Scaling Relationships in Catalysis: From Fundamental Limitations to Advanced Optimization Strategies

Abstract

Scaling relationships, the linear correlations between adsorption energies of reaction intermediates, represent a fundamental paradigm and a central limitation in catalytic science. This article provides a comprehensive exploration for researchers and drug development professionals, covering the foundational principles of these relationships and their impact on catalytic activity, particularly in reactions like the oxygen evolution reaction (OER). It delves into advanced computational and experimental methods for identifying intermediates and quantifying scaling relations, reviews innovative strategies for breaking these constraints through dynamic structural regulation and multi-site cooperation, and addresses the critical challenge of reconciling experimental data variability. By synthesizing insights from foundational theory to cutting-edge optimization techniques, this review serves as a strategic guide for designing next-generation catalysts with enhanced efficiency and selectivity, with profound implications for energy conversion and pharmaceutical development.

The Foundation of Scaling Relationships: Understanding the Energetic Links That Govern Catalysis

Defining Linear Scaling Relationships (LSRs) and Key Intermediates

Table of Contents

• Introduction to Linear Scaling Relationships (LSRs)
• The Fundamental Challenge of LSRs in Catalysis
• Experimental and Computational Methodologies
• Emerging Strategies to Break LSRs
• Research Reagent Solutions
• Conclusion and Future Perspectives

In multi-step catalytic reactions, the adsorption energies of different reactive intermediates often correlate linearly with one another; these correlations are known as Linear Scaling Relationships (LSRs) [1]. They arise from fundamental chemical principles, such as bond order conservation and electron counting rules, which dictate how adsorbates interact with catalytic surfaces [2]. For instance, on transition metal surfaces, the adsorption energies of intermediates like *OH, *O, and *OOH in the oxygen evolution reaction (OER) are typically bound by these linear relationships, meaning they cannot be adjusted independently [1]. While LSRs simplify the prediction of catalytic activity trends and help construct volcano plots for catalyst screening, they also impose an intrinsic limitation on the maximum achievable catalytic activity and selectivity [1] [3] [2]. The top of the activity volcano represents the best possible compromise dictated by these scaling relations, making it impossible to optimize the binding strength of all intermediates simultaneously on a conventional single-site catalyst [4].

The Fundamental Challenge of LSRs in Catalysis

The core problem posed by LSRs is the constraint they place on a catalyst's ability to independently modulate the free energy of each reaction intermediate. In many key energy conversion reactions, this leads to a significant overpotential or limiting potential.

Table 1: Key Intermediates and Scaling Relations in Selected Catalytic Reactions

Reaction	Key Intermediates	Scaling Relationship	Catalytic Consequence
Oxygen Evolution Reaction (OER) [1]	*OH, *O, *OOH	Linear scaling between *OOH and *OH	Limits the minimum theoretical overpotential [1].
CO2 Electroreduction (C1 Products) [4]	*CO, *COOH	Linear scaling between *COOH and *CO	Limits the potential for CO formation and subsequent reduction [4].
CO2 Electroreduction (C2 Products) [3]	*OCHO, *OCHO, *OCHOH	Linear scaling between intermediates	Hinders C-C coupling by making it difficult to optimize all intermediate adsorption strengths simultaneously [3].
Hydrogen Evolution Reaction (HER) [5]	*H	N/A (Single intermediate)	Linear scaling relationships are less relevant for single-intermediate reactions.

The following diagram illustrates how LSRs couple the energies of different intermediates on a single-site catalyst, creating a fundamental barrier to enhanced activity.

Experimental and Computational Methodologies

Advancing the field requires sophisticated techniques to probe active sites under working conditions and compute reaction pathways.

Computational Screening with Density Functional Theory (DFT)

Purpose: To predict adsorption energies, identify active sites, and map scaling relationships across many materials [3] [4] [5].

• Protocol:
  • Model Construction: Build atomic-scale models of catalyst surfaces, such as single-atom catalysts (SACs), double-atom catalysts (DACs), or high-entropy alloy (HEA) nanoparticles [3] [4].
  • Geometry Optimization: Use DFT to relax all atom positions until forces are minimized (e.g., < 0.05 eV/Ã…) [5].
  • Adsorption Energy Calculation: For each intermediate (e.g., *COOH, *OCHO), compute the adsorption energy (Eads) as Eads = E(catalyst+adsorbate) - Ecatalyst - Eadsorbategas [2].
  • Scaling Analysis: Plot E_ads of one intermediate against another (e.g., *OOH vs. *OH) to establish linear scaling relationships [1] [2].
  • Activity Prediction: Use the adsorption energies as descriptors to plot activity volcanoes and identify promising catalysts that deviate from scaling lines [4].

Operando X-ray Absorption Fine Structure (XAFS)

Purpose: To determine the dynamic electronic and geometric structure of active sites during electrochemical operation [1].

• Protocol:
  • Cell Design: Use an electrochemical flow cell with X-ray transparent windows (e.g., Kapton film).
  • In Situ Electrochemical Activation: Apply a controlled potential (e.g., CV between 1.1 and 1.65 V vs. RHE) to the catalyst to induce transformation to its active state [1].
  • Data Collection: Simultaneously measure XANES and EXAFS spectra at the metal absorption edge (e.g., Ni K-edge) under reaction conditions.
  • Data Analysis: Fit EXAFS spectra to extract coordination numbers and bond distances, revealing the formation of active sites like O-bridged Ni-Fe trimers [1].

The workflow below integrates these computational and experimental methods to discover and validate catalysts that break LSRs.

Emerging Strategies to Break LSRs

Recent research has moved beyond simple single-site catalysts to more complex structures that provide the spatial and electronic flexibility needed to circumvent LSRs.

Table 2: Strategies for Breaking Linear Scaling Relationships

Strategy	Mechanism	Key Example(s)
Dynamic Dual-Site Cooperation [1]	A second site (e.g., Ni) dynamically changes coordination during the cycle, electronically perturbing the primary active site (e.g., Fe) to favorably adjust energies of different steps.	Ni-Fe molecular complex for OER, where Ni-adsorbate coordination alters adjacent Fe site, simultaneously lowering energy barriers for O-H cleavage and O-O formation [1].
Multi-Component Alloy Sites [4]	The unique local environment in high-entropy alloys creates "special sites" that stabilize non-traditional adsorbate binding modes, decoupling intermediate energies.	AgAuCuPdPt HEA nanoparticles with Au sites neighbored by Cu atoms stabilize bidentate COOH/*CHO, breaking the *CO-CHO scaling relation for CO2RR [4].
Supported Dual-Atom Catalysts (DACs) [3]	Two adjacent metal atoms provide distinct binding sites that can selectively stabilize different intermediates, overcoming the limitations of single-atom catalysts.	Fe-Ni pairs supported on hexagonal boron nitride (h-BN) selectively tune adsorption of *OCHO vs. *OCHOH, breaking LSRs for CO2RR to C2 products [3].
Transition Metal Carbides (TMCs) [2]	Surface sites with mixed ionic/covalent character and varied adsorption geometries interact differently with diverse adsorbates, disrupting universal scaling.	TiC, MoC surfaces and supported metal clusters (e.g., Au/TiC) show deviations from traditional transition metal scaling relations for species like CO2 and CHx [2].

The following diagram illustrates the powerful mechanism of dynamic dual-site cooperation, where the active site is not static but evolves during the catalytic cycle.

Research Reagent Solutions

Table 3: Essential Research Reagents and Materials for LSR Studies

Reagent / Material	Function in Research	Specific Example(s)
Single-Atom Pre-catalysts	Serves as a well-defined precursor for constructing more complex active sites via in situ electrochemical transformation.	Ni single atoms on holey graphene nanomesh (Ni-SAs@GNM) for constructing Ni-Fe molecular complexes [1].
Metal Salt Precursors	Introduces dopant or secondary metal atoms to create bimetallic active sites.	Fe(OH)₄⁻ anions from Fe salts in KOH electrolyte for incorporation into Ni-SAs@GNM [1].
h-BN Substrate	A polar, chemically stable 2D material used as a support for anchoring single and dual metal atoms, providing a tunable coordination environment.	Used as a substrate for M@BN and M1M2@BN catalysts in CO2RR studies [3].
High-Entropy Alloy (HEA) Nanoparticles	Provides a vast landscape of unique local atomic environments to discover sites that deviate from scaling relations.	AgAuCuPdPt HEA nanoparticles for CO2 electroreduction [4].
Transition Metal Carbides (TMCs)	act as catalytic supports or direct catalysts with distinct electronic properties that can break traditional scaling relationships.	TiC, MoC, WC substrates and supported metal clusters (e.g., Au/TiC) [2].
Purified Electrolyte	Ensures that trace metal impurities do not inadvertently influence catalyst structure or activity, which is critical for reproducible in situ studies.	Use of purified Fe-free 1 M KOH electrolyte in OER studies [1].

The study of Linear Scaling Relationships has evolved from merely understanding a fundamental limitation to actively devising strategies to overcome it. The paradigm is shifting from static, single-site catalysts to dynamic and multi-component systems. The key enabling insights come from combining advanced operando characterization techniques with high-throughput computational screening, revealing that dynamic site evolution [1] and complex local environments in alloys [4] and supports [2] can effectively decouple the energies of key intermediates. Future research will likely involve the accelerated discovery of such complex catalysts using machine learning [4], guided by a deeper atomic-level understanding of reaction pathways under operational conditions. This multi-faceted approach is essential for designing the next generation of highly efficient catalysts for energy conversion and sustainable chemical synthesis.

In multi-step catalytic reactions, the universal existence of linear scaling relationships (LSRs) creates a fundamental thermodynamic constraint that intrinsically limits maximum catalytic performance. These relationships describe the linear correlation between the adsorption energies of different reactive intermediates on conventional single-site catalysts [1]. While LSRs simplify performance prediction and elucidate activity trends, they inevitably place intrinsic limitations on simultaneously optimizing the adsorption of every intermediate to achieve peak activity and/or selectivity [1]. This thermodynamic sabotage manifests across diverse catalytic processes, from electrochemical oxygen evolution to emissions control, imposing an inescapable performance ceiling that researchers must confront.

The core problem resides in the electronic structure of active sites. For reactions involving chemically similar oxygenated intermediates (such as *OH, *O, and *OOH in the oxygen evolution reaction), the adsorption energies of these species are linearly correlated and cannot be adjusted independently [1]. This correlation-imposed constraint makes achieving maximal catalytic performance exceptionally challenging, creating what is essentially a thermodynamic sabotage of catalytic efficiency.

The Molecular Origin and Consequences of LSRs

Theoretical Foundation of Scaling Relationships

The breakthrough in understanding LSRs came from the recognition that binding energies of partially hydrogenated adsorbates (AHâ‚“) scale linearly with the binding energy of their corresponding atomic adsorbates (A) across transition metal surfaces [6]. This relationship follows:

ΔE(AHₓ) = mᴇ × ΔE(A) + bᴇ

where the slope mᴇ depends on the valencies of A and AHₓ [6]. These relationships enable estimation of thermodynamic properties for reaction intermediates, facilitating computational catalyst screening. Fundamentally, LSRs arise because the adsorption energy (ΔEadₛ) comprises both sp-band (ΔEₛₚ) and d-band (ΔE_d) contributions, with the scaling slope attributed primarily to the coupling of the adsorbate's s and p orbitals with the transition metal's d-band [6].

Vibrational scaling relationships (VSRs) represent a related phenomenon, where the squares of vibrational frequencies of adsorbates scale linearly across different metal surfaces [6]. This relationship follows:

ν²(AHₓ) = mᵥ × ν²(A) + bᵥ

These vibrational relationships significantly impact zero-point energies and temperature contributions to Gibbs free energy, which are crucial for accurate prediction of reaction rate and equilibrium constants [6].

Catalytic Performance Limitations in Practice

The detrimental impact of LSRs extends beyond theoretical constructs to practical applications across energy and environmental technologies. In automotive catalytic converters, for instance, LSR-influenced degradation patterns directly affect pollutant conversion efficiency. Research demonstrates that low-odometer catalysts exhibit uniform light-off temperatures for CO, HC, and NO that increase by approximately 20% compared to new catalysts, with thermal deterioration of the alumina washcoat identified as the dominant deactivation mechanism under normal operating conditions [7].

The consequences of these limitations become more severe in high-odometer catalysts, where activity loss varies significantly based on thermal exposure history [7]. The best-performing high-mileage catalysts show activity similar to low-odometer groups, while the worst performers become completely ineffective due to extreme thermal exposure [7]. This degradation directly illustrates how LSR-imposed constraints manifest in real-world systems, where induced thermal deactivation causes substantial loss of hydrocarbon conversion activity and narrowing of the lambda window [7].

Table 1: Performance Deterioration in Automotive Catalytic Converters

Catalyst Type	Light-Off Temperature Increase	Dominant Deactivation Mechanism	Performance Variation
Low Odometer	~20% for CO, HC, and NO	Baseline thermal washcoat deterioration	Minimal variation
High Odometer	Highly variable	Severe thermal exposure	Best: matches low odometer; Worst: completely ineffective

Case Studies: LSR Impacts Across Catalytic Systems

Oxygen Evolution Reaction (OER)

The oxygen evolution reaction represents a prime example where LSRs impose severe performance limitations. The adsorption energies of intermediates *OH, *O, and *OOH participating in the adsorbate evolution mechanism (AEM) are linearly correlated on single active sites [1]. Specifically, the ubiquitous adsorption-energy scaling relationship between *OOH and *OH creates a fundamental constraint that limits OER electrocatalyst performance [1]. This thermodynamic limitation affects numerous energy technologies, including hydrogen generation, COâ‚‚ reduction, and other processes where OER serves as the ideal anodic reaction for providing electrons and protons [1].

Metal-Air Batteries

In aprotic metal-air batteries, LSR-influenced reaction pathways enable parasitic release of singlet oxygen, significantly reducing efficiency and cycle life. The superoxide disproportionation reaction represents a key step controlling peroxide formation upon discharge and enabling singlet oxygen release [8]. This reaction follows:

O₂⁻ + O₂⁻ → O₂²⁻ + O₂

The presence of group 1A cations (Li⁺, Na⁺, K⁺) weakens the Coulomb repulsion between superoxides and allows the reaction to proceed, with the energy landscape revealing competing pathways that facilitate singlet oxygen release [8]. This side reaction not only decreases efficiency but creates reactive neutral solvated species that can initiate additional parasitic chemistries due to their well-known reducing properties [8].

Table 2: LSR-Mediated Limitations in Different Catalytic Systems

Catalytic System	Key Scaling Relationship	Performance Impact
Oxygen Evolution Reaction	*OOH vs. *OH adsorption energies	Limits maximum achievable activity; constrains catalyst design
Metal-Air Batteries	Superoxide disproportionation pathways	Enables parasitic singlet oxygen release; reduces efficiency
Automotive Converters	Thermal degradation vs. activity relationships	Reduces pollutant conversion; narrows operational lambda window
Ammonia Oxidation	N-containing intermediate adsorption energies	Constraints on selectivity and conversion efficiency

Breaking the Scaling Relationship Barrier

Dynamic Structural Regulation of Active Sites

Conventional strategies to circumvent LSRs have focused on engineering catalyst heterogeneity through confining intermediates within nanoscopic channels, introducing proton acceptors, or creating multifunctional surfaces and interfacial sites [1]. However, a groundbreaking approach involves dynamic structural regulation of active sites during the catalytic cycle itself [1].

Research on a Ni-Fe₂ molecular catalyst demonstrates that dynamic evolution of Ni-adsorbate coordination driven by intramolecular proton transfer can effectively alter the electronic structure of adjacent Fe active centers during oxygen evolution [1]. This dynamic dual-site cooperation simultaneously lowers the free energy change associated with both O–H bond cleavage and O–O bond formation, thereby disrupting the inherent scaling relationship in OER [1]. The dynamic coordination between the Ni site and adsorbates (OH and H₂O) plays a crucial role in modulating the electronic structure of the adjacent Fe active site, contributing to circumventing the LSR limitations [1].

Advanced Catalyst Design Strategies

Beyond dynamic regulation, several innovative design principles show promise for mitigating LSR limitations:

• Multifunctional Surfaces: Creating heterostructured catalysts with multiple functionality sites can enable different reaction steps to occur on optimized distinct sites, though this approach poses challenges in identifying precise active site structures [1].

• Single-Atom and Molecular Complexes: Precise engineering of well-defined molecular catalysts, such as the Ni-Feâ‚‚ complex created via in situ electrochemical activation, provides atomic-level control over active sites [1]. These systems offer superior opportunities for mechanistic studies and rational design compared to conventional heterogeneous catalysts.

• Biohybrid and Enzyme-Inspired Systems: Integrating biological and inorganic components creates catalytic systems that leverage nature's solutions to similar challenges. Machine learning-assisted biocatalysis combines AI-based stability predictions with protein language models to design enzymes functional under extreme conditions [9].

Experimental Methodologies for LSR Investigation

Catalyst Synthesis and Activation Protocols

Ni-Feâ‚‚ Molecular Catalyst Preparation [1]:

• Pre-catalyst Synthesis: Create a 3D Ni(OH)â‚‚/graphene hydrogel by sealing an aqueous graphene oxide suspension in a Ni vessel at 80°C. After freeze-drying, thermally anneal the resultant aerogel at 700°C under Ar atmosphere to produce Ni single atoms trapped in holey graphene nanomesh (Ni-SAs@GNM).
• Electrochemical Activation: Employ a standard three-electrode system with Ni-SAs@GNM loaded onto a glassy carbon working electrode. Perform activation using cyclic voltammetry between 1.1 and 1.65 V versus RHE in purified 1 M KOH electrolyte with deliberate addition of 1 ppm Fe ions.
• Characterization: Verify structural transformation from Ni monomer to O-bridged Ni-Feâ‚‚ trimer during activation using operando X-ray absorption fine structure (XAFS) measurements.

Thermal Deactivation Simulation [7]:

• Induced Thermal Exposure: Subject low-odometer catalysts to prolonged high-temperature excursion conditions to simulate extended vehicle operation.
• Performance Assessment: Measure hydrocarbon conversion activity and lambda window characteristics before and after thermal treatment using engine dynamometer and laboratory tests.

Computational and Analytical Approaches

Multireference Ab Initio Methods [8]:

• Electronic Structure Calculation: Employ multiconfigurational computational methods to correctly evaluate systems with multiple determinant electronic configurations.
• Thermodynamic Profiling: Sample reaction thermodynamics by calculating energy variations along dissociation pathways using correlated multiconfiguration approaches.
• Geometric Optimization: Utilize MP2-level geometry optimization for superoxide complexes and peroxide species.

Vibrational Scaling Analysis [6]:

• DFT Normal Mode Computation: Calculate harmonic frequencies for adsorbed species based on normal mode analysis across transition metal surfaces.
• VSR Validation: Test linear correlations between squared vibrational frequencies of AHâ‚“ and atomic A species.
• Thermodynamic Impact Assessment: Evaluate how vibrational frequencies affect zero-point energies and temperature contributions to Gibbs free energy.

Table 3: Essential Research Reagent Solutions for LSR Studies

Reagent/Material	Function in Experimental Protocol	Key Characteristics
Graphene Oxide (GO) Suspension	Forms 3D support structure for single-atom catalysts	Provides high surface area; enables metal atom stabilization
Purified KOH Electrolyte (1 M)	Electrochemical activation medium	Fe-free base electrolyte; allows controlled Fe addition
Fe Ions (1 ppm)	Active site precursor for molecular complexes	Enables in situ formation of Ni-Feâ‚‚ structures
Transition Metal Surfaces	Substrate for adsorption energy studies	Enables LSR validation across different metals
Zeolite-based Carriers	Catalyst support for emissions control	Provides high surface area; enables study of thermal degradation

The Scientist's Toolkit: Research Reagent Solutions

Linear scaling relationships represent a fundamental form of thermodynamic sabotage that intrinsically limits catalytic performance across diverse chemical processes. While these relationships impose challenging constraints on catalyst design, emerging strategies—particularly dynamic structural regulation of active sites—offer promising pathways to circumvent these limitations. The development of Ni-Fe₂ molecular catalysts that leverage dynamic dual-site cooperation demonstrates that LSRs are not insurmountable barriers but rather design challenges requiring innovative approaches.

Future progress in overcoming LSR limitations will likely integrate multiple strategies, including machine learning-assisted catalyst discovery, advanced operando characterization techniques, and bio-inspired design principles. As research increasingly focuses on dynamic catalyst behavior under operational conditions rather than static structures, new opportunities will emerge to design catalytic systems that can adaptively optimize intermediate adsorption energies throughout the catalytic cycle, ultimately defeating the thermodynamic sabotage imposed by linear scaling relationships.

In multi-step catalytic reactions, scaling relationships refer to the linear correlations between the adsorption energies of different reactive intermediates on catalyst surfaces [10]. These relationships emerge because the adsorption energies of chemically similar intermediates, such as the oxygenated species *OH, *O, and *OOH in the Oxygen Evolution Reaction (OER), are often governed by the same underlying bonding principles with the catalyst surface [10] [11]. Consequently, these energies cannot be adjusted independently. For the OER, which is crucial for technologies like water electrolyzers and metal-air batteries, this creates a fundamental constraint: the adsorption energy of *OOH is invariably approximately 3.2 eV higher than that of *OH on conventional single-site catalysts [11]. This fixed energy difference defines a thermodynamic overpotential that limits the maximum achievable activity for any catalyst following the adsorbate evolution mechanism (AEM), creating a "volcano plot" relationship where catalyst performance peaks at an optimal, intermediate binding strength [10] [11]. Overcoming these universal constraints represents one of the most significant challenges in advancing electrocatalysis for renewable energy applications.

Quantitative Scaling Relationships in OER

Theoretical and experimental studies have quantified the linear scaling relationships between OER intermediates. The most problematic relationship is between *OOH and *OH, which imposes the primary thermodynamic limitation on the OER overpotential.

Table 1: Experimentally and Theoretically Determined Scaling Relationships for OER Intermediates

Intermediate Pair	Typical Energy Relationship	Impact on Theoretical Overpotential	Key Supporting References
OOH vs. OH	ΔGOOH ≈ ΔGOH + 3.2 ± 0.2 eV	Defines the minimum theoretical overpotential (~0.37 V) for conventional AEM [11].	Koper et al. [11]
O vs. OH	ΔGO ≈ 2ΔGOH + Constant	Influences the position of the volcano peak; value of (ΔGO - ΔGOH) is a common activity descriptor [11].	Montoya et al. [11]
General Scaling	Eads(XHn) ∝ Eads(X)	A general principle where the adsorption energy of a species (XHn) scales with the adsorption energy of the corresponding atom (X) [10].	Abild-Pedersen et al. [10]
DOTA-(t-Butyl)3-PEG5-azide	DOTA-(t-Butyl)3-PEG5-azide, MF:C40H76N8O12, MW:861.1 g/mol	Chemical Reagent	Bench Chemicals
Taurochenodeoxycholate-3-sulfate	Taurochenodeoxycholate-3-sulfate, CAS:67030-59-5, MF:C26H45NO9S2, MW:579.8 g/mol	Chemical Reagent	Bench Chemicals

These scaling relationships are not merely theoretical constructs; they have been experimentally validated and observed across a wide range of catalyst materials, from pure metals to metal oxides. The constant 3.2 eV difference between *OOH and *OH adsorption free energies is particularly significant because it makes either the formation of *OOH from *O or the deprotonation of *OOH to form O2 the potential-limiting step, depending on which side of the volcano plot the catalyst resides [11].

Mechanisms and Pathways for Circumventing Scaling Relations

The inherent limitations imposed by scaling relationships have driven the exploration of alternative OER mechanisms and advanced catalyst designs that can circumvent or "break" these linear constraints.

Established OER Mechanisms and Their Limitations

Three primary OER mechanisms have been identified, each with a different relationship to the scaling problem.

Table 2: Comparison of Primary Oxygen Evolution Reaction Mechanisms

Mechanism	Key Steps & Intermediates	Relationship to Scaling	Advantages & Disadvantages
Adsorbate Evolution Mechanism (AEM)	Four concerted proton-electron transfers: *OH → *O → *OOH → O2 [11].	Governed by linear scaling relations between *OH, *O, and *OOH [11].	Advantage: Well-understood.Disadvantage: Fundamental overpotential limit (~370 mV) [11].
Lattice Oxygen Mechanism (LOM)	*O couples with lattice oxygen (Olattice) to form O2. Involves oxygen vacancies [11].	Bypasses OOH formation, circumventing the *OOH-OH scaling relation [11].	Advantage: Higher activity possible.Disadvantage: Lattice oxygen loss causes structural instability and catalyst dissolution [11].
Oxide Path Mechanism (OPM)	Direct O–O radical coupling between adjacent *O or *OH species [11].	Involves only *OH and *O, completely avoiding the *OOH intermediate and its scaling relation [11].	Advantage: No scaling relations; high stability.Disadvantage: Requires precise geometric arrangement of two active sites [11].

Diagram 1: OER mechanisms and scaling relationship impact. The AEM pathway is constrained by the scaling relationship between *OOH and *OH, while LOM and OPM offer pathways to circumvent this limitation.

Classification of Strategies for Manipulating Scaling Relations

A comprehensive framework has been proposed to categorize the various strategies for dealing with scaling relations in oxygen electrocatalysis [10]. These strategies represent a chronological and conceptual evolution in the field's approach to this fundamental problem.

• Tuning: The classical approach of adjusting the adsorption energy of intermediates, typically by ligand or strain effects, to move the catalyst along the scaling line toward the top of the volcano plot. This strategy works within the existing scaling relations [10].
• Breaking: Selectively stabilizing one intermediate relative to another to change the slope of the scaling relation. This can be achieved by introducing hydrogen bonding to *OOH (e.g., via spectator species or proton acceptors) or by creating multifunctional sites that interact differently with various intermediates [10] [1].
• Switching: Changing the reaction mechanism to one that avoids the most problematic intermediate entirely. For OER, this means switching from AEM to LOM or OPM to bypass the formation of *OOH [10] [11].
• Pushing: A combined approach that involves both switching to an alternative mechanism and applying stabilising interactions, effectively "pushing" the catalyst beyond the limits defined by the original scaling relation [10].
• Bypassing: A more radical strategy that employs two distinct electronic states to decouple the adsorption energies entirely, potentially eliminating all scaling relations [10].

Experimental Protocols and Methodologies

Protocol: Constructing a Molecular Catalyst to Break Scaling Relations via Dynamic Coordination

A recent groundbreaking study demonstrated the breaking of scaling relations in a Ni-Fe molecular catalyst [1]. The following protocol details the key experimental steps.

Table 3: Protocol for Constructing and Testing a Ni-Fe Molecular OER Catalyst

Step	Procedure Description	Key Parameters & Techniques	Purpose & Rationale
1. Pre-catalyst Synthesis	Synthesize Ni Single Atoms on Holey Graphene Nanomesh (Ni-SAs@GNM).	- Assemble Ni(OH)2/graphene hydrogel in a Ni vessel at 80°C.- Freeze-dry and anneal at 700°C under Ar.- Acid treatment to remove nanoparticles [1].	Create a well-defined, atomically dispersed Ni pre-catalyst on a conductive, high-surface-area support.
2. Electrochemical Activation	Convert Ni-SAs@GNM to the active Ni-Fe complex in situ.	- Use Fe-free 1 M KOH with 1 ppm Fe ions added.- Perform CV scanning between 1.1 and 1.65 V vs. RHE.- Characterize via operando XAFS [1].	Drive the formation of an O-bridged Ni-Fe2 trimer, the true active site. Fe(OH)4− anions anchor to Ni sites.
3. Operando Characterization	Probe the dynamic local structure of active sites during OER.	- Collect Ni K-edge XANES and EXAFS spectra under reaction conditions.- Analyze oxidation states and coordination geometry [1].	Verify the formation of the Ni-Fe complex and monitor the dynamic coordination evolution of the Ni site during catalysis.
4. Electrokinetic Analysis	Determine reaction orders and rate-determining steps.	- Measure current density as a function of OH− concentration and potential.- Fit data to kinetic models [1].	Provide experimental evidence for the proposed mechanism and the involvement of OH− in the rate-determining step.
5. DFT & AIMD Simulations	Model the reaction pathway and free energy landscape.	- Use DFT with Hubbard U correction for transition metals.- Perform AIMD to simulate solvation and dynamic behavior.- Calculate free energy changes for each OER step [1].	Elucidate the atomic-level mechanism, confirm the dynamic structural change, and quantify the breaking of the scaling relation.

Key Findings and Validation

The combination of these techniques revealed an unconventional dynamic dual-site cooperation mechanism. During the OER cycle, the coordination of the Ni site to adsorbates (OH and H₂O) dynamically changes, which in turn modulates the electronic structure of the adjacent Fe active center [1]. This dynamic regulation simultaneously lowers the free energy barriers for both O–H bond cleavage and O–O bond formation, two steps that are typically mutually competing in conventional AEM. This simultaneous optimization is impossible under the constraints of classic scaling relationships, demonstrating a genuine breaking of the LSR [1].

The Scientist's Toolkit: Essential Reagents and Materials

The following table catalogues key reagents, materials, and characterization tools essential for research aimed at understanding and manipulating scaling relationships in OER.

Table 4: Essential Research Reagents and Tools for OER Scaling Relation Studies

Category	Item	Specific Examples / Characteristics	Primary Function in Research
Catalyst Precursors	Single-Atom Pre-catalysts	Ni-SAs@GNM, Fe-SAs@GNM [1].	Well-defined starting materials to construct dual-atom or molecular complex catalysts via in situ activation.
	Metal Salts	Fe(NO3)3, Ni(NO3)2, K2IrCl6, RuCl3 [1] [12].	Source of metal cations for doping, forming heterostructures, or in situ electrodeposition.
Electrolyte Components	High-Purity Alkali	Fe-free KOH (Purified) [1].	Standardized alkaline OER environment; purity is critical to avoid Fe contamination which can active sites.
	Proton Acceptors / Modifiers	Phosphate buffer, Camphorsulfonic acid (CSA) [13] [14].	Induce local spin bias or facilitate proton transfer steps to selectively stabilize intermediates and break scaling relations.
Characterization Tools	Operando XAFS	XANES and EXAFS [1].	Probe the oxidation state and local coordination geometry of active sites under operating conditions.
	AIMD Simulations	DFT with Hubbard U correction [15] [1].	Model reaction pathways at solid-liquid interfaces, identify key intermediates, and calculate free energies.
	Electrokinetic Analysis	Measurement of reaction orders (e.g., with respect to OH− and potential) [1].	Deduce the OER mechanism and identify the rate-determining step from experimental current-potential data.
Advanced Concepts	Spin-Polarization Sources	External magnetic fields (Global Spin Bias), Chiral molecules (CISS effect) [13].	Manipulate the spin state of reaction intermediates to enhance OER kinetics towards triplet O2 formation.
	Regeneration Pathways	Guyard reaction (4Mn2+ + Mn7+ → 5Mn3+) in MnOx systems [14].	Incorporate self-healing redox cycles to maintain catalyst stability under harsh, fluctuating operation conditions.
9,10,16-Trihydroxyhexadecanoic acid	Aleuritic Acid	High-purity Aleuritic Acid for research. A key precursor in perfumery, pharmaceutical, and polymer studies. For Research Use Only. Not for human use.	Bench Chemicals
Butyrylcholine chloride	Butyrylcholine chloride, CAS:2963-78-2, MF:C9H20NO2.Cl, MW:209.71 g/mol	Chemical Reagent	Bench Chemicals

The existence of linear scaling relationships between OH, *O, and *OOH intermediates presents a universal constraint for the Oxygen Evolution Reaction, imposing a fundamental ceiling on the performance of catalysts operating via the conventional adsorbate evolution mechanism. However, as detailed in this review, this constraint is not insurmountable. The development of advanced strategies—including tuning, breaking, switching, pushing, and bypassing—provides a systematic framework for moving beyond these limitations. Critical to this progress is the shift from studying static catalyst models to understanding *dynamic active sites under operational conditions, as exemplified by the Ni-Fe molecular catalyst where dynamic coordination breaks the scaling relation [1]. Furthermore, the exploration of alternative mechanisms like OPM, which avoids the critical OOH intermediate altogether, offers a promising path forward for designing next-generation OER catalysts that combine high activity and stability [11]. Future research will likely focus on precisely controlling the geometric and electronic structure of active sites at the atomic level, leveraging *operando characterization and advanced simulation techniques to rationally design catalysts that transcend historical limitations.

In the field of electrocatalysis, the period since the early 2000s has been defined by the recognition and systematic study of scaling relations—correlations between the adsorption energies of different reactive intermediates on catalyst surfaces. These relations represent a fundamental limitation in multi-step catalytic reactions, particularly for the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER), which underpin energy conversion technologies such as fuel cells, metal-air batteries, and electrolyzers [16] [17]. The existence of these linear scaling relationships (LSRs) means that the adsorption energies of chemically similar oxygenated intermediates (such as *OH, *O, and *OOH in OER) cannot be adjusted independently on conventional single-site catalysts [1]. This correlation-imposed constraint creates an inherent thermodynamic limitation on catalytic performance, making it challenging to simultaneously optimize the binding strength of all reaction intermediates to achieve maximum activity [1].

The physical origin of these scaling relations lies in the fundamental principles of bond order conservation, electron counting rules, and local coordination numbers [2]. Importantly, these relations are not externally imposed constraints but rather emerge naturally from the underlying physical laws governing the interactions between nuclei and electrons [2]. While scaling relations have simplified the prediction of catalyst performance and helped elucidate catalytic activity trends through the construction of volcano plots, they inevitably place intrinsic limitations on optimally adjusting the adsorption of every intermediate, thereby restricting the maximum achievable activity and selectivity [1]. This review examines the historical development of our understanding of these relations, summarizes breakthrough strategies for circumventing them, and provides detailed experimental protocols for studying them, all within the context of a broader thesis on scaling relationships between reaction intermediates in catalysis research.

