N6-Coordinated Dual-Atom Catalysts

• N6-coordinated DACs are atomically dispersed catalysts with dual-metal centers embedded in nitrogen-rich frameworks, optimizing reactivity and selectivity.
• The six nitrogen ligands modify metal d-orbital energies, enhancing adsorption properties for reactions like HER, NRR, and ORR through controlled electron transfer.
• High formation barriers and ML-guided screening ensure stability and scalability, paving the way for applications in electrocatalysis, photocatalysis, and energy conversion.

N6-coordinated dual-atom catalysts (DACs) are a class of atomically dispersed catalytic systems in which two metal atoms are embedded into a nitrogen-rich, typically hexagonal framework such as nitrogen-doped graphene, graphite carbon nitride (g-CN), or related lattice substrates. Both metal centers are spatially confined, each coordinated by six nitrogen ligands in the local atomic structure. This motif enables direct control over geometric and electronic parameters that influence catalytic reactivity, selectivity, and stability across critical energy conversion and chemical synthesis processes.

1. Coordination Environment and Structural Features

N6-coordination refers to the embedding of each metal atom in an environment where six nitrogen atoms serve as ligands, typically substituting for carbon sites in the substrate lattice. In studied systems such as Co–Pt@N-doped graphene, the active site consists of two metals (Cobalt and Platinum) lodged in a graphene quadra-vacancy surrounded by six nitrogen atoms—a configuration denoted CoPt@N6V4 (Hunter et al., 2019). Similarly, dual-metal centers (e.g., Fe–Fe in Fe₂/g–CN (Lv et al., 2020), Co–Co or Fe–Fe in M₂N₆ motifs (Lv et al., 2023)) are coordinated by nitrogen within g-CN’s hexagonal cavities.

Full N6 coordination, compared to lower nitrogen concentrations (N₁–N₃), leads to significant modification of the electronic structure and binding environment of the metal centers. Nitrogen ligands, due to their higher electronegativity relative to carbon, weaken the metal–substrate bond, allow higher oxidation states, and tune the magnetic and charge distribution on the metals, thereby optimizing catalytic activation and intermediate stabilization.

The geometric configuration is further stabilized by high formation and diffusion energy barriers (e.g., 4.75–12.32 eV (Shu et al., 2023)), making dissociation into single atoms or aggregation into clusters thermodynamically and kinetically unfavorable. SYSTEMATIC screen datasets cover multiple quadra-vacancy patterns and bimetal compositions, as evidenced by the LOCAL framework on 611,648 structures (Yin et al., 25 Mar 2025).

2. Electronic Structure and Descriptor Engineering

N6-coordination profoundly impacts metal d-orbital energies, resulting in effective electronic interactions with adsorbates. Orbital population analysis and integrated crystal orbital Hamilton population (ICOHP) calculations are used to quantify the strength and character of metal–adsorbate bonds (Lv et al., 2020, Shu et al., 2023).

For nitrogen activation and reduction (NRR), optimal electron transfer from Fe sites to adsorbed N₂ is signaled by a linear correlation between excess electrons (upon the N₂ molecule) and ICOHP values: $$\text{ICOHP} \approx a \times (\text{excess electrons}) + b$$ where more negative ICOHP implies stronger (and more activating) TM-N coupling for N≡N bond cleavage (Lv et al., 2020).

In oxygen reduction reactions (ORR), the local spin magnetic moment ($M_s$) of the metal center is found to be a unified descriptor for activity in DACs. For Co₂N₆, the binding free energy of *OH follows a linear relation: $G_{(*OH)} = -0.47\,M_s + 1.62 \quad (R^2=0.87)$ and the ORR overpotential displays a volcano-type dependence on $M_s$, indicating optimal catalytic states arise within a specific spin regime (Lv et al., 2023).

For sulfur reduction and polysulfide adsorption in Li–S batteries, d-band alignment and charge polarization between dual-metal centers (e.g., Fe–Ni, Fe–Pt) enable effective S–S bond breaking and Li₂S decomposition (Kumar et al., 17 Oct 2025). ML descriptor sets for rapid property prediction include ICOHP, bond lengths, electronegativity differences, and inter-metal electron affinity (IEA),

$$\text{IEA} = |(N_{\mathrm{Fe}} - N_{\mathrm{M}})\cdot(X_{\mathrm{Fe}} - X_{\mathrm{M}})\cdot r_{\mathrm{Fe}}/X_{\mathrm{S}}\cdot I_M|$$

with ML regression yielding $R^2=0.85$ and MAE 0.26 eV for Gibbs free energies across DAC candidates (Kumar et al., 17 Oct 2025).

3. Catalytic Performance Across Key Reactions

Hydrogen Evolution Reaction (HER)

N6-coordinated Co–Pt dual-atom sites, as the N-ligand concentration increases from N₁ to N₆, show a dramatic shift of H adsorption free energy ($\Delta G_{H^*}$) from non-spontaneous ($>0$) to spontaneous ($<-0.36$ to $-0.42$ eV), rivaling Pt(111) ($\approx -0.09$ eV) (Hunter et al., 2019). Full N6-doping amplifies the activity of Co for H adsorption, while Pt helps stabilize the dimer and fine-tune Co's electronic properties.

