Machine learning assembly landscapes from particle tracking data

Andrew W. Long, Jie Zhang, Steve Granick, Andrew L. Ferguson

Research output: Contribution to journalArticlepeer-review


Bottom-up self-assembly offers a powerful route for the fabrication of novel structural and functional materials. Rational engineering of self-assembling systems requires understanding of the accessible aggregation states and the structural assembly pathways. In this work, we apply nonlinear machine learning to experimental particle tracking data to infer low-dimensional assembly landscapes mapping the morphology, stability, and assembly pathways of accessible aggregates as a function of experimental conditions. To the best of our knowledge, this represents the first time that collective order parameters and assembly landscapes have been inferred directly from experimental data. We apply this technique to the nonequilibrium self-assembly of metallodielectric Janus colloids in an oscillating electric field, and quantify the impact of field strength, oscillation frequency, and salt concentration on the dominant assembly pathways and terminal aggregates. This combined computational and experimental framework furnishes new understanding of self-assembling systems, and quantitatively informs rational engineering of experimental conditions to drive assembly along desired aggregation pathways.

Original languageEnglish (US)
Pages (from-to)8141-8153
Number of pages13
JournalSoft Matter
Issue number41
StatePublished - 2015

ASJC Scopus subject areas

  • Chemistry(all)
  • Condensed Matter Physics


Dive into the research topics of 'Machine learning assembly landscapes from particle tracking data'. Together they form a unique fingerprint.

Cite this