Theoretical Foundations and Historical Development

The Origin and Nature of Scaling Relations

Scaling relations in catalysis find their theoretical foundation in the observation that the adsorption energies of various intermediates on transition metal surfaces scale linearly with each other. These dependencies arise from the fact that molecules bound to surfaces through similar elements and bonding configurations experience similar variations in adsorption strengths across different catalytic materials [2]. For oxygen electrocatalysis, this is particularly evident in the relationship between *OOH and *OH adsorption energies, where the scaling relation emerges because both species bind to the surface through oxygen atoms with similar coordination environments [1].

The mathematical formulation of these relationships typically follows linear correlations, where the adsorption energy of one intermediate (e.g., OOH) can be expressed as a linear function of another (e.g., *OH): ΔEOOH = αΔEOH + β, where α and β are constants. This linear dependence dramatically reduces the degrees of freedom in complex multi-step reactions, allowing researchers to describe catalytic activity using only one or two descriptors in volcano plot constructions [2]. While this simplification has enabled high-throughput computational screening of catalysts, it has also revealed the fundamental limitations imposed by these relations, with the theoretical overpotential for OER being constrained to a minimum of approximately 0.37 V due to the *OOH-OH scaling relation [1].

Evolution of Understanding Over Two Decades

The systematic recognition of scaling relations as a central paradigm in electrocatalysis began in the early 2000s, with initial studies focusing on transition metal surfaces. Over the past twenty years, research has evolved from simply identifying these relations to developing comprehensive strategies for manipulating and overcoming them [16]. The field has witnessed a paradigm shift from viewing catalysts as static structures to recognizing them as dynamic systems that can undergo significant structural evolution under reaction conditions [1].

Early work primarily utilized scaling relations as a predictive tool and for understanding activity trends across different families of materials. The pioneering studies of Jens K. Nørskov and colleagues established d-band center theory as a fundamental electronic descriptor for adsorption strengths, enabling the rationalization of scaling relations across transition metal catalysts [18]. This theoretical framework allowed researchers to trace activation barriers back to the energy of one or more intermediates through Brønsted-Evans-Polanyi relationships, further simplifying catalyst optimization [2].

As the field matured, attention shifted toward understanding the limitations imposed by these relations and developing strategic approaches to circumvent them. This evolution in thinking represents a journey from descriptive understanding to strategic manipulation, driven by the urgent need to advance energy conversion technologies beyond the performance limits dictated by conventional scaling relations [16] [17].

Quantitative Analysis of Scaling Relations in Electrocatalysis

Thermodynamic Limitations and Activity Descriptors

The quantitative impact of scaling relations on catalytic performance can be visualized through volcano plots, which relate catalytic activity to descriptor variables such as the adsorption energy of key intermediates. The apex of these volcanoes represents the optimal balance of adsorption strengths for all reaction intermediates, while the descending limbs illustrate how deviations from this optimum degrade performance due to either too-weak or too-strong binding of key species.

Table 1: Key Adsorption Energy Scaling Relations in Oxygen Electrocatalysis

Scaling Relation	Slope (α)	Intercept (β, eV)	Theoretical Overpotential Limit	Primary Limitation
*OOH vs. *OH	~1.0	~3.2 ± 0.2	~0.37 V	Limits OER/ORR activity
*O vs. *OH	~0.5	~1.6 ± 0.1	-	Affects oxide formation
*OOH vs. *O	~1.5	~-1.6 ± 0.2	-	Constrains intermediate stabilization

For multi-electron molecular catalysts, quantitative analysis has revealed that the logarithm of rate constants for catalytic steps often follows a linear relationship with the average formal potentials of charge transfer processes. In the study of chlorate electroreduction by a molybdenium polyoxometalate catalyst, which involves ten charge transfer steps and three different two-electron catalytic processes, researchers observed precisely such linear free energy relationships [19]. This quantitative understanding enables the prediction of catalytic performance across related materials systems and provides a foundation for designing improved catalysts.

Quantitative Assessment of Breaking Scaling Relations

Recent advances in breaking scaling relations have yielded quantifiable improvements in catalytic performance. The construction of a Ni-Feâ‚‚ molecular catalyst through in situ electrochemical activation has demonstrated a notable intrinsic OER activity that surpasses the limitations predicted by conventional scaling relations [1]. Theoretical calculations and electrokinetic studies revealed that the dynamic evolution of Ni-adsorbate coordination, driven by intramolecular proton transfer, effectively alters the electronic structure of the adjacent Fe active center during the catalytic cycle [1].

Table 2: Performance Comparison of Catalysts With and Without Scaling Relation Manipulation

Catalyst System	Strategy for Breaking LSRs	Overpotential (mV)	Stability	Improvement Factor
Conventional Ni-Fe catalysts	None (subject to LSRs)	350-450	Moderate	Baseline
Ni-Feâ‚‚ molecular catalyst	Dynamic dual-site cooperation	~250	High	~2x activity enhancement
Transition metal carbides	Diverse adsorption sites	280-320	Variable	1.5-2x vs. parent metals
TM@TMC supported systems	Metal-support interactions	220-300	High	2-3x vs. extended surfaces

The data in Table 2 illustrates how strategic manipulation of scaling relations can lead to substantial improvements in catalytic performance. For transition metal carbides (TMCs), high-throughput screening based on density functional theory has shown that these materials break the limitations imposed by linear scaling relations on transition metals [2]. The diversity of adsorption sites in TMCs, combined with their complex electronic structure resulting from metallic, covalent, and ionic bonding contributions, enables them to interact differently with various adsorbates, thereby circumventing the constraints of conventional scaling relations [2].

Strategic Approaches to Overcome Scaling Relations

Dynamic Structural Regulation of Active Sites

One of the most promising approaches for circumventing scaling relations involves the dynamic structural regulation of active sites under reaction conditions. This strategy recognizes that catalysts are not static entities but can undergo significant structural evolution during catalysis. In the case of the Ni-Feâ‚‚ molecular catalyst, operando X-ray absorption fine structure (XAFS) measurements verified the structural transformation from a Ni monomer to an O-bridged Ni-Feâ‚‚ trimer during the electrochemical activation process [1]. This dynamic evolution enables a dual-site cooperation mechanism where the Ni center directly participates in the catalytic process to induce intramolecular proton transfer and trigger coordination evolution.

The existence of dynamic coordination between the Ni site and adsorbates (OH and H₂O) plays a key role in modulating the electronic structure of the adjacent Fe active site during the OER cycle. Theoretical calculations combined with ab initio molecular dynamics (AIMD) simulations demonstrate that such dynamic regulation simultaneously lowers the free energy required for the mutually competing steps of O–H bond cleavage and *OOH formation, effectively surmounting the LSRs in OER [1]. This mechanism represents a significant departure from conventional static catalyst models and highlights the importance of understanding and harnessing dynamic structural changes under operational conditions.

Materials Engineering Strategies

Transition Metal Carbides and Supported Clusters

Transition metal carbides (TMCs) represent another powerful platform for breaking linear scaling relationships. High-throughput screening based on density functional theory shows that TMCs break the limitations that linear scaling relations impose on transition metals [2]. The (001) facet of face-centered cubic TMCs contains 50% C atoms and 50% metal atoms in the surface layer, exhibiting a higher variety of adsorption sites compared to the (111) surface of face-centered cubic transition metals [2]. Additionally, surface C atoms are negatively charged while surface metal atoms are positively charged, meaning that possible adsorption sites might interact very differently depending on the nature of the adsorbed species, making them more likely to break linear scaling relations [2].

Supporting small clusters of transition metals on TMCs (TM@TMCs) adds further complexity and functionality. In the last decade, TMCs have been shown to be excellent substrates for dispersing metallic particles, as they polarize the electron density of the supported particles in ways that can significantly enhance catalytic activity [2]. For instance, small Au, Cu, and Ni particles supported on TiC display very high activity for COâ‚‚ hydrogenation, orders of magnitude higher compared to extended Au(100), Cu(100), or Ni(100) surfaces [2]. The diversity of adsorption sites in TM@TMCs skyrockets due to their much more complex structure, creating more opportunities to circumvent conventional scaling relations.

Spin Manipulation and Magnetic Field Effects

Beyond geometric and electronic effects, spin manipulation has emerged as a promising strategy for enhancing catalytic performance. Magnetic catalysts can utilize spin alignment to promote specific reaction pathways, particularly for reactions involving triplet-state molecules like oxygen (Oâ‚‚) [20]. Recent work has extended this spin catalysis mechanism to singlet-state molecules like COâ‚‚, demonstrating that the parallel/antiparallel alignment of catalyst spins corresponds to bond breaking/forming processes in singlet molecules [20].

Experimental studies have shown that applying external magnetic fields can significantly enhance catalytic performance. For example, in the electrocatalytic coupling of CO₂ and NO₃⁻ to synthesize urea, an external magnetic field directly eliminated the plateau in the linear sweep voltammetry curve, resulting in a 2.2-fold increase in urea yield at lower overpotentials and nearly doubled Faradaic efficiency [20]. Similarly, the application of an alternating magnetic field to Cu-ZnMg ultrathin metal-organic framework (MOF) catalysts induced spin flip through spin-lattice relaxation, generating transient local energy that drove the cooperative proton-electron transfer kinetics in CO₂ reduction [20].

Data-Driven Catalyst Discovery

The integration of machine learning with high-throughput virtual screening represents a paradigm shift in catalyst discovery beyond scaling relation limitations. Active learning algorithms have been successfully applied to discover stable iridium oxide polymorphs for the oxygen evolution reaction, with the algorithm achieving at least a 2x higher discovery rate compared to random searches of the candidate space [18]. This approach has identified previously unknown structures, such as an FeF₃-type structure termed α-IrO₃, which was found to be globally stable under acidic OER conditions, replacing the stability of rutile IrO₂ [18].

For COâ‚‚ reduction reaction (COâ‚‚RR), researchers have developed a data-driven workflow combining binding energy prediction machine learning models with a COâ‚‚RR selectivity map to discover active and selective catalysts [21]. This approach evaluated the potential-dependent activity and selectivity of COâ‚‚RR for 465 binary combinations without performing time-consuming density functional theory calculations and surface structure modeling [21]. The method successfully predicted previously unreported promising behavior of Cu-Ga and Cu-Pd alloys, which were subsequently validated experimentally [21].

Experimental Characterization and Methodologies

Advanced Electrochemical Techniques

Understanding and manipulating scaling relations requires sophisticated experimental methodologies to probe the electrochemical interface and catalyst structure under operating conditions. Key techniques include:

• Cyclic Voltammetry (CV): Used to study redox processes and determine double-layer capacitance by scanning potential in regions with only non-Faradaic charging current. The double-layer capacitance can be extracted from the linear relationship between log(current density) and log(scan rate) at specific potentials [22].

• AC Voltammetry (ACV): Utilizes superimposed DC and AC signals to distinguish between double-layer structure (non-Faradaic component) and interfacial electrochemical reactions (Faradaic component). Fourier Transform ACV (FT-ACV) can intuitively display the response of Faradaic and non-Faradaic components to different frequency AC signals, enabling separation of these processes [22].

• Electrochemical Impedance Spectroscopy (EIS): A quasi-steady-state process that changes AC frequency at fixed electrode potential to obtain comprehensive information for accurate determination of double-layer capacitance. EIS can represent specific adsorption, Faradaic reactions, diffusion, and non-Faradaic processes as different circuit elements, clarifying the contribution of each process to the electrocatalytic mechanism [22].

The following diagram illustrates the key experimental workflows for characterizing scaling relations and catalyst performance:

In Situ and Operando Spectroscopic Techniques

To overcome the limitations of conventional electrochemical measurements, which primarily provide macroscopic electrical signals, researchers increasingly rely on in situ and operando spectroscopic techniques that can provide direct microscopic evidence of interfacial processes:

• In Situ Infrared Spectroscopy: Techniques including infrared reflection absorption spectroscopy (IRRAS), attenuated total reflection infrared spectroscopy (ATR-IR), and surface-enhanced infrared absorption spectroscopy (SEIRAS) utilize the characteristic infrared vibration signals of interfacial species to qualitatively identify the type, configuration, and dynamic changes of target species in the double layer under different voltages and reaction environments [22].

• In Situ Raman Spectroscopy: Particularly surface-enhanced Raman spectroscopy (SERS), can elucidate the types, structures, and orientations of target molecules in the double layer, as well as study the dynamic changes of target species with surface potential, charge state, and time. SERS has a wide wavenumber range (approximately 10-4000 cm⁻¹), giving it unique advantages in characterizing M-X bonds (M is metal, X is other atom/group) on catalyst surfaces [22].

• Operando X-ray Absorption Fine Structure (XAFS): Provides critical information about the local coordination environment and electronic structure of active sites under reaction conditions. For the Ni-Feâ‚‚ molecular catalyst, operando XAFS measurements verified the structural transformation from Ni monomer to O-bridged Ni-Feâ‚‚ trimer during the activation process, providing direct evidence of dynamic structural changes [1].

The following diagram illustrates the key signaling pathways and logical relationships in scaling relation manipulation strategies:

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Key Research Reagent Solutions for Studying Scaling Relations

Reagent/Material	Function/Application	Key Characteristics	Representative Examples
Single-Atom Precursors	Foundation for constructing well-defined active sites	High purity, controlled coordination environment	Ni-SAs@GNM (Ni single atoms trapped in graphene nanomesh) [1]
Metal Ion Dopants	Introducing secondary metals to create multi-functional sites	Controlled concentration, specific speciation	Fe(OH)₄⁻ in KOH electrolyte for Ni-Fe molecular complexes [1]
Transition Metal Carbides	Breaking scaling relations through diverse adsorption sites	Specific facet control, surface stoichiometry	TiC(001), ZrC(001), VC(001) with 50% C/50% metal surface atoms [2]
Supported Cluster Systems	Enhancing activity through metal-support interactions	Controlled cluster size, strong metal-support interaction	Au, Cu, Ni clusters on TiC(001) substrates [2]
Magnetic Catalyst Systems	Exploiting spin effects to enhance catalytic rates	Specific magnetic moments, field responsiveness	Cu-In₂O₃ dilute magnetic oxides, Cu-ZnMg MOF with isolated Cu sites [20]
Purified Electrolytes	Minimizing unintended contamination during activation	Ultra-low impurity levels, controlled ionic strength	Fe-free 1 M KOH with deliberate Fe addition at ppm levels [1]
Nitrobenzylthioinosine 5'-monophosphate	Nitrobenzylthioinosine 5'-monophosphate, CAS:65199-10-2, MF:C17H18N5O9PS, MW:499.4 g/mol	Chemical Reagent	Bench Chemicals
Nitrilotriacetic acid-d9	Nitrilotriacetic acid-d9, CAS:807630-34-8, MF:C6H9NO6, MW:200.19 g/mol	Chemical Reagent	Bench Chemicals

Over the past twenty years, our understanding of scaling relations in oxygen electrocatalysis has evolved from fundamental recognition to sophisticated manipulation strategies. The field has progressed from viewing these relations as fundamental limitations to developing innovative approaches for circumventing them through dynamic site regulation, materials design, and advanced characterization. The emerging paradigm recognizes that optimal catalytic performance requires moving beyond static catalyst models toward systems that can adapt their structure and properties under reaction conditions to differentially stabilize reaction intermediates.

Future research directions will likely focus on several key areas: First, the deliberate design of dynamic catalytic systems that can reconfigure their active sites during the catalytic cycle to overcome the limitations imposed by static scaling relations. Second, the integration of machine learning and automated experimentation to accelerate the discovery of materials that can break conventional scaling relations. Third, the exploration of multi-dimensional optimization strategies that simultaneously control electronic, geometric, and spin degrees of freedom to achieve unprecedented catalytic performance. Finally, the development of advanced operando characterization techniques with higher spatial, temporal, and energy resolution will provide deeper insights into the dynamic structural evolution of catalysts under operating conditions.

As we look toward the future, the systematic manipulation of scaling relations through dynamic structural regulation, materials engineering, and spin control represents a promising path for developing next-generation electrocatalysts that surpass the fundamental limitations currently imposed by conventional scaling relationships. These advances will be crucial for enabling the widespread adoption of energy conversion technologies essential for a sustainable energy future.

In catalytic reactions involving multiple intermediates, a fundamental constraint known as a linear scaling relationship (LSR) often governs the interaction energies between reactive species and catalytic active sites [1]. These relationships arise because the adsorption energies of different intermediates (such as *OH, *O, and *OOH in the oxygen evolution reaction) are typically correlated on conventional single-site catalysts [1]. While LSRs simplify performance prediction and help elucidate activity trends, they place intrinsic limitations on optimally adjusting the adsorption of every intermediate simultaneously to achieve maximum activity and/or selectivity [1]. The uniform and isolated active sites of single-site catalysts fall short in catalyzing complex chemical processes that simultaneously involve multiple intermediates with different optimal binding energy requirements [23] [24]. This review comprehensively examines the origin of these energetic correlations and the emerging strategies to circumvent them through multi-site catalytic architectures, providing both theoretical foundations and experimental methodologies for researchers working at the forefront of catalytic design.

Theoretical Foundations of Energetic Correlations

The Electronic Origin of Linear Scaling Relationships

Linear scaling relationships fundamentally originate from the similar chemical nature of catalytic intermediates that bind to the same active site. In multi-step reactions, the adsorption energies of chemically similar oxygenated intermediates (e.g., *OH, *O, and *OOH) demonstrate linear correlations because they interact with the catalytic center through comparable electronic mechanisms [1]. For instance, in the oxygen evolution reaction (OER) following the widely accepted adsorbate evolution mechanism, the universal linear scaling relationship between *OOH and *OH adsorption energies creates an inherent thermodynamic overpotential ceiling [1].

The electronic structure of the active site dictates these correlations. In single-site catalysts (SSCs), the uniform coordination environment produces a single type of active site with characteristic binding properties. While this uniformity offers advantages in fundamental studies and specific applications, it creates the scaling relationship limitation [25] [23]. The adsorption energy correlations emerge because each intermediate interacts with the same local electronic environment, making it nearly impossible to independently optimize the binding strength for all required intermediates in a complex reaction pathway [1].

Table 1: Fundamental Scaling Relationships in Key Catalytic Reactions

Reaction	Intermediates Involved	Scaling Relationship Impact	Theoretical Overpotential Limit
Oxygen Evolution Reaction (OER)	*OH, *O, *OOH	Correlated *OOH and *OH adsorption	~0.37 V [1]
Oxygen Reduction Reaction (ORR)	*OOH, *O, *OH	Similar to OER, reverse pathway	Activity volcano relationships [24]
COâ‚‚ Reduction (CO2RR)	*COOH, *CO, *CHO	*COOH vs *CO scaling	Selectivity limitations [24]
Hydrogen Evolution (HER)	*H	Minimal intermediate scaling	Lower impact from LSRs [26]

Quantum Mechanical Underpinnings

The electronic interactions governing scaling relationships have profound quantum mechanical origins. Catalytic active sites with open-shell orbital configurations exhibit distinctive quantum behaviors, including non-weak (strong) electronic correlations and various electronic orders such as spin-orbital interactions [27]. These quantum catalysts demonstrate properties that cannot be fully described by classical interactions or mean-field approximations [27]. The rivalry between different quantum interactions—specifically quantum spin exchange interaction (QSEI) and quantum excitation interactions (QEXI)—forms the electronic background that explains the properties of quantum materials and their catalytic behavior [27].

In single-site catalysts, the uniform coordination environment creates a consistent quantum interaction landscape with adsorbates, leading to the characteristic scaling relationships. Breaking these relationships requires introducing heterogeneity in these quantum interactions, which can be achieved through multi-site catalytic designs where different intermediates can interact with distinct electronic environments [23] [27].

Single-Site Catalysts: Strengths and Limitations

Coordination Engineering in Single-Site Catalysts

Single-site catalysts (SSCs) represent an important class of catalytic materials characterized by isolated, uniform active sites. Coordination engineering has emerged as an efficient approach for editing the local microenvironment of SSCs to optimize their activity, selectivity, and stability [25]. Through strategies including size control, ligand modification, post-synthesis treatment, surface modification, and deliberate coordination environment design, researchers can systematically regulate the physicochemical properties of SSCs [25]. The precisely defined active sites in SSCs facilitate mechanistic studies and structure-performance correlations, bridging the gap between homogeneous and heterogeneous catalysis [26].

The construction of high-performance SSCs often relies on enhancing metal-support interactions. For precious metals, this typically involves introducing dopants or defects to the supports or confining single atoms inside porous supports [26]. For instance, hierarchical nitrogen-doped carbon nanocages (hNCNC) with coexisting micro-meso-macro pore structures and high nitrogen content have demonstrated exceptional capability for stabilizing single platinum atoms through the synergistic effect of micropore trapping and nitrogen anchoring [26].

Inherent Limitations from Scaling Relationships

Despite their high atomic utilization and well-defined structures, SSCs face fundamental limitations in complex multi-step reactions due to linear scaling relationships. The uniform active sites in SSCs are unfavorable for multi-elementary reactions because they display different adsorption energies for multiple intermediates [24]. While they may accelerate one specific step of a reaction (potentially the rate-determining step), they cannot independently optimize all steps simultaneously [24]. This limitation becomes particularly pronounced in reactions involving multiple intermediates with different chemical characteristics, such as OER and CO2RR [1] [24].

Table 2: Performance Comparison of Single-Site vs. Multi-Site Catalysts

Catalyst Characteristic	Single-Site Catalysts	Multi-Site Catalysts
Active Site Uniformity	High [26]	Variable/Low [23]
Atomic Efficiency	Maximum [26]	High [24]
Scaling Relationship Behavior	Follow LSRs [1]	Can break LSRs [1] [23]
Multi-Intermediate Reaction Capability	Limited [24]	Enhanced [23] [24]
Structural Characterization Complexity	Moderate [26]	High [1]
Typical Synthesis Approach	Impregnation-adsorption, pyrolysis [26]	In situ electrochemical activation [1]

Multi-Site Catalysis: Disrupting Energetic Correlations

Integrative Catalytic Pairs and Dynamic Regulation

An emerging class of catalysts with adjacent binary active centers, termed integrative catalytic pairs (ICPs), demonstrates the ability to overcome the limitations of SSCs through site-to-site electronic interactions and synergistic catalytic effects [23]. Unlike conventional single-site catalysts or their derivative dual-atom catalysts (DACs), ICPs can accommodate multi-interactive intermediates that overcome kinetic barriers, adjust reaction pathways, and break universal linear scaling relationships [23].

A groundbreaking demonstration of this approach involves the dynamic structural regulation of active sites in a Ni-Fe₂ molecular catalyst during oxygen evolution reaction [1]. Theoretical calculations and electrokinetic studies revealed that the dynamic evolution of Ni-adsorbate coordination, driven by intramolecular proton transfer, effectively alters the electronic structure of the adjacent Fe active center during the catalytic cycle [1]. This dynamic dual-site cooperation simultaneously lowers the free energy change associated with O–H bond cleavage and O–O bond formation, thereby disrupting the inherent scaling relationship in OER [1].

Diagram 1: Dynamic dual-site cooperation mechanism in Ni-Fe integrative catalytic pairs, showing how proton transfer triggers coordination evolution that enhances electronic coupling between sites.

Synergistic Composite Catalysts

Combining single atoms with clusters or nanoparticles represents another effective strategy to design efficient electrocatalysts that overcome scaling relationships [24]. These synergistic composite catalysts leverage both the high atomic utilization of single-atom sites and the complementary functionality of clusters or nanoparticles. The interaction between different catalytic components enables electron transfer and modulation of electronic structures, leading to enhanced catalytic activity, longevity, and improved reaction dynamics while maintaining high atomic dispersion characteristics [24].

For example, integrating Fe-Nₓ atomic centers with Fe/Fe₃C nanoparticles creates a catalyst where the nanoparticles serve as electronic regulators that modify the coordination environment of the single-atom sites, optimizing energy barriers in the reaction pathway [24]. Similarly, coupling Co single atoms with Co nanoparticles in nanofiber structures has demonstrated exceptional oxygen reduction reaction performance, with the nanoparticles serving as accelerators for the single-atom sites [24].

Experimental Methodologies and Characterization

Synthesis Protocols for Advanced Catalytic Systems

Protocol 1: Construction of Ni-Fe Molecular Complex Catalyst via In Situ Electrochemical Activation

• Pre-catalyst Synthesis: Prepare single-atom Ni pre-catalyst by sealing an aqueous suspension of graphene oxide (GO) in a Ni vessel at 80°C to spontaneously assemble a 3D Ni(OH)â‚‚/graphene hydrogel [1].
• Freeze-Drying: Subject the resultant aerogel to freeze-drying to preserve the porous structure [1].
• Thermal Annealing: Anneal the dried material at 700°C under Ar atmosphere to reduce Ni(OH)â‚‚ to Ni/NiO nanoparticles and simultaneously etch graphene nanosheet into holey graphene nanomesh (GNM) [1].
• Acid Treatment: Apply acid treatment to remove nanoparticles and obtain Ni single atoms trapped in GNM (Ni-SAs@GNM) [1].
• Electrochemical Activation: Employ a standard three-electrode system with Ni-SAs@GNM pre-catalyst loaded onto a glassy carbon working electrode. Perform activation using cyclic voltammetry between 1.1 and 1.65 V versus RHE in purified 1 M KOH electrolyte with deliberate addition of 1 ppm Fe ions [1].
• Alternative Activation Methods: Anodic chronopotentiometry and chronoamperometry can also be employed for activation if CV is not suitable [1].

Protocol 2: Synergistic Single Atom-Cluster Catalyst via Impregnation and Pyrolysis

• Support Preparation: Synthesize spherical superstructure carbon nanorods (SSCNRs) derived from superstructure MOF nanorods with high nitrogen content and rich defect structure [24].
• Metal Impregnation: Introduce manganese species via impregnation of metal precursors onto the SSCNR support [24].
• Secondary Pyrolysis: Perform high-temperature pyrolysis (typically 700-900°C in inert atmosphere) to form MnSA/MnAC-SSCNRs with both single atoms and atomic clusters [24].
• Characterization: Confirm the formation of both single atoms and clusters through aberration-corrected HAADF-STEM and X-ray absorption spectroscopy [24].

Advanced Characterization Techniques

Operando characterization techniques are essential for understanding the dynamic structural evolution of catalytic sites under working conditions. Operando X-ray absorption fine structure (XAFS) measurements have proven particularly valuable for probing the local structures of metal atoms and incorporation dynamics during activation processes [1]. For the Ni-Fe molecular complex catalyst, operando XAFS verified the structural transformation from Ni monomer to O-bridged Ni-Feâ‚‚ trimer during the electrochemical activation process [1].

Complementary techniques including synchrotron-based X-ray fluorescence (SXRF) spectroscopy, aberration-corrected high-angle annular dark-field scanning TEM (HAADF-STEM), and X-ray photoelectron spectroscopy (XPS) provide comprehensive information about elemental distribution, atomic dispersion, and oxidation states [1] [26]. Electrokinetic studies combined with density functional theory (DFT) and ab initio molecular dynamics (AIMD) simulations offer insights into reaction mechanisms and energy landscapes [1].

Diagram 2: Integrated experimental-computational workflow for characterizing single-site and multi-site catalysts, showing the relationship between synthesis, characterization techniques, and theoretical validation.

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Key Research Reagent Solutions for Catalyst Synthesis and Evaluation

Reagent/Material	Function in Research	Application Examples
Graphene Oxide (GO) Suspension	3D framework formation	Support for single-atom pre-catalysts [1]
Metal-Nitrogen-Carbon Precursors (ZIF-8, ZIF-67)	Single-atom catalyst supports	Pyrolysis to form M-N-C catalysts [24]
Fe-free KOH Electrolyte	Electrochemical activation	Purified electrolyte for controlled Fe incorporation [1]
Fe Ion Solutions (1-10 ppm)	Deliberate heteroatom introduction	In situ formation of bimetallic sites [1]
Polydopamine (PDA) Coating	Surface functionalization	Enhanced metal adsorption on supports [24]
H₂PtCl₆ and Similar Metal Salts	Single-atom precursors	Impregnation-adsorption synthesis [26]
Mesoporous SiOâ‚‚ Templates	Controlled porosity creation	Hard templates for hierarchical structures [24]
Isopropylidenylacetyl-marmesin	Isopropylidenylacetyl-marmesin, CAS:35178-20-2, MF:C19H20O5, MW:328.4 g/mol	Chemical Reagent
Ciproxifan hydrochloride	Ciproxifan hydrochloride, CAS:1049741-81-2, MF:C16H19ClN2O2, MW:306.79 g/mol	Chemical Reagent

Computational Approaches and Machine Learning

Modeling Dynamic Catalytic Processes

Computational methods play an indispensable role in understanding and predicting the behavior of single-site and multi-site catalysts. Density functional theory (DFT) calculations provide insights into electronic structures, adsorption energies, and reaction pathways [1] [26]. For dynamic catalytic systems, ab initio molecular dynamics (AIMD) simulations can capture structural evolution and proton transfer processes under reaction conditions [1].

Advanced neural network potentials (NNPs) have emerged as powerful tools that bridge the gap between computational accuracy and efficiency. Frameworks like EMFF-2025 enable molecular dynamics simulations of complex catalytic systems with DFT-level accuracy while being more efficient than traditional quantum chemical methods [28]. These approaches are particularly valuable for studying the thermal stability and reaction mechanisms of catalytic materials under realistic conditions [28].

Data-Driven Catalyst Design

Machine learning approaches integrated with computational and experimental data are accelerating the discovery and optimization of advanced catalysts. Transfer learning strategies leverage existing data to reduce the need for extensive training, accelerating learning and improving predictive performance [28]. Graph neural network (GNN)-based approaches effectively enhance accuracy and extrapolation capabilities by incorporating physical symmetries such as translation, rotation, and periodicity [28].

These data-driven methods are particularly valuable for exploring the complex chemical space of multi-site catalysts, where multiple active centers and their interactions create a high-dimensional parameter space that challenges traditional experimental approaches [28] [24]. Principal component analysis (PCA) and correlation heatmap analysis can help identify intrinsic relationships and formation mechanisms of structural motifs in catalytic materials [28].

The fundamental distinction between single-site and multi-site catalysis lies in their ability to manipulate the energetic correlations that govern multi-step catalytic reactions. While single-site catalysts offer exceptional uniformity and atomic efficiency, they are inherently constrained by linear scaling relationships that prevent independent optimization of all steps in complex reactions [1] [23]. Multi-site catalytic systems, including integrative catalytic pairs and synergistic composite catalysts, disrupt these scaling relationships through dynamic structural regulation, site-to-site electronic interactions, and complementary functions [1] [23] [24].

Future advancements in catalytic design will likely focus on precisely engineered multi-site architectures that leverage dynamic processes under reaction conditions. The integration of advanced operando characterization techniques with machine learning-assisted computational methods will enable unprecedented understanding and control of these complex catalytic systems [1] [28]. As these approaches mature, we anticipate the development of next-generation catalysts that transcend traditional scaling relationship limitations, enabling more efficient energy conversion processes and sustainable chemical synthesis pathways.

Methodologies for Mapping Reaction Pathways: From Computational Predictions to Experimental Validation

Density Functional Theory (DFT) has established itself as a cornerstone computational method in catalysis research, providing an atomic-level understanding of reaction mechanisms that are often difficult to probe experimentally [29]. Its exceptional value lies in the optimal compromise between computational cost and accuracy, enabling researchers to map complex reaction pathways by predicting the structures and energies of intermediate species [30]. This capability is fundamental to the development and refinement of scaling relationships—theoretical constructs that relate the binding energies of diverse catalytic intermediates across different catalyst surfaces [31]. These relationships, derived from bond order conservation principles, allow for the prediction of catalytic activity and the rational design of new catalysts [31]. This guide details the application of DFT for determining the intermediate structures and energetics that underpin these essential scaling relationships.

Theoretical Foundations of DFT in Catalysis

Basic Principles

DFT is a first-principles electronic structure method that uses the electron density, ρ(r), as its fundamental variable instead of the many-electron wavefunction [30]. This approach is computationally efficient because the electron density depends on only three spatial coordinates. The foundation of modern DFT rests on the Hohenberg-Kohn theorems, which state that the ground-state electron density uniquely determines all properties of a system, including its energy [30]. The practical implementation of DFT typically uses the Kohn-Sham scheme, which introduces a fictitious system of non-interacting electrons that has the same electron density as the real system [30].

Key Concepts for Intermediate Analysis

For catalysis research, several DFT-derived properties are particularly crucial for establishing scaling relationships:

• Adsorption Energies: The energy change when an intermediate binds to a catalyst surface serves as a primary descriptor of catalytic activity [30] [31]. Accurate calculation of adsorption energies for various intermediates allows researchers to construct scaling relationships that predict binding energies across different catalyst surfaces.

• Reaction Energies and Barriers: DFT enables calculation of the energy changes between intermediates along a reaction pathway and the activation barriers that must be overcome [30]. According to the Brønsted-Evans-Polanyi (BEP) relation, these energy barriers often scale approximately linearly with the adsorption energies of key molecules [30].

• Electronic Descriptors: Properties such as the d-band center for transition metal catalysts have been proven effective as descriptors for rationalizing and predicting electrocatalytic activity [30].

Computational Protocols and Methodologies

Best-Practice DFT Protocols

Selecting appropriate computational parameters is crucial for obtaining reliable results. The field has moved beyond outdated method combinations like B3LYP/6-31G*, which suffer from inherent errors including missing London dispersion effects and significant basis set superposition error (BSSE) [32]. Contemporary best practices include:

• Robust Functional/Basis Set Combinations: Modern approaches emphasize method combinations that provide an optimal balance of accuracy, robustness, and efficiency [32]. Composite methods such as B3LYP-3c, r2SCAN-3c, and B97M-V/def2-SVPD have been developed to eliminate systematic errors without substantially increasing computational cost [32].

• Dispersion Corrections: Proper accounting for London dispersion forces is essential for accurate description of adsorption phenomena, typically achieved through established dispersion corrections [32].

• Solvation Models: Incorporating solvation effects through implicit or explicit solvation models is critical for modeling electrocatalytic and liquid-phase reactions [32].

