Framework Library Curation
Standardize experimental and hypothetical CIFs; remove duplicates; validate cell, topology, density, coordination, and pore accessibility; and document source, chemistry, and preprocessing decisions.
Zeolites Screening and Design
Zeolites Screening and Design
AI-guided screening of framework topology, chemical composition, adsorption performance, and transport properties for validation-ready zeolite adsorbent candidates.
Start Your ProjectOverview
Translate zeolite design space into testable adsorbent choices
Zeolite performance is controlled by coupled variables: framework topology, accessible window and cage dimensions, framework composition, Si/Al ratio, aluminum siting, extra-framework cations, defects, crystal size, binder, activation history, water loading, pressure, temperature, and feed composition. A structure that ranks well for an ideal dry single-gas calculation may lose selectivity or working capacity in a humid multicomponent cycle.
Our AI for Porous Adsorbent Materials Screening and Design workflow combines validated structural data, pore-network analysis, molecular simulation, physics-informed machine learning, process metrics, and synthesis-aware down-selection. Each result is tied to its assumptions, applicability domain, and recommended experimental confirmation.
Core Services
Zeolite-specific modeling from topology to process fit
Geometric and Topological Screening
Compute pore-limiting diameter, largest cavity diameter, accessible volume, surface area, channel dimensionality, ring descriptors, and shape constraints using probe-appropriate definitions.
Adsorption and Selectivity Modeling
Use Henry-regime calculations and GCMC where appropriate to estimate isotherms, heats of adsorption, multicomponent uptake, selectivity, working capacity, and regenerability over defined conditions.
Diffusion and Transport Assessment
Evaluate molecular accessibility, intracrystalline diffusion, kinetic selectivity, loading dependence, and transport bottlenecks using molecular dynamics or reduced models matched to the decision.
Composition and Cation Design
Explore Si/Al ratio, feasible Al siting, charge-balancing cations, ion exchange, defect and hydration scenarios while respecting electroneutrality and clearly separating idealized from realizable models.
ML and Multi-Objective Ranking
Train interpretable surrogate models with scaffold- or topology-aware validation, uncertainty estimates, out-of-domain flags, and Pareto ranking across performance, stability, cost, and validation burden.
Integrated Workflow
From target separation to validation-ready shortlist
| Stage | Key Activities | Decision Output |
|---|---|---|
| 1. Project scoping | Define adsorbates, impurities, humidity, pressure–temperature envelope, equilibrium or kinetic objective, cycle, pellet constraints, baseline, and success criteria. | Fit-for-purpose screening brief |
| 2. Structure and data audit | Curate CIFs and adsorption data; check topology, disorder, accessibility, units, activation state, duplicate structures, provenance, and measurement comparability. | Analysis-ready library and gap map |
| 3. Tiered virtual screening | Apply geometric exclusion, descriptors, Henry calculations, ML surrogates, then higher-fidelity GCMC/MD to progressively smaller candidate sets. | Ranked candidates with confidence tiers |
| 4. Realism and robustness tests | Stress-test mixture, humidity, framework flexibility, cation or defect scenarios, force-field sensitivity, heat effects, and relevant operating windows. | Robustness map and risk register |
| 5. Process and feasibility screen | Translate isotherms and kinetics into working capacity, productivity, purity/recovery proxies, regeneration energy, stability, synthesis precedent, and shaping risk. | Pareto shortlist and process fit |
| 6. Validation and update | Specify synthesis or procurement, activation, PXRD, composition, porosity, single- and mixed-gas tests, cycling, moisture challenge, and model recalibration. | Executable validation plan |
Fidelity increases only where it changes a decision
Low-cost topology and pore filters remove inaccessible frameworks first. Molecular simulation and targeted calculations are then reserved for candidates that survive chemistry, condition, and process constraints.
Experimental feedback is used to diagnose force-field, structure, activation, and sample-quality discrepancies rather than treating a single measured uptake as universal validation.
Deliverables
Traceable outputs for screening and experimental action
Curated Structure Package
Standardized CIFs, topology identifiers, preprocessing records, accessibility results, descriptors, provenance, exclusions, and known structure limitations.
