Related Experiment Video
Updated: May 1, 2026

Phase Diagram Characterization Using Magnetic Beads as Liquid Carriers
Published on: September 4, 2015
Machine Learning-Assisted Phase Diagram Determination in Aqueous Two-Phase Systems
Nidhin Thomas1, Adam Witmer2, Hyung Kae Lee3
1Theoretical Division, Theoretical Biology and Biophysics (T-6), Los Alamos National Laboratory, Los Alamos, New Mexico 87545, United States.
This study introduces an automated method using microscopy and machine learning for accurate phase diagram construction in aqueous two-phase systems (ATPSs). This overcomes limitations of traditional methods, enabling efficient characterization of phase separation.
Area of Science:
- Physical Chemistry
- Materials Science
- Biotechnology
Background:
- Accurate phase diagrams for aqueous two-phase systems (ATPSs) are crucial for downstream applications.
- Conventional turbidimetric titration methods for ATPS phase diagram construction are labor-intensive and prone to inaccuracies.
Purpose of the Study:
- To develop an automated methodology for precise phase diagram construction in ATPSs.
- To overcome the limitations of traditional experimental techniques for characterizing phase separation.
Main Methods:
- Integration of optical microscopy with machine learning for automated phase separation detection.
- Utilizing a segmentation model (SAM 2) to analyze thousands of images and quantify emulsion droplets.
- Employing supervised machine learning on droplet features for robust classification of one- or two-phase regions.
Main Results:
- Successfully identified two-phase regions for a challenging ATPS (poly(ethylene glycol)/dextran) with slow phase separation kinetics.
- Demonstrated accurate phase boundary definition, overcoming issues with dim or occluded droplet edges.
- Validated the efficiency and robustness of the automated platform for phase diagram construction.
Conclusions:
- The developed automated strategy offers a significant advancement over conventional methods for ATPS phase diagram construction.
- This approach provides a reliable and efficient platform for characterizing phase-separating systems.
- Potential for broad applications in various fields requiring precise phase behavior analysis.
More Related Videos
Related Concept Videos
Phase Diagrams of Ternary Systems
Phase Diagram
Phase Diagram
Phase Diagrams
Two Components: Liquid–Liquid Systems
Solid–Solid Solutions