Table 1: Recommended DFT Methodologies for Different Catalytic Systems

System Type	Recommended Functional	Basis Set	Key Considerations
Homogeneous Catalysts	B3LYP-D3, ωB97X-D	def2-TZVP	Include solvation models; verify single-reference character [32]
Metallic Surfaces	BEEF-vdW, RPBE	Plane waves (400-500 eV cutoff)	Use appropriate k-point sampling; model sufficient slab layers [33]
Single-Atom Catalysts	PBE-D3, SCAN	def2-SVP for atoms, plane waves for support	Account for support effects; check for multi-reference character [34]
Oxide Catalysts	HSE06, PBE+U	Plane waves (500-600 eV cutoff)	Apply Hubbard U correction for transition metals; model surface termination [31]

Model System Selection

Constructing an appropriate model of the catalytic system requires careful consideration:

• Active Site Representation: The model must accurately capture the essential chemistry of the active site while remaining computationally tractable. For homogeneous catalysts, this includes the complete coordination sphere, while for surfaces, a slab model with sufficient layers and vacuum is necessary [30].

• Periodic vs. Cluster Models: Periodic models using plane-wave basis sets are typically preferred for extended surfaces, while atomic-centered basis sets with cluster models may be suitable for molecular systems or localized active sites [30].

• Size Considerations: Models must be large enough to avoid finite-size effects, particularly for reactions where the interaction between adjacent intermediates or defects plays a significant role [30].

Workflow for Intermediate Structure Optimization

A systematic approach to determining intermediate structures involves:

• Initial Structure Generation: Based on chemical intuition, analogous structures, or automated sampling methods.

• Geometry Optimization: Energy minimization with respect to nuclear coordinates until forces converge below a threshold (typically 0.01-0.05 eV/Ã…).

• Frequency Analysis: Calculation of vibrational frequencies to confirm stationary points as minima (all real frequencies) and to provide zero-point energy and thermal corrections.

• Electronic Energy Analysis: Single-point energy calculations on optimized structures to obtain precise energies using higher-level methods if necessary.

• Reaction Pathway Validation: Ensuring intermediates connect appropriately through transition states via intrinsic reaction coordinate (IRC) calculations.

The following workflow diagram illustrates the process of using DFT-derived data to establish scaling relationships in catalysis research:

Application to Scaling Relationships

Establishing Scaling Relationships

Scaling relationships are powerful tools that emerge from the systematic DFT analysis of reaction intermediates across different catalytic surfaces [31]. The process typically involves:

• Systematic Binding Energy Calculations: Computing adsorption energies for key intermediates across a range of similar catalyst materials (e.g., different transition metals) [31].

• Correlation Analysis: Identifying linear relationships between the binding energies of different intermediates, which often arises because the bonding mechanism (e.g., through carbon, oxygen, or nitrogen atoms) remains consistent across surfaces [31].

• Descriptor Identification: Finding the key electronic or geometric properties (e.g., d-band center, coordination number) that correlate with the binding energies, enabling predictions for new materials [30] [31].

For example, in the oxygen evolution reaction (OER) on single-atom catalysts, new scaling relationships have been developed that connect the energies of different intermediates, allowing for more accurate predictions of catalytic activity [34].

Volcano Plots and Activity Optimization

Scaling relationships naturally lead to the construction of volcano plots, which relate catalytic activity to descriptor variables:

• Plot Construction: The rate of a catalytic reaction is plotted against a descriptor variable (typically the binding energy of a key intermediate) [31].

• Optimization Principle: The left side of the volcano is typically limited by over-strong binding, while the right side is limited by over-weak binding, with the peak representing the optimal binding energy [31].

• Design Guidance: Volcano plots identify promising catalyst materials and reveal the theoretical limits of catalyst classes, guiding the search for new materials that overcome these limitations [31].

Table 2: Key Intermediate Descriptors for Common Catalytic Reactions

Reaction	Key Intermediate	Primary Descriptor	Scaling Relationship
Oxygen Evolution	O, OH, OOH*	Binding energy of O*	ΔGOOH* = ΔGOH* + constant [34]
CO2 Reduction	COOH, CO, CHO*	Binding energy of CO*	ΔGCOOH* vs. ΔGCO* [35]
Syngas Conversion	CH, CH2, CH3*	Binding energy of C*	Linear correlations across metals [33]
Nitrate Reduction	NO, NOH, NH*	Binding energy of N*	Depends on active site [36]

Advanced Applications and Combinatorial Approaches

DFT with Machine Learning

The integration of DFT with machine learning (ML) techniques has emerged as a powerful strategy for addressing complexity in catalytic reaction networks:

• Surrogate Models: ML models trained on DFT data can predict adsorption energies and reaction barriers for similar systems without performing full DFT calculations, dramatically accelerating screening processes [33] [35].

• Feature Engineering: Selection of appropriate input features (e.g., elemental properties, geometric descriptors, electronic structure parameters) is crucial for ML model accuracy [35].

• Uncertainty Quantification: Advanced frameworks like those using Bayesian error estimation functionals (BEEF-vdW) provide uncertainty estimates for DFT calculations, which can be propagated through ML models to assess prediction reliability [33].

In CO2 reduction reaction studies on transition metal-doped r-GeSe monolayers, combining DFT with XGBoost machine learning achieved excellent prediction accuracy (R² = 0.92) for intermediate free energies, enabling rapid screening of single-atom catalysts [35].

Microkinetic Modeling

DFT-derived parameters feed into microkinetic models that simulate overall reaction rates and selectivity:

• Parameter Integration: DFT-calculated activation barriers, reaction energies, and adsorption strengths serve as inputs for microkinetic models [33].

• Mechanism Validation: Comparing model predictions with experimental kinetics helps validate proposed reaction mechanisms and identify rate-determining steps [33].

• Selectivity Analysis: Microkinetic models can predict how changes in catalyst composition affect not only activity but also product distribution [36].

For complex reaction networks such as syngas conversion on Rh(111), this approach has demonstrated reductions of 60% in intermediate calculations and 95% in transition-state calculations while maintaining predictive accuracy [33].

The Scientist's Toolkit

Table 3: Essential Computational Tools for DFT-Based Catalysis Research

Tool Category	Specific Examples	Function	Application Notes
DFT Codes	VASP, Quantum ESPRESSO, Gaussian, ORCA	Electronic structure calculation	Periodic codes for surfaces; molecular codes for complexes [30]
Structure Analysis	ASE, pymatgen, ChemCraft	Structure manipulation and analysis	Interface between different code formats; automated workflow management
Transition State Search	Dimer, NEB, CI-NEB	Locating saddle points on PES	CI-NEB is standard for surface reactions; verify with frequency calculations [33]
Machine Learning	Scikit-learn, XGBoost, Gaussian Processes	Building surrogate models	Accelerate screening; quantify uncertainty [33] [35]
Data Analysis	matplotlib, pandas, NumPy	Data visualization and analysis	Create volcano plots; analyze scaling relationships [31]
Fingolimod Hydrochloride	Fingolimod Hydrochloride, CAS:162359-56-0, MF:C19H34ClNO2, MW:343.9 g/mol	Chemical Reagent	Bench Chemicals
11-Oxahomoaminopterin	11-Oxahomoaminopterin, CAS:78520-72-6, MF:C20H21N7O6, MW:455.4 g/mol	Chemical Reagent	Bench Chemicals

DFT calculations provide the fundamental data on intermediate structures and energetics necessary to establish scaling relationships in catalysis research. Following best-practice protocols for functional selection, model construction, and computational workflow ensures the reliability of these data. The integration of DFT with machine learning and microkinetic modeling creates a powerful framework for elucidating complex reaction networks and accelerating catalyst design. As these computational approaches continue to evolve, they will enable increasingly accurate predictions of catalytic behavior and contribute to the development of more efficient and selective catalysts for sustainable energy and chemical processes.

Microkinetic Modeling (MKM) and Energy Barrier Analysis for Simulating Catalytic Activity

Microkinetic analysis plays an indispensable role in modern catalyst design because it provides critical insight into the fundamental surface chemistry that controls catalyst performance, bridging the gap between quantum mechanical calculations and experimental observables [37]. This computational approach allows researchers to deconstruct complex catalytic reactions into their constituent elementary steps, then simulate the system's time evolution under realistic working conditions [38]. When coupled with scaling relationships—mathematical correlations that describe how the energies of adsorbed intermediates and transition states vary with catalyst composition—microkinetic modeling (MKM) transforms from a descriptive tool to a predictive framework for rational catalyst design [37]. These scaling relations enable researchers to systematically explore vast catalyst design spaces by reducing the number of independent parameters needed to describe catalytic performance, creating powerful volcano-shaped activity plots that guide the search for optimal catalysts [37] [39].

The integration of density functional theory (DFT) with microkinetic modeling has proven particularly valuable for understanding and optimizing catalytic processes relevant to sustainability and clean energy [38] [39]. For instance, in CO₂ hydrogenation over Ni catalysts—a process crucial for converting greenhouse gases into valuable fuels—microkinetic modeling has helped resolve conflicting mechanistic interpretations by revealing how surface coverage and lateral interactions influence the dominant reaction pathways [38]. Similarly, in methane conversion over doped nanocarbon catalysts, MKM has identified the most efficient edge decorations and predicted their resistance to carbon poisoning, guiding the development of metal-free catalysts for hydrogen production [39]. This review examines the theoretical foundations, methodological framework, and practical applications of microkinetic modeling, with particular emphasis on its integration with scaling relationships and energy barrier analysis for accelerating catalyst discovery.

Theoretical Foundations of Microkinetic Modeling

Fundamental Principles and Equations

Microkinetic modeling operates on the principle that complex catalytic reactions can be understood as networks of elementary steps, including adsorption, surface reaction, and desorption processes [38]. The core mathematical framework relies on transition state theory (TST) to calculate rate constants for each elementary step based on energetic parameters typically derived from DFT calculations [38]. For a generic surface reaction A* + B* → C* + D*, where * denotes a surface site, the rate expression takes the form:

[ r = k \thetaA \thetaB ]

where ( \thetaA ) and ( \thetaB ) represent surface coverages of species A and B, and k is the rate constant determined by TST [38]. The mean-field approximation, which assumes uniform distribution of adsorbates across the catalyst surface, significantly simplifies these calculations by neglecting local spatial correlations between adsorbed species [38]. This approximation works well for systems with low surface coverages where adsorbate-adsorbate interactions are minimal, but becomes less accurate under high-coverage conditions where lateral interactions significantly influence reaction kinetics [38].

A critical requirement for developing physically meaningful microkinetic models is maintaining stoichiometric and thermodynamic consistency throughout the reaction network [37]. This ensures that all possible reaction pathways obey microscopic reversibility and that the simulated system converges to the correct thermodynamic equilibrium under appropriate conditions. The growing integration of scaling relationships into microkinetic models has further enhanced their predictive power by establishing mathematical correlations between adsorption energies of different intermediates, thereby reducing the parameter space that must be explored in catalyst optimization studies [37].

Comparison with Kinetic Monte Carlo Methods

While microkinetic modeling employs a mean-field approach that averages surface coverages across the catalyst, kinetic Monte Carlo (kMC) simulations provide a spatially resolved alternative that explicitly tracks the arrangement of adsorbates on discrete surface sites [38]. This fundamental difference becomes particularly significant when modeling systems with high surface coverages, where adsorbate-adsorbate interactions and the topological arrangement of species strongly influence reaction rates [38].

Comparative studies of COâ‚‚ hydrogenation on Ni(111) reveal that for systems with low coverage of adsorbed species, MKM provides results very similar to the more computationally intensive kMC simulations but with significantly lower computational cost [38]. However, as surface coverage increases, the mean-field approximation becomes increasingly inadequate, and kMC simulations become essential for accurately capturing the system's behavior [38]. Precisely, the use of kMC simulations is of paramount importance when dealing with systems with high coverages in which adsorbate-adsorbate interactions and the topological arrangement of the adsorbates are more influential [38]. Hybrid approaches that incorporate lateral interactions into MKM frameworks can partially bridge this gap, yielding results that align more closely with kMC simulations, though local considerations in kMC still lead to differences in predicted macroscopic properties and dominant reaction mechanisms [38].

Computational Workflow for Microkinetic Analysis

Data Generation via Density Functional Theory

The foundation of any reliable microkinetic model lies in accurate energetic parameters for all relevant surface species and transition states, typically obtained through density functional theory calculations [39]. These quantum mechanical computations enable researchers to determine adsorption energies, reaction barriers, and vibrational frequencies for each elementary step in the catalytic cycle. For instance, in studying methane conversion over edge-decorated nanocarbons, DFT calculations reveal how heteroatom doping (N, B, P, Si) alters the electronic structure of active sites and modifies energy barriers for C-H bond activation [39]. The p-band center of surface oxygen atoms serves as a particularly useful descriptor for predicting catalytic activity, with nitrogen-substituted edges shifting this center closer to the Fermi level and significantly lowering activation barriers for rate-limiting steps [40].

Table 1: Key DFT-Calculated Parameters for Microkinetic Modeling

Parameter Type	Specific Properties	Application Example
Adsorption Energies	Binding strengths of intermediates on active sites	COâ‚‚ on Ni(111), CHâ‚“ on doped nanocarbons [38] [39]
Transition State Energies	Activation barriers for elementary steps	C-H bond cleavage in methane, CO hydrogenation [39]
Electronic Properties	p-band or d-band centers, Bader charges	p-band center of oxygen in MXenes [40]
Geometric Parameters	Bond lengths, active site coordination	M-X and M-O bond distances in MXenes [40]
Vibrational Frequencies	Entropic contributions to free energy	Pre-exponential factors in rate constants [38]

Model Construction and Simulation Techniques

Constructing a microkinetic model involves integrating DFT-derived parameters into a system of differential equations that describe the time evolution of surface coverages and reaction rates [38]. The open-source Cantera toolkit has emerged as a powerful platform for implementing these models, recently extended to include surface diffusion between facets in multifaceted nanoparticles—a critical advancement for modeling structure-sensitive reactions like CO₂ hydrogenation on Ni catalysts [41]. This framework allows researchers to simulate temperature-programmed desorption profiles, steady-state reaction rates, and product distributions under various operating conditions [41].

Degree of rate control (DRC) analysis represents a particularly valuable technique within the microkinetic modeling toolkit, as it identifies which transition states and adsorbed intermediates exert the greatest influence on the overall reaction rate [39]. By systematically varying the activation barriers for each elementary step while maintaining thermodynamic consistency, researchers can pinpoint the kinetic bottlenecks that limit catalyst performance and guide targeted improvements [37]. This approach naturally leads to the derivation of general reaction kinetics rate expressions in terms of changes in binding energies of the relevant transition states and intermediates, facilitating catalyst optimization through scaling relationships [37].

Case Studies in Catalytic Energy Barrier Analysis

COâ‚‚ Hydrogenation on Ni Surfaces

The hydrogenation of CO₂ to value-added chemicals represents a strategically important reaction for sustainable energy storage and greenhouse gas mitigation. Microkinetic modeling has proven instrumental in unraveling the complex reaction mechanism over Ni catalysts, where structure sensitivity and operating conditions dramatically influence product selectivity [38] [41]. Comparative studies employing both microkinetic modeling and kinetic Monte Carlo simulations on Ni(111) have revealed that the apparent discrepancies between earlier theoretical studies—where one predicted methane formation and another only CO production—stemmed primarily from different mechanistic assumptions rather than the kinetic methods themselves [38]. When surface coverages remain low, both methods yield similar predictions, but under high-coverage conditions typical of industrial operation, the mean-field approximation breaks down, and spatially resolved kMC simulations become necessary to accurately capture the system's behavior [38].

Recent advancements in structure-dependent microkinetic modeling have further demonstrated the critical importance of considering multiple crystal facets and surface diffusion in Ni nanoparticles [41]. By extending the Cantera toolkit to include surface diffusion between facets, researchers developed a thermodynamically consistent microkinetic model for COâ‚‚ temperature-programmed desorption that revealed Ni(110) facets, despite contributing minimally to the total surface area, dominate the desorption pattern due to their unique catalytic properties [41]. This multifaceted approach, coupled with rigorous uncertainty quantification, achieved excellent agreement with experimental desorption profiles and highlighted the essential role of surface diffusion in distributing adsorbates between facets with different catalytic activities [41].

Table 2: Microkinetic Analysis of COâ‚‚ Hydrogenation on Ni Catalysts

Catalyst System	Key Findings from Microkinetic Analysis	Methodological Insights
Ni(111) facet	Low-coverage: MKM and kMC yield similar results; Mechanism sensitive to allowed products [38]	MKM suitable for low coverage; kMC needed for high coverage [38]
Multifacet Ni nanoparticle	Ni(110) dominates COâ‚‚ desorption despite small surface area; Surface diffusion crucial [41]	Inclusion of surface diffusion in Cantera significantly improves model agreement [41]
Ni-based catalysts under methanation conditions	High coverages alter dominant reaction pathways; Lateral interactions critical [38]	MKM with lateral potential aligns more closely with kMC results [38]

Methane Conversion on Doped Nanocarbon Catalysts

The direct conversion of methane to hydrogen and light hydrocarbons represents an attractive alternative to steam reforming for valorizing natural gas resources while reducing COâ‚‚ emissions. Microkinetic modeling combined with DFT calculations has revealed the exceptional potential of edge-decorated nanocarbons (EDNCs) as metal-free catalysts for this demanding transformation [39]. Systematic investigation of nitrogen-, boron-, phosphorus-, and silicon-doped zigzag graphene edges demonstrated that nitrogen decoration (N-EDNC) delivers outstanding performance for Hâ‚‚ production at temperatures above 900 K, followed by P-EDNC, Si-EDNC, and B-EDNC [39]. The microkinetic analysis further uncovered a shift from Langmuir-Hinshelwood to Eley-Rideal mechanisms for hydrogen formation under certain conditions, with phosphorus-doped edges exhibiting particularly high activity in the Eley-Rideal regime [39].

Degree of rate control analysis identified C-H bond cleavage in methane as the rate-determining step for most EDNCs, while the combined microkinetic and DFT approach enabled prediction of coke resistance—a critical factor for catalyst longevity [39]. The analysis revealed that N-EDNC and P-EDNC active sites display strong resistance to carbon poisoning, whereas Si-EDNC shows higher propensity to regenerate its active sites at temperatures exceeding 1100 K [39]. These insights demonstrate how microkinetic modeling moves beyond identifying active catalysts to predicting operational stability and regeneration behavior under realistic process conditions.

Propane Dehydrogenation on MXene Catalysts

Propane dehydrogenation (PDH) to propylene has gained significant attention as a specialized process for bridging the "propylene gap" between supply and demand in the petrochemical industry. Recent research combining DFT with microkinetic simulations has demonstrated how strategic substitution of the X-layer in MXenes (transition metal carbides, nitrides, and carbonitrides) precisely tunes their catalytic performance for PDH [40]. Nitrogen substitution in V₃N₂O₂ MXene was found to shift the p-band center of surface oxygen atoms closer to the Fermi level, significantly lowering the energy barrier for the first C-H bond activation in propane [40]. Microkinetic analysis confirmed that nitrogen-substituted MXenes exhibit the highest turnover frequencies for both propane conversion and propylene formation, outperforming their carbon-based counterparts [40].

The microkinetic models further revealed that modifying the X-layer composition simultaneously affects both geometric parameters (M-X and M-O bond distances) and electronic structure (charge distribution at active sites) [40]. This dual influence creates opportunities for fine-tuning catalytic performance by controlling the vertical stacking sequence of carbon and nitrogen layers along the c-axis, with TM-N/C (nitrogen-over-carbon) configurations generally outperforming TM-C/N (carbon-over-nitrogen) arrangements [40]. These findings establish X-layer substitution as a powerful strategy for optimizing MXene-based catalysts and highlight how microkinetic modeling enables rational design of two-dimensional materials for selective alkane dehydrogenation.

Experimental Protocols and Computational Methodologies

Density Functional Theory Calculations

Surface Models and Computational Parameters First-principles calculations typically employ periodic slab models with sufficient vacuum spacing (typically 15-20 Ã…) to prevent spurious interactions between periodic images [39]. For metal surfaces like Ni(111), a 3-4 layer slab model with the bottom layers fixed in their bulk positions adequately represents the catalytic surface while maintaining computational efficiency [38]. For two-dimensional materials like graphene edges or MXenes, single-layer models with appropriate supercell sizes capture the essential chemistry while accommodating adsorbates without excessive computational cost [39] [40]. The Perdew-Burke-Ernzerhof (PBE) functional or similar generalized gradient approximation (GGA) functionals provide reasonable accuracy for surface energies and reaction barriers, though van der Waals corrections are often necessary for physisorbed species [39].

Transition State Optimization Locating transition states represents one of the most critical aspects of generating parameters for microkinetic modeling. The climbing image nudged elastic band (CI-NEB) method has emerged as the standard approach for identifying first-order saddle points on potential energy surfaces [39]. This technique involves optimizing a series of images (typically 5-7) along the reaction coordinate between initial and final states, with the highest energy image providing a robust estimate of the transition state geometry and energy [39]. Subsequent frequency calculations confirm the presence of a single imaginary frequency corresponding to the reaction coordinate, while intrinsic reaction coordinate (IRC) calculations verify the transition state connects the appropriate reactants and products [39].

Microkinetic Model Implementation

Reaction Network Specification Constructing a comprehensive reaction network requires enumerating all possible elementary steps, including adsorption/desorption processes, surface reactions, and diffusion events [38]. For COâ‚‚ hydrogenation on Ni(111), this involves 19 intermediate species, 7 gas-phase species, and 86 elementary processes that encompass multiple pathways to various products including CO, methane, methanol, and formaldehyde [38]. Maintaining thermodynamic consistency throughout this network ensures the model converges to the correct equilibrium composition and obeys microscopic reversibility for all reversible steps [37].

Numerical Integration and Steady-State Solution The system of ordinary differential equations describing coverage evolution typically takes the form:

[ \frac{d\thetai}{dt} = \sumj \nu{ij} rj ]

where ( \thetai ) represents the coverage of species i, ( \nu{ij} ) the stoichiometric coefficient of species i in reaction j, and ( rj ) the rate of reaction j [38]. Stiff ODE solvers such as SUNDIALS or MATLAB's ode15s handle the numerical integration efficiently, particularly when surface coverages span multiple orders of magnitude [38]. Steady-state solutions, where ( \frac{d\thetai}{dt} = 0 ) for all species, represent the primary interest for continuous catalytic processes, found either through time integration until coverages stabilize or via root-finding algorithms applied directly to the rate equations [38].

Table 3: Essential Computational Tools and Their Applications in Microkinetic Modeling

Tool/Resource	Primary Function	Application in Catalysis Research
DFT Software (VASP, Quantum ESPRESSO)	Electronic structure calculations	Determining adsorption energies, transition states, and electronic properties of catalysts [39] [40]
Transition State Theory (TST)	Rate constant calculation	Calculating rate constants for elementary steps from DFT energies [38]
Microkinetic Modeling Platforms (Cantera)	Reaction network simulation	Simulating coverage evolution, turnover frequencies, and product distributions [41]
Kinetic Monte Carlo (kMC)	Spatially-resolved kinetics	Modeling systems with high coverages and strong lateral interactions [38]
Scaling Relations	Energetic correlations	Reducing parameter space for catalyst screening and optimization [37]
Degree of Rate Control (DRC) Analysis	Rate-limiting step identification	Pinpointing transition states and intermediates that control overall reaction rate [39]
Uncertainty Quantification	Parameter sensitivity analysis	Identifying critical parameters and assessing model reliability [41]

Microkinetic modeling, firmly grounded in DFT calculations and enhanced by scaling relationships, has evolved into an indispensable tool for unraveling complex catalytic mechanisms and guiding the rational design of advanced catalysts [37]. The integration of these approaches has proven particularly valuable for sustainable catalytic processes, including COâ‚‚ hydrogenation [38], methane conversion [39], and propane dehydrogenation [40], where understanding and optimizing catalytic performance contributes to closing carbon cycles and reducing greenhouse gas emissions. As demonstrated across multiple case studies, microkinetic analysis provides unique insights into rate-determining steps, surface coverage effects, and structure-activity relationships that transcend the limitations of conventional kinetic analysis.

Future advancements in microkinetic modeling will likely focus on several key frontiers: First, improved integration with machine learning approaches will accelerate the exploration of catalyst design spaces and enhance uncertainty quantification [41]. Second, more sophisticated treatments of coverage effects and lateral interactions will bridge the gap between mean-field models and spatially resolved simulations, potentially through hybrid MKM-kMC approaches [38]. Third, increased emphasis on multi-facet and multi-site models will better represent real catalyst nanoparticles and their structure-sensitive behavior [41]. Finally, the development of user-friendly software platforms will democratize access to advanced microkinetic analysis, enabling broader adoption across the catalysis research community. As these methodological improvements converge with growing computational power and more accurate electronic structure methods, microkinetic modeling will continue to strengthen its role as a cornerstone of computational catalyst design.

Comprehending dynamic catalytic processes is crucial for optimizing reactions and developing innovative catalytic systems central to the energy and chemical industries [42]. Conventional ex-situ characterizations provide a fundamental understanding of initial and final states but fall short of capturing the true nature of electrocatalytic processes, which are not state functions [43]. Electrocatalysts frequently undergo significant surface and bulk reconstructions, as well as compositional changes, under operating conditions, making characterization under reaction-resembling conditions essential for achieving a genuine understanding of the chemical processes involved [43].

Operando spectroscopy represents the pinnacle of this approach, facilitating real-time observations of dynamic changes occurring during catalysis while simultaneously measuring catalytic activity [42] [44]. This capability is particularly valuable for investigating scaling relationships in catalysis, where the binding energies of reaction intermediates are often linearly correlated, creating fundamental limitations in catalytic optimization [34] [45]. By directly detecting and characterizing transient intermediates under actual working conditions, operando techniques provide the experimental data necessary to validate, refine, or overcome these scaling relationships through innovative catalyst design.

This technical guide provides an in-depth examination of three powerful operando spectroscopic techniques—X-ray Absorption Fine Structure (XAFS), Infrared (IR) spectroscopy, and Nuclear Magnetic Resonance (NMR) spectroscopy—focusing on their application in detecting transient intermediates and elucidating reaction mechanisms within the framework of catalytic scaling relationships.

Technical Foundation: Operando Spectroscopy Principles and Methodologies

Defining Operando Spectroscopy and Its Distinction from In-Situ Approaches

Within catalytic research, a critical distinction exists between in-situ and operando techniques. In-situ techniques are performed on a catalytic system under simulated reaction conditions (e.g., elevated temperature, applied voltage, immersed in solvent, presence of reactants), whereas operando techniques probe the catalyst under the same, or as close as possible, conditions while simultaneously measuring its activity [44]. This simultaneous measurement is crucial for directly correlating observed structural or chemical changes with catalytic function.

The primary advantage of operando methodologies lies in their ability to capture the true active state of catalysts, which often differs significantly from their pre- or post-reaction structure. For instance, in operating anion exchange membrane fuel cells (AEMFCs), Mn valence states within spinel Mn₃O₄/C increase to above 3+, adopting an octahedral coordination devoid of Jahn-Teller distortions—a structural transformation that would be missed in ex-situ analysis [43].

The Challenge of Scaling Relationships in Catalysis

Scaling relationships describe the linear correlations between the adsorption energies of different intermediates on catalytic surfaces [34] [45]. In multi-step reactions like the oxygen reduction reaction (ORR), the binding energies of intermediates such as *Oâ‚‚, *OOH, and *O are intrinsically linked, creating an fundamental limitation known as the "thermodynamic overpotential" or "scaling relationship bottleneck" [45]. This relationship means that strengthening the binding of one intermediate inevitably strengthens the binding of others, preventing independent optimization of each reaction step.

Operando spectroscopy provides a direct experimental approach to investigate these relationships by enabling the detection and characterization of these intermediates under actual working conditions, thereby offering insights into how catalyst design might circumvent these inherent limitations.

Table 1: Core Concepts in Operando Spectroscopy and Scaling Relationships

Concept	Definition	Implication for Catalyst Design
Scaling Relationships	Linear correlations between adsorption energies of different intermediates on catalytic surfaces	Creates fundamental limitations in catalytic optimization; "thermodynamic overpotential"
Transient Intermediates	Short-lived species formed during catalytic cycles that determine reaction pathways and kinetics	Direct detection enables mechanistic understanding and rational design
Active State	The actual structural and electronic state of a catalyst under working conditions	Often differs significantly from pre- or post-reaction structure; essential for performance
Dual-site Mechanism	Reaction pathway where two adjacent active sites simultaneously interact with an intermediate	Can bypass traditional scaling relationships by altering adsorption geometries

Operando X-ray Absorption Fine Structure (XAFS) Spectroscopy

Fundamental Principles and Experimental Protocols

Operando X-ray Absorption Fine Structure (XAFS) spectroscopy, which includes both XANES (X-ray Absorption Near Edge Structure) and EXAFS (Extended X-ray Absorption Fine Structure), is exceptionally suited for determining the local electronic and geometric structure of catalysts under reaction conditions [44]. XANES provides information about oxidation states and electronic configuration, while EXAFS yields bond distances, coordination numbers, and species identities in the immediate vicinity of the absorbing atom.

For investigating oxygen reduction reaction (ORR) catalysts, researchers have developed specialized fuel cell devices for operando XAS studies. The cell typically comprises a conventional fuel cell flow field and current collector on one side, and a titanium flow field plate featuring an X-ray window on the opposite side, which also serves as a current collector [43]. Using synchrotron-based X-rays, operando XAS studies capture applied potential-dependent changes in the catalyst metal center valence and coordination in an operating AEMFC.

Application Case Study: Tracking Mn Valence States in Oxygen Reduction Reaction

In a seminal study investigating spinel oxide electrocatalysts for ORR, operando XAS revealed profound structural transformations under working conditions. While ex-situ characterization showed Mn₃O₄/C with a tetragonal spinel structure with Jahn-Teller distortion, operando measurements demonstrated that in the AEMFC environment, the tetragonal Mn₃O₄ spinel nanocrystals transformed to a structure with Mn occupying octahedral sites devoid of Jahn-Teller distortions, with the average Mn valence state increasing to above 3+ [43]. This structural transformation resulted in Mn₃O₄/C performing at the same level as Co₁.₅Mn₁.₅O₄/C in the AEMFC, despite significantly lower performance in RDE tests where tetrahedral Mn²⁺ sites were present.

These findings strongly suggest that octahedrally coordinated Mn³⁺ sites are more active than tetrahedral sites toward the ORR in alkaline media, directing rational catalyst design toward stabilizing this specific coordination environment [43].

Protocol for Operando XAFS in Fuel Cell Applications

• Cell Design: Fabricate a custom fuel cell with an X-ray transparent window (e.g., titanium flow field plate with Kapton or polyimide window) allowing X-ray transmission while maintaining gas flow and electrical connectivity [43].

• Catalyst Preparation: Synthesize catalyst materials (e.g., spinel oxide nanocrystals supported on carbon) with controlled loading (typically 40 wt% verified by TGA) [43].

• Membrane Electrode Assembly (MEA): Incorporate catalyst into a standard MEA architecture using appropriate ionomers and membranes.

• Data Collection: Perform XAFS in fluorescence mode using synchrotron radiation across a range of applied potentials (typically from open-circuit voltage to operating voltage).

• Reference Standards: Collect spectra of well-characterized reference compounds with known oxidation states and coordination geometries for linear combination fitting.

• Data Analysis: Process EXAFS data using standard software (e.g., Athena, Artemis) to extract structural parameters (bond distances, coordination numbers, disorder parameters).

Table 2: Key XAFS Spectral Features and Their Structural Interpretation

Spectral Feature	Spectral Region	Structural Information	Case Study: Mn Spinel Oxides
Edge Position	XANES	Oxidation state	Shift to higher energy indicates Mn valence >3+ [43]
Pre-edge Features	XANES	Coordination geometry/symmetry	Intensity changes indicate transition from tetrahedral to octahedral coordination [43]
White Line Intensity	XANES	Electron density/occupancy	Changes in 3d orbital occupancy with potential application
EXAFS Oscillations	EXAFS	Bond distances, coordination numbers, disorder	Loss of Jahn-Teller distortion signatures [43]
Fourier Transform Peaks	EXAFS	Radial distribution around absorber	Appearance of Mn-O-Mn scattering paths in operating cell

Operando Infrared (IR) Spectroscopy

Technical Approaches and Implementation Methodologies

Operando infrared spectroscopy provides exceptional sensitivity for detecting molecular vibrations associated with reaction intermediates adsorbed on catalyst surfaces. Two primary implementations include transmission IR, where the infrared beam passes through the catalyst layer, and attenuated total reflection (ATR) configurations, where the infrared beam interacts with the catalyst deposited on an internal reflection element.

For electrocatalytic systems, specialized spectroelectrochemical cells are required that incorporate working, counter, and reference electrodes while allowing infrared access to the electrode-electrolyte interface. These cells must control potential/current application while maintaining appropriate mass transport conditions to replicate realistic catalytic environments [44].

Application Case Study: Direct Observation of ORR Intermediates via SR-FTIR

In the investigation of Pt-Fe atomic-scale bimetal assembly (ABA) catalysts for ORR, researchers employed in situ synchrotron radiation Fourier transform infrared spectroscopy (SR-FTIR) to monitor the formation of key intermediate states under operating conditions [45]. The technique successfully identified the formation of a Pt-O-O-Fe transition state, providing direct evidence that the ORR follows a dual-sites mechanism that bypasses the formation of *OOH intermediates.

This finding is particularly significant in the context of scaling relationships, as the conventional ORR pathway on single sites involves *OOH intermediates whose binding energy is intrinsically linked to other oxygen-containing intermediates through scaling relationships [45]. The dual-site mechanism circumvents this limitation by enabling direct O-O bond breakage without *OOH formation, resulting in dramatically enhanced ORR kinetics—the Pt = N₂ = Fe ABA catalyst exhibited nearly two orders of magnitude higher kinetic current density compared to commercial Pt/C [45].

Protocol for Operando SR-FTIR in ORR Studies

• Catalyst Deposition: Prepare thin, uniform catalyst layers on appropriate IR-transparent or reflective substrates (e.g., Si, ZnSe, or diamond ATR crystals).

• Cell Assembly: Construct a spectroelectrochemical flow cell with controlled electrolyte delivery and gas purging capabilities.

• Potential Control: Utilize a potentiostat to apply controlled potentials while collecting spectral data.

