The dashed line indicates the Fermi energy level.
In Figs. S.19, S.20, S.21 and S.22, we present the PDOS of the -orbitals and the -orbital of the atoms for each B/M/B system in the vicinity of the Fermi level.

Figure S.19. The PDOS of and -orbitals for B/B and B/M/B for M = Ti, V and Cr, where the lines blue is -orbital, green is -orbital, purple is -orbital, black is -orbital, cyan is -orbital, and the red line is -orbital from C atoms. The dashed line indicates the Fermi energy level.

Figure S.20. The PDOS of and -orbitals for B/M/B for M = Mn, Fe, Co and Ni, where the lines blue is -orbital, green is -orbital, purple is -orbital, black is -orbital, cyan is -orbital, and the red line is -orbital from C atoms. The dashed line indicates the Fermi energy level.

Figure S.21. The PDOS of and -orbitals for B/M/B for M = Cu, Nb, Ru and W, where the lines blue is -orbital, green is -orbital, purple is -orbital, black is -orbital, cyan is -orbital, and the red line is -orbital from C atoms. The dashed line indicates the Fermi energy level.

Figure S.22. The PDOS of and -orbitals for B/M/B for M = Os and Pt, where the lines blue is -orbital, green is -orbital, purple is -orbital, black is -orbital, cyan is -orbital, and the red line is -orbital from C atoms. The dashed line indicates the Fermi energy level.
References
References
- [1]M. Cococcioni and S. de Gironcoli, Linear response approach to the calculation of the effective interaction parameters in the LDA + U method, Phys. Rev. B 71, 035105 (2005).
- [2]G. Henkelman, A. Arnaldsson, and H. Jónsson, A fast and robust algorithm for bader decomposition of charge density, Computational Materials Science 36, 354 (2006).DOI
- [3]Y. Luo, C. Ren, Y. Xu, J. Yu, S. Wang, and M. Sun, A first principles investigation on the structural, mechanical, electronic, and catalytic properties of biphenylene, Scientific Reports 11, 10.1038/s41598-021-98261-9 (2021).
- [4]H. Xu, D. Cheng, D. Cao, and X. C. Zeng, Revisiting the universal principle for the rational design of single-atom electrocatalysts, Nature Catalysis 7, 207 (2024).DOI
- [5]K. Mathew, V. S. C. Kolluru, and R. G. Hennig, Vasp sol: Implicit solvation and electrolyte model for density-functional theory, https://github.com/henniggroup/VASPsol (2018).
- [6]K. Mathew, R. Sundararaman, K. Letchworth-Weaver, T. A. Arias, and R. G. Hennig, Implicit solvation model for density-functional study of nanocrystal surfaces and reaction pathways., J. Chem. Phys. 140, 084106 (2014).
- [7]K. Mathew, V. S. C. Kolluru, S. Mula, S. N. Steinmann, and R. G. Hennig, Implicit self- consistent electrolyte model in plane-wave density-functional theory., J. Chem. Phys. 151, 234101 (2019).
- [8]H.-T. Chen and D. K. Dhanthala Chittibabu, A new carbon allotrope: Biphenylene as promising anode materials for li-ion and lio2 batteries, Solid State Ionics 395, 116214 (2023).DOI