Nitrogen Reduction Reaction (NRR)

Fe₂/g–CN demonstrates an optimal compromise in N₂H and NH₂ adsorption, yielding a low limiting potential ($U_L$) of $-0.13$ V and a Faradaic efficiency (FE) of 100% for NH₃, outperforming conventional Ru(0001) ($U_L \sim -1.08$ V) (Lv et al., 2020). The synergistic effect is linked to tailored electron donation and d-orbital coupling, as reflected in ICOHP and excess electron descriptors.

Mixed-metal alloys (Co–Mo, as on p-InP) promote N₂ activation under photoelectrochemical conditions. Co, by lowering N₂ adsorption energy, cooperatively facilitates the stepwise formation of Mo≡N and Mo–NH$_x$ intermediates, as verified by XPS post-PEC dinitrogen activation (Kaur et al., 2023).

Oxygen Reduction Reaction (ORR)

Co₂N₆ and Fe₂N₆ UHD-DACs reach peak ORR activities with overpotentials of 0.31 and 0.33 V, respectively. Ultra-high-density DACs ($>$40 wt% metal loading) retain or enhance activity through spin-state control and maintain stability under acidic operating conditions, outperforming low-density analogues (Lv et al., 2023). Metal spacing modulates spin moments and prevents quenching found in denser configs.

Nitrate Reduction (NO3RR)

FeMo@g–CN and CrMo@g–CN sites achieve limiting potentials of $-0.34$ and $-0.39$ V, respectively, for NO3RR (Shu et al., 2023). Orbital coupling between dimer d-orbitals and NO₃⁻ antibonding orbitals directly activates the NO₃⁻, while dual-metal interaction ensures kinetic stability and prevents dimer dissociation.

Sulfur Reduction Reaction (SRR) for Li–S Batteries

N6-DACs such as Fe–Ni and Fe–Pt demonstrate superior polysulfide adsorption energies ($-1.0$ to $-2.3$ eV), low conversion free-energy barriers ($\leq 0.4$ eV for Li$_2$S$_2$ → Li$_2$S) and facile Li₂S decomposition ($\leq 1.0$ eV), underpinned by dual-metal induced S–S bond polarization and optimal d-band alignment (Kumar et al., 17 Oct 2025).

4. Stability, Scalability, and Machine Learning Frameworks

Stable incorporation of metal dimers in N6-coordinated motifs is achieved via high formation/diffusion barriers ($\sim$5–12 eV), as shown for g–CN-supported DACs (Shu et al., 2023). The LOCAL framework uses hierarchical GCN models—POS2COHP (bond strength prediction via ICOHP) and Graph2E (global stability via transformer-based pooling and MLP regression)—to analyze over 0.6 million DAC@NG structures, spanning 38 elements and multiple coordination environments (Yin et al., 25 Mar 2025).

Phase diagrams generated for bimetal systems (Co–Ni, Fe–Ni, Fe–Mn, Ag–Ni) by varying chemical potentials match experimental stability trends, while ML error metrics (MAE 0.145 eV) support robust screening and prediction. These scalable, data-driven approaches are extensible to alternate heteroatom doping and electronic structure mapping.

5. Optoelectronic Properties and Photocatalytic Potential

Dual-atom decoration on g–CN substrates significantly narrows the bandgap from a pristine value ($\sim$3.23 eV) to as low as 0.41–0.52 eV (Shu et al., 2023). Enhanced visible-light absorption expands the photocatalytic applicability for solar-driven reactions. Multi-orbital metal–N coordination extends the bandwidth and stabilizes charge-carrier dynamics critical for efficient light-induced catalysis, as demonstrated by superior photoconversion efficiency in NO₃RR and NRR contexts.

6. Design Principles and Future Directions

Empirical and computational evidence supports several key principles for N6-DACs:

• Maximizing nitrogen coordination (N6) uniformly enhances catalytic activity by modulating metal charge states and binding energies.
• Dual-metal synergy is critical for selective catalysis, enabling multi-functional reaction pathways and intermediate stabilization.
• Electronic descriptors (ICOHP, spin moment, d-band alignment) and ML-predicted Gibbs energies (GBR model: R^2=0.85, MAE=0.26 eV (Kumar et al., 17 Oct 2025)) guide rational screening and optimization.
• High-density assembly and modular ligand architectures facilitate scalability and mass activity per catalytic site.
• Advanced ML–DFT workflows (PACE, LOCAL) enable exploration of vast configuration spaces with quantitative fidelity, accelerating catalyst discovery across electrocatalysis, photocatalysis, and battery technologies.

A plausible implication is the further employment of N6-coordinated DACs in industrially relevant ammonia and hydrogen production, renewable energy storage (Li–S batteries), and selective environmental remediation. Extension to ternary or higher-order multi-metal motifs, alternate lattice dopants, and integration into device architectures are logical next steps. Common misconceptions regarding dual-atom instability are addressed by clear kinetic/thermodynamic analyses, supporting experimental realization and deployment.

---

N6-coordinated dual-atom catalysts represent a convergence of atomic-level design, electronic structure engineering, and scalable computational screening, offering a robust platform for catalytic enhancement across a spectrum of chemical transformations.