• Spectral Acquisition: Collect interferograms at various applied potentials using rapid-scan or step-scan modes, with sufficient spectral resolution (typically 4-8 cm⁻¹) to resolve key vibrational features.

• Reference Spectra: Acquire background spectra at a potential where no Faradaic processes occur.

• Isotope Labeling: Employ ¹⁸Oâ‚‚ to confirm vibrational assignments through predictable isotopic shifts.

• Data Processing: Process raw spectra to compute difference spectra (ΔR/R) highlighting potential-dependent features.

Operando Nuclear Magnetic Resonance (NMR) Spectroscopy

Technical Innovations in SEC-NMR Methodology

Operando NMR presents unique technical challenges due to the interference of metal components in standard electrochemical cells with magnetic fields and the difficulty in achieving adequate sensitivity for detecting low-concentration intermediates. Recent advances in spectroelectrochemical NMR (SEC-NMR) have enabled operando characterization of molecular species during organometallic electrocatalysis by addressing these design incompatibilities [46].

Modern SEC-NMR platforms integrate microfluidic electrochemical cells within NMR spectrometers, allowing for simultaneous potential control and spectral acquisition. These systems typically feature optimized cell geometries that minimize magnetic field distortions while maintaining controlled electrochemical conditions.

Application Case Study: Mapping Rh Complex Electrocatalysis

In a groundbreaking study of a molecular rhodium(I) diphosphine complex, researchers utilized multinuclear SEC-NMR to characterize competing pathways for proton and COâ‚‚ reduction [46]. The technique enabled direct observation of the bielectronic reduction leading to a highly reactive low-valent rhodium(-I) intermediate and subsequent protonation into a Rh-hydride complex in a time-resolved manner.

Deuterium labeling combined with ex-situ NMR analysis after SEC-NMR electrolysis revealed unexpected proton sources, including Hofmann elimination of the nBu₄NPF₆ electrolyte in addition to the acetonitrile solvent [46]. This level of mechanistic insight is crucial for understanding and optimizing catalytic systems, particularly when multiple competing pathways exist.

Protocol for Operando SEC-NMR of Molecular Catalysts

• Cell Design: Utilize a specialized NMR cell with integrated electrodes (typically Pt or carbon working and counter electrodes with a pseudo-reference) that minimizes field distortions.

• Solvent/Electrolyte Selection: Choose solvents and supporting electrolytes that provide adequate conductivity while minimizing interfering background signals (e.g., deuterated solvents for lock signal).

• Potential Application: Apply controlled potentials or current using a potentiostat compatible with the NMR environment.

• Rapid Data Acquisition: Employ pulse sequences with minimal repetition delays to enhance temporal resolution for capturing transient species.

• Multinuclear Detection: Acquire data for relevant nuclei (¹H, ¹³C, ³¹P, etc.) to obtain comprehensive mechanistic information.

• Quantification: Utilize electronic references or internal standards for quantitative analysis of species concentrations.

• Supplementary Techniques: Correlate with ex-situ NMR analysis of electrolysis products and computational modeling of NMR parameters.

Integrated Approaches and Best Practices

Multi-Technique Integration for Comprehensive Mechanistic Understanding

No single operando technique provides a complete picture of complex catalytic mechanisms. Integration of multiple complementary techniques is essential for obtaining comprehensive insights. For instance, combining XAFS with IR spectroscopy allows correlation of electronic and geometric structural changes (from XAFS) with the appearance of molecular intermediates (from IR). Similarly, coupling electrochemical mass spectrometry with spectroscopic techniques enables direct correlation between detected intermediates and product formation rates.

The integration of operando spectroscopies with transient analysis is particularly powerful for acquiring quantitative dynamic information beyond steady-state kinetics [42]. Transient experiments examine catalysts' responses to and recovery from dynamic conditions, providing insights into the kinetics of elementary steps and the lifetimes of reactive intermediates.

Reactor Design Considerations for Valid Operando Studies

A crucial challenge in operando spectroscopy is the design of experimental cells that simultaneously provide optimal conditions for both spectroscopic characterization and realistic catalytic performance [44]. Key considerations include:

• Mass Transport: Many operando reactors are designed for batch operation with planar electrodes, which can create significant differences in species transport compared to practical reactors employing electrolyte flow or gas diffusion electrodes [44].

• Signal Detection Path: The path through which a spectroscopic beam travels and the path length between a reaction event and the spectroscopic probe significantly impact response time and signal-to-noise ratio [44].

• Window Materials: Appropriate X-ray, IR, or NMR-transparent window materials must be selected that withstand reaction conditions while minimizing signal attenuation.

• Proximity to Real Devices: Whenever possible, operando cells should approach the complexity of practical catalytic devices to ensure mechanistic relevance [43] [44].

Correlation with Theoretical Modeling

Operando spectroscopic data provides essential experimental validation for theoretical models of catalytic mechanisms. Density functional theory (DFT) calculations can predict the structures, energies, and spectroscopic signatures of proposed intermediates, which can then be compared with experimental observations [45]. This iterative process of experimental measurement and theoretical modeling significantly strengthens mechanistic proposals and provides deeper fundamental understanding.

For investigating scaling relationships, this combined approach is particularly valuable, as computations can rapidly screen potential catalyst modifications to identify those most likely to break traditional scaling relationships, while operando spectroscopy can experimentally verify successful strategies.

Research Reagent Solutions and Essential Materials

Table 3: Essential Research Reagents and Materials for Operando Spectroscopy

Category	Specific Materials	Function/Application	Technical Considerations
Catalyst Materials	Spinel oxide nanocrystals (Mn₃O₄, Co₁.₅Mn₁.₅O₄) [43]	ORR electrocatalysis	Controlled size (5-8 nm), carbon support, defined crystal structure
Molecular Catalysts	Rh(I) diphosphine complexes [46]	COâ‚‚ and proton reduction	Defined coordination sphere, redox-active metal center
Support Materials	Amino-functionalized carbon nanoflakes [45]	Single-atom catalyst support	Nitrogen-rich surface for metal anchoring, defined functionalization
Electrode Materials	Pt, carbon, specialized alloys [45]	Working/counter electrodes	Purity, defined surface area, compatibility with spectroscopy
Membranes & Electrolytes	Anion exchange membranes, nBu₄NPF₆ [43] [46]	Ionic conduction	Purity, stability under reaction conditions, minimal background signals
Isotopic Labels	¹⁸O₂, D₂O [46] [45]	Mechanistic studies	Isotopic purity, handling considerations
Reference Compounds	CdO, metal foils, well-characterized complexes [43] [47]	Spectroscopic calibration	High purity, defined structure and composition
Cell Components	X-ray transparent windows (Kapton) [43]	Spectroelectrochemical cells	Mechanical stability, chemical resistance, transmission properties

Operando spectroscopic techniques represent powerful tools for elucidating catalytic mechanisms and addressing fundamental challenges in catalysis, particularly the limitations imposed by scaling relationships between reaction intermediates. XAFS provides unparalleled insights into the electronic and geometric structure of catalytic sites under working conditions; IR spectroscopy offers sensitive detection of molecular intermediates at catalyst surfaces; and NMR enables detailed mechanistic studies of molecular catalysts and solution species.

The continuing advancement of these techniques—through improved time resolution, enhanced sensitivity, better cell designs, and more sophisticated data analysis—will further expand their capabilities for detecting increasingly transient intermediates with higher spatial and temporal resolution. As these methodologies mature, their integration with theoretical modeling and multi-technique approaches will provide increasingly comprehensive understanding of catalytic mechanisms, guiding the rational design of next-generation catalysts that overcome traditional scaling relationship limitations.

For the field of catalysis research, these advanced spectroscopic approaches are not merely characterization tools but essential components of a fundamental strategy to understand and optimize complex chemical transformations central to sustainable energy and chemical production.

The rational design of next-generation electrocatalysts for sustainable energy conversion is fundamentally limited by scaling relationships between the adsorption energies of key reaction intermediates. These linear correlations often dictate that weakening the binding of one intermediate inevitably strengthens the binding of another, creating a thermodynamic bottleneck that constrains overall catalytic efficiency. While computational studies have extensively modeled these relationships, a critical experimental gap has persisted in directly measuring the binding energies of reactive intermediates under operational conditions. Electroadsorption analysis has emerged as a powerful technique to bridge this divide, enabling the experimental determination of intermediate energetics on well-defined electrocatalyst surfaces.

This technical guide details how electroadsorption measurements, when applied to precisely synthesized nanocatalysts, provide quantitative access to the reaction energetics of multi-electron transfer processes. By examining the potential-dependent adsorption of intermediates, researchers can directly construct scaling relationships, validate theoretical models, and identify catalyst compositions that circumvent traditional thermodynamic limitations. The integration of this experimental methodology with controlled nanocrystal synthesis represents a paradigm shift in electrocatalyst design, moving beyond trial-and-error approaches toward rational optimization based on fundamental energetic principles.

Theoretical Foundation of Electroadsorption

Principles of Electrosorption

Electroadsorption describes the potential-dependent adsorption of species at electrode-electrolyte interfaces with accompanying electron transfer. This phenomenon differs fundamentally from physical adsorption through its potential dependence and electron transfer component, characterized by the electrosorption valency (γ). The process is governed by the formation of the electrochemical double layer and is highly sensitive to the chemical composition and structure of both the electrode surface and the electrolyte.

The theoretical framework for quantifying electrosorption leverages grand canonical density functional theory (GC-DFT) combined with thermodynamic cycles to compute potential-dependent adsorption free energies [48]. This approach properly accounts for the electrochemical potential of electrons and avoids systematic errors inherent in earlier methodologies. The adsorption free energy (ΔΩad) varies linearly with applied potential (U) according to:

ΔΩad(U) = -γF(U - U0)

where γ represents the electrosorption valency (extent of electron transfer), F is Faraday's constant, and U0 is the standard equilibrium adsorption potential [48]. This linear relationship enables the experimental determination of binding energies from electrochemical measurements.

Connecting Electroadsorption to Intermediate Binding Energies

In electrocatalysis, reaction intermediates undergo electron transfer steps that make their surface coverage inherently potential-dependent. For example, in the oxygen evolution reaction (OER), the intermediates *OH, *O, and *OOH exhibit distinct electroadsorption signatures that can be deconvoluted through careful electrochemical analysis [49]. The binding energies of these intermediates directly influence their surface coverage at relevant operating potentials, creating a bridge between experimental electroadsorption profiles and theoretical binding energy calculations.

The potential-dependent coverage of intermediates follows a Langmuir-type isotherm modified for electrochemical interfaces:

θ(U) = 1 / {1 + exp[ΔΩad(U)/RT]}

where θ represents the surface coverage at applied potential U [48]. By analyzing how coverage evolves with potential, researchers can extract quantitative information about intermediate binding energies and their scaling relationships across different catalyst compositions and surface structures.

Experimental Methodologies

Synthesis of Well-Defined Nanocrystal Surfaces

Precise catalyst synthesis is foundational to meaningful electroadsorption analysis. Controlled fabrication of nanocrystals with specific crystallographic terminations eliminates the heterogeneity that obscures structure-activity relationships in conventional catalysts.

Table 1: Molten Salt Synthesis Method for Well-Defined Oxide Nanocrystals

Component	Function	Example Composition
NaCl Matrix	High-temperature molten salt medium	4.55 g
Transition Metal Precursors	Forms ternary oxide structure	50 μL of 400 mM metal salt in 2M HCl
Na2SO4	Mineralizer promoting crystal growth	500 μL of 400 mM solution
Thermal Treatment	Controls crystallization & facet expression	700°C for 1 hour (RuO2, M-RuOx)

A representative synthesis for ternary transition metal ruthenium oxide (M-RuOx) nanocrystals involves combining NaCl, RuCl3, Na2SO4, and specific transition metal salt precursors in a ceramic crucible [49]. The mixture is heated at 700°C for one hour in a box furnace, then cooled naturally. The resulting nanocrystals are purified through a series of washing steps including water, 2M HCl at 90°C, and isopropanol to remove residual salts and unstable oxides [49]. This method produces nanocrystals with well-defined (110) and (111)/(112) crystallographic facets essential for facet-dependent energy scaling relation studies.

Electrochemical Characterization Protocols

Electroadsorption analysis requires carefully controlled electrochemical measurements to probe intermediate binding without interference from faradaic processes.

Table 2: Standard Electrochemical Setup for Electroadsorption Studies

Component	Specification	Purpose
Working Electrode	Glassy carbon RDE (5 mm diameter) with catalyst ink	Provides standardized, reproducible surface
Reference Electrode	Ag/AgCl (calibrated to RHE)	Potential control & reporting
Counter Electrode	Platinum wire/mesh	Completes circuit without contamination
Electrolyte	1.0 M HClO4 (acidic OER studies)	Non-coordinating anions minimize interference
Instrumentation	Potentiostat with rotating electrode control	Mass transport control & precise potential application

Catalyst inks are prepared by suspending nanopowder in a mixture of H2O, isopropanol, and Nafion (83.5:35.5:6 ratio per 1 mg catalyst) followed by 40 minutes of sonication [49]. The ink is drop-cast onto a polished glassy carbon rotating disk electrode operated at 2000 rpm to maintain consistent mass transport. All measurements should be conducted in oxygen-saturated electrolyte at 25°C to standardize conditions.

Cyclic voltammetry in non-Faradaic potential regions provides the primary data for electroadsorption analysis. The current-potential profiles reveal adsorption/desorption peaks corresponding to specific intermediate species. For OER catalysts, the potential range typically spans 0.3-1.5 V versus RHE, carefully avoiding the onset of oxygen evolution. Integration of current peaks after double-layer correction yields charge values directly related to intermediate surface coverage.

Diagram 1: Experimental workflow for electroadsorption analysis, showing the sequence from nanocrystal synthesis to catalyst design.

Data Analysis and Interpretation

The transformation of electrochemical data into binding energy information requires careful processing. First, the measured current must be corrected for non-adsorptive contributions, primarily the double-layer charging. This is typically achieved by measuring the current in a supporting electrolyte without active species or by extrapolating the double-layer capacitance from potential regions without Faradaic processes.

The coverage of each intermediate (θi) is calculated from the integrated charge (Qi) under the corresponding adsorption peak after double-layer correction:

θi = Qi / (nFAΓ)

where n is the number of electrons transferred, F is Faraday's constant, A is the electroactive surface area, and Γ is the total site density. The electroactive surface area is determined independently through methods such as underpotential deposition of copper or lead, or from double-layer capacitance measurements.

The binding energy (ΔGi) is then related to the potential at half-coverage (U50):

ΔGi = -FU50 + C

where C is a constant that depends on the reference potential and the standard states. The potential of half-coverage is determined from the coverage-potential isotherm constructed from the integrated charges at various potentials.

For OER intermediates, the binding energies of *OH, *O, and *OOH can be determined from their characteristic adsorption potentials, enabling the construction of experimental scaling relationships that can be directly compared with theoretical predictions [49].

Case Study: Oxygen Evolution Reaction on M-RuOxNanocrystals

Experimental Scaling Relationships

Application of electroadsorption analysis to a series of ternary first-row transition metal ruthenium oxide nanocrystals (M-RuOx, where M = V, Cr, Mn, Fe, Co, Ni, Cu, Zn) revealed fundamental insights into OER energetics [49]. The well-defined (110) and (111)/(112) facets enabled deconvolution of facet-dependent effects, demonstrating that distinct facet-dependent energy scaling relations govern the OER pathway on different crystallographic terminations.

The binding energies of *OH, *O, and *OOH intermediates followed linear scaling relationships consistent with theoretical predictions, but with system-specific deviations. Notably, only Mn-RuOx exhibited substantially improved OER activity compared to pure RuO2 on an electrochemically-active surface-area basis [49]. The electroadsorption analysis enabled quantification of both intermediate surface coverage and absolute reaction energetics for all M-RuOx compositions.

Rational Catalyst Design

The experimental scaling relationships provided a blueprint for designing improved OER catalysts. Based on the energetics observed across the M-RuOx series, a quaternary FeMn-RuOx electrocatalyst was predicted to exhibit superior activity [49]. Experimental validation confirmed an 876% increase in mass activity compared to RuO2 and a 309% improvement over the best-performing ternary Mn-RuOx catalyst [49]. This successful rational design highlights the power of electroadsorption analysis in guiding catalyst development.

The activity enhancement originated from optimized intermediate binding energies that balanced formation and desorption of all three OER intermediates, moving closer to the theoretical thermodynamic optimum. The FeMn-RuOx catalyst achieved this through synergistic electronic effects and possibly through modified reaction pathways involving lattice oxygen evolution, as suggested by related studies on rutile oxide surfaces [50].

Diagram 2: Logical relationship between theoretical scaling relationships, experimental electroadsorption analysis, and rational catalyst design.

Advanced Applications and Methodological Extensions

Beyond OER: Applications to Other Electrocatalytic Reactions

While this guide has focused on OER, electroadsorption analysis applies broadly across electrocatalysis. In nitrogen reduction reaction (NRR), scaling relationships between *N2H and *NH/ *NH2 intermediates limit catalytic efficiency [51]. Similar constraints exist in carbon dioxide reduction reaction (CO2RR), where scaling relationships couple the stabilities of *COOH and *CHO intermediates to CO adsorption [52].

Breaking these scaling relationships represents a central challenge in electrocatalysis. Strategies include using high-entropy alloys to create unique local environments that stabilize intermediates through non-linear effects [52], or introducing hydrogen bonding and confinement effects in layered structures to differentially stabilize transition states [51]. Electroadsorption analysis provides the experimental verification needed to confirm when these strategies successfully deviate from traditional scaling relationships.

Integrating Machine Learning and High-Throughput Screening

The combination of electroadsorption analysis with machine learning accelerates catalyst discovery. For instance, a two-tier machine learning and DFT workflow screened millions of local environments in AgAuCuPdPt high-entropy alloy nanoparticles for CO2RR [52]. An ultralight linear-regression surrogate model predicted CO adsorption energy, enabling rapid identification of optimal compositions that break conventional scaling relationships.

Similar approaches can be applied to electroadsorption data, using experimentally determined binding energies as training sets for predictive models. This integration enables rational design rules based on both composition and local coordination environment, moving beyond traditional trial-and-error approaches.

The Scientist's Toolkit

Table 3: Essential Research Reagent Solutions for Electroadsorption Analysis

Category	Specific Materials	Function/Purpose
Supporting Electrolytes	HClO4, H2SO4, KOH	Provide ionic conductivity with minimal specific adsorption
Nanocrystal Precursors	RuCl3, MnCl2, FeCl3	Synthesis of well-defined oxide nanocrystals
Molten Salt Media	NaCl, Na2SO4	High-temperature matrix for controlled crystal growth
Electrode Materials	Glassy carbon RDE, Pt counter electrode, Ag/AgCl reference	Standardized three-electrode electrochemical cell
Dispersion Agents	Nafion solution, isopropanol	Form stable catalyst inks for electrode preparation
Surface Characterization	Cu UPD solutions, Fe(CN)6^3-/4-^	Determination of electroactive surface area
Computational Tools	GC-DFT codes, thermodynamic databases	Prediction and interpretation of electrosorption valencies
11-Thiohomoaminopterin	11-Thiohomoaminopterin, CAS:74163-10-3, MF:C20H21N7O5S, MW:471.5 g/mol	Chemical Reagent
7-Methyl-6-mercaptopurine	7-Methyl-6-mercaptopurine, CAS:3324-79-6, MF:C6H6N4S, MW:166.21 g/mol	Chemical Reagent

Electroadsorption analysis represents a transformative methodology for experimentally probing intermediate binding energies on well-defined surfaces, directly addressing the challenge of scaling relationships in catalysis research. By integrating precisely synthesized nanocrystals with sophisticated electrochemical characterization, this approach provides quantitative energetic information that bridges computational predictions and experimental observations. The case study on M-RuOx nanocrystals demonstrates how electroadsorption-guided design can lead to catalysts with dramatically improved activity, validating the power of this methodology. As electrocatalysis advances toward increasingly complex materials systems, electroadsorption analysis will play an essential role in deconvoluting composition-structure-activity relationships and enabling the rational design of next-generation catalysts for sustainable energy conversion.

Scaling relations, the linear correlations between adsorption energies of reactive intermediates, represent a fundamental limitation in multi-step electrocatalytic reactions. This review examines strategies for manipulating these relations, focusing on the oxygen evolution reaction (OER) as a representative case. We present a comprehensive analysis of how the FeMn–RuOx catalyst system successfully circumvents traditional scaling constraints through precise electronic structure modulation and dynamic site cooperation. Experimental validation demonstrates an 876% increase in mass activity compared to benchmark RuO₂, achieving an unprecedented degradation rate of 53 μV h⁻¹ during 1300-hour operational stability tests. This systematic exploration of scaling relation manipulation strategies provides a framework for rational design of next-generation electrocatalysts.

Fundamental Concepts and Historical Context

Scaling relations describe the correlations between adsorption energies of different reaction intermediates on catalytic surfaces [10]. First identified in 2005 when Rossmeisl et al. discovered linear relations between OH and OOH adsorption energies versus O on metal surfaces, these relationships fundamentally limit catalyst optimization by preventing independent adjustment of intermediate binding strengths [10]. The term "scaling relations" was formally introduced in 2007 by Abild-Pedersen et al., who demonstrated that adsorption energies of OH, CHâ‚“, and NHâ‚“ scale proportionally with their corresponding atomic species (O, C, N) [10].

In oxygen electrocatalysis, scaling relations create intrinsic thermodynamic limitations. For the oxygen evolution reaction (OER), the adsorption energies of critical intermediates (*OH, *O, *OOH) are linearly correlated, resulting in a theoretical minimum overpotential of approximately 0.37 V [10] [1]. This constraint emerges because *OOH and *OH intermediates bind through similar molecular motifs, with their adsorption energy difference remaining nearly constant across different catalyst materials [1].

Impact on Catalyst Design

The pervasive nature of scaling relations affects numerous energy technologies, including:

• Fuel cells and metal-air batteries limited by oxygen reduction reaction (ORR) kinetics [10] [53]
• Water electrolyzers constrained by OER overpotentials [54] [1]
• COâ‚‚ and nitrogen reduction systems hampered by correlated intermediate binding [10]

The recognition of these fundamental limitations has prompted two decades of research focused on understanding, manipulating, and potentially circumventing scaling relations to advance electrocatalytic technologies [10] [17].

Theoretical Framework: Scaling Relation Manipulation Strategies

Classification of Manipulation Approaches

Research has identified five primary strategies for addressing scaling relations, each with distinct mechanisms and implications for catalyst design [10]:

Table 1: Strategic Approaches to Scaling Relation Manipulation

Strategy	Fundamental Approach	Key Mechanism	Representative Systems
Tuning	Adjusting absolute adsorption energies within existing scaling relations	Modifying electronic structure to shift all intermediates along scaling line	Pt-based catalysts with adjusted OH adsorption [10]
Breaking	Selectively stabilizing specific intermediates	Introducing hydrogen bonding, electrostatic stabilization, or spatial confinement	Spectator groups forming hydrogen bonds with OOH [10]
Switching	Employing alternative reaction mechanisms	Avoiding problematic intermediates through different pathways	Catalysts utilizing lattice oxygen mechanism (LOM) [54]
Pushing	Combining mechanism switching with stabilizations	Simultaneously employing alternative mechanisms and stabilizing interactions	Dynamic dual-site systems [1]
Bypassing	Completely decoupling adsorption energies	Utilizing two distinct sites to independently control intermediate binding	Dual-atom catalysts with complementary active centers [10]

Geometric and Electronic Considerations

The emerging paradigm in scaling relation manipulation recognizes that both activity and selectivity depend on two adsorption sites being "neither too far nor too close" geometrically [10]. This geometric principle complements the well-established Sabatier principle that optimal binding should be "neither too strong nor too weak" [10].

Recent theoretical advances highlight the importance of:

• Local coordination environments influencing d-band centers and orbital overlap [55] [56]
• Spin state configurations affecting orbital electron distribution and intermediate binding [56]
• Dynamic structural evolution during catalysis enabling transient active sites [1]

The FeMn–RuOx Case Study: Experimental Realization

Catalyst Design Rationale

The FeMn–RuOx system was designed specifically to manipulate OER scaling relations through precise control of the ruthenium-oxygen bonding environment [49]. Theoretical predictions indicated that incorporating both iron and manganese into the RuO₂ lattice would create optimal Ru charge states (approximately +1.52) for simultaneously stabilizing multiple OER intermediates [49].

Density functional theory (DFT) calculations predicted that the FeMn–RuOx combination would:

• Moderate Ru-O bond covalency to prevent lattice oxygen participation [54]
• Optimize *OOH binding energy through charge redistribution [49]
• Maintain high Ru vacancy formation energy (ΔG_VRu) to enhance stability [54]

Synthesis Methodology

Table 2: Molten Salt Synthesis Protocol for FeMn–RuOx Nanocrystals

Step	Reagents/Parameters	Function/Purpose	Critical Notes
Precursor Preparation	4.55 g NaCl, 250 μL 80 mM RuCl₃, 500 μL 400 mM Na₂SO₄, 15-50 μL 400 mM MnCl₂, 0-50 μL 400 mM FeCl₃	Creates homogeneous reaction environment with controlled stoichiometry	Metal precursors dissolved in 2M HCl; total volume adjusted with deionized water
Thermal Treatment	500°C for 1 hour in box furnace (20°C min⁻¹ heating rate)	Forms rutile oxide structure with controlled faceting	Lower temperature than binary M-RuOx (700°C) enhances ternary incorporation
Purification Process	Series of washes with H₂O, 2M HCl (90°C, 1 hour), isopropyl alcohol	Removes salt matrix and unstable oxide phases	Acid treatment critical for removing poorly incorporated species
Final Processing	Drying in vacuum desiccator	Prepares catalyst for ink formulation	Complete dryness essential for reproducible electrochemical performance

The synthetic approach produces nanocrystals with well-defined (110) and (111)/(112) crystallographic facets and minimal structural defects, enabling precise correlation of structure-function relationships [49].

Structural and Electronic Characterization

Advanced characterization techniques confirm the successful formation of the designed active sites:

X-ray Absorption Spectroscopy reveals charge transfer from Mn to Ru, creating the optimal Ru charge state of approximately +1.52 [49]

Electroadsorption Analysis enables experimental determination of intermediate binding energies, confirming modified scaling relations between *OH, *O, and *OOH [49]

HAADF-STEM shows atomic dispersion of metal elements without nanoparticle formation, indicating homogeneous solid solution formation [49]

The catalyst exhibits a specific surface area of 266.9 m² g⁻¹ with mesoporous structure facilitating mass transport [1]

Performance Evaluation and Mechanistic Insights

Enhanced Catalytic Activity

The FeMn–RuOx system demonstrates exceptional OER performance metrics:

Table 3: Quantitative Performance Comparison of FeMn–RuOx

Performance Metric	FeMn–RuOx	Benchmark RuO₂	Improvement	Testing Conditions
Mass Activity	2360 A g⁻¹ at 1.48 V	270 A g⁻¹ (equivalent)	876% increase	1 M HClO₄, RDE configuration [49]
Turnover Frequency	0.63 s⁻¹	0.07 s⁻¹ (estimated)	~9× enhancement	AEM pathway, per Ru site [54]
Stability	53 μV h⁻¹ degradation	>500 μV h⁻¹ (typical)	Order of magnitude improvement	1300 h at 1 A cm⁻² in PEMWE [54]
Low-Loading Performance	26.8 mV increase over 250 h	Complete degradation (typical)	Exceptional stability	41.65 μgcat cm⁻², 100 mA cm⁻² [54]

Operando Mechanistic Studies

Operando X-ray absorption fine structure (XAFS) measurements provide direct evidence of the dynamic structural evolution during OER [1]. The data reveals:

• Transformation to O-bridged trimer structures during electrochemical activation [1]
• Dynamic coordination changes between Ni site and adsorbates (OH and Hâ‚‚O) modulating adjacent Fe centers [1]
• Intramolecular proton transfer triggering coordination evolution that regulates electronic structure [1]

Electrokinetic studies confirm an unconventional dual-site-cooperated OER mechanism where the dynamic coordination environment simultaneously lowers the free energy required for O–H bond cleavage and *OOH formation [1].

Scaling Relation Analysis

Electroadsorption studies of well-defined M–RuOx nanocrystals enable experimental determination of OER intermediate energetics [49]. Analysis reveals:

• Facet-dependent scaling relations with distinct linear trends for (110) versus (111)/(112) surfaces [49]
• Modified OOH-OH energy difference of approximately 3.2 eV for FeMn–RuOx compared to 3.5 eV for RuOâ‚‚ [49]
• Intermediate surface coverage optimization that balances active site availability and intermediate stabilization [49]

The experimental scaling relations confirm that FeMn–RuOx achieves simultaneous stabilization of multiple intermediates, effectively circumventing the traditional limitation where strengthening *OOH binding invariably weakens *O binding [49].

Research Toolkit: Essential Methodologies and Reagents

Critical Research Reagents

Table 4: Essential Research Reagents for Scaling Relation Studies

Reagent Category	Specific Examples	Function in Catalyst Research	Technical Notes
Metal Precursors	RuCl₃, FeCl₃, MnCl₂	Source of active metal components	Dissolve in 2M HCl for improved stability and incorporation
Structure-Directing Agents	NaCl, Na₂SO₄	Molten salt medium for controlled crystallization	NaCl creates ionic environment; SO₄²⁻ controls facet expression
Support Materials	Holey graphene nanomesh (GNM)	High-surface-area conductive support	~20-60 nm pores with density ~6.2×10⁹ per cm² [1]
Electrochemical Components	Purified KOH, HClOâ‚„, Nafion binder	Electrolyte and electrode preparation	Fe-free KOH essential for controlled Fe incorporation studies
Characterization Standards	Ag/AgCl reference electrode, Pt counter electrode	Electrochemical measurements calibration	RHE calibration critical for comparative analysis

Essential Experimental Protocols

Electroadsorption Analysis for Energetics Determination

• Utilize cyclic voltammetry in non-Faradaic potential regions
• Analyze current-potential profiles to determine surface coverage of intermediates
• Extract absolute binding energies through thermodynamic analysis of adsorption isotherms [49]

Operando XAFS for Structural Analysis

• Collect XANES and EXAFS spectra during electrochemical operation
• Monitor changes in oxidation state and local coordination environment
• Correlate structural evolution with catalytic activity [1]

DFT Computational Workflow

• Build slab models with appropriate surface terminations
• Calculate adsorption energies for *OH, *O, and *OOH intermediates
• Determine theoretical overpotential from free energy diagrams [54] [49]

The FeMn–RuOx case study demonstrates the successful application of scaling relation manipulation principles to achieve exceptional catalytic performance. By combining precise synthesis of well-defined nanocrystals with advanced operando characterization and theoretical modeling, this approach enables rational design of next-generation electrocatalysts.

Key insights for future catalyst development include:

• Ru charge state serves as an intuitive descriptor bridging activity and stability [54]
• Dynamic structural evolution during catalysis provides opportunities for transient active sites that circumvent static scaling relations [1]
• Dual-site cooperation mechanisms enable simultaneous optimization of multiple reaction steps [1] [56]

The experimental and theoretical methodologies established in this study provide a roadmap for systematically addressing scaling relations across diverse catalytic transformations, ultimately accelerating the development of efficient energy conversion technologies.

FeMn-RuOx Design Strategy Map: This diagram illustrates how multiple scaling relation manipulation strategies and structural features combine to enable the exceptional performance of the FeMn-RuOx catalyst system.

Breaking the Scaling Limits: Innovative Strategies to Overcome Energetic Constraints

In multi-step catalytic reactions, linear scaling relationships (LSRs) represent a fundamental thermodynamic constraint that interlinks the adsorption energies of various reactive intermediates [1] [10]. For reactions involving oxygenated species, such as the oxygen evolution reaction (OER), the adsorption energies of key intermediates like *OH, *O, and *OOH are linearly correlated on conventional single-site catalysts [1] [57]. This correlation imposes an intrinsic limitation on optimally adjusting the adsorption strength for every intermediate simultaneously, thereby establishing a minimum theoretical overpotential and limiting maximum achievable catalytic activity [1] [10] [58]. The most problematic scaling relation in oxygen electrocatalysis is widely recognized as the one between *OOH and *OH intermediates, which dictates a theoretical overpotential of at least ~0.4 V even for ideal catalysts [10] [58].

The recognition of these constraints has prompted the development of strategic approaches to manipulate scaling relations. A comprehensive analysis published in Chemical Society Reviews (2025) categorizes these approaches into five general strategies: tuning (adjusting adsorption energies within scaling relation constraints), breaking (altering the scaling relation through stabilizing interactions), switching (adopting alternative reaction mechanisms), pushing (combining mechanism switching with stabilizing interactions), and bypassing (using two states to decouple adsorption energies completely) [10]. Among these, the dynamic structural regulation of active sites has emerged as a particularly promising approach for circumventing LSRs without sacrificing catalytic stability.

Mechanisms of Dynamic Structural Regulation

Dynamic Dual-Site Cooperation

Recent breakthroughs in OER catalysis have demonstrated that dynamic structural regulation of active sites can effectively circumvent linear scaling relationships through a process of dynamic dual-site cooperation. In a landmark study published in Nature Communications, researchers constructed a model Ni-Feâ‚‚ molecular catalyst via in situ electrochemical activation that demonstrated remarkable OER activity [1]. Theoretical calculations and electrokinetic studies revealed that the dynamic evolution of Ni-adsorbate coordination, driven by intramolecular proton transfer during the catalytic cycle, effectively altered the electronic structure of the adjacent Fe active center [1].

This dynamic cooperation mechanism operates through a sophisticated sequence of structural adaptations:

• Continuous coordination changes between Ni sites and adsorbates (OH and Hâ‚‚O) during catalysis
• Intramolecular proton transfer that triggers electronic modulation of adjacent metal centers
• Simultaneous optimization of the free energy changes associated with both O–H bond cleavage and O–O bond formation

The significance of this mechanism lies in its ability to independently regulate the adsorption strengths of different oxygen intermediates that are normally constrained by scaling relationships, thereby enabling simultaneous facilitation of mutually competing reaction steps [1].

