Mehrdad Asgari

Benchmark tables are used to justify descriptor choice, model superiority, and candidate prioritization in MOF machine learning, yet the reproducibility of those conclusions is rarely examined as training data are added.

Machine learning can screen millions of metal–organic frameworks (MOFs), but the crucial question for chemists is not only which model is accurate on average; it is which individual predictions are sufficiently supported to act on.

Transparent descriptors support first-pass MOF stability triage – with domain awareness.

Defect engineering in Zr-based metal–organic frameworks (Zr-MOFs) has focused primarily on missing-linker defects. However, recent studies suggest that node dehydroxylation–which creates distortions and coordinatively unsaturated Zr sites (Zrcus)–may have a more significant impact on properties. The present work uses pair distribution function (PDF) and thermogravimetric analysis coupled with systematic defect manipulation to study the effect of node dehydroxylation and missing-linker defects in UiO-66. By employing rapid heat treatment (RHT) under humid flow, we tracked the transition from high-symmetry [Zr6O4(OH)4]12+ to distorted [Zr6O6]12+ nodes. This structural evolution significantly improves As(V) uptake, whereas increasing the number of missing linkers–via chemical treatment or RHT of mixed-ligand frameworks–fails to enhance performance. Crucially, our detection of distorted nodes in as-synthesized UiO-66 also raises the possibility that these defects were silently present in many earlier studies that span various applications, where their role in governing performance may have been inadvertently overlooked. The present study challenges the prevailing “missing-linker” paradigm and establishes cluster dehydroxylation as a defect-engineering strategy to enhance Lewis-acidic performance in Zr-MOFs.

Machine learning is accelerating the design of functional dyes, but reported errors are often dominated by random train–test splits that mix related chromophores, solvents and literature sources. This makes it unclear whether models can support discovery beyond familiar dyesolvent neighborhoods. Here we ask how optical-property predictors perform when evaluation is made chemically honest. We harmonized Deep4Chem and ChemFluor into deduplicated solution-phase benchmarks for absorption and emission wavelengths, preserving dye, solvent and provenance structure. Classical fingerprint/descriptor models were then tested under random, scaffold-family, Murcko-scaffold, scaffold+solvent and article/provenance splits. Random splits gave apparently excellent mean absolute errors of 12.8 nm for absorption and 18.9 nm for emission. The same benchmark became far harder under scaffold-family testing (39.2 and 46.4 nm) and the scaffold+solvent-stress protocol (50.6 and 59.3 nm). Cross-dataset transfer remained substantially poorer than within-dataset interpolation, while error atlases localized failures in red/NIR regimes, macrocyclic families and solvent-sensitive environments. These results reframe dye-property prediction as a benchmark-metrology problem: useful models must report not only low random-split error, but also scaffold, solvent and dataset-shift robustness.

Machine-learning-guided MIL-100(Fe) synthesis reveals that methylene blue removal is maximized by application-specific functionality, not by BET surface area alone.

Defect engineering in Zr-based metal-organic frameworks (Zr-MOFs) has focused primarily on missing-linker defects. However, recent studies suggest that node dehydroxylation–which creates distortions and coordinatively unsaturated Zr sites (Zr _cus )–may have a more significant impact on properties. The present work uses pair distribution function (PDF) analysis and systematic defect manipulation to decouple node dehydroxylation from missing-linker defects in UiO-66. By employing rapid heat treatment (RHT) under humid flow, we tracked the transition from high-symmetry [Zr6O4(OH)4] ¹²⁺ to distorted [Zr6O6] ¹²⁺ nodes. This structural evolution significantly improves As(V) uptake, whereas increasing the number of missing linkers–via chemical treatment or RHT of mixed-ligand frameworks–fails to enhance performance. Crucially, our detection of distorted nodes in as-synthesized UiO-66 also raises the possibility that these defects were silently present in many earlier studies that span various applications, where their role in governing performance may have been inadvertently overlooked. The present study challenges the prevailing ”missing-linker” paradigm and establishes cluster dehydroxylation as a defect-engineering strategy to enhance Lewis-acidic performance in Zr-MOFs.

fo Discover chemically novel MOFs using CHEM-AD, an unsupervised machine learning pipeline. We leverage autoencoders for structural anomaly detection to expand reticular design and enhance material dataset reliability.