Powered by 7 MOF Databases · 138,158 Reference Structures

Identify
MOF Structures
from XRD Data

Upload XRD data files (CSV, Excel, XY, XRDML, CIF and 8 more formats) or XRD pattern images. Our engine detects peaks via Bragg's law, enforces MOF crystallinity criteria, and matches against 138,158 reference structures from CSD, CoRE MOF, QMOF, and hMOF databases.

Start Analysis → See how it works
138,158MOF References
98.6%Classifier Accuracy
13File Formats
7Source Databases

Upload & Analyse

Upload your XRD data file or pattern image. We'll match peaks against 138,158 MOF reference structures from 7 databases.

📂
Drop your XRD data file here or click to browse
.csv .xlsx .xls .txt .xy .asc .jcp .udf .x00 .xrdml .dat .raw .cif
🖼️
Drop your XRD pattern image here or click to browse
.png .jpg .jpeg .tiff .bmp .webp

📦 Download Demo Files for Testing

Try these real XRD data files from PANalytical instruments to test the analyzer:

📄
UiO-66_demo.xy Zr-BDC MOF (Zirconium-based)
📄
HKUST-1_demo.xy Cu-BTC MOF (Copper-based)
📄
ZIF-8_demo.xy Zn-Imidazolate MOF (Zinc-based)
📊
BiMOF.csv PANalytical EMPYREAN CSV format
🗂️
BiMOF.xrdml PANalytical XRDML XML format
📝
BMM.asc Tab-separated ASCII format
Don't have a file?
⚛️ Upload a file, then click Analyse to select metals

How It Works

A four-step pipeline from raw XRD data to MOF identification against 138,158 reference structures

01
📤

Upload Data

Upload XRD data files in any of 13 formats (CSV, Excel, XY, XRDML, UDF, JCP, X00, ASC, CIF, DAT, TXT, RAW, image) — no format conversion needed.

02
⚛️

Select Metal Atoms

Choose your MOF's metal node(s) from the interactive periodic table. The engine filters 138,158 references to chemically relevant matches only.

03
📈

Peak Detection

Peaks found via local maxima detection. d-spacing computed via Bragg's law. MOF validity rule enforced: sharp peaks (FWHM < 0.3°) below 15° 2θ are required.

04
🔎

Database Matching

Peaks matched against 138,158 MOF structures from CSD, CoRE MOF, QMOF, and hMOF databases (±0.25° tolerance). Top 10 matches returned with confidence scores.

138,158 MOF Structures 7 Source Databases Google Drive + Apps Script Bragg's Law 13 File Formats Image Digitization Vercel Hosting

About the Database

The reference database contains 138,158 MOF structures with simulated XRD fingerprints, compiled from seven public crystallographic databases: the Cambridge Structural Database (CSD) MOF subset, CoRE MOF 2019 & 2024, CoRE MOF 2014, the QMOF database (DFT-optimised), the hypothetical MOF (hMOF) library (51,163 structures), and the Northwestern mofdb.

Each structure has simulated 2θ peak positions (Cu Kα, λ = 1.5406 Å), primary d-spacings, unit cell parameters (a, b, c, α, β, γ), space group, metal node identity, and cell volume. The CSD-derived subset of 35,216 structures additionally includes DDEC6 partial charges and class labels.

A Random Forest classifier trained on structural features achieves 98.55% classification accuracy (AUC = 0.982) in distinguishing true MOFs, charge-reconstructed variants, and non-MOF structures. DDEC6 anomaly analysis reveals that 61.7% of charge-reconstructed CSD variants carry anomalous partial charges — a finding relevant to molecular simulation and machine learning force field development.

All MOF-validity criteria are enforced scientifically: patterns must show sharp peaks (FWHM < 0.3°) below 15° 2θ to be confirmed as MOF. Broad or high-angle-only patterns are correctly classified as non-MOF.

Try the Tool
138,158 Total MOF structures
7 Source databases
61.7% DDEC6 anomaly rate
98.6% Classifier accuracy