Milad Abolhasani and colleagues demonstrated Artificial Chemist 2.0, an autonomous flow chemistry platform for quantum dot synthesis
In December 2020, Milad Abolhasani's group at North Carolina State University published Artificial Chemist 2.0, an autonomous flow chemistry system combining machine learning with robotic synthesis to navigate a space of approximately 20 million quantum dot formulations and produce a target material within roughly 30 minutes of initiating a search.
Background
Making quantum dots, tiny semiconductor crystals that emit light at precise wavelengths, is less straightforward than it sounds. The colour a quantum dot emits depends on its size and composition, and getting both exactly right means adjusting several chemical variables at once: precursor concentrations, temperature, flow rate, reaction time. Even small changes can shift the result considerably. In practice, researchers would make a batch, measure it, adjust something, make another batch, and repeat. Weeks could pass before a formulation met the target specification.
Automated chemistry had started to reduce some of that burden. Flow chemistry systems, which push reagents through narrow tubes and mix them continuously rather than in flasks, allowed reactions to be adjusted more quickly than traditional batch methods. But the search for the right conditions still relied heavily on human intuition. A researcher would decide what to try next based on experience and judgement, which meant the process could only move as fast as a person could reason through the results and plan the next experiment.
The deeper problem was the sheer number of possible formulations. For quantum dots used in LED displays and solar cells, the relevant chemical space runs to tens of millions of distinct combinations. No human-directed campaign could cover more than a tiny fraction of it.
What happened
In December 2020, Milad Abolhasani’s group at North Carolina State University published their work on Artificial Chemist 2.0 in the journal Matter. The system combined a machine learning layer with a robotic flow chemistry platform in a closed loop: the algorithm would decide what experiment to run, the hardware would run it, and the measured result would feed straight back into the next decision.
The machine learning component used Bayesian optimisation, a method that builds a probabilistic model of a search space and chooses each new experiment to give the most useful information given what is already known. This matters because Bayesian optimisation can work without a pre-existing dataset to train on. The system started from scratch each time and still found its way to a target formulation efficiently.
The space it was searching across was about 20 million possible quantum dot formulations. Starting from nothing, Artificial Chemist 2.0 could identify and produce a material meeting a given optical specification in roughly 30 minutes. The paper described results across several quantum dot compositions, with the platform hitting target photoluminescence properties that would have taken a human-led effort far longer to locate. The work appeared in Matter, published by Cell Press, and the NC State Department of Chemical and Biomolecular Engineering described it at the time as a step toward industrial-scale autonomous materials synthesis.
Why it mattered
Artificial Chemist 2.0 demonstrated that a closed-loop machine learning system could replace human-directed trial-and-error in materials synthesis, compressing weeks of laboratory work into under an hour. By coupling Bayesian optimisation with automated microfluidic reactors, the platform made high-dimensional chemical search spaces tractable without requiring a pre-existing dataset. This represented a concrete step toward fully autonomous materials discovery pipelines with direct industrial relevance in display and photovoltaic manufacturing.
People
Organisations
North Carolina State University
Sources
- Artificial Chemist 2.0: Toward Autonomous Synthesis of Colloidal Nanomaterials.Matter (Cell Press).Primary source
- The Artificial Chemist 2.0: From Lab Concept to Industrial Relevance.NC State University, Department of Chemical and Biomolecular Engineering.Official
Cite this page
AI Achievements. (2020). Milad Abolhasani and colleagues demonstrated Artificial Chemist 2.0, an autonomous flow chemistry platform for quantum dot synthesis. Retrieved 2026-08-22, from https://achievements.ai/milestone/milad-abolhasani-artificial-chemist-2-0
@misc{achievements_milad_abolhasani_artificial_chemist_2_0,
title = {Milad Abolhasani and colleagues demonstrated Artificial Chemist 2.0, an autonomous flow chemistry platform for quantum dot synthesis},
author = {{AI Achievements}},
year = {2020},
url = {https://achievements.ai/milestone/milad-abolhasani-artificial-chemist-2-0}
}