Oxide Path Mechanism (OPM) for Direct O–O Coupling

An alternative pathway for circumventing LSRs involves structural arrangements that enable the oxide path mechanism, which facilitates direct O–O bond coupling without forming the *OOH intermediate [57] [58]. Research on bowl-like Co-MOFs with compressive strain has demonstrated that dynamically evolved CoOOH-like phases with contractive Co–Co distances (2.72 Å versus 2.92 Å in conventional CoOOH) can promote direct O–O coupling over cobalt atom pairs (Co–O–O–Co) [57].

This mechanism effectively bypasses the formation of *OOH intermediates, thus completely avoiding the most challenging scaling relationship between *OOH and *OH [57] [58]. The OPM pathway requires specific structural configurations:

• Appropriate interatomic distances between metal sites (typically <2.8 Ã…)
• Symmetric bimetallic positions that enable O–O radical coupling
• Ferromagnetic coupling to accelerate the release of triplet-state oxygen

The successful implementation of this mechanism in CoFeSâ‚“ nanoclusters has demonstrated how dual-site segmentally synergistic mechanisms can break scaling relationships without sacrificing stability, achieving remarkable operational durability exceeding 633 hours [58].

Lattice Oxygen-Mediated Mechanisms

A third approach to circumventing LSRs involves activating lattice oxygen participation through the lattice-oxygen-mediated mechanism (LOM) [57] [58]. This strategy requires careful engineering of the catalyst's electronic structure to raise the energy of O 2p orbitals near the Fermi level, thereby enhancing the covalency of metal-oxygen bonds and enabling lattice oxygen to participate directly in the reaction [58].

While LOM can effectively overcome scaling relationship constraints, it often faces challenges related to excessive oxygen vacancy formation and consequent structural instability during operation [58]. Advanced characterization techniques have revealed that catalysts following LOM mechanisms typically undergo significant structural reconstruction under operational conditions, which can lead to irreversible changes and performance degradation over extended periods [57].

Table 1: Comparative Analysis of Mechanisms for Circumventing Linear Scaling Relationships

Mechanism	Key Feature	Intermediates Involved	Advantages	Limitations
Dynamic Dual-Site Cooperation [1]	Dynamic coordination changes during catalysis	All conventional OER intermediates	Simultaneously optimizes multiple steps; High stability	Requires precise atomic arrangement
Oxide Path Mechanism (OPM) [57] [58]	Direct O–O coupling without *OOH	*OH, *O only	Completely avoids problematic *OOH intermediate	Requires specific interatomic distances
Lattice Oxygen-Mediated Mechanism (LOM) [57] [58]	Lattice oxygen participation	*OH, *O, lattice oxygen	High activity; Circumvents adsorption scaling	Structural instability; Oxygen vacancy formation

Experimental Methodologies for Investigating Dynamic Active Sites

Advanced In Situ/Operando Characterization Techniques

The investigation of dynamically evolving active sites requires sophisticated characterization methodologies that can probe structural and electronic changes under actual reaction conditions. The following table summarizes key experimental techniques employed in recent groundbreaking studies:

Table 2: Essential Characterization Techniques for Dynamic Active Site Analysis

Technique	Key Application	Information Obtained	Experimental Considerations
Operando XAFS [1] [57]	Local structure analysis of metal atoms	Oxidation state, coordination environment, interatomic distances	Requires specialized electrochemical cells; Synchrotron radiation source
Electrochemical TERS [59]	Single-site structural evolution	Geometric and electronic changes at individual active sites	Ultra-high spatial resolution (~8 nm); Complex tip fabrication
In Situ SRIR [57]	Intermediate identification	Detection of reaction intermediates and bonding configurations	Synchrotron-based IR for enhanced sensitivity
In Situ Raman [57]	Phase transformation tracking	Chemical bonding changes, phase evolution, reaction intermediates	Surface-enhanced variants needed for low concentrations
HAADF-STEM [1] [58]	Atomic structure imaging	Atomic dispersion, cluster size distribution, elemental mapping	Requires aberration correction; Electron-beam-sensitive materials

Catalyst Synthesis Protocols

The preparation of catalysts capable of dynamic structural regulation often requires sophisticated synthesis approaches that enable precise control at the atomic level:

Ni-Feâ‚‚ Molecular Complex Catalyst [1]:

• Pre-catalyst Preparation: Synthesize Ni single atoms on holey graphene nanomesh (Ni-SAs@GNM) through thermal annealing of Ni(OH)â‚‚/graphene hydrogel at 700°C under Ar atmosphere, followed by acid treatment to remove nanoparticles
• Electrochemical Activation: Perform cyclic voltammetry between 1.1 and 1.65 V vs. RHE in purified 1 M KOH electrolyte with deliberate addition of 1 ppm Fe ions
• Structural Validation: Confirm formation of O-bridged Ni-Feâ‚‚ trimer structure through operando XAFS measurements

Bowl-like Co-MOFs with Compressive Strain [57]:

• Morphology Control: Introduce polyoxometalates (hexaammonium molybdate) during hydrothermal synthesis to compete with organic ligands and weaken MOF rigidity, creating open-mouthed bowl-like structure
• Strain Induction: Utilize the competing coordination effect to generate intrinsic compressive strain within the framework
• Phase Transformation: Activate under OER conditions to form reconstructed CoOOH-like active phase with contractive Co–Co distance (2.72 Ã…)

CoFeSâ‚“ Nanoclusters on CNT [58]:

• Support Functionalization: Pre-clean carbon nanotubes to create oxygen functionalities for metal precursor anchoring
• Adsorption-Reduction: Prepare amorphous CoFeOâ‚“ nanosheets on CNT (CFO-p/CNT) via NaBHâ‚„ reduction process
• Hydrothermal Sulfidation: Convert to CoFeSâ‚“ nanoclusters using thioacetamide decomposition to generate acidic environment during hydrothermal treatment

Diagram 1: Experimental Workflow for Dynamic Active Site Investigation

Key Research Reagent Solutions and Materials

The experimental investigation of dynamically evolving active sites requires specialized materials and reagents that enable precise control and characterization at the atomic scale:

Table 3: Essential Research Reagents and Materials for Dynamic Active Site Studies

Reagent/Material	Function/Application	Key Characteristics	Example Usage
Fe-free KOH electrolyte [1]	Electrochemical activation	High purity (Fe < ppb level)	Baseline control for Fe incorporation studies
Controlled Fe additions (ppm level) [1]	Precise dopant introduction	Controlled concentration (e.g., 1 ppm)	In situ formation of Ni-Feâ‚‚ molecular complexes
Thioacetamide [58]	Sulfur source for nanocluster formation	Controlled decomposition to Hâ‚‚S	Formation of CoFeSâ‚“ nanoclusters on CNT supports
Polyoxometalates [57]	Structure-directing agents	Compete with organic ligands for metal coordination	Inducing compressive strain in bowl-like MOFs
NaBHâ‚„ [58]	Reducing agent	Strong reduction capability at room temperature	Preparation of amorphous CoFeOâ‚“ nanosheets
Holey graphene nanomesh [1]	Support material	High surface area (≈267 m² g⁻¹), abundant edge sites	Anchoring Ni single atoms for pre-catalyst synthesis

Structural and Electronic Basis for Circumventing LSRs

The fundamental principle enabling dynamic active sites to circumvent linear scaling relationships lies in their ability to decouple the adsorption energies of different intermediates that are normally correlated on static catalytic surfaces. This decoupling occurs through several sophisticated structural and electronic mechanisms:

Electronic Structure Modulation

Dynamic structural changes during catalysis induce significant electronic modulation of active centers. In the Ni-Feâ‚‚ system, the continuous evolution of Ni-adsorbate coordination alters the electronic structure of adjacent Fe centers, creating a situation where the effective adsorption strength for different intermediates can be optimized independently [1]. This electronic modulation operates through:

• Intramolecular charge transfer between metal centers with different electronic configurations
• Dynamic orbital coupling that adjusts during the reaction cycle
• Redistribution of electron density at active sites in response to adsorbate binding

Similar effects have been observed in Ce-doped Co₃O₄ catalysts, where f-d-p gradient orbital coupling redistributes electron density at Co and O sites, independently modulating the adsorption strengths of *NO and *H intermediates in nitrate reduction reactions [60].

Geometric Factor Optimization

The geometric arrangement of active sites plays a crucial role in determining their ability to circumvent scaling relationships. Successful systems typically feature:

• Optimal interatomic distances (typically 2.7-2.9 Ã…) that enable direct O–O coupling [57]
• Asymmetric atomic structures that create heterogeneous charge distributions [61]
• Dual-site arrangements with complementary functions for different reaction steps [58]

In asymmetric atomic structures, the introduction of heteroelements, vacancies, or clusters into symmetric configurations (e.g., M-Nâ‚„) creates asymmetric coordination environments (such as M-Nâ‚“, M-Nâ‚“-S/B/O) that substantially alter internal electronic structure and trigger charge redistribution [61]. This asymmetric charge distribution enables independent optimization of adsorption strengths for different intermediates, effectively breaking the linear scaling relationship [61].

Diagram 2: Conceptual Framework: Static vs. Dynamic Active Sites

Performance Metrics and Comparative Analysis

Catalysts employing dynamic structural regulation strategies have demonstrated remarkable performance improvements compared to conventional materials constrained by linear scaling relationships:

Table 4: Performance Comparison of Catalysts with Dynamic Active Sites

Catalyst System	Reaction	Performance Metrics	Stability	Reference
Ni-Feâ‚‚ molecular complex [1]	OER	Notable intrinsic activity enhancement	-	Nat Commun (2025)
Bowl-like Co-MOFs with compressive strain [57]	OER	259 mV @ 100 mA cm⁻²; 275 mV @ 300 mA cm⁻²	80 h @ large current density	Appl Catal B (2024)
CoFeSₓ nanoclusters on CNT [58]	OER	Mass activity: 12516 A g⁻¹; TOF: 6888 h⁻¹	633 h without significant loss	Nat Commun (2024)
Ce-doped Co₃O₄ [60]	NO₃⁻ reduction to NH₃	FE: 97.8%; Yield: 3423.0 µg h⁻¹ cm⁻²	Excellent cycling stability	Angew Chem (2025)
Commercial RuO₂ (reference) [58]	OER	Mass activity: 8.2 A g⁻¹; TOF: 10.2 h⁻¹	-	-

The performance data reveals extraordinary improvements achieved through dynamic structural regulation strategies, with the CoFeSₓ nanocluster system demonstrating approximately 1526-fold higher mass activity and 675-fold higher TOF compared to commercial RuO₂ benchmark catalysts [58]. Similarly, the bowl-like Co-MOFs with compressive strain achieve remarkably low overpotentials at industrially relevant current densities (300 mA cm⁻²), highlighting their potential for practical applications [57].

The dynamic structural regulation of active sites represents a paradigm shift in catalyst design that effectively circumvents the fundamental limitations imposed by linear scaling relationships in multi-step catalytic reactions. Through mechanisms such as dynamic dual-site cooperation, oxide path mechanisms, and lattice oxygen activation, these sophisticated catalytic systems enable independent optimization of adsorption energies for different intermediates that are normally constrained by thermodynamic scaling relationships.

The experimental methodologies outlined in this review—particularly advanced operando characterization techniques like XAFS, electrochemical TERS, and in situ vibrational spectroscopy—provide powerful tools for investigating these dynamic processes under actual reaction conditions. The synthesis approaches for creating pre-catalysts capable of undergoing beneficial structural evolution under operation conditions continue to advance in sophistication, enabling precise control at the atomic level.

Future research directions will likely focus on extending these principles to a broader range of catalytic reactions beyond oxygen electrochemistry, including COâ‚‚ reduction, nitrogen reduction, and organic synthesis reactions. The integration of machine learning approaches with dynamic catalyst design, along with the development of even more sophisticated multi-modal operando characterization platforms, will accelerate the discovery and optimization of next-generation catalysts that transcend traditional scaling relationship limitations.

As the field progresses, the strategic design of catalysts with deliberately engineered dynamic capabilities will undoubtedly play an increasingly central role in overcoming fundamental thermodynamic constraints, ultimately enabling more efficient and sustainable chemical transformations for energy conversion and chemical production.

Recent advancements in catalytic mechanisms have identified dual-site cooperative catalysis as a powerful strategy to overcome the persistent challenge of scaling relationships between reaction intermediates. By enabling segmentally synergistic mechanisms, intramolecular proton transfer, and electronic modulation via spin-state engineering, dual-site catalysts effectively decouple the adsorption energies of key intermediates that are typically linearly correlated in conventional single-site catalysts. This in-depth technical guide synthesizes current research to elucidate the fundamental principles, experimental methodologies, and characterization techniques underlying these mechanisms, providing researchers with a comprehensive framework for designing next-generation catalysts that balance high activity with stability across various electrochemical transformations.

The development of efficient electrocatalysts is fundamentally limited by scaling relationships—linear correlations between the adsorption energies of different reaction intermediates on catalytic surfaces. These relationships, particularly evident in reactions involving multiple proton-coupled electron transfers like the oxygen evolution reaction (OER), impose a theoretical minimum overpotential of approximately 0.4 eV even for optimal materials [62]. Conventional catalytic mechanisms face inherent trade-offs: catalysts following the adsorbate evolution mechanism (AEM) are constrained by these scaling relationships between *OOH and *OH intermediates, while those operating via the lattice-oxygen-mediated mechanism (LOM) often suffer from structural instability due to lattice oxygen migration [62].

Dual-site cooperative catalysis has emerged as a transformative approach to circumvent these limitations. By strategically distributing reaction steps across different metal sites with complementary electronic properties, these catalysts break the conventional scaling relationships without sacrificing structural stability. This guide examines three prominent manifestations of this principle: the dual-site segmentally synergistic mechanism (DSSM) in OER, Brønsted base-enhanced interfacial proton transfer in biomass electrooxidation, and spin-state modulation for nitrate-to-ammonia conversion, providing a comprehensive technical foundation for researchers working at the forefront of catalytic design.

Mechanistic Frameworks for Dual-Site Cooperation

Dual-Site Segmentally Synergistic Mechanism (DSSM) for Oxygen Evolution

The DSSM represents a paradigm shift in OER catalysis by leveraging complementary active sites to bypass traditional scaling relationships. In CoFeSx nanoclusters supported on carbon nanotubes, Co³⁺ (low-spin, t₂g⁶eg⁰) provides optimal OH* adsorption, while Fe³⁺ (medium-spin, t₂g⁴eg¹) favors O* adsorption [62]. This specialization enables a reaction pathway where these dual sites synergistically produce Co-O-O-Fe intermediates, directly accelerating the release of triplet-state oxygen (↑O=O↑) without requiring the formation of *OOH intermediates that create scaling relationship constraints [62].

Table 1: Performance Metrics of DSSM-Based OER Catalysts

Catalyst Material	Overpotential	Stability	Mechanistic Features	Comparative Performance
CoFeSx/CNT nanoclusters	Lower than commercial IrOâ‚‚	~633 hours without significant loss [62]	Co-O-O-Fe intermediates, triplet-state Oâ‚‚ release	Breaks scaling relationship between *OOH and *OH
Conventional AEM catalysts	Theoretical minimum ~0.4 eV	Typically high	*OOH and *OH intermediates constrained by scaling	Limited by inherent scaling relationships
LOM-based catalysts	Potentially lower	Limited due to oxygen vacancies	Lattice oxygen participation, structural reconstruction	Stability compromised despite higher activity

Brønsted Base-Mediated Interfacial Proton Transfer

In electrochemical biomass conversion, the sluggish kinetics of interfacial proton transfer often limit reaction rates. Recent research demonstrates that functionalizing cobalt-based catalysts with Brønsted bases (specifically phosphate, Pi groups) dramatically enhances proton transfer during 5-hydroxymethylfurfural electrooxidation (HMFOR) to 2,5-furandicarboxylic acid (FDCA) [63]. The Pi groups on cobalt surfaces shorten the distance between proton donor (HMF molecules) and acceptor (catalyst surface), effectively promoting interfacial proton transfer during the dehydrogenation of HMF [63].

This approach yields remarkable performance improvements: Pi-functionalized cobalt nanowire catalysts exhibit a 6.5-fold increase in current density compared to unmodified cobalt catalysts and achieve nearly 100% selectivity for FDCA, reaching a current density of 1000 mA cm⁻² at just 1.41 V_RHE [63]. This strategy specifically addresses the proton transfer bottleneck that typically limits rates in multi-step electrochemical transformations involving proton-coupled electron transfers.

Table 2: Quantitative Performance Enhancement via Brønsted Base Modification

Performance Metric	Pi-Co Catalyst	Unmodified Co Catalyst	Enhancement Factor
Current Density at 1.40 V_RHE	6.5× higher	Baseline reference	6.5× [63]
FDCA Selectivity	98.8%	16.1%	6.1× improvement [63]
FDCA Faradaic Efficiency	98.5%	50.3%	1.96× improvement [63]
FDCA Formation Rate	0.0274 μmol s⁻¹	0.00387 μmol s⁻¹	7.1× improvement [63]
Potential for 1000 mA cm⁻²	1.41 V_RHE	1.65 V_RHE	240 mV reduction [63]

Spin-State Modulation for Nitrate Reduction to Ammonia

Electronic modulation through spin-state engineering represents another dimension of dual-site cooperation. In Cu/Co dual-active sites embedded in Co₃O₄, the introduction of single-atom Cu creates asymmetric active sites that modify the coordination environment and manipulate spin states [64]. This configuration establishes a connected spin channel that optimizes the adsorption of intermediates while providing substantial *H for deep hydrogenation of nitrogen-containing intermediates [64].

The resulting D-Cu-Co₃O₄ catalyst achieves exceptional performance in electrochemical nitrate reduction to ammonia, with a Faradaic efficiency of 96.4 ± 2.8% at -0.4 V versus RHE and an ammonia yield rate of 1574.84 ± 46.70 μg h⁻¹ mg_cat⁻¹ [64]. This represents a biomimetic approach mirroring enzymatic mechanisms in biological nitrate reduction, where dual active sites enable efficient multi-electron, multi-proton transfer processes that are typically constrained by scaling relationships between intermediate adsorption energies.

Experimental Protocols and Methodologies

Synthesis of CoFeSx Nanoclusters on CNT Support

Objective: Fabricate single-domain ferromagnetic CoFeSx nanoclusters on carbon nanotubes for DSSM OER studies [62].

Procedure:

• Precursor Preparation: Adsorb cobalt and iron precursors onto pre-cleaned CNTs through strong interactions with oxygen functionalities on the CNT surface.
• Reduction Step: Treat with NaBHâ‚„ solution to form amorphous CoFeOâ‚“ nanosheets supported on CNTs (CFO-p/CNT).
• Sulfur Incorporation: Subject CFO-p/CNT to hydrothermal treatment with thioacetamide (TAA) as sulfur source. The acidic environment generated by TAA decomposition facilitates the transformation into CoFeSx nanoclusters while maintaining cluster size below 8 nm to preserve single-domain ferromagnetic properties [62].
• Purification: Wash thoroughly with deionized water and ethanol, then dry under vacuum at 60°C for 12 hours.

Critical Parameters: Control cluster size by regulating TAA concentration and hydrothermal duration (typically 6-12 hours at 120-180°C) to maintain single-domain ferromagnetic properties essential for spin-polarized OER [62].

Phosphate Group Functionalization of Cobalt Nanowires

Objective: Introduce Brønsted base (Pi) groups on cobalt nanowire surfaces to enhance interfacial proton transfer kinetics [63].

Procedure:

• Nanowire Growth: Synthesize Co nanowire precursor on nickel foam (NF) substrate via hydrothermal treatment at 105°C for 6 hours using cobalt chloride and urea precursors.
• Phosphating Treatment: Subject Co nanowires to phosphating process using sodium hypophosphite (NaHâ‚‚POâ‚‚) as phosphorus source at 300°C for 1-2 hours under argon atmosphere.
• Electrochemical Activation: Activate the Pi-functionalized catalyst through cyclic voltammetry scanning in 1.0 M KOH electrolyte (typically 20-50 cycles between 0.8-1.6 V versus RHE at 50 mV s⁻¹) [63].
• Quantification: Determine Pi content (typically ~2.77 mmol%) using ion chromatography [63].

Characterization Validation: Confirm successful functionalization through XPS analysis showing P-O-Co linkage at 531.2 eV in O 1s spectra and P 2p peaks at 133.6 eV (P 2p₃/₂) and 134.4 eV (P 2p₁/₂) corresponding to PO₄³⁻ groups [63].

Preparation of Cu/Co Dual-Active Sites on Co₃O₄

Objective: Create atomically dispersed Cu sites on Co₃O₄ to form dual-active sites for spin-state modulation [64].

Procedure:

• Co₃Oâ‚„ Support Synthesis: Mix cobalt chloride (CoCl₂·6Hâ‚‚O, 0.432 g) and urea (CO(NHâ‚‚)â‚‚, 0.216 g) in 30 mL water. Stir vigorously for 15 minutes, then heat at 105°C for 6 hours. Centrifuge to collect pink precipitate, dry overnight in vacuum oven, and calcine at 300°C for 3.5 hours to obtain black Co₃Oâ‚„ powder [64].
• Cu Deposition: Disperse 50 mg Co₃Oâ‚„ in ethanol via ultrasonication for 1 hour. Add quantified copper acetate solution with desired Cu loading (typically 0.5-5 wt%). Stir for 12 hours, then evaporate solvent and calcine at 300°C for 2 hours [64].
• Optimization: For single-atom dispersion (D-Cu-Co₃Oâ‚„), maintain Cu loading below 1.5 wt% to prevent nanoparticle formation and ensure atomic-level interaction with Co sites [64].

Structural Confirmation: Verify Cu single-atom dispersion through aberration-corrected HAADF-STEM imaging and maintain spinel structure integrity via XRD analysis [64].

Visualization of Dual-Site Mechanisms

Diagram 1: Three Pathways for Overcoming Catalytic Scaling Relationships

Diagram 2: Brønsted Base-Enhanced Interfacial Proton Transfer Mechanism

The Scientist's Toolkit: Essential Research Reagents

Table 3: Key Research Reagents for Dual-Site Catalyst Development

Reagent/Category	Function/Application	Specific Examples	Technical Notes
Metal Precursors	Provide source metals for active sites	Cobalt chloride (CoCl₂·6H₂O), Iron salts, Copper acetate [62] [63] [64]	Purify to >99% to prevent contamination; control oxidation state during synthesis
Support Materials	Provide high-surface-area support for nanoclusters	Carbon nanotubes (CNTs), Nickel foam (NF) [62] [63]	Pre-clean CNTs with acid treatment to introduce oxygen functionalities
Structure-Directing Agents	Control morphology and dispersion	Urea (CO(NHâ‚‚)â‚‚), Thioacetamide (TAA) [62] [64]	TAA decomposition creates acidic environment for cluster formation
Dopants/Functionalizers	Introduce secondary functionality	Sodium hypophosphite (NaHâ‚‚POâ‚‚) for Pi groups [63]	Post-synthetic functionalization preserves base catalyst structure
Characterization Standards	Validate catalyst composition and structure	ICP-OES standards, XPS calibration materials [62] [63]	Regular calibration essential for quantitative analysis

Advanced Characterization Techniques

Validating dual-site cooperation requires sophisticated characterization approaches to confirm structure-property relationships:

X-ray Photoelectron Spectroscopy (XPS): Essential for determining chemical states and coordination environments. Critical for identifying P-O-Co linkages at 531.2 eV in Pi-functionalized catalysts and confirming Co³⁺ (779.9 eV) and Fe³⁺ states in CoFeSx nanoclusters [62] [63].

High-Angle Annular Dark-Field STEM: Provides atomic-resolution imaging to verify nanocluster dispersion (CFS-ACs < 8 nm) and confirm single-atom Cu deposition in D-Cu-Co₃O₄ catalysts [62] [64].

In Situ Fourier Transform Infrared Spectroscopy: Monitors reaction intermediates in real-time, particularly valuable for identifying Co-O-O-Fe intermediates in DSSM and tracking hydrogenation intermediates in nitrate reduction [64].

Electrochemical Surface Area (ECSA) Determination: Quantifies active surface area through double-layer capacitance (Cdl) measurements, enabling normalization of activity for intrinsic property comparison [63].

Dual-site cooperative catalysis represents a fundamental advancement in overcoming the persistent challenge of scaling relationships in heterogeneous catalysis. The three mechanisms detailed in this guide—segmentally synergistic catalysis, Brønsted base-enhanced proton transfer, and spin-state modulation—provide complementary strategies for designing next-generation catalysts that simultaneously achieve high activity and stability. As research progresses, the integration of machine learning for predictive catalyst design, advanced operando characterization techniques, and biomimetic approaches inspired by enzymatic systems will further accelerate the development of efficient catalytic processes for clean energy conversion and sustainable chemical synthesis.

The pursuit of advanced catalytic systems has increasingly focused on engineering the microenvironments in which reactions occur, moving beyond the design of isolated active sites. The concepts of multi-functional surfaces and confined nanoscopic environments represent a paradigm shift in catalytic design, enabling unprecedented control over reaction pathways, selectivity, and stability. These approaches are particularly valuable for addressing one of the most fundamental challenges in catalysis: the scaling relationships between reaction intermediates.

Scaling relationships describe the linear correlations between the adsorption energies of different reaction intermediates on catalytic surfaces [34] [45]. These relationships arise because the binding energies of various intermediates (e.g., *O, *OH, *OOH in oxygen reduction reactions) are intrinsically linked to the properties of the catalyst material [45]. While these relationships provide valuable insights into catalytic behavior, they also create an insurmountable thermodynamic bottleneck for multi-step reactions, limiting the overall catalytic efficiency that can be achieved on conventional surfaces [45].

This technical guide explores how engineered heterogeneity through confined environments provides innovative strategies to circumvent these limitations. By creating tailored nanospaces that modify the physical and electronic properties of reactants and intermediates, confined catalysis offers a pathway to break conventional scaling relationships and achieve catalytic performance beyond traditional thermodynamic constraints [65] [66] [67].

Theoretical Foundations: Scaling Relationships and the Confinement Effect

The Fundamental Challenge of Scaling Relationships in Catalysis

Scaling relationships create a fundamental constraint in heterogeneous catalysis by linking the adsorption energies of different reaction intermediates. In the oxygen reduction reaction (ORR), for instance, the binding energies of *OOH, *O, and *OH species are linearly correlated, which establishes a thermodynamic overpotential ceiling that even ideal catalysts cannot overcome [45]. This relationship means that strengthening the binding of one intermediate inevitably strengthens the binding of others, preventing all elementary steps from being optimized simultaneously.

The origin of these relationships lies in the similar bonding configurations of different intermediates on the same catalytic surface. For example, both *O and *OH typically bind to the same surface sites through similar electronic interactions. This intrinsic limitation has been described as "the bottleneck for accelerating kinetics" in multi-step catalytic reactions [45].

Nanoconfinement as a Strategic Solution

Nanoconfined catalysis utilizes spatial constraints at the nanoscale to modify the physical and chemical properties of molecules and reaction intermediates [65] [66]. Professor Xinhe Bao pioneered this concept, defining it as a "catalytic mechanism that modulates catalyst electronic structures via nanoscale spatial or interfacial confinement effects" [66]. The confinement effect can be conceptually understood as a spatial phenomenon where nanocavities impose geometric constraints on guest molecules, while electronic interactions between the nanocavity wall and the guest can regulate their atomic and electronic structures [66].

The core principle leverages quantum effects and microenvironment synergies to stabilize active sites and optimize intermediate adsorption energy, thereby markedly enhancing catalytic efficiency and selectivity [66]. Research has demonstrated that confined spaces can significantly alter multiple aspects of catalytic processes:

• Structure and behavior of water molecules [65]
• Adsorption energy of oxidants and their bond lengths [65]
• Reaction energy barriers [65]
• Reaction paths [65]

Table 1: Classification of Confined Catalysts by Dimensionality

Dimensional Classification	Structural Characteristics	Representative Materials	Primary Confinement Effects
0D Confinement	Nanocavities spatially confined in three dimensions	Zeolites, Metal-Organic Frameworks (MOFs)	Nest effect, quantum confinement effect [67]
1D Confinement	Nanocavities spatially confined in two dimensions	Carbon nanotubes (CNTs)	Altered molecular orbital energy, changed activity of confined molecules [67]
2D Confinement	Planar microenvironment between surfaces and 2D material overlayers	Graphene/Pt(111), h-BN/metal systems	Geometric constraint and confinement field effects [67]
3D Confinement	Core-shell structures and complex porous networks	Core-shell nanoparticles, hierarchical porous materials	Synergistic combination of multiple confinement effects [65]

Experimental Realization of Confined Catalytic Systems

Design and Fabrication Strategies for Confined Environments

The construction of effective confined catalysts requires precise control over nanoscale architecture. Several strategic approaches have been developed to achieve this:

3.1.1 Encapsulation within Porous Materials This approach involves confining catalytic nanoparticles within structured porous materials such as zeolites, metal-organic frameworks (MOFs), or mesoporous silicas. For example, embedding the confined catalyst within a membrane addresses the challenges associated with the recovery and reuse of traditional powdered catalysts [65]. A specific demonstration involved the confinement of AuCu bimetallic nanoparticles within EDTA-modified UiO-66 (a typical MOF), utilizing MOF channels to restrict metal growth. This strategy significantly reduced the particle size and enriched the oxygen vacancies, thereby enhancing the catalytic activity for CO2 hydrogenation to methanol with a 3.21-fold improvement in methanol space-time yield [66].

3.1.2 2D Material Overlayers Creating 2D spaces between catalyst surfaces and 2D material overlayers like graphene or hexagonal boron nitride (h-BN) provides a well-defined model system to explore confined catalysis [67]. These systems enable the "catalysis under cover" concept, where the 2D cover creates a unique microenvironment that modulates surface reactions [67]. Experimental and theoretical studies on graphene/Pt(111) systems have demonstrated that adsorption of atoms and molecules on the Pt(111) surface is consistently weakened under monolayer graphene, attributed to both geometric constraint and confinement field effects in the 2D space [67].

3.1.3 Core-Shell Nanostructures Fabricating core-shell structures with catalytic materials encapsulated within protective or functional shells creates confined environments that enhance stability and selectivity. For instance, Ag@Cu2O cascade nanoreactors with tunable shell thickness were designed for CO2 reduction, demonstrating that moderate shell thickness optimizes CO diffusion and confinement effects to boost C2 product selectivity [66].

Detailed Experimental Protocol: Construction of N-bridged Pt = N2 = Fe Atomic-Scale Bimetal Assembly

The following protocol details the synthesis of a confined catalytic system specifically designed to circumvent scaling relationships in the oxygen reduction reaction [45]:

Diagram 1: Atomic-scale bimetal assembly synthesis workflow

Materials and Equipment:

• Pyrene (C16H10) feedstock
• Nitric acid (HNO3), concentrated
• FeCl3·6H2O and H2PtCl6 as metal precursors
• Glycol solvent
• Freeze-dryer
• Tube furnace with inert gas capability
• Characterization equipment: STEM-HAADF, XPS, XAFS

Step-by-Step Procedure:

• Synthesis of Amino-Functionalized Carbon Nanoflakes (CNF-NH2):

  • Nitrate pyrene (C16H10) into trinitropyrene using hot HNO3 under reflux at 80°C for 6 hours.
  • Replace the nitro groups (-NO2) with amino groups (-NH2) through aminofunctionalization using ammonium hydroxide at 60°C for 12 hours to obtain CNF-NH2 [45].

• Preparation of Bimetallic Precursor Solution:

  • Dissolve FeCl3·6H2O and H2PtCl6 in glycol solvent to promote effective affinity between the hetero-electric metal groups [45].
  • Maintain a molar ratio of Pt:Fe at approximately 2:1 based on the final catalyst composition of 2.93 wt% Pt and 1.34 wt% Fe [45].

• Chelation and Freeze-Drying:

  • Mix the bimetallic precursor solution with CNF-NH2, allowing the metal cations to chelate with the functionalized amine groups (-NH2).
  • Freeze-dry the mixture to obtain a homogeneous solid precursor [45].

• Pyrolysis and Catalyst Formation:

  • Heat the precursor to 700°C in an inert atmosphere (Ar or N2) for 2 hours.
  • During pyrolysis, the metal atoms become atomically dispersed and form the N-bridged Pt = N2 = Fe structure with precisely controlled intermetallic spacing of approximately 2.83-2.91 Ã… [45].

Characterization and Validation:

• Perform Cs-corrected STEM-HAADF imaging to confirm atomic dispersion and interatomic distances.
• Conduct XPS analysis to verify chemical states and absence of metal oxide species.
• Use in situ synchrotron radiation XAFS to monitor the formation of the Pt-O-O-Fe intermediate during ORR [45].

Research Reagent Solutions for Confined Catalyst Development

Table 2: Essential Research Reagents for Confined Catalyst Synthesis

Reagent/Chemical	Function in Synthesis	Specific Application Example
Aminopropyltriethoxysilane (APTES)	Surface functionalization agent; provides amine groups for metal anchoring	Functionalization of supports for single-atom catalysts [45] [68]
Metal Precursors (H2PtCl6, FeCl3·6H2O, AgNO3)	Source of catalytic metal centers	Formation of atomic-scale bimetal assemblies (Pt-Fe) [45] or nanoparticles [68]
2D Materials (Graphene, h-BN)	Creation of 2D confined spaces	Overlayer on metal surfaces to modify adsorption energies [67]
Porous Supports (Zeolites, MOFs, SBA-15)	Providing structured confinement environments	Encapsulation of catalytic nanoparticles (e.g., Ni in SBA-15) [65] [66]
Tridecafluoroctyl triethoxysilan (FAS)	Hydrophobic surface modification	Creating self-cleaning features in multifunctional coatings [68]
Reducing Agents (Na ascorbate, trisodium citrate)	Nanoparticle formation and stabilization	Synthesis of silver nanoparticles by Turkevich method [68]

Mechanistic Insights: How Confinement Modifies Catalytic Fundamentals

Electronic and Geometric Effects in Nanoconfined Spaces

The remarkable properties of confined catalysts arise from fundamental modifications to both electronic structures and geometric constraints:

Electronic Effects: Confined environments significantly alter the electronic structures of both catalysts and reactants. Studies on carbon nanotube (CNT)-confined catalysis revealed that shortened catalyst bond lengths lead to weakened binding energy and a downshift of the d-band center, elucidating the mechanism behind activity enhancement or suppression [66]. In the case of the Pt = N2 = Fe atomic-scale bimetal assembly, the N-bridged structure enables a unique electronic configuration that promotes direct O-O radical breakage without the formation of redundant *OOH intermediates, thereby circumventing conventional scaling relationships [45].

