Back to Services

Phase Identification

Overview

Automatically match your experimental diffraction pattern against crystallographic reference databases. MatPilot searches both the Crystallography Open Database (COD) API and a curated local database to identify crystalline phases present in your sample.

Scientific Background

Phase identification in XRD relies on matching the observed peak positions and relative intensities against reference patterns from known crystalline materials. Each crystal structure produces a unique diffraction fingerprint determined by its unit cell dimensions, space group, and atomic positions. MatPilot uses multiple similarity metrics including cosine similarity, figure of merit (FOM), and RMSE of peak positions to rank candidate phases.

Workflow

1

Run the analysis pipeline to detect peaks from your experimental pattern

2

Configure search parameters: query text, element constraints, database selection

3

Search the COD API for formula or element matches across 500,000+ structures

4

Query the local reference database for common materials

5

Generate theoretical diffraction patterns from CIF crystal data using pymatgen

6

Compare theoretical patterns against experimental data using cosine similarity

7

Rank candidates by match score, confidence, and number of matched peaks

Input Formats

Detected peaks from pipeline
Manual peak lists
Element or formula queries

Expected Outputs

Ranked list of candidate phases with match scores
Confidence levels (High / Medium / Low)
Matched peak count and peak fraction statistics
CIF files for top candidates
Theoretical diffraction patterns overlaid on experimental data

Benefits

Searches both online COD and local offline database simultaneously
Automatic theoretical pattern generation from crystal data
Multiple similarity metrics provide robust matching
Confidence scoring helps prioritize candidate selection
CIF download for downstream Rietveld refinement

Frequently Asked Questions

What databases does MatPilot search?
MatPilot searches the Crystallography Open Database (COD) via its REST API, and a local curated database of 50+ common crystalline materials. The local database serves as an offline fallback when the COD API is unreachable.
How accurate is phase identification?
Accuracy depends on pattern quality, preferred orientation effects, and database coverage. MatPilot uses multiple scoring metrics and reports confidence levels to help assess reliability.
Can I upload my own reference patterns?
Yes, you can upload CIF files directly to an experiment. These are parsed and used alongside database results for phase identification.
Try This Service