AI for Catalyst Discovery and Optimization

Electrocatalyst Discovery

Prioritize experimentally testable electrocatalysts by combining traceable data, electrochemical descriptors, physics-aware machine learning, density functional theory (DFT), and active learning.

Discuss Your Project Contact Us

Overview

Discover Better Electrocatalysts with Fewer Experimental Cycles

Electrocatalyst discovery is rarely a single-property ranking exercise. Practical candidates must satisfy activity, selectivity, durability, precious-metal loading, electrolyte compatibility, synthesis feasibility, and test-protocol constraints at the same time. A catalyst that looks excellent under one potential, pH, loading, or product metric can fail once selectivity, stability, or scale-up constraints are added.

CD ComputaBio provides an AI-guided electrocatalyst discovery workflow that combines curated experimental data, DFT-derived descriptors, graph and structure-aware models, active learning, and expert electrochemical interpretation. The goal is not only to output a list of predicted materials, but to generate a defensible, experiment-ready candidate portfolio with uncertainty, rationale, and validation priorities.

Typical input materials

Target reaction, electrolyte, potential range, baseline catalyst and success thresholds
Internal activity/selectivity/stability data or literature datasets for harmonization
Candidate chemistry constraints, excluded elements, cost and synthesis preferences
Available validation budget, analytical methods and acceptable turnaround time

Core Services

From fragmented evidence to testable decisions

Each module can be commissioned independently or integrated into a staged program, from rapid feasibility assessment to closed-loop optimization.

Data foundation and harmonization

Extract and normalize composition, support, active-site structure, electrolyte, potential scale, loading, product analysis, stability protocol and provenance across literature and client datasets.

Descriptor and surrogate modeling

Build interpretable descriptors and ML models for adsorption, limiting potential, selectivity windows, activity trends and uncertainty; benchmark against transparent baselines.

DFT-informed candidate screening

Run targeted DFT or higher-level calculations for decision-critical candidates, including adsorption energetics, reaction free energies, scaling-relation deviations and solvation or field corrections where needed.

Active learning and design of experiments

Select the next most informative calculations or experiments by expected improvement, uncertainty reduction, diversity and practical feasibility.

Multi-objective ranking

Prioritize candidates by Pareto performance across activity, selectivity, stability, abundance, synthesis accessibility and IP-sensitive chemistry constraints.

Validation-ready reporting

Deliver ranked candidates, reasoning, model limits, test matrix, controls and go/no-go criteria that electrochemistry teams can act on directly.

Workflow

A gated, uncertainty-aware project plan

The workflow is organized as a decision table so scope, evidence quality and next actions remain clear throughout the project.

Project stage	Key activities	Decision output
1. Decision framing	Define target reaction, product metric, baseline, electrolyte, potential, durability threshold and forbidden chemistries.	A scoped discovery question with measurable success criteria.
2. Data build	Curate public and client data; harmonize units, RHE conversion, catalyst loading, test geometry, product quantification and quality flags.	Auditable dataset and evidence map.
3. Model development	Train descriptor, graph, ensemble or hybrid models; evaluate leakage, calibration, uncertainty and applicability domain.	Validated model card and candidate search space.
4. Candidate triage	Apply feasibility filters, Pareto ranking, diversity clustering and targeted DFT verification for high-impact candidates.	Shortlist with rationale and risk labels.
5. Validation design	Specify synthesis/test conditions, controls, replicate strategy, expected signals and stopping rules.	Experiment-ready plan for the next discovery cycle.

Deliverables

Files your scientists can inspect and reuse

Deliverable	What is included
Data package	Versioned tables, schema, units, provenance, exclusions, RHE conversions, protocol notes and quality-control summary.
Model package	Model card, features, split strategy, metrics, calibration plots, applicability domain and reproducible scripts or notebooks where appropriate.
Candidate portfolio	Ranked and Pareto candidates with uncertainty intervals, chemistry rationale, diversity clusters and experimental feasibility flags.
Validation dossier	DFT inputs/outputs or experiment-ready test matrix, recommended controls, expected product analysis and go/no-go rules.
Decision review	Scientist-led review of model confidence, limitations, next-batch recommendation and options for a closed-loop follow-up.

Applications

Representative project contexts

Oxygen reduction and evolution (ORR/OER)Hydrogen evolution and oxidation (HER/HOR)CO2 and CO electroreductionNitrogen and nitrate reductionAlcohol and biomass electro-oxidationPaired electrolysis and bifunctional catalysts

Scientific Evidence

Open-access examples informing the workflow

These studies illustrate how interpretable descriptors, high-throughput screening and AI-assisted experiments can reduce expensive calculations and focus validation on higher-value candidates.

Descriptor-driven electrocatalyst discovery workflow — Descriptor-driven workflow for dual-atom electrocatalyst discovery and validation.¹

High-throughput OER experiment and AI workflow — High-throughput experimental and AI-assisted workflow for oxygen evolution electrocatalysts.²

References

Lin, X. et al. Machine learning-assisted dual-atom sites design with interpretable descriptors unifying electrocatalytic reactions. Nature Communications 15, 8169 (2024). doi:10.1038/s41467-024-52519-8. Open access, CC BY 4.0.
Xu, S. et al. Developing new electrocatalysts for oxygen evolution reaction via high throughput experiments and artificial intelligence. npj Computational Materials 10, 194 (2024). doi:10.1038/s41524-024-01386-4. Open access, CC BY 4.0.
Mok, D. H. et al. Data-driven discovery of electrocatalysts for CO2 reduction using active motifs-based machine learning. Nature Communications 14, 7303 (2023). doi:10.1038/s41467-023-43118-0. Open access, CC BY 4.0.

Online Inquiry

Submit your project details below, and our team will respond within 24 hours.