Geometric Constraints: The physical confinement in nanospaces creates substantial geometric constraints that affect molecular orientation and reaction pathways. Research on graphene/Pt(111) systems demonstrated that the van der Waals interaction between the graphene overlayer and Pt substrate induces strong geometric constraint on trapped interlayer molecules [67]. This constraint was quantified by calculating the interaction between the Gr layer and CO molecule, showing that the equilibrium distance (dGr-CO) shifts from 3.0 Ã… in free space to 2.7 Ã… in the confined space between Gr and Pt(111) [67].

Breaking Scaling Relationships Through Dual-Site Mechanisms

The most significant advantage of confined catalysis is its ability to circumvent traditional scaling relationships through innovative reaction mechanisms. The Pt = N2 = Fe system exemplifies this approach by implementing a dual-site mechanism that bypasses the formation of *OOH intermediates [45].

Traditional Single-Site Mechanism: O2 → *O2 → *OOH → *O → *OH → H2O (This pathway is constrained by scaling relationships between *OOH, *O, and *OH) [45]

Dual-Site Mechanism in Confined Catalysis: O2 → Pt-O-O-Fe → *O + *O → *OH + *OH → H2O (This pathway avoids the *OOH intermediate entirely) [45]

In situ synchrotron radiation Fourier transform infrared spectroscopy (SR-FTIR) has directly monitored the formation of the key intermediate state (Pt-O-O-Fe) in the Pt = N2 = Fe ABA catalysts under operating conditions, confirming that the ORR follows this dual-sites mechanism without the production of *OOH intermediates [45]. This mechanism enables nearly two orders of magnitude enhanced kinetic current density at the half-wave potential of 0.95 V relative to commercial Pt/C and an almost 99% efficiency of 4-electron pathway selectivity [45].

Diagram 2: Comparison of ORR mechanisms with and without confinement

Advanced Characterization Techniques for Confined Systems

Understanding the mechanisms of confined catalysis requires specialized characterization approaches:

In Situ/Operando Spectroscopy: Advanced techniques such as in situ synchrotron radiation X-ray absorption fine structure (XAFS) spectroscopy and Fourier transform infrared spectroscopy (FTIR) are essential for probing the structure and behavior of catalysts under actual reaction conditions [45]. These methods have directly revealed the formation of key intermediates such as the Pt-O-O-Fe state in confined catalysts [45].

Surface Science Methods: Well-defined 2D confined environments between catalyst surfaces and 2D material overlayers are particularly amenable to investigation by surface science techniques [67]. These systems provide ideal model environments for fundamental studies of confinement effects.

Computational Modeling: Density functional theory (DFT) calculations have been instrumental in understanding confinement phenomena at the atomic level [67] [69]. Studies using random forest models combined with Grand Canonical Monte Carlo simulations have provided insights into how confinement affects catalyst bond lengths and binding energies [66].

Performance Metrics and Applications

Quantitative Assessment of Confined Catalyst Performance

The efficacy of confined catalytic systems is demonstrated through significant enhancements in key performance metrics:

Table 3: Performance Comparison of Confined vs. Traditional Catalysts

Catalytic System	Reaction	Key Performance Metrics	Improvement Over Conventional Catalysts
Pt = N2 = Fe Atomic Bimetal Assembly	Oxygen Reduction Reaction (ORR)	Kinetic current density: 5.83 mA cm⁻² at 0.95 V; 4-electron pathway selectivity: ~99% [45]	Nearly two orders of magnitude higher kinetic current than Pt/C [45]
Sub-2 nm Ni in SBA-15	Dry Methane Reforming	Stability against sintering; enhanced active site exposure [66]	Addresses Ni catalyst deactivation via sintering [66]
AuCu in EDTA-modified UiO-66	CO2 Hydrogenation to Methanol	Methanol space-time yield [66]	3.21-fold improvement [66]
Ag@Cu2O Cascade Nanoreactors	CO2 Reduction to C2 Products	C2 product selectivity [66]	Enhanced selectivity through optimized CO diffusion and confinement [66]
C/Cu/C Sandwich Structure	CO2 to C2 Products Electrocatalysis	C2 product selectivity [66]	Doubled selectivity through 2D nanoconfinement effect [66]

Application in Multi-Functional Surface Systems

The principles of confined catalysis extend beyond traditional chemical transformations to enable sophisticated multi-functional surfaces. Recent research has demonstrated the development of hybrid inorganic-organic materials that combine antibacterial properties, hydrophobic character, and fluorescent features in a single platform [68]. These systems exemplify how nanoconfinement strategies can be applied to create surfaces with tailored multifunctionality.

For instance, silver nanoparticles embedded together with inorganic and organic surface coatings and silicon quantum dots create symbiotic antibacterial character and UV-excited visible light fluorescent features [68]. When combined with fluorosilane materials to add hydrophobic functionality (achieving water contact angles around 120°), these systems demonstrate self-cleaning capabilities [68]. Such integrated approaches showcase the potential of confined environment engineering to address complex application requirements that demand multiple simultaneous functionalities.

Computational Approaches and Future Directions

Computational Tools for Confined Catalyst Design

The design and optimization of confined catalysts increasingly relies on computational methods:

Density Functional Theory (DFT): DFT remains the cornerstone for computational studies of confined catalytic systems, despite challenges in selecting appropriate exchange-correlation functionals and accounting for van der Waals interactions in confined spaces [69]. Popular packages for periodic DFT calculations include VASP, Quantum Espresso, and CASTEP, which are essential for modeling extended periodic structures such as crystal surfaces and 2D material overlayers [69].

Machine Learning and Advanced Sampling: The high-dimensional nature of catalyst design problems makes them suitable for machine learning approaches. Recent efforts have focused on developing both forward and inverse catalyst mapping tools that are less data-dependent and more transferable [69]. These approaches are particularly valuable for navigating the complex parameter space of confined catalytic systems.

Challenges and Future Perspectives

Despite significant progress, several challenges remain in the development and implementation of confined catalytic systems:

Mechanistic Understanding: The mechanisms governing confined catalysis remain partially obscure, necessitating more in-depth and meticulous studies to elucidate them [65]. Particularly needed are advanced in situ characterization techniques that can probe the internal mechanisms of confined catalysis under operational conditions [65].

Synthesis Control: Precise control over the structure of confined environments at the atomic scale remains challenging. Recent investigations demonstrate that reducing the confinement space to the size of angstrom can further enhance performance, presenting an innovative strategy for advancing water treatment technology and other applications [65]. However, reliably synthesizing such precisely controlled structures at scale requires further development.

Economic Viability: A thorough analysis of the economic feasibility of confined materials is urgently needed [65]. While confined catalysts often demonstrate exceptional stability, selectivity, and economic viability when addressing complex environments, comprehensive techno-economic assessments are necessary to guide industrial implementation [65].

Future advancements will likely focus on integrating machine learning, in situ characterization, and modular design as essential pathways to establish structure-activity correlations and accelerate the industrial implementation of confined catalytic systems [66]. These integrated approaches promise to bridge fundamental mechanisms with practical applications to advance carbon-neutral energy systems and other critical technologies [66].

This technical guide explores the integral role of proton acceptors and bimetallic synergy in enhancing the catalytic performance of Ni-Fe molecular catalysts. Framed within the context of scaling relationships between reaction intermediates, this review demonstrates how strategic catalyst design can circumvent fundamental limitations in catalytic efficiency. By examining proton-coupled electron transfer mechanisms and structural configurations in bimetallic systems, we provide a comprehensive analysis of how Ni-Fe catalysts achieve remarkable performance in applications ranging from COâ‚‚ methanation to water oxidation. The synthesis methodologies, advanced characterization techniques, and mechanistic insights presented herein offer researchers a roadmap for developing next-generation catalytic systems that transcend conventional scaling relationship constraints.

The Scaling Relationship Challenge in Catalytic Reactions

Scaling relationships describe the linear correlations between adsorption energies of different reaction intermediates on catalytic surfaces, creating fundamental limitations to catalytic efficiency. These relationships arise from the similar bonding nature of intermediates sharing common atoms and emerge naturally from the physical laws governing interactions between nuclei and electrons [2]. For multi-step reactions involving multiple intermediates, these linear dependencies reduce the degrees of freedom available for optimization, effectively creating an "volcano plot" where catalytic activity reaches an inherent maximum. The challenging requirement of bringing reaction intermediates in close proximity often serves as the main obstacle in critical reactions such as oxygen evolution and COâ‚‚ reduction [70].

The limitations imposed by scaling relationships become particularly evident in reactions involving proton-coupled electron transfer (PCET), where proton transfer (PT) and electron transfer (ET) can occur in either a stepwise (PTET or ETPT) or concerted proton electron transfer (CPET) pathway. Compared to the concerted pathway, the stepwise pathway requires higher energy to generate reaction intermediates, making it energetically unfavorable [71]. This energy landscape is profoundly affected by proton donor/acceptor characteristics, including concentration, pKa, and steric structure, which collectively influence potential response, reaction rate, and mechanism.

Bimetallic Systems as a Strategic Solution

Bimetallic catalysts present a promising strategy for overcoming scaling relationship limitations through the introduction of additional active sites with distinct electronic properties. The strategic combination of metals enables optimization of adsorption/desorption energies for different reaction intermediates across distinct metal centers, potentially breaking the linear energy relationships that constrain monometallic catalysts [72]. In particular, Ni-Fe bimetallic systems have emerged as a versatile platform for investigating and exploiting these effects across diverse catalytic applications, from thermal and plasma-catalytic COâ‚‚ methanation to water oxidation electrocatalysis [73] [70].

Table 1: Performance Comparison of Monometallic and Bimetallic Ni-Fe Catalysts in COâ‚‚ Methanation

Catalyst Configuration	Synthesis Method	Methane Productivity	Key Characteristics
10%Ni WI (monometallic)	Wet impregnation	Baseline (0%)	Reference performance
10%Ni3%Fe WI	Wet co-impregnation	+78% increase	Segregated Janus-like structure
10%Ni3%Fe DP	Co-deposition precipitation	-90% reduction	Fe-excessively decorated Ni
10%Ni DP	Co-deposition precipitation	Not reported	Fe-encapsulated Ni structure

Proton Transfer Mechanisms in Ni-Fe Catalytic Systems

Metal-Hydroxyl Groups as Intramolecular Proton Transfer Relays

In heterogeneous Ni-Fe systems, adjacent metal-hydroxyl groups function as intramolecular proton-electron transfer relays to enhance reaction kinetics. Research has demonstrated that in well-defined molecular platforms with aza-fused π-conjugated microporous polymers coordinating molecular Ni or Ni-Fe sites, both anions in solution and adjacent Ni³⁺–OH sites act as proton transfer relays [70]. This facilitation is particularly crucial for O–O bond formation, leading to pH-dependent water oxidation kinetics.

The proton relay function operates through a mechanism where high-valent Fe⁴⁺ species serve as active sites for water nucleophilic attack (WNA), while adjacent OH-bridged Ni³⁺ species function as intramolecular proton transfer (IPT) sites that accelerate deprotonation kinetics during O–O bond formation. This cooperative effect represents a significant advancement in understanding the controversial synergistic interaction between Ni and Fe centers in heterogeneous (oxy)hydroxide-based NiFe catalysts [70].

Proton Donor/Acceptor Effects on Reaction Pathways

Proton donor/acceptor species can directly act as redox electroactive species or provide interactive proton transfer species for PCET reactions. These species, typically originating from the electrolyte or electrocatalyst (solvent, buffer, extra added species, and functional groups of electrocatalyst), modulate electrochemical PCET reactions through their concentration, pKa, and steric hindrance [71].

In Ni-Fe catalytic systems, the strategic placement of intramolecular proton transfer sites near metal centers in the secondary coordination sphere markedly accelerates proton transfer and stabilizes charged intermediates through the intramolecular atom–proton transfer (APT) mechanism. This arrangement is particularly effective in overcoming the rate limitations imposed by stepwise proton-electron transfers in conventional catalytic cycles [70].

Structural Configurations and Synthesis of Ni-Fe Catalysts

Controlled Synthesis Methods for Tailored Bimetallic Configurations

The spatial arrangement of Ni and Fe atoms within bimetallic catalysts profoundly influences their catalytic performance by modulating metal-metal interactions and active site accessibility. Two primary synthesis methods have been developed to achieve distinct bimetallic configurations:

Wet Co-Impregnation (WI) produces catalysts with segregated Janus-like structures or Fe-decorated Ni configurations where the interaction between Fe and Ni influences the nature of Ni species within the deposit. Sufficient intimacy between Ni-Fe or a sufficiently large Fe presence, evidenced by an increased fraction of Ni(OH)â‚‚ in the deposits, can boost methane productivity by 78% relative to monometallic Ni catalysts [73].

Co-Deposition Precipitation (DP) tends to create structures where Fe excessively decorates or encapsulates the Ni deposits. In these configurations, the benefit of bimetallic intimacy is lost, resulting in a 90% reduction in methane productivity relative to optimal Ni-Fe ratios [73]. This performance reduction highlights the critical importance of controlled synthesis in achieving productive bimetallic synergy rather than simple physical mixture of metals.

Atomic-Scale Bimetal Assemblies with Regulated Intermetallic Distances

For electrochemical reactions such as oxygen reduction, precisely controlled intermetallic distances in atomic-scale bimetal assemblies (ABA) can catalyze direct O–O radical breakage without forming redundant *OOH intermediates, thereby regulating inherent linear scaling relationships. Nitrogen-bridged Pt = N₂ = Fe assemblies with intermetallic distances of approximately 2.83–2.91 Å promote the generation of a key intermediate state (Pt–O–O–Fe) during the ORR process, resulting in high reaction kinetics and selectivity [45].

This design principle is translatable to Ni-Fe systems, where appropriate interatomic spacing between two adjacent metallic active sites is mandatory for the dissociation of *O₂ and the triggering of O–O radical breakage (M–O–O–M) to demonstrably suppress M–OOH production and limit two-electron reaction path selectivity [45].

Table 2: Structural and Performance Characteristics of Ni-Fe Catalysts by Synthesis Method

Synthesis Parameter	Wet Co-Impregnation (WI)	Co-Deposition Precipitation (DP)
Typical Structure	Segregated Janus-like or Fe-decorated Ni	Fe-encapsulated or excessively decorated Ni
Metal-Metal Interaction	Moderate, regulated by composition	High intimacy, potentially excessive
Ni(OH)â‚‚ Fraction	Increases with optimal Fe presence	Variable, depends on encapsulation degree
Methane Productivity	Up to +78% at optimal Ni:Fe ratios	Up to -90% with excessive Fe encapsulation
Recommended Applications	Plasma-catalytic COâ‚‚ methanation, thermal catalysis	Systems requiring intimate metal contact

Advanced Characterization of Ni-Fe Catalysts and Intermediates

Operando Spectroscopy for Mechanistic Insights

Understanding the dynamic structural evolution and reaction mechanisms of Ni-Fe catalysts under operating conditions requires advanced characterization techniques. Operando X-ray absorption near edge structure (XANES) spectroscopy reveals valence states of metal centers during catalysis, showing that Ni centers in Ni-Fe sites reach only +2.5 oxidation state under 1.7 V, implying the absence of Ni⁴⁺ during catalysis, in contrast to monometallic Ni systems that form Ni⁴⁺ species [70].

Sensitive in situ synchrotron radiation Fourier transform infrared spectroscopy (SR-FTIR) can monitor the formation of key intermediate states (Pt–O–O–Fe) in bimetal assembly catalysts under operating conditions, demonstrating reaction pathways that follow dual-sites mechanisms without the production of *OOH intermediates [45]. This direct observational capability provides critical evidence for mechanism validation beyond theoretical predictions.

Computational Modeling of Electronic Structure and Reaction Pathways

Density functional theory (DFT) calculations provide essential insights into the electronic structure modifications induced by bimetallic synergy in Ni-Fe catalysts. Studies of Fe/Ni–N–C materials for electrochemical CO₂ reduction reveal that the electrical properties and catalytic action of Fe atoms are largely determined by spin orientation, with high-spin states potentially increasing CO₂ activation by offering more accessible electron states [74].

In Ni-Fe systems, the incorporation of electron-withdrawing Fe³⁺ modifies the d-band center of Ni sites, optimizing the adsorption/desorption energies of intermediates and potentially breaking linear scaling relationships. These electronic structure modifications enable distinctive reaction pathways not accessible to monometallic counterparts [72] [74].

Applications and Performance of Ni-Fe Bimetallic Catalysts

Plasma-Catalytic COâ‚‚ Methanation

The integration of Ni-Fe catalysts with non-thermal plasma (NTP) enables COâ‚‚ methanation at lower temperatures than thermal catalysis alone, simultaneously electrifying the chemical process. In these systems, the arrangement of individual metals within catalyst deposits significantly influences plasma-catalyst synergy, with optimal configurations boosting methane productivity by 78% relative to monometallic Ni catalysts [73].

The Ni-Fe synergy appears translatable between thermal-catalytic and plasma-catalytic environments, underscoring the potential for adapting bimetallic benefits across different energy input modes. This translatability is particularly valuable for industrial applications seeking to leverage renewable electricity for chemical synthesis through plasma processes [73].

Electrochemical and Photochemical COâ‚‚ Reduction

In electrochemical CO₂ reduction reactions (CO₂RR), Ni-Fe bimetallic catalysts exhibit enhanced performance compared to their monometallic counterparts due to synergistic effects between the metals. Fe/Ni–N–C electrocatalysts achieve impressive faradaic efficiency for CO production (FECO) of 92.9% at -0.677 V versus reversible hydrogen electrode (RHE), maintaining over 89% efficiency after 40 hours of continuous operation [74].

DFT studies reveal that the binary metal combination effect increases charge transfer rates, resulting in favorable kinetics and long-term electrochemical performance. The pairing of Ni with Fe optimizes the adsorption energies of key intermediates (*COOH, *CO), enabling more efficient progression through the COâ‚‚ reduction pathway [74].

Water Oxidation Catalysis

Ni-Fe catalysts demonstrate exceptional performance in the oxygen evolution reaction (OER), a critical process for water splitting and renewable energy storage. Well-defined molecular Ni–Fe sites in aza-fused π-conjugated microporous polymers exhibit outstanding turnover frequencies (TOFs), with TOF₍redox-active₎ reaching 18.7 s⁻¹ at an overpotential of 300 mV [70].

This performance surpasses most nickel-iron-oxy-hydroxide-based catalysts and is comparable to the TOF₍surface₎ of NiFeOₓHᵧ nanoparticles (6.2 s⁻¹, η = 300 mV) and Fe:NiOOH (10.4 s⁻¹, η = 300 mV) [70]. The enhanced activity stems from the cooperative mechanism between Ni and Fe sites, where metal-hydroxyl groups mediate intramolecular proton transfer during the critical O–O bond formation step.

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Essential Research Reagents and Materials for Ni-Fe Catalyst Studies

Reagent/Material	Function/Application	Technical Specifications
Nickel(II) nitrate hexahydrate	Ni precursor for catalyst synthesis	≥98.5% purity (Sigma-Aldrich)
Iron(III) nitrate nonahydrate	Fe precursor for catalyst synthesis	≥98% purity (Sigma-Aldrich)
γ-Al₂O₃ support	High-surface-area catalyst support	<50 nm, ≥98% purity (Sigma-Aldrich)
Urea	Precipitation agent in DP synthesis	≥98% purity (Sigma-Aldrich)
Aza-fused π-conjugated polymer	Molecular platform for defined sites	Phenanthroline-like coordination sites
Carbon nanoflakes (CNF–NH₂)	Support for atomic-scale assemblies	Amino-functionalized for metal anchoring

The strategic integration of proton acceptors and bimetallic synergy in Ni-Fe molecular catalysts presents a powerful approach for overcoming fundamental scaling relationship limitations in catalysis. Through controlled synthesis methods that regulate metal-metal interactions and spatial configurations, researchers can optimize proton-coupled electron transfer pathways to achieve enhanced activity, selectivity, and stability across diverse applications from COâ‚‚ conversion to water oxidation.

Future research directions should focus on further elucidating the dynamic structural evolution of Ni-Fe catalysts under operating conditions, developing more precise synthetic control over intermetallic distances and coordination environments, and expanding the application of these design principles to other bimetallic combinations. The continued integration of advanced operando characterization with computational modeling will accelerate the discovery and optimization of next-generation catalysts that transcend conventional scaling relationship constraints.

As the field advances, the translation of these fundamental insights to industrial applications will be crucial for addressing global challenges in renewable energy storage and carbon utilization, ultimately contributing to the development of sustainable chemical processes.

The pursuit of efficient catalysts is fundamentally guided by structure-activity relationships, which have long relied on the analysis of catalyst structures before and after reactions. However, this static perspective often fails to capture the true nature of catalytic systems, as it overlooks the profound structural transformations that occur during operation. The limitation of this approach becomes particularly evident when considering linear scaling relationships (LSRs), which create inherent thermodynamic constraints in multi-step catalytic reactions. These relationships linearly correlate the adsorption energies of reaction intermediates on conventional single-site catalysts, placing intrinsic limitations on optimally adjusting the adsorption of every intermediate simultaneously to achieve maximum activity. This is especially problematic for reactions like the oxygen evolution reaction (OER), where the adsorption energies of intermediates such as *OH, *O, and *OOH are linearly correlated and cannot be adjusted independently on a static active site [1]. However, emerging research reveals that these scaling relationships are not absolute constraints but rather artifacts of examining catalysts under static, non-operative conditions. This whitepaper synthesizes recent advances demonstrating that active sites in heterogeneous and molecular catalysis undergo significant structural evolution under reaction conditions, providing a strategic pathway to circumvent traditional scaling limitations and design next-generation catalytic systems.

The Theoretical Framework: Scaling Relationships and Their Limitations

The Fundamental Challenge of Linear Scaling Relationships

Linear scaling relationships establish that the adsorption energies of different reactive intermediates in multi-step catalytic reactions are correlated on conventional single-site catalysts [1]. For instance, in the OER following the adsorbate evolution mechanism, the adsorption energies of *OOH and *OH scale linearly with each other across various catalyst materials. This relationship arises because these intermediates bind to the same active site through chemically similar interactions. The practical consequence is a fundamental thermodynamic limitation: strengthening the binding of one intermediate inevitably weakens the binding of another, creating an unavoidable compromise that caps the maximum achievable catalytic activity. This manifests as a constrained overpotential that cannot be improved beyond a certain point for a given class of static catalysts.

Dynamic Active Sites as a Pathway Beyond Scaling Limitations

The recognition that catalytic sites are not static but dynamically evolve under reaction conditions provides an innovative strategy to circumvent LSRs. When active sites undergo structural changes during the catalytic cycle, they can effectively alter their electronic structure and binding properties at different stages of the reaction pathway. This dynamic regulation enables simultaneous optimization of the adsorption strengths for multiple intermediates, breaking the constraints imposed by static scaling relationships [1]. The key mechanism involves structural evolution that decouples the binding energies of different intermediates, allowing catalysts to achieve intermediate binding strengths that would be thermodynamically inaccessible in static configurations. For example, in bimetallic systems, the dynamic coordination between metal centers and adsorbates can selectively stabilize certain transition states while destabilizing others, effectively creating a "time-averaged" active site that outperforms any static configuration.

Experimental Evidence: Documented Cases of Active Site Evolution

Structural Evolution in Single-Atom and Cluster Catalysts

Metal single-atom catalysts (SACs), particularly those with M–N–C configurations, exemplify the dynamic nature of active sites under operational conditions. These systems demonstrate remarkable structural plasticity, with single atoms aggregating into clusters and clusters re-dispersing into single atoms in response to reaction environments:

• Copper Single-Atom Systems: In Cu–N–C catalysts during the nitrate reduction reaction (NO3RR), operando X-ray absorption spectroscopy (XAS) and identical-location electron microscopy revealed the potential-driven transformation of Cu single atoms with Cu–N4 coordination into Cu clusters when the applied potential switched from 0.00 to -1.00 V vs. RHE. This structural evolution correlated with a dramatic enhancement in NH3 production, achieving rates of 4.5 mg cm−2 h−1 with 84.7% faradaic efficiency, confirming that the dynamically formed Cu clusters served as the true active species [75].

• Platinum Single-Atom Systems: Pt single atoms stabilized on alumina supports in O2 atmosphere undergo dynamic structural evolution to form Pt clusters in either H2 or a mixture of CO + O2 atmosphere during CO oxidation. This reversible transformation does not deactivate the catalyst but creates the actual active sites that promote efficient chemical reactions [75].

• Rare-Earth and Noble Metal Systems: Under pyrolysis conditions, clusters of rare-earth, transition, and noble metals dynamically evolve into single atoms stabilized by neighboring N-dopants, forming thermodynamically stable M–N–C structures. This cluster-to-single-atom transformation demonstrates the bidirectional nature of structural evolution in catalytic systems [75].

Table 1: Documented Cases of Structural Evolution in Catalytic Systems

Catalyst System	Reaction	Structural Evolution	Impact on Performance
Cu–N–C	NO3RR and CO2RR	Cu SAs → Cu clusters under negative potential	Enhanced NH3 and C2H5OH production [75]
Pt/Alumina	CO oxidation	Pt SAs → Pt clusters in H2 or CO+O2	Creation of actual active sites [75]
Ni-Fe molecular complex	OER	Ni monomer → O-bridged Ni-Fe2 trimer	Breaks OER scaling relationships [1]
Marcasite CoSe2	Acidic HER	Surface corrosion → disordered Se-Co-Se moieties	Maintains structural integrity and activity [76]
Marcasite CoSe2	Alkaline HER	Surface oxidation → metallic Se-Co-Co-Se moieties	Generation of true active species [76]

Dynamic Sites in Oxygen Evolution Catalysis

The oxygen evolution reaction has served as a critical testbed for studying dynamic active site evolution due to its technological importance and well-characterized scaling relationships:

• Ni-Fe Molecular Complex: A model Ni-Fe2 molecular catalyst constructed via in situ electrochemical activation demonstrates how dynamic structural regulation circumvents OER scaling relationships. Operando XAFS analysis verified the structural transformation from Ni monomer to O-bridged Ni-Fe2 trimer during the activation process. Theoretical calculations and electrokinetic studies revealed that the dynamic evolution of Ni-adsorbate coordination, driven by intramolecular proton transfer, modulates the electronic structure of the adjacent Fe active center during the catalytic cycle. This dynamic dual-site cooperation simultaneously lowers the free energy change for O–H bond cleavage and O–O bond formation, thereby disrupting the inherent scaling relationship in OER [1].

• Cobalt-Based Chalcogenides: Operando spectroelectrochemical monitoring of CoSe2 catalysts reveals that the in situ formation of highly disordered Co(IV) species is a common denominator for OER activity across different chalcogenide catalysts. This transformation occurs regardless of the initial catalyst structure, indicating that the pre-catalyst structure may differ significantly from the true active phase under operational conditions [76].

pH-Dependent Restructuring in Water Electrolysis

The dynamic behavior of active sites exhibits significant dependence on operational conditions, particularly pH, as demonstrated in cobalt diselenide (CoSe2) catalysts for overall water electrolysis:

• Acidic vs. Alkaline HER: Marcasite-type CoSe2 undergoes distinct restructuring processes under different pH conditions. In acidic environments, it experiences slight surface corrosion, producing disordered Se-Co-Se moieties that catalyze the hydrogen evolution reaction (HER). In alkaline media, however, it undergoes potential-driven restructuring from an initially reconstructed O-rich surface into metallic Se-Co-Co-Se moieties that serve as the true active species. Notably, these pH-dependent restructuring phenomena are absent in pristine or heteroatom-substituted pyrite CoSe2, highlighting how catalyst phase and composition influence dynamic behavior [76].

Table 2: pH-Dependent Structural Evolution in CoSe2 Catalysts

Catalyst Phase	pH Conditions	Structural Evolution	Active Species
Marcasite CoSe2 (o-CoSe2)	Acidic HER	Surface corrosion	Disordered Se-Co-Se moieties
Marcasite CoSe2 (o-CoSe2)	Alkaline HER	Potential-driven restructuring	Metallic Se-Co-Co-Se moieties
Pyrite CoSe2 (c-CoSe2)	Acidic and Alkaline HER	Minimal restructuring	Original surface sites
S-substituted Pyrite CoSe2 (c-S-CoSe2)	Acidic and Alkaline HER	Minimal restructuring	Original surface sites

Methodological Approaches: Probing Dynamic Active Sites

Operando Characterization Techniques

Accurately identifying structural evolution during catalytic reactions requires advanced characterization techniques that can probe catalyst structures under operational conditions:

• Operando X-ray Absorption Fine Structure (XAFS): This technique provides element-specific information about the oxidation state and local coordination environment of metal centers during catalysis. For the Ni-Fe molecular complex catalyst, operando XAFS measurements at the Ni and Fe K-edges enabled researchers to track the transformation from Ni monomers to O-bridged Ni-Fe2 trimers during electrochemical activation, revealing the dynamic formation of the true active site [1].

• Identical-Location Electron Microscopy: This approach involves examining the exact same catalyst region before and during reaction conditions, enabling direct visualization of structural changes. Researchers applied this method to track the potential-driven aggregation of Cu single atoms into clusters in Cu–N–C catalysts during NO3RR [75].

• Operando Raman Spectroscopy: Combined with other techniques, Raman spectroscopy provides insights into molecular vibrations and surface species under reaction conditions, helping identify reaction intermediates and structural phases. This method has been particularly valuable in studying the formation of CoOOH-related species during the OER process in chalcogenide catalysts [76].

• Synchrotron-based X-ray Fluorescence (SXRF): This technique offers elemental mapping with high sensitivity, enabling detection of trace metal incorporation in catalytic materials. For the Ni-Fe catalyst, SXRF confirmed the incorporation of Fe species into Ni-SAs@GNM during electrochemical activation, with an atomic Ni/Fe ratio of approximately 5.2:1 [1].

Computational Modeling of Dynamic Processes

Computational approaches provide the theoretical framework to interpret experimental observations and understand the fundamental forces driving structural evolution:

• Ab Initio Molecular Dynamics (AIMD): These simulations model the dynamic behavior of atoms and electrons under realistic reaction conditions, capturing time-dependent structural fluctuations. For Au/oxide systems, AIMD simulations combined with microkinetic modeling demonstrated that Au single atoms detach from clusters to catalyze CO oxidation and subsequently reintegrate after the reaction [75].

• Constant-Potential DFT Calculations: This approach incorporates the effect of applied potential on catalyst structure and reactivity, essential for understanding electrochemical systems. In studies of Cu–N–C catalysts, constant-potential calculations revealed the leaching barriers of Cu atoms at operational potentials and monitored the stability of Cu–N bonds during dynamic evolution [75].

• Microkinetic Modeling: Combining theoretical calculations with kinetic analysis helps bridge the gap between identified active sites and observed catalytic rates. This approach has been instrumental in establishing that dynamically formed sites, rather than the initial structures, govern the catalytic performance in systems such as Ni-Fe OER catalysts [1].

The following diagram illustrates the integrated experimental and computational approach required to probe dynamic active sites:

The Scientist's Toolkit: Essential Reagents and Methodologies

Table 3: Essential Research Reagent Solutions for Studying Dynamic Active Sites

Reagent/Technique	Function in Research	Key Applications
Graphene Oxide (GO) Suspension	Support material for constructing single-atom catalysts	Preparation of Ni-SAs@GNM pre-catalyst [1]
Fe-free KOH Electrolyte with Controlled Fe Impurities (1 ppm)	Electrochemical activation medium	In situ construction of Ni-Fe molecular complex catalysts [1]
ZIF-67 Precursors	Template for hierarchical catalyst structures	Synthesis of CoSe2 catalysts with controlled phases [76]
Heteroatom Dopants (S, P)	Modifiers of electronic structure and phase composition	Tuning dynamic behavior of CoSe2 catalysts [76]
Operando XAFS Cell	Enables X-ray absorption measurements under reaction conditions	Tracking local structure changes in Ni, Fe, Co centers [1] [76]
Identical-Location TEM Grids	Allows direct visualization of same catalyst area pre/post-reaction	Monitoring aggregation of single atoms into clusters [75]
Ab Initio Molecular Dynamics Software	Models time-dependent structural evolution	Simulating dynamic behavior of active sites [75] [1]
Constant-Potential DFT Code	Incorporates electrochemical potential in calculations	Predicting potential-driven structural changes [75]

Future Perspectives and Research Directions

The recognition of dynamic active site evolution presents both challenges and opportunities for catalysis research. Future progress will require developing more sophisticated operando characterization techniques with higher temporal and spatial resolution to capture rapid structural transitions during catalysis. The integration of multi-modal operando approaches—combining spectroscopic, microscopic, and scattering techniques—will provide complementary insights into different aspects of structural evolution. From a theoretical perspective, advancing computational methods to simulate longer timescales and more complex interfaces will be essential for predicting and rationalizing dynamic behavior. For practical catalyst design, strategies that intentionally harness dynamic processes rather than resisting them offer promising pathways to break scaling relationships. This might involve creating catalysts with designed "structural flexibility" or developing activation protocols that systematically transform pre-catalysts into highly active dynamic states. As these approaches mature, the traditional paradigm of optimizing static catalyst structures will increasingly give way to a new philosophy: designing systems that dynamically evolve into highly active states under operational conditions, then revert to stable configurations during idle periods, thus balancing both activity and stability requirements for industrial applications.

