Autonomous laboratories typically rely on data-driven decision-making, occasionally with human-in-the-loop oversight to inject domain expertise. Fully leveraging AI agents, however, requires tightly coupled, collaborative workflows spanning hypothesis generation, experimental planning, execution, and interpretation. To