Adsorption and Transport Dataset
Condition-tagged isotherms, Henry coefficients, heats, mixture metrics, diffusion estimates, force fields, simulation settings, and convergence records.
Ranked Candidate Portfolio
Shortlist and Pareto front with performance drivers, confidence tiers, applicability-domain status, stability considerations, and elimination rationale.
Robustness and Sensitivity Maps
Ranking sensitivity to humidity, composition, temperature, pressure, cation or defect model, flexibility assumptions, and uncertain model parameters.
Model and Audit Report
Data lineage, methods, validation splits, residuals, uncertainty calibration, extrapolation warnings, limitations, and reproducible computational settings.
Experimental Validation Plan
Prioritized samples, activation protocol, characterization, mixture and cycling tests, acceptance logic, and data needed for the next model update.
Applications
Adsorption challenges where confinement and chemistry interact
Carbon Capture and Gas Upgrading
CO₂/N₂, CO₂/CH₄, acid-gas removal, humid feeds, pressure or vacuum swing operation, and regeneration-aware ranking.
Hydrogen and Light-Gas Purification
Trace impurity capture, PSA-relevant working capacity, competitive adsorption, kinetic effects, and thermal management considerations.
Hydrocarbon Separations
Shape-selective separation of paraffins, olefins, aromatics, and isomers where window geometry and diffusion can dominate equilibrium selectivity.
Water and VOC Management
Humidity tolerance, water harvesting or removal, volatile organic capture, breakthrough risk, desorption temperature, and cyclic durability.
Catalyst and Molecular-Sieve Design
Reactant access, product egress, confinement, acid-site environment, cation state, and diffusion constraints for adsorptive or catalytic use.
Thermal Energy and Heat-Pump Media
Water uptake windows, heat of adsorption, hydrothermal stability, cycle conditions, and candidate–process matching for adsorption heat systems.
Scientific Evidence
Open research supporting multiscale and AI-enabled modeling
Open-access studies show how machine-learned potentials can extend first-principles sampling for zeolite proton dynamics and for chemically diverse, water-loaded acidic zeolites. These examples support targeted use of ML where training coverage, transfer tests, and uncertainty are made explicit—not the unrestricted replacement of higher-fidelity calculations.1,2
The relevant model domain is defined by framework topology, chemical composition, loading, temperature, and reactive configuration space. For adsorption projects, predicted ranking is therefore stress-tested against the actual feed and cycle before experimental down-selection.
The IZA Structure Commission database and topology-checking tools provide authoritative framework identifiers and structural context for curated zeolite libraries.3
1 Bocus, M.; Goeminne, R.; Lamaire, A.; et al. Nuclear Quantum Effects on Zeolite Proton Hopping Kinetics Explored with Machine Learning Potentials and Path Integral Molecular Dynamics. Nature Communications 2023, 14, 1008. https://doi.org/10.1038/s41467-023-36666-y. Distributed under Open Access license CC BY 4.0, with modification.
2 Erlebach, A.; Šípka, M.; Saha, I.; et al. A Reactive Neural Network Framework for Water-Loaded Acidic Zeolites. Nature Communications 2024, 15, 4215. https://doi.org/10.1038/s41467-024-48609-2. Distributed under Open Access license CC BY 4.0.
3 Baerlocher, Ch.; Brouwer, D.; Marler, B.; McCusker, L. B. Database of Zeolite Structures, Structure Commission of the International Zeolite Association. https://www.iza-structure.org/databases/. Publicly accessible structural resource; consulted 5 August 2026. No database drawing reproduced.
Project Strategy
Models designed around the next defensible experiment
We keep framework version, adsorbent chemistry, activation state, feed composition, thermodynamic conditions, model assumptions, uncertainty, and validation handoffs visible throughout the project. To discuss a zeolite library, target separation, adsorption cycle, or internal dataset, please Contact Us or submit the Online Inquiry below.
Related Services
Connected porous adsorbent capabilities
Online Inquiry
Submit your project details below, and our team will respond within 24 hours.
x
Need help getting the data you need?
Talk to our technical team about your project!
I Want To Talk