Bridging Theory and Experiment: Validating Mechanisms and Reconciling Data Discrepancies

In catalysis research, significant kinetic data variations for the same catalyst and reaction across different laboratories have long posed a challenge for model validation and active site identification. This review explores how catalyst structure sensitivity serves as a fundamental source of this experimental variance, using examples from methane oxidation on Pt nanoparticles and COâ‚‚ electroreduction on Cu surfaces. We demonstrate how structure-descriptor-based microkinetic modeling can rationalize most observed variations through differences in nanoparticle coordination environments. Furthermore, we examine this phenomenon within the broader theoretical framework of scaling relationships between reaction intermediates, which impose intrinsic limitations on catalytic activity. The dynamic restructuring of catalyst surfaces under reaction conditions and strategies for manipulating scaling relations are discussed to provide a comprehensive understanding of structure-performance relationships in heterogeneous catalysis.

Literature kinetics data for catalytic reactions often exhibit significant discrepancies among laboratories, even when studying identical catalysts and reactions. These variations have remained poorly understood, complicating model validation and active site identification [77]. For multi-step catalytic reactions, a fundamental challenge arises from scaling relationships – correlations between adsorption energies of different reaction intermediates that limit optimal adjustment of every intermediate's adsorption energy simultaneously [10] [1]. The structure sensitivity of catalytic reactions, where activity and selectivity depend on the atomic-scale structure of the catalyst surface, provides a powerful framework for understanding and reconciling these experimental variances [77].

The universal linear scaling relationships between adsorption energies of reactive intermediates create intrinsic performance limitations in multi-step catalytic reactions [1]. When combined with structure sensitivity, these relationships explain why different catalyst preparations and treatments – yielding different nanoparticle structures, step densities, or kink sites – produce divergent kinetic data. This review integrates these concepts to demonstrate how structure sensitivity manifests across different catalytic systems, how it can be systematically investigated, and how understanding this phenomenon enables more rational catalyst design while explaining experimental variances.

Core Concepts: Structure Sensitivity and Scaling Relationships

Structure Sensitivity in Heterogeneous Catalysis

Structure sensitivity refers to the phenomenon where a catalyst's activity and selectivity depend on the arrangement of atoms on its surface. This occurs because different crystal facets, step edges, and kink sites possess atoms with varying coordination numbers, which directly influences their ability to adsorb and transform reactants [77] [78].

Table 1: Key Concepts in Catalyst Structure Sensitivity

Concept	Description	Impact on Catalytic Performance
Coordination Number	Number of atoms directly bonded to a surface atom	Determines adsorption strength; optimum values exist for specific reactions [77]
Planar Surfaces	Flat crystal surfaces like Cu(111) and Cu(100)	Often exhibit low activity due to weak intermediate binding [78]
Step Edges	Linear defects between terraces of different heights	Enhance activity by providing undercoordinated sites [78]
Kink Sites	Point defects where step directions change	Provide highly undercoordinated atoms with unique properties [78]
Restructuring	Surface rearrangement under reaction conditions	Generates active sites in situ; driven by adsorbate binding [78]

In COâ‚‚ electroreduction on copper, for instance, perfect planar Cu(111) and Cu(100) surfaces exhibit extremely low CO coverage and consequently low activity for multi-carbon products. Instead, steps and kinks serve as the primary active sites, with the strong binding of CO on defective sites acting as a thermodynamic driving force for surface restructuring under reaction conditions [78].

Scaling Relationships Between Reaction Intermediates

Scaling relationships represent correlations between adsorption energies of different reaction intermediates on catalytic surfaces. First recognized in the 2000s, these relationships emerge because adsorption energies of chemically similar intermediates (e.g., *OH, *OOH, and *O in oxygen electrocatalysis) are often linearly correlated and cannot be adjusted independently on conventional single-site catalysts [10] [1].

For the oxygen evolution reaction (OER), the scaling relationship between *OOH and *OH adsorption energies presents a particular challenge. These correlations inevitably place intrinsic limitations on optimally adjusting adsorption of every intermediate simultaneously to achieve maximum activity and/or selectivity [1]. The manifestation of these scaling relationships varies with catalyst structure, contributing significantly to observed experimental variances across different catalyst preparations.

Interplay Between Structure Sensitivity and Scaling Relationships

The coordination environment of catalyst active sites directly influences adsorption energies of reaction intermediates, thereby affecting how scaling relationships manifest. For complete methane oxidation on Pt nanoparticles, a volcano-like relationship emerges between activity and coordination number, with an optimum coordination number providing the best performance [77]. Smaller particles, despite having more undercoordinated sites, can exhibit very low reactivity due to poisoning effects like carbon deposition [77].

This interplay creates a complex landscape where both the structure of the catalyst and the inherent scaling relationships between intermediates collectively determine the observed kinetics. The following diagram illustrates how these factors contribute to experimental variance in catalytic studies:

Methodological Approaches: Investigating Structure Sensitivity

Structure-Descriptor-Based Microkinetic Modeling

A powerful methodology for rationalizing structure-dependent kinetic variations involves building structure-descriptor-based microkinetic models. This approach investigates relationships between nanoparticle structure and reaction kinetics using literature data mining and computational modeling [77]. For complete methane oxidation on Pt, such models have demonstrated that most data variation can be successfully traced to structure sensitivity, with the methodology serving as both a predictive tool for kinetic performance and a data quality assessment for identifying experimental outliers [77].

Table 2: Experimental Methods for Studying Structure Sensitivity

Method	Principle	Application in Structure Sensitivity
Ultra-high Vacuum Surface Preparation	Creates atomically clean, well-defined surfaces	Reveals intrinsic activity of specific facets without defects [78]
Electrochemical Scanning Tunneling Microscopy	Provides atomic-scale surface imaging under reaction conditions	Direct observation of surface restructuring [78]
Operando X-ray Absorption Fine Structure	Probes local electronic structure and coordination	Identifies dynamic changes in active sites during reaction [1]
Scanning Electrochemical Cell Microscopy	Correlates local activity with surface structure	Directly links step/kink density with reactivity [78]
Grand Canonical Density Functional Theory	Models adsorption at realistic potentials	Predicts CO coverage and activity on different surfaces [78]

Experimental Protocols for Ultraclean Surface Studies

To isolate the intrinsic structure sensitivity of catalysts from restructuring effects, researchers have developed protocols for preparing and characterizing ultraclean surfaces:

• Ultra-high vacuum (UHV) surface preparation: Copper single crystals are prepared under UHV conditions to create well-ordered atomically clean Cu(111) and Cu(100) surfaces without electropolishing treatments that introduce defects [78].

• Controlled defect introduction: Surfaces are systematically modified with steps and kinks through Ar+ bombardment or other treatments, with defect density quantified before reaction studies [78].

• In situ electrochemical activation: Pre-catalysts such as Ni single atoms on graphene nanomesh are activated via cyclic voltammetry in purified electrolytes with controlled impurity additions (e.g., 1 ppm Fe ions) to construct well-defined active sites [1].

• Operando characterization: Techniques like X-ray absorption fine structure (XAFS) are employed during reaction conditions to track structural transformations and identify true active sites [1].

Case Studies: Structure Sensitivity Across Catalytic Systems

Complete Methane Oxidation on Pt Nanoparticles

For complete methane oxidation on platinum, a structure-descriptor-based microkinetic model reveals a volcano-like relationship between reaction rate and coordination number of surface atoms [77]. The model demonstrates that:

• An optimum coordination number exists that maximizes catalytic activity
• Unlike common expectations, smaller particles with very low coordination numbers exhibit poor reactivity due to carbon poisoning
• Most literature data variation can be rationalized through differences in nanoparticle structure
• The methodology enables rapid prediction of kinetic performance and active site determination

This case illustrates how structure sensitivity explains divergent experimental results across different laboratories studying the same nominal reaction and catalyst.

COâ‚‚ Electroreduction on Copper Surfaces

Copper represents the most promising metal catalyst for COâ‚‚ electroreduction to multi-carbon products, yet the structure sensitivity of this reaction and catalyst stability under operating conditions remain controversial [78]. Key findings include:

• COâ‚‚ reduction does not occur significantly on perfect planar Cu(111) and Cu(100) surfaces
• Steps and kinks serve as primary active sites, with square motifs adjacent to defects identified as particularly active for C-C coupling
• Planar surfaces restructure under reaction conditions to form active stepped surfaces
• Strong CO binding on defective sites provides the thermodynamic driving force for restructuring
• The active surface under reaction conditions often differs substantially from the initial catalyst structure

This system demonstrates the critical importance of considering in situ restructuring when interpreting kinetic data and structure-activity relationships.

Oxygen Evolution Reaction via Dynamic Structural Regulation

The oxygen evolution reaction (OER) faces limitations from linear scaling relationships between *OOH, *O, and *OH intermediates. A recent study demonstrated how these relationships can be circumvented through dynamic structural regulation of active sites [1]:

• A Ni-Feâ‚‚ molecular catalyst constructed via in situ electrochemical activation delivers notable intrinsic OER activity
• Dynamic evolution of Ni-adsorbate coordination, driven by intramolecular proton transfer, alters the electronic structure of adjacent Fe active centers
• This dynamic dual-site cooperation simultaneously lowers free energy changes for O-H bond cleavage and O-O bond formation
• The approach effectively disrupts the inherent scaling relationship in OER

This case illustrates strategies for moving beyond limitations imposed by scaling relationships through careful design of dynamic catalytic systems.

Research Reagent Solutions and Essential Materials

Table 3: Essential Research Materials for Studying Structure Sensitivity

Material/Reagent	Function in Research	Application Example
Ultra-pure Metal Single Crystals	Provide well-defined surface structures	Cu(111), Cu(100) for COâ‚‚RR studies [78]
Purified Electrolytes with Controlled Impurities	Enable precise construction of active sites	KOH with 1 ppm Fe ions for Ni-Fe catalyst formation [1]
Graphene Oxide Support	Anchor single atoms and create defined coordination environments	Preparation of Ni-SAs@GNM pre-catalyst [1]
Argon Ion Sputtering Source	Introduce controlled defects on single crystal surfaces	Creating steps and kinks on Cu surfaces [78]
Reference Electrodes	Maintain precise potential control	RHE for electrochemical measurements [1]

Strategies for Manipulating Scaling Relationships

The recognition of scaling relations as fundamental limitations in catalysis has led to the development of systematic strategies for their manipulation [10]:

• Tuning: Adjusting adsorption energy within the limits of existing scaling relations to approach the "volcano top" of optimal activity

• Breaking: Selectively stabilizing specific intermediates through strategic design, such as creating hydrogen bonds with OOH intermediate

• Switching: Avoiding problematic intermediates altogether by shifting to alternative reaction mechanisms

• Pushing: Combining mechanism switching with stabilising interactions

• Bypassing: Employing two distinct states to decouple adsorption energies and eliminate scaling relations entirely

These strategies provide a systematic framework for addressing the limitations imposed by scaling relationships while accounting for structure sensitivity in catalyst design.

Catalyst structure sensitivity represents a fundamental source of experimental variance in kinetic data across different laboratories. When understood within the framework of scaling relationships between reaction intermediates, this phenomenon provides a powerful explanation for divergent results while offering pathways toward more rational catalyst design. Methodologies combining structure-descriptor-based microkinetic modeling with carefully controlled experiments on well-defined surfaces enable researchers to reconcile apparently contradictory data and identify true active sites. Furthermore, recognizing the dynamic nature of catalyst surfaces under reaction conditions – and developing strategies to manipulate scaling relationships – provides a roadmap for advancing catalytic science beyond current limitations. As these approaches mature, the research community moves closer to predictive catalyst design while developing more sophisticated interpretations of kinetic data that account for structural effects.

In heterogeneous catalysis, the existence of linear scaling relationships (LSRs) between the adsorption energies of reaction intermediates often dictates catalyst activity and selectivity, creating fundamental limitations for multi-step catalytic reactions [1]. These relationships introduce systematic patterns in catalytic data, but real-world data frequently exhibits significant scatter around these trends, complicating prediction and optimization efforts. Structure-descriptor-based microkinetic modeling, enhanced by Graph Convolutional Networks (GCNs), provides a powerful framework to rationalize this scatter and develop more accurate predictive models. This technical guide examines how GCNs utilize structural descriptors of catalysts and adsorbates to predict energetic properties, enabling more reliable microkinetic simulations that account for the complexities underlying apparent data scatter in catalytic datasets.

Theoretical Foundation: Scaling Relationships and Their Limitations

Linear Scaling Relationships in Catalysis

Linear scaling relationships establish correlations between the adsorption energies of different intermediates on catalytic surfaces. For instance, in oxygen evolution reaction (OER), the adsorption energies of *OOH and *OH intermediates typically scale linearly across different catalyst materials [1]. These relationships emerge from similar bonding patterns across different catalyst surfaces and create fundamental constraints on catalytic performance by linking the energetics of different reaction steps.

The scaling relationship between *OOH and *OH adsorption energies imposes a theoretical minimum overpotential for OER, creating what is often termed a "thermodynamic overpotential" [1]. This limitation arises because optimizing the adsorption strength of one intermediate inevitably misaligns the energetics of other steps in the reaction mechanism.

Origins of Data Scatter in Catalytic Relationships

While LSRs provide valuable simplifying approximations, significant data scatter occurs due to several factors:

• Local coordination environments: Variations in surface atom coordination and bonding geometries
• Electronic effects: Ligand and strain effects that modify electronic structure
• Dynamic structural changes: Reconstruction of catalytic sites under reaction conditions [1]
• Multifunctional sites: Cooperative effects between different catalytic centers [1]

This scatter represents both noise and valuable physical information about deviations from simple scaling patterns. GCN-based approaches can capture the structural origins of these deviations to improve prediction accuracy.

Graph Convolutional Networks for Catalytic Property Prediction

Molecular Graph Representations

GCNs operate on graph-based representations of catalytic structures, where atoms constitute nodes and chemical bonds form edges. For catalytic applications, several specialized graph representations have been developed:

• Crystal Graph Convolutional Neural Network (CGCNN): Uses crystal graph representations generated from crystal structures to predict material properties [79]
• Modified molecular graph representations: Ignore atomic distances of neighboring atoms in CGCNN to predict adsorption energies [79]
• "Virtual bond" approach: Provides unique molecular graphs for intermediates and transition states by connecting reacting parts in transition states [79]

These representations enable GCNs to learn from both local atomic environments and global structural patterns that influence catalytic properties.

Architecture and Training

GCNs for catalytic applications typically employ multiple graph convolutional layers that update atom features by aggregating information from neighboring atoms. This is followed by global pooling and fully connected layers that generate final predictions. The models are trained on DFT-calculated energies of intermediates and transition states, learning to map structural features to energetic properties while implicitly accounting for deviations from simple scaling relationships.

Table: GCN Performance for Adsorption Energy Prediction

Model Architecture	Representation Type	Typical RMSE (eV)	Applicable Systems
CGCNN	Crystal graphs	~0.2-0.3	Bulk catalysts
Modified molecular graph	Molecular graphs	~0.2-0.3	Adsorbates on surfaces
Virtual bond approach	Reaction graphs	~0.2-0.3	Intermediates & TS

Microkinetic Modeling with GCN-Predicted Energetics

Integration Framework

The integration of GCN predictions with microkinetic modeling follows a structured workflow:

• Structure representation: Convert catalyst and adsorbate structures to graph representations
• Energy prediction: Use trained GCN models to predict energies of all intermediates and transition states
• Mechanism enumeration: Generate possible reaction pathways connecting intermediates
• Rate calculation: Compute reaction rates using microkinetic equations with predicted energies
• Performance evaluation: Calculate turnover frequencies, selectivities, and other kinetic parameters

Error Propagation Analysis

A critical consideration in GCN-enhanced microkinetic modeling is error propagation from energy predictions to kinetic properties [79]. Studies demonstrate that:

• Root Mean Square Error (RMSE) of energy prediction models alone cannot guarantee accuracy of kinetic property evaluations [79]
• Errors in individual energies of key species disproportionately affect mechanistic predictions
• The sequence of activation barriers rather than specific global error values often determines the preferred reaction pathway [79]
• Even energy errors within DFT intrinsic error (~0.2 eV) can significantly impact predicted reaction rates and mechanism identification [79]

This error propagation analysis highlights the importance of accuracy in predicting relative energetics between competing pathways rather than just overall error metrics.

Case Study: Ethanol Steam Reforming

Application Scope

The application of GCN-enhanced microkinetic modeling to ethanol steam reforming (ESR) provides a compelling case study due to the reaction's complex network of approximately 90 C2 species, including reaction intermediates and transition states [79]. This complexity creates significant challenges for traditional descriptor-based approaches and highlights the value of GCN methods.

Methodology and Implementation

In a recent ESR study, researchers employed three graph-based ML models (GNN, graph convolution, and Weave) to predict energies across species and metal catalysts (Rh(111), Pd(111), Pt(111), and Ir(111)) [79]. The experimental protocol involved:

Diagram: GCN-Microkinetic Modeling Workflow for ESR

Key Findings and Limitations

The ESR case study revealed several important insights:

• Preferred reaction pathways varied significantly between models using different GCN architectures, despite similar overall prediction accuracy [79]
• This pathway variability resulted from errors in energetics of competing steps rather than global error metrics [79]
• Accurate prediction of reaction rates required precise relative ordering of activation barriers for competing pathways [79]
• The performance in reaction pathway identification cannot be straightforwardly predicted from the RMSE of energy prediction models [79]

Advanced Strategies: Breaking Scaling Relationships

Dynamic Structural Regulation

Recent research demonstrates that dynamic structural regulation of active sites can effectively break linear scaling relationships [1]. In OER, a Ni-Feâ‚‚ molecular catalyst constructed via in situ electrochemical activation shows how dynamic evolution of metal-adsorbate coordination can alter the electronic structure of active centers during the catalytic cycle [1].

This dynamic dual-site cooperation simultaneously lowers free energy changes for both O–H bond cleavage and O–O bond formation, disrupting the inherent scaling relationship in OER [1]. For GCN models to capture such effects, they must incorporate representations of dynamic coordination changes under reaction conditions.

Single-Atom Catalysts and New Scaling Relationships

Single-atom catalysts (SACs) exhibit distinct scaling relationships that differ from traditional extended surfaces [34]. The localized nature of active sites in SACs creates different electronic environments that modify adsorption energy correlations. GCN approaches can capture these unique relationships through appropriate graph representations of coordination environments in single-atom systems.

Table: Research Reagent Solutions for GCN-Enhanced Microkinetic Modeling

Reagent/Software	Function	Application Context
VASP	DFT calculations for training data	Energy computation for intermediates and transition states [79]
BEEF-vdW functional	Accounts for van der Waals interactions in adsorption	Improved accuracy for weakly-bound species [79]
CGCNN framework	Crystal graph convolutional neural network	Material property prediction across crystal structures [79]
"Virtual bond" approach	TS representation in graph models	Distinguishing transition states from stable intermediates [79]
Microkinetic simulation code	Rate calculation and mechanism analysis	Evaluating catalytic performance from predicted energies [79]

Implementation Protocols

DFT Calculation Methodology

The foundation for accurate GCN models lies in high-quality DFT calculations:

Computational Parameters:

• Use VASP software with BEEF-vdW functional for van der Waals corrections [79]
• Apply projector-augmented wave (PAW) formalism with energy cutoff of 500 eV [79]
• Employ Methfessel-Paxton smearing method with 0.1 eV broadening [79]
• Optimize adsorption geometries using force-based conjugate gradient algorithm [79]
• Locate transition states with constrained minimization approach [79]

Reference Systems:

• Use gaseous CO, CHâ‚„, and Hâ‚‚ as references for formation energy calculations [79]
• Ensure consistent referencing across all intermediates and transition states

GCN Training Protocol

Data Preparation:

• Convert all intermediates and transition states to molecular graph representations
• For transition states, implement "virtual bonds" connecting reacting parts [79]
• Normalize energy values relative to established references

Model Training:

• Implement multiple GCN architectures (GNN, GC, Weave) for comparison [79]
• Use k-fold cross-validation to assess model robustness
• Evaluate performance using RMSE, MAE, and relative energy ordering metrics

Error Analysis Framework

A comprehensive error analysis protocol is essential for reliable microkinetic modeling:

Diagram: Error Propagation in GCN-Microkinetic Modeling

Structure-descriptor-based microkinetic models enhanced by Graph Convolutional Networks represent a significant advancement in rationalizing data scatter in catalytic relationships. By capturing the structural origins of deviations from simple scaling patterns, GCNs enable more accurate prediction of catalytic properties across diverse materials and reaction conditions. The integration of these approaches with microkinetic modeling provides a powerful framework for catalyst design, though careful attention to error propagation remains essential.

Future development should focus on incorporating dynamic structural changes, modeling complex multi-functional sites, and improving representations for single-atom catalysts. As these methodologies mature, they will increasingly enable the design of catalysts that break conventional scaling relationships to achieve unprecedented catalytic performance.

{}

Comparative Analysis: Facet-Dependent Energy Scaling Relations on (110) vs. (111) Surfaces

Scaling relations, the linear correlations between the adsorption energies of different reaction intermediates on catalytic surfaces, are a fundamental concept in catalysis research. These relationships arise because the binding strengths of various intermediates often depend on a common descriptor, typically the binding energy of a key atom like carbon or oxygen. While these relations powerfully simplify the computational screening of catalysts, they also impose fundamental limitations on catalytic performance, creating a "volcano plot" where the highest activity is constrained to a narrow peak. This analysis explores a critical strategy for breaking these limitations: the exploitation of crystallographic facet dependency. Specifically, we investigate how the distinct atomic arrangements and electronic structures of (110) and (111) surfaces lead to divergent scaling relations and catalytic behaviors. By comparing these facets across a range of material systems—from face-centered cubic (fcc) metals to complex metal oxides—this guide provides a technical framework for designing next-generation catalysts that circumvent traditional activity-scaling constraints.

Theoretical Foundations of Scaling Relations and Facet Dependency

Scaling relations are most commonly observed for adsorbates that bind to the surface through a particular type of atom, such as C, O, H, N, or S. The binding energy of a complex intermediate is often correlated with the binding energy of the central binding atom or a simpler molecule. This correlation arises from the similar nature of the chemical bonds being formed. For instance, the adsorption energies of *OH, *OCH3, and *OCH might all scale linearly with the adsorption energy of *O, as they all involve oxygen binding to the surface. These linear relationships allow for the parameterization of complex reaction networks using a single descriptor, dramatically increasing the efficiency of computational catalyst screening [80].

However, the simplicity of scaling relations also introduces a constraint. The adsorption energies of different intermediates are linked, making it difficult to independently optimize the binding strength of each intermediate along a reaction pathway. This often results in a trade-off, where strengthening the binding of one intermediate inevitably over-binds another, positioning most catalysts on the sides of the volcano plot rather than at its peak. The pursuit of catalysts that break scaling relations is therefore a central theme in modern catalysis research.

The atomic-level structure of a catalyst surface is a primary factor governing adsorption energies. Key surface properties that vary with crystallographic orientation include:

• Surface Atom Coordination Number: Lower coordination of surface atoms often leads to stronger adsorbate binding.
• Atomic Arrangement and Symmetry: The geometry and spacing of surface atoms create distinct adsorption sites (e.g., atop, bridge, hollow).
• Electronic Structure: The local density of states and band structure can vary significantly between facets.
• Surface Energy and Stability: Different facets have different stabilities, influencing reconstruction and segregation phenomena.

The (111) and (110) facets of fcc crystals represent two common but structurally distinct surfaces. The (111) surface is typically the most close-packed, with high symmetry and atomic coordination, often leading to smoother potential energy landscapes. In contrast, the (110) surface is more open, featuring rows of atoms with troughs between them, which results in lower coordination numbers and a higher density of under-coordinated sites [81]. These inherent differences predispose the two facets to exhibit different adsorption energetics and, consequently, different scaling relations.

Structural and Electronic Disparities Between (110) and (111) Surfaces

Metallic Systems (FCC)

In face-centered cubic (fcc) metals, the (111) and (110) facets exhibit starkly different physical structures. The (111) surface is the most densely packed, exhibiting a hexagonal symmetry where each surface atom has nine nearest neighbors: six within the same surface layer and three in the layer immediately below. This results in a high coordination number of 9 and a relatively smooth, stable surface [81].

Conversely, the fcc (110) surface has a more open, anisotropic structure characterized by atomic rows separated by troughs. This arrangement exposes atoms in the underlying second layer. A surface atom on the fcc (110) facet has a coordination number of just 7 [81]. This lower coordination number often leads to stronger interaction with adsorbates and a higher intrinsic surface energy compared to the (111) facet. The surface energies for a Ni monolayer on a Cu substrate further illustrate this trend, with the (110) orientation consistently exhibiting higher energy than the (111) facet [82].

Table 1: Comparison of Low-Index FCC Metal Surfaces

Surface Facet	Atomic Arrangement	Coordination Number	Relative Surface Energy	Common Adsorption Sites
(111)	Close-packed, hexagonal	9	Low	On-top, Hollow (fcc/hcp)
(110)	Open, row-and-trough	7	High	On-top, Short-bridge, Long-bridge, Hollow
(100)	Square, less dense than (111)	8	Medium	On-top, Bridge, Hollow (4-fold)

Metal Oxide and Perovskite Systems

The facet-dependent structural differences extend to metal oxides and perovskites, where they significantly influence catalytic mechanism and performance. For instance, in LaNiO₃−δ perovskite, the (111) facet is notably more active for the Oxygen Evolution Reaction (OER) than the (110) and (001) facets, exhibiting an overpotential approximately 30-60 mV lower [83]. This enhanced activity is linked to a more pronounced surface transformation under reaction conditions. The (111) facet facilitates the formation of a highly active oxyhydroxide-like NiOO(H) layer, and the structural match between this transformed layer and the underlying perovskite lattice is superior for the (111) orientation compared to others [83].

Similarly, studies on α-MnO₂ nanorods have demonstrated that facets with a higher abundance of unsaturated metal atoms, such as the {310} facet, exhibit stronger Lewis acidity and a superior ability to adsorb and hydrolyze organophosphate esters (OPEs) like 4-nitrophenyl phosphate (pNPP) compared to the {110} and {200} facets [84]. This highlights how facet-dependent Lewis acidity, driven by atomic coordination, can dictate catalytic performance in hydrolysis reactions.

Catalytic Performance Across Material Systems

The structural and electronic disparities between (110) and (111) surfaces manifest in distinct catalytic activities across diverse reactions.

Oxygen Evolution Reaction (OER)

The OER is a cornerstone reaction for renewable energy technologies. The facet-dependent activity of electrocatalysts is a critical design parameter.

• LaNiO₃−δ: As noted, the (111) facet is the most active, with a lower overpotential and Tafel slope than the (110) facet [83].
• Taâ‚‚Oâ‚…: Computational studies reveal a clear facet-dependent OER activity trend: (200) > (120) > (001). This activity was correlated with the Ta density on the topmost atomic layer and the moderate density of states at the Fermi level. The overpotential for the pristine (200) facet was calculated to be 0.61 V [85].

Adsorption and Hydrodechlorination

The adsorption of chlorinated organic contaminants (COCs) on Pd nanoparticles is highly facet-dependent. Molecular dynamics simulations show that the under-coordinated Pd {110} surface exhibits anomalous adsorption behavior for contaminants like trichloroethylene (TCE) compared to the {111} and {100} surfaces [86]. This strong, facet-specific adsorption is a crucial first step for efficient hydrodechlorination, a key water remediation pathway. The efficiency of dechlorination is directly influenced by the molecular orientation and packing geometry of the adsorbates, which are dictated by the atomic structure of the facet [86].

Hydrolysis Reactions

The hydrolysis of organophosphate esters (OPEs) catalyzed by α-MnO₂ is highly sensitive to the exposed crystal facet. The catalytic efficiency follows the order {310} > {110} > {200} [84]. This trend is directly linked to the density of unsaturated Mn atoms on each surface, which governs the surface Lewis acidity. Facets with higher Lewis acidity, such as {310}, promote the adsorption and activation of the phosphate ester bond, leading to more efficient catalytic hydrolysis [84].

Table 2: Experimental Catalytic Performance Metrics by Facet

Material	Reaction	Most Active Facet	Key Performance Metric	Rationale for Enhanced Activity
LaNiO₃−δ	OER	(111)	Overpotential ~30-60 mV lower than (110)	Better structural match of transformed active layer [83]
Taâ‚‚Oâ‚…	OER	(200)	Overpotential = 0.61 V (pristine)	Optimal Ta site density and DOS at Fermi level [85]
α-MnO₂	OPE Hydrolysis	{310}	Highest catalytic hydrolysis rate	Highest density of unsaturated Mn atoms & Lewis acidity [84]
Pd	COC Adsorption	{110}	Anomalous adsorption strength	Undercoordinated sites in open surface structure [86]

Breaking Scaling Relations with Facet Engineering and Inverse Catalysts

The constraints imposed by linear scaling relations can be overcome by designing active sites whose geometry or electronic structure deviates from the typical surfaces used to establish these relations. Inverse catalysts—metal oxide nanoparticles supported on a metal substrate—are a promising class of materials for this purpose. The complex, asymmetric structures at the metal-oxide interface create a wide variety of unique active sites [87].

Machine learning explorations of InᵧOₓ/Cu(111) inverse catalysts for CO₂ hydrogenation to methanol have revealed that the active sites at the interface can indeed break linear scaling relations. This breaking effect is a key reason for the superior catalytic performance of inverse catalysts compared to their conventional counterparts. The workflow involved training a machine learning interatomic potential to efficiently probe a vast array of active sites at the interface, identifying transition state geometries that do not conform to the scaling relations observed on flat metal surfaces [87].

Facet engineering provides another direct route. Since (110) and (111) surfaces possess different coordination environments and electronic structures, the scaling relations between adsorption energies of intermediates (e.g., *O vs. *OH) will have different slopes and intercepts on each surface. A catalyst nanoparticle exposing a combination of both facets, or a surface engineered to have step edges mimicking the (110) structure on a (111) terrace, can create unique adsorption sites that fall off the standard scaling line, thereby optimizing the binding energies of all intermediates simultaneously.

Experimental and Computational Methodologies

Synthesis of Facet-Controlled Nanomaterials

Epitaxial Thin Film Growth (for Perovskites):

• Method: Pulsed Laser Deposition (PLD) is a common technique for growing high-quality, epitaxial thin films with controlled facet orientation.
• Protocol: A polycrystalline LaNiO₃−δ target is ablated with a high-energy pulsed laser in a vacuum chamber containing an oxygen background pressure. The ablated material is deposited onto a single-crystal substrate (e.g., SrTiO₃) with the desired orientation [(001), (110), or (111)]. The substrate temperature is carefully controlled to ensure epitaxial, two-dimensional growth.
• Validation: The facet orientation and epitaxial quality are confirmed using in-situ Reflection High-Energy Electron Diffraction (RHEED), X-ray Diffraction (XRD) to observe characteristic film peaks and Laue fringes, and Atomic Force Microscopy (AFM) to verify smooth, step-terraced surfaces [83].

Hydrothermal/Solvothermal Synthesis (for Metal Oxides):

• Method: This is widely used for synthesizing metal oxide nanomaterials with controlled exposed facets, such as α-MnOâ‚‚ nanorods.
• Protocol: For α-MnOâ‚‚-110, a typical synthesis involves hydrothermal treatment of a mixture of KMnOâ‚„ and (NHâ‚„)â‚‚SOâ‚„ in deionized water at a specific temperature (e.g., 140°C) for several hours. The resulting precipitate is collected, washed, and dried. Variations in the precursor salts and ratios (e.g., using KMnOâ‚„ and MnSOâ‚„, or KMnOâ‚„ and HCl) can selectively promote the growth of different dominant facets, such as {310} or {200} [84].
• Validation: The exposed facets are characterized using XRD, Scanning Electron Microscopy (SEM), Transmission Electron Microscopy (TEM), and Raman spectroscopy.

Computational Analysis of Scaling Relations

Density Functional Theory (DFT) Calculations:

• Method: DFT is the standard computational tool for calculating adsorption energies and establishing scaling relations.
• Protocol:
  • Surface Slab Model: A slab model of the catalytic surface is created, with a sufficient vacuum layer to avoid periodic interactions.
  • Geometry Optimization: The structure of the clean slab and the slab with adsorbed intermediates (e.g., O, *OH, *OOH for OER) is optimized until the forces on atoms are below a threshold (e.g., 0.05 eV/Ã…).
  • Energy Calculation: The adsorption energy (Eads) of an intermediate A is calculated as: Eads = E(slab+A) - Eslab - EA, where EA is the energy of the free A molecule in the gas phase.
  • Scaling Plot: The adsorption energies of different intermediates are plotted against each other (e.g., EOOH vs. E*OH) or against a common descriptor. A linear regression reveals the scaling relation [80] [87].
• Advanced Modeling: For systems with strong electron correlations (e.g., perovskites), the DFT+U method is employed. Van der Waals corrections (e.g., DFT-D3) are often added for a more accurate description of dispersion forces [85] [83].

Diagram 1: DFT Scaling Relation Workflow (Title: Computational Screening Workflow)

In-Situ/Operando Characterization

Understanding surface transformations under operating conditions is crucial.

• In-Situ Electrochemical ATR-FTIR: Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy can be used to monitor the formation of reaction intermediates on the catalyst surface during electrolysis, providing mechanistic insights [84].
• In-Situ X-ray Spectroscopy: Techniques like X-ray Photoelectron Spectroscopy (XPS) can track changes in the oxidation state of surface atoms (e.g., Ni in LaNiO₃) under reaction conditions, confirming facet-dependent transformations [83].

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Key Reagents and Computational Tools for Facet-Dependent Catalysis Research

Item Name	Function/Brief Explanation	Example Application/Context
Single-Crystal Substrates (e.g., SrTiO₃)	Provides an atomically flat, crystallographically defined base for epitaxial growth of oriented thin films.	Synthesizing (001)-, (110)-, and (111)-oriented LaNiO₃ thin films for OER studies [83].
Metal-Organic Precursors (e.g., KMnO₄)	Used in hydrothermal synthesis as the source of metal cations for the growth of metal oxide nanocrystals.	Controlling the exposed facets of α-MnO₂ nanorods by varying precursor salts and ratios [84].
DFT+U Code (e.g., Quantum ESPRESSO, GPAW)	Software for electronic structure calculations that includes a Hubbard U parameter to better describe strongly correlated electrons.	Accurately modeling the electronic properties of Ta₂O₅ and LaNiO₃ surfaces [85] [83] [87].
Machine Learning Interatomic Potential (MLIP)	A machine-learned force field trained on DFT data, enabling rapid exploration of complex catalyst structures.	High-throughput screening of transition states at complex inverse catalyst interfaces [87].
Palladium Nanoparticles	Catalytically active material known for hydrodechlorination reactions; facet control enhances adsorption efficiency.	Studying the adsorption dynamics of chlorinated organic contaminants on {111}, {100}, and {110} facets [86].

The comparative analysis of (110) and (111) surfaces unequivocally demonstrates that crystallographic facet engineering is a powerful strategy for manipulating adsorption energetics and breaking the limitations imposed by linear scaling relations in heterogeneous catalysis. The intrinsic structural and electronic differences between these facets—such as coordination number, atomic packing, and Lewis acidity—dictate their respective catalytic properties across a wide range of materials and reactions, from OER on perovskites to hydrolysis on metal oxides. The emergence of sophisticated synthetic methods for facet-controlled nanomaterials, combined with advanced computational screening techniques and operando characterization, provides a comprehensive toolkit for the rational design of next-generation catalysts. By deliberately targeting specific facets or creating interfaces that generate unique active sites, researchers can continue to discover catalytic systems that transcend conventional activity limits, driving progress in energy conversion and environmental remediation.

The design of artificial enzymes represents one of the most ambitious goals in modern catalysis research. Theozymes—computational models of catalytic sites that incorporate key residues responsible for transition state stabilization—have emerged as powerful tools for this endeavor. These simplified active site models enable researchers to study enzyme mechanisms using quantum chemical methods while maintaining the essential features of enzymatic catalysis [88]. The ultimate objective is to bridge the gap between computational design and experimental realization, creating artificial enzymes with catalytic efficiencies rivaling their natural counterparts.

This pursuit must be contextualized within the fundamental framework of scaling relations in catalysis. In electrocatalysis, scaling relations refer to the linear correlations between adsorption energies of different reaction intermediates, which create inherent limitations on catalytic efficiency [10]. For instance, in oxygen reduction reactions (ORR), the scaling relation between OOH and OH intermediates creates a thermodynamic constraint that limits the overall catalytic activity. Similar fundamental relationships govern all catalytic processes, where the binding energies of different intermediates are often correlated, preventing independent optimization of each reaction step. Understanding and manipulating these scaling relations is essential for advancing catalyst design, whether for artificial metalloenzymes or computational enzyme designs [89] [10].

The validation of theozyme designs therefore requires not only demonstrating catalytic activity but also characterizing how these designs interact with and potentially overcome the limitations imposed by scaling relations. This technical guide examines the integrated computational and experimental framework required to validate theozyme designs, with particular emphasis on recent breakthroughs that have achieved unprecedented catalytic efficiencies through sophisticated computational methodologies.

Computational Design Methodologies

Theozyme Construction and Active Site Modeling

Theozyme construction begins with identifying the quantum mechanical requirements for catalysis. The fundamental approach involves:

• Catalytic Constellation Definition: Specifying the essential residues and functional groups necessary for transition state stabilization based on quantum mechanical calculations of the reaction mechanism [88]. For Kemp eliminase designs, this typically includes a nucleophilic base (Asp or Glu) for proton abstraction and an aromatic sidechain for Ï€-stacking interactions with the substrate transition state [90].

• Fragment-Based Assembly: Generating backbone diversity using fragments from homologous natural proteins to create structural variation within the active site pocket [90]. This approach increases the probability of obtaining foldable backbones that position the theozyme in a catalytically competent configuration.

• Geometric Matching: Implementing algorithms to position the theozyme within each generated backbone structure, followed by optimization of the entire active site using atomistic calculations [90]. This step effectively mutates all active-site positions to create an optimal environment for catalysis.

Recent advances have demonstrated that excluding certain constraints from the theozyme definition can enhance design success. For example, omitting the requirement for a polar group to stabilize the developing negative charge in the Kemp elimination transition state—allowing water molecules to fulfill this role—resulted in more effective designs [90].

Overcoming Limitations Through Advanced Workflows

Traditional computational enzyme design has been limited by low catalytic rates and extensive requirement for experimental optimization. A fully computational workflow recently demonstrated overcoming these limitations through several key innovations:

Table 1: Components of Advanced Computational Workflows for Theozyme Design

Workflow Component	Function	Impact on Design Quality
Combinatorial backbone assembly	Generates structural diversity from natural protein fragments	Increases probability of catalytically competent active sites
Fuzzy-logic optimization	Balances conflicting objectives (e.g., low system energy vs. high desolvation)	Enables simultaneous optimization of stability and activity
Full-protein stabilization	Applies stability design across entire protein structure	Creates robust scaffolds that tolerate active site mutations
Unrestricted active site optimization	Uses atomistic energy as sole optimization criterion	Discovers novel catalytic configurations beyond natural homology

This integrated workflow resulted in Kemp eliminase designs with catalytic efficiencies exceeding 12,700 M⁻¹ s⁻¹ and catalytic rates (kcat) of 2.8 s⁻¹, surpassing previous computational designs by two orders of magnitude [90]. Further optimization of a residue considered essential in all previous designs increased efficiency to >10⁵ M⁻¹ s⁻¹, achieving parameters comparable to natural enzymes [90].

Workflow Visualization

The complete computational pipeline for designing high-efficiency enzymes involves multiple integrated steps, from initial backbone generation to final characterization:

Experimental Characterization and Validation

Structural Validation Techniques

Validating computational designs requires rigorous structural characterization to confirm that the engineered proteins adopt the intended conformation:

• X-ray Crystallography: Provides high-resolution structural data to verify overall fold and active site geometry compared to design models [91].

• Nuclear Magnetic Resonance (NMR) Spectroscopy: Offers solution-state structural information and insights into protein dynamics and flexibility [91].

• Electron Microscopy: Useful for characterizing larger synthetic enzyme complexes or those incorporating nanomaterial components [91].

These techniques are essential for identifying structural distortions that often undermine catalytic efficiency in designed enzymes. Even small deviations of a few degrees or tenths of an Ã…ngstrom from the design conception can reduce catalytic efficiency by orders of magnitude [90].

Functional Assays and Kinetic Characterization

Comprehensive kinetic analysis is crucial for evaluating catalytic performance against natural enzymes and previous designs:

Table 2: Key Kinetic Parameters for Artificial Enzyme Validation

Parameter	Measurement Purpose	Benchmark Values	Recent Advancements
kcat (turnover number)	Chemical transformation rate	Natural enzymes: ~10 s⁻¹ [90]	Recent designs: 2.8-30 s⁻¹ [90]
kcat/KM (catalytic efficiency)	Overall catalytic proficiency	Natural enzymes: ~10⁵ M⁻¹ s⁻¹ [90]	Optimized designs: >10⁵ M⁻¹ s⁻¹ [90]
KM (Michaelis constant)	Substrate binding affinity	Lower values indicate tighter binding	Can be optimized independently of kcat
Thermal stability (Tm)	Structural robustness	Natural enzymes: variable	Designs achieving >85°C [90]

Functional assays should be conducted under varied conditions (pH, temperature, solvent composition) to evaluate the robustness of synthetic enzymes compared to their natural counterparts [91]. For metalloenzyme designs, kinetic characterization should include metal dependence studies to verify proper incorporation and function of metal ions [89].

Validation Workflow

The experimental validation of computationally designed enzymes follows a systematic process from expression to detailed kinetic analysis:

The Scientist's Toolkit: Essential Research Reagents and Materials

Successful validation of theozyme designs requires specific reagents and methodologies:

Table 3: Essential Research Reagents for Theozyme Validation

Reagent/Material	Function	Application Example
Plasmid expression systems	Protein production	High-yield expression of designed enzymes [90]
Chromatography resins	Protein purification	Isolation of pure functional enzymes (e.g., IMAC, SEC) [91]
Spectroscopic substrates	Activity assays	Kemp elimination with 5-nitrobenzisoxazole [90]
Stable isotope-labeled compounds	NMR characterization	Protein dynamics and binding studies [91]
Crystallization screens	Structural analysis	Optimization of crystal formation for X-ray studies [91]
Metal cofactors	Metalloenzyme function	Incorporation of Zn²⁺, Cu²⁺, Fe²⁺ in synthetic active sites [89]
Quantum chemistry software	Mechanism analysis	DFT calculations on theozyme models [88]

Case Study: Validating Kemp Eliminase Designs

A recent landmark study demonstrates the complete workflow for designing and validating highly efficient artificial enzymes [90]. The validation process for these Kemp eliminase designs included:

• Expression and Solubility Screening: 66 of 73 designs were solubly expressed in E. coli, with 14 showing cooperative thermal denaturation, indicating proper folding.

• Initial Activity Screening: Three designs showed measurable Kemp elimination activity, with the top two designs (Des27 and Des61) exhibiting kcat/KM values of 130 and 210 M⁻¹ s⁻¹, respectively.

• Computational Optimization: Application of the FuncLib method to active-site positions resulted in variants with dramatically improved activity, achieving catalytic efficiencies of 3,600 M⁻¹ s⁻¹ and ultimately >10⁵ M⁻¹ s⁻¹ after further optimization.

• Stability Characterization: The most efficient design demonstrated exceptional thermal stability (>85°C), confirming that the computational stabilization methods had successfully created robust protein scaffolds.

This case study illustrates how iterative computational design and careful experimental validation can overcome traditional limitations in artificial enzyme development, resulting in catalysts that rival natural enzymes in both efficiency and stability.

The validation of theozyme designs must be interpreted within the conceptual framework of scaling relations. Successful artificial enzymes effectively manage the energy landscapes of their catalytic cycles to minimize the limitations imposed by inherent correlations between intermediate states [10]. The most advanced designs appear to achieve this through several strategies:

• Transition State Stabilization: Precisely positioning functional groups to stabilize key transition states without over-stabilizing ground states or intermediates.

• Dynamic Control: Utilizing protein dynamics to create transient interactions that circumvent static scaling relations [88].

• Multi-State Design: Engineering active sites that can access multiple conformational states to decouple adsorption energies of different intermediates [10].

As the field progresses, validation methodologies must increasingly incorporate techniques that probe these subtler aspects of catalytic function, including advanced spectroscopic methods, computational simulations of dynamics, and single-molecule approaches. By integrating rigorous experimental characterization with theoretical understanding of scaling relations, the design and validation of theozymes will continue to advance toward the creation of artificial enzymes with customized functions and unprecedented catalytic capabilities.

The quest for efficient and sustainable energy conversion processes hinges on the development of high-performance catalysts. For reactions central to these technologies, such as the oxygen evolution reaction (OER) and oxygen reduction reaction (ORR), first-row transition metal oxides represent a promising class of earth-abundant electrocatalysts. However, their performance is fundamentally governed by linear scaling relationships (LSRs) that create inherent thermodynamic limitations. This whitepaper synthesizes recent advances in benchmarking the activity of these materials, specifically examining Mn, Fe, Co, and Ni-based oxides. We present a structured analysis of quantitative performance data, detailed experimental methodologies for activity assessment, and the mechanistic role of electronic structure in overcoming scaling relationships to guide the rational design of next-generation catalysts.

In multi-step catalytic reactions, the adsorption energies of different reactive intermediates are often correlated by linear scaling relationships (LSRs) [1]. While LSRs help elucidate broad activity trends, they place an intrinsic ceiling on catalytic performance by making it impossible to independently optimize the binding strength of every intermediate involved in the reaction cycle [1]. This is particularly consequential for the oxygen evolution reaction (OER) and oxygen reduction reaction (ORR), which are pivotal for water splitting, fuel cells, and metal-air batteries.

The widespread LSR between *OOH and *OH intermediates on numerous electrocatalysts constrains the achievable overpotential [1]. Consequently, a primary objective in modern catalysis research is to identify material systems and design strategies that can circumvent these relationships. First-row transition metal oxides (Mn, Fe, Co, Ni) serve as ideal model systems to explore this challenge due to their tunable electronic structures, cost-effectiveness, and promising activity, especially in alkaline environments [92]. This review establishes a structured framework for benchmarking their performance, thereby illuminating the path toward breaking scaling relationships.

Experimental Benchmarks and Quantitative Activity Trends

Performance of First-Row Transition Metal Antimonates

Systematic studies on metal antimonates MSb₂O₆ (M = Mn, Fe, Co, Ni) reveal distinct trends in ORR and OER activity, closely tied to the identity of the transition metal cation. High-phase purity powders of these antimonates are typically synthesized via solid-state reactions at 1000 °C [93].

Table 1: Electrocatalytic Performance of First-Row Transition Metal Antimonates (MSb₂O₆)

Material	ORR Activity (4e⁻ pathway)	Primary ORR Intermediate	OER Activity	Key Electronic Feature
MnSb₂O₆	High	*O₂ (bond breaking)	Moderate	Populated Mn 3d states
FeSb₂O₆	Low	N/A	Low	-
CoSb₂O₆	Low	N/A	High	Vacant (holes) Co 3d states
NiSb₂O₆	Low	N/A	Moderate	-

Electrochemical studies in alkaline electrolytes demonstrate that MnSb₂O₆ is the sole antimonate in this series capable of catalyzing the ORR via a direct four-electron transfer pathway [93]. This activity is linked to its populated Mn 3d states, which possess the unique configuration required to break the O–O bond. In contrast, CoSb₂O₆ exhibits the highest activity towards the OER, facilitated by its vacant Co 3d states that promote the formation of surface-bound OH radicals in the reaction's early stages [93]. Normalizing kinetic currents by the changes in d-state population allows for a more intrinsic benchmarking of catalytic sites, revealing parallels with other oxide families like lanthanide perovskites.

Dynamic Bimetallic Systems: Breaking the Scaling Relations

The limitations of static, single-site catalysts can be overcome by designing dynamic, multi-functional active sites. A notable example is a Ni-Feâ‚‚ molecular complex catalyst, constructed via in situ electrochemical activation, which delivers notable OER performance [1].

Table 2: Performance of a Dynamic Ni-Fe Molecular Catalyst vs. Conventional Scaling Limits

Feature	Conventional Single-Site Catalyst	Dynamic Ni-Feâ‚‚ Molecular Catalyst
Scaling Relationship	Obeys LSR between *OOH and *OH	Breaks inherent LSR
Active Site Structure	Static	Dynamic, coordination-evolving
Mechanism	Adsorbate Evolution Mechanism (AEM)	Dynamic dual-site cooperation
Electronic Effect	Fixed during cycle	Ni-adsorbate coordination modulates adjacent Fe site
Energetic Outcome	Theoretical overpotential limit	Simultaneously lowers barriers for O–H cleavage and *OOH formation

This system undergoes a dynamic structural evolution during the catalytic cycle, where the coordination environment of the Ni site changes, triggering intramolecular proton transfer. This, in turn, modulates the electronic structure of the adjacent Fe active center [1]. Theory and electrokinetics confirm that this cooperation simultaneously lowers the free energy change for O–H bond cleavage and O–O bond formation, thereby disrupting the inherent scaling relationship and enabling superior activity [1].

Essential Research Reagent Solutions and Methodologies

The accurate benchmarking of catalyst performance requires a suite of specialized materials and analytical techniques. The following toolkit is essential for researchers in this field.

Table 3: The Scientist's Toolkit for Catalyst Benchmarking

Research Reagent / Material	Function in Catalysis Research
High-Purity Transition Metal Precursors (e.g., Mn, Fe, Co, Ni salts)	Synthesis of catalyst powders with controlled stoichiometry.
Solid-State Synthesis Furnaces	High-temperature (e.g., 1000°C) synthesis of phase-pure oxide materials [93].
Rotating Ring-Disk Electrode (RRDE)	Electrochemical assessment of ORR/OER activity and determination of electron transfer pathway (e.g., 4e⁻ vs. 2e⁻) [93].
Alkaline Electrolytes (e.g., KOH, NaOH)	Standard medium for testing OER/ORR activity of transition metal oxides [93] [1].
Fe-doped KOH Electrolyte	In situ formation of bimetallic active sites (e.g., Ni-Fe complexes) during electrochemical activation [1].
X-ray Photoelectron Spectroscopy (XPS)	Determining surface elemental composition and oxidation states of metal cations [93].
Operando X-ray Absorption Fine Structure (XAFS)	Probing the local atomic structure and dynamic evolution of active sites under actual reaction conditions [1].
Aberration-Corrected HAADF-STEM	Imaging individual atomic sites and confirming single-atom dispersion in advanced catalyst designs [1].

Detailed Experimental Protocols for Catalyst Synthesis and Evaluation

Synthesis of Metal Antimonate Powders (MSb₂O₆)

• Solid-State Synthesis: Prepare stoichiometric mixtures of high-purity Sbâ‚‚O₃ and the respective transition metal oxide (MnO, Feâ‚‚O₃, Co₃Oâ‚„, or NiO).
• Calcination: Subject the mixed powders to a high-temperature treatment in a muffle furnace at 1000 °C under an air atmosphere for several hours to achieve high-phase purity [93].
• Verification: Characterize the resulting powders using X-ray diffraction (XRD) to confirm phase purity and the absence of unreacted precursors or secondary phases.

Electrode Preparation and Electrochemical Characterization

• Ink Formulation: Prepare a catalyst ink by dispersing the synthesized catalyst powder in a mixture of solvent (e.g., water/isopropanol) and a binder (e.g., Nafion ionomer) via sonication.
• Electrode Fabrication: Deposit a precise volume of the homogeneous ink onto a polished glassy carbon rotating ring-disk electrode (RRDE) and allow it to dry [93].
• ORR/OER Testing: Perform electrochemical measurements in a standard three-electrode cell filled with an oxygen-saturated alkaline electrolyte (e.g., 0.1 M or 1 M KOH). The counter electrode is typically a Pt wire or graphite rod, and the reference is a Hg/HgO or Ag/AgCl electrode.
• Data Collection:
  • Record cyclic voltammograms (CVs) in both nitrogen and oxygen-saturated electrolytes.
  • Perform linear sweep voltammetry (LSV) with the electrode rotating at various speeds (e.g., 400 to 1600 rpm).
  • For ORR, the ring electrode can be used to detect peroxide species, allowing for the calculation of electron transfer numbers [93].
• Activation for Dynamic Catalysts: For systems like the Ni-Fe pre-catalyst, an in situ activation is required. This involves applying a potential cycle (e.g., between 1.1 and 1.65 V vs. RHE) in a KOH electrolyte containing trace amounts (e.g., 1 ppm) of Fe ions to form the active Ni-Fe molecular complex [1].

Advanced Characterization of Active Sites

• Operando XAFS: Acquire X-ray absorption spectra at the metal K-edges (e.g., Ni and Fe) while the catalyst is under operating conditions in an electrochemical cell. This technique can directly observe the formation of O-bridged trimer structures (e.g., Ni-Feâ‚‚) from single-atom precursors [1].
• Microscopy: Use aberration-corrected High-Angle Annular Dark-Field Scanning Transmission Electron Microscopy (HAADF-STEM) to confirm the atomic dispersion of metal atoms on the support material [1].

Mechanistic Insights: Electronic Structure and Beyond

The activity trends across first-row transition metal oxides are not arbitrary but are deeply rooted in their electronic configurations.

• ORR Activity and d-State Population: The unique capability of MnSbâ‚‚O₆ to facilitate the 4e⁻ ORR pathway is attributed to its populated Mn 3d states, which provide the optimal electronic structure for Oâ‚‚ bond scission [93].
• OER Activity and d-State Holes: Conversely, high OER activity in CoSbâ‚‚O₆ is linked to vacant (holes) Co 3d states. These holes promote the interaction with electron-donating OH⁻ ligands, stabilizing key intermediates like surface-bound OH radicals [93].
• Dynamic Site Modulation: The most advanced strategy to break scaling relationships involves moving beyond static surfaces. In the Ni-Feâ‚‚ system, the dynamic evolution of the Ni site's coordination sphere during the catalytic cycle electronically tunes the adjacent Fe active center. This dual-site cooperation allows for the independent optimization of energetics for different steps, a feat impossible on a single, static site [1].

The following diagram illustrates the fundamental challenge of scaling relationships and the dynamic strategy used to overcome it.

Diagram 1: Scaling relationships and dynamic breaking mechanism. Conventional single-site catalysts are constrained by LSRs, imposing a theoretical overpotential limit. A dynamic dual-site system, where one site modulates the other, can break this relationship.

Benchmarking the performance of first-row transition metal oxides reveals a clear hierarchy in activity for oxygen electrocatalysis, intrinsically connected to the electronic structure of the metal cation. While Mn-based oxides show particular promise for ORR, Co-based materials are leading candidates for OER. The standardization of synthesis protocols and electrochemical evaluation methods is critical for making direct and meaningful comparisons between material systems.

However, the future of high-performance catalyst design lies in strategically overcoming linear scaling relationships. The emergence of dynamic catalysts, where the active site structure evolves during the reaction to enable multi-site cooperation, represents a paradigm shift. These systems, exemplified by the Ni-Feâ‚‚ molecular complex, demonstrate that the limitations once thought to be fundamental can be broken. Future research should focus on elucidating such dynamic mechanisms with advanced operando techniques and leveraging this understanding to engineer a new generation of oxide catalysts with activities surpassing current thermodynamic limits.

Conclusion

Scaling relationships are no longer seen as insurmountable physical laws but as design challenges that can be addressed through innovative catalytic strategies. The synthesis of knowledge from foundational principles, advanced characterization methodologies, and disruptive optimization techniques—such as dynamic active sites and multi-center cooperation—provides a powerful toolkit for surpassing traditional catalytic limits. The successful application of these principles, exemplified by the design of highly active FeMn–RuOx and dynamically regulated Ni-Fe catalysts, demonstrates a clear path forward. For biomedical and clinical research, these advances promise more efficient catalytic processes for pharmaceutical synthesis, including the selective activation of substrates and design of enzyme mimetics. Future research must focus on elucidating dynamic reaction environments with atomic precision and developing multi-functional catalysts that intelligently manage reaction energetics, ultimately enabling transformative applications in drug development and sustainable chemical synthesis.

References

• 1. Breaking linear scaling relationships in oxygen evolution ... [https://www.nature.com/articles/s41467-024-55150-9]
• 2. Breaking linear scaling relationships with transition metal ... [https://pubs.rsc.org/en/content/articlehtml/2023/cy/d3cy00863k]
• 3. Breaking the linear scaling relationship in BN-supported ... [https://www.sciencedirect.com/science/article/abs/pii/S1005030224010661]
• 4. Breaking scaling relations in AgAuCuPdPt high-entropy ... [https://pubs.rsc.org/en/content/articlelanding/2025/mh/d5mh01064k]
• 5. Linear Scaling Relationships to Predict pKa's and ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8753599/]
• 6. Scaling relationships and theory for vibrational frequencies ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5705602/]
• 7. Deterioration of Automotive Catalytic Converters (Part 2) [https://trid.trb.org/View/1791637]
• 8. Reactive pathways toward parasitic release of singlet ... [https://www.nature.com/articles/s41524-021-00597-3]
• 9. Current Challenges in Catalysis [https://www.azonano.com/article.aspx?ArticleID=6408]
• 10. scaling relations in oxygen electrocatalysis and beyond [https://pubs.rsc.org/en/content/articlehtml/2025/cs/d5cs00597c]
• 11. Advances in Oxygen Evolution Reaction Electrocatalysts ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11881674/]
• 12. Molecular Mechanisms of Oxygen Evolution Reactions for ... [https://www.mdpi.com/2673-9801/3/4/27]
• 13. Synergistic Spin-Mediated Catalysis for the Oxygen ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12636008/]
• 14. Oxygen evolution electrocatalysis resilient to voltage ... [https://www.nature.com/articles/s41893-025-01665-y]
• 15. Identifying key intermediates for the oxygen ... [https://www.nature.com/articles/s41467-024-54796-9]
• 16. scaling relations in oxygen electrocatalysis and beyond [https://pubs.rsc.org/doi/d5cs00597c]
• 17. Scaling Relations in Oxygen Electrocatalysis and Beyond [https://chemrxiv.org/engage/chemrxiv/article-details/67ed469081d2151a02b33a98]
• 18. H因子190，催化大佬Jens K. Nørskov教授2020成果汇总 [https://www.huasuankeji.com/news/?p=137304]
• 19. Quantitative analysis of the electrochemical performance ... [https://www.sciencedirect.com/science/article/pii/S0021951722002846]
• 20. 张闯课题组在自旋增强CO2催化转化方面取得新进展 [http://www.iccas.ac.cn/news/kyjz/202510/t20251022_7995183.html]
• 21. Data-driven discovery of electrocatalysts for CO2 reduction ... [https://www.nature.com/articles/s41467-023-43118-0]
• 22. 什么是界面双电层？电催化表征体系的系统梳理与关键参数 ... [https://www.huasuankeji.com/news/?p=359710]
• 23. Integrative catalytic pairs for efficient multi-intermediate ... [https://pubmed.ncbi.nlm.nih.gov/39103451/]
• 24. Advances of Synergistic Electrocatalysis Between Single ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11233490/]
• 25. Coordination engineering in single-site catalysts [https://www.sciencedirect.com/science/article/pii/S2451929423004138]
• 26. The simplest construction of single-site catalysts by ... [https://www.nature.com/articles/s41467-019-09596-x]
• 27. Quantum catalysts - EES Catalysis (RSC Publishing) DOI ... [https://pubs.rsc.org/en/content/articlehtml/2025/ey/d4ey00288a]
• 28. EMFF-2025: a general neural network potential for ... [https://www.nature.com/articles/s41524-025-01809-w]
• 29. Density functional theory for catalyst development and ... [https://www.nature.com/articles/s44359-025-00087-5]
• 30. Density functional theory methods applied to ... [https://pubs.rsc.org/en/content/articlehtml/2024/cp/d4cp00266k]
• 31. Scaling Relationships and Computational Catalyst Design [https://pubmed.ncbi.nlm.nih.gov/27088666/]
• 32. Best‐Practice DFT Protocols for Basic Molecular ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9826355/]
• 33. To address surface reaction network complexity using ... [https://www.nature.com/articles/ncomms14621]
• 34. New scaling relationships for the oxygen evolution reaction ... [https://www.sciencedirect.com/science/article/pii/S0920586123004339]
• 35. Prediction of CO2 reduction reaction intermediates and ... [https://arxiv.org/abs/2504.15710]
• 36. Mechanistic insights and rational catalyst design in NO x ... [https://pubs.rsc.org/en/content/articlehtml/2025/nr/d5nr01682g]
• 37. Microkinetic Analysis and Scaling Relations for Catalyst ... [https://pubmed.ncbi.nlm.nih.gov/29641915/]
• 38. CO 2 hydrogenation on Ni(111): microkinetic modelling vs ... [https://pubs.rsc.org/en/content/articlehtml/2025/su/d5su00240k]
• 39. First-Principles Microkinetic Modeling Unravelling the ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9923683/]
• 40. Tailoring the catalytic performance of MXenes in propane ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12483649/]
• 41. Structure-dependent microkinetic modeling of the CO2 ... [https://www.sciencedirect.com/science/article/abs/pii/S0021951725004737]
• 42. Integrating operando spectroscopies and transient analysis ... [https://link.springer.com/article/10.1007/s11426-024-2455-1]
• 43. Operando X-ray absorption spectroscopic investigation of ... [https://www.nature.com/articles/s41467-025-57177-y]
• 44. Best practices for in-situ and operando techniques within ... [https://www.nature.com/articles/s41467-025-57563-6]
• 45. Regulating the scaling relationship for high catalytic ... [https://www.nature.com/articles/s41467-022-34169-w]
• 46. Mapping Proton and Carbon Dioxide Electrocatalytic ... [https://pubs.rsc.org/en/Content/ArticleLanding/2025/SC/D5SC05744B]
• 47. Phase-interface-anchored cadmium single-atom catalysts ... [https://www.nature.com/articles/s41467-025-63060-7]
• 48. Predicting competitive anion electrosorption on late transition ... [https://pubs.rsc.org/en/content/articlehtml/2025/sc/d5sc03757c]
• 49. Understanding the reaction energetics of oxygen-evolving ... [https://pubs.rsc.org/en/content/articlehtml/2025/ee/d5ee02196k]
• 50. Lattice oxygen evolution in rutile Ru [https://www.sciencedirect.com/science/article/pii/S0013468624008077]
• 51. Breaking linear scaling relationships in NRR with hydrogen ... [https://www.sciencedirect.com/science/article/abs/pii/S246802302400378X]
• 52. Breaking scaling relations in AgAuCuPdPt high-entropy alloy ... [https://pubs.rsc.org/en/content/articlehtml/2025/mh/d5mh01064k]
• 53. Rational Design of Atomically Dispersed Metal Site ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10104677/]
• 54. Customized reaction route for ruthenium oxide towards ... [https://www.nature.com/articles/s41467-023-36380-9]
• 55. Performance of exchange-correlation approximations to ... [https://www.sciencedirect.com/science/article/abs/pii/S0927025625001806]
• 56. Ru-Mn pair-site triggers key oxygen intermediate for ... [https://www.sciencedirect.com/science/article/abs/pii/S1385894724062156]
• 57. Tracking the role of compressive strain in bowl-Like Co ... [https://www.sciencedirect.com/science/article/abs/pii/S0926337324004284]
• 58. Dual-site segmentally synergistic catalysis mechanism [https://pmc.ncbi.nlm.nih.gov/articles/PMC10897303/]
• 59. Visualizing the structural evolution of individual active sites ... [https://www.nature.com/articles/s41929-024-01148-x]
• 60. Circumventing Scaling Relations via Gradient Orbital ... [https://pubmed.ncbi.nlm.nih.gov/40458913/]
• 61. Oxygen Reduction Reactions of Catalysts with Asymmetric ... [https://www.mdpi.com/2073-4344/15/7/615]
• 62. Dual-site segmentally synergistic catalysis mechanism [https://www.nature.com/articles/s41467-024-45700-6]
• 63. Brønsted bases promote interfacial proton transfer for ... [https://pubs.rsc.org/en/content/articlehtml/2025/sc/d5sc04035c]
• 64. Modulate the strength of *H adsorption by changing ... [https://www.sciencedirect.com/science/article/abs/pii/S0926337325004783]
• 65. Advances and challenges in the application of confined ... [https://www.sciencedirect.com/science/article/abs/pii/S095965262500112X]
• 66. Multidimensional Engineering of Nanoconfined Catalysis [https://www.mdpi.com/2073-4344/15/5/477]
• 67. Confined catalysis under two-dimensional materials - PMC [https://pmc.ncbi.nlm.nih.gov/articles/PMC5468632/]
• 68. Multifunctional, Biocompatible Hybrid Surface Coatings ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12349159/]
• 69. The design and optimization of heterogeneous catalysts using ... [https://pubs.rsc.org/en/content/articlehtml/2024/cy/d3cy01160g]
• 70. Metal-hydroxyls mediate intramolecular proton transfer in ... [https://www.nature.com/articles/s41557-025-01993-8]
• 71. Proton donor/acceptor effects on electrochemical ... [https://www.sciencedirect.com/science/article/abs/pii/S2451910323001709]
• 72. Synergetic bimetallic catalysts: A remarkable platform for ... [https://www.sciencedirect.com/science/article/abs/pii/S2095495623004485]
• 73. The impact bimetallic – Ni deposit configuration has on accessing... Fe [https://pubs.rsc.org/en/content/articlehtml/2025/cy/d5cy00036j]
• 74. Advances in the synthesis of Fe -based bimetallic electrocatalysts for... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11915088/]
• 75. Structural evolution of metal single-atoms and clusters in ... [https://pubs.rsc.org/en/content/articlehtml/2025/sc/d5sc01221j]
• 76. Dynamics and control of active sites in cobalt diselenide ... [https://www.nature.com/articles/s41467-025-65210-3]
• 77. Reconciling experimental catalytic data stemming from ... [https://pubs.rsc.org/en/content/articlelanding/2023/sc/d2sc06819b]
• 78. Structure Sensitivity and Catalyst Restructuring for CO2 ... [https://www.nature.com/articles/s41467-025-59267-3]
• 79. Microkinetic modeling with machine learning predicted ... [https://www.sciencedirect.com/science/article/abs/pii/S2468823123000275]
• 80. Scaling relations between adsorption energies for computational screening and design of catalysts - Catalysis Science & Technology (RSC Publishing) [https://pubs.rsc.org/en/content/articlelanding/2014/cy/c4cy00335g]
• 81. 1.3: Surface Structures- fcc Metals [https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Surface_Science_(Nix)/01%3A_Structure_of_Solid_Surfaces/1.03%3A_Surface_Structures-_fcc_Metals]
• 82. Calculated properties of surface and subsurface nickel ... [https://www.sciencedirect.com/science/article/abs/pii/S0039602899007487]
• 83. Crystal-facet-dependent surface transformation dictates the ... [https://www.nature.com/articles/s41467-023-43901-z]
• 84. Mechanistic insights into facet-dependent catalytic ... [https://www.sciencedirect.com/science/article/abs/pii/S1383586624046793]
• 85. Facet- and graphene-support-dependent activity of Ta 2 O 5 ... [https://pubs.rsc.org/en/content/articlehtml/2025/lf/d5lf00195a]
• 86. Frontiers | Non-reactive facet specific adsorption as a route to remediation of chlorinated organic contaminants [https://www.frontiersin.org/journals/catalysis/articles/10.3389/fctls.2023.1116867/full]
• 87. Breaking scaling relations with inverse catalysts [https://arxiv.org/html/2504.16493v1]
• 88. Using the Theozyme Model to Study the Dynamical ... [https://www.mdpi.com/1420-3049/29/23/5518]
• 89. Artificial enzyme catalytic capabilities improved, say ... [https://www.drugtargetreview.com/news/57121/artificial-enzyme-catalytic-capabilities-improved-say-researchers/]
• 90. Complete computational design of high-efficiency Kemp ... [https://www.nature.com/articles/s41586-025-09136-2]
• 91. Synzymes: The Future of Modern Enzyme Engineering - PMC [https://pmc.ncbi.nlm.nih.gov/articles/PMC12568871/]
• 92. First-row transition metal oxide oxygen evolution ... [https://pubs.rsc.org/en/content/articlelanding/2020/se/d0se01087a]
• 93. Composition-activity relationship of first-row transition ... [https://www.sciencedirect.com/science/article/pii/S0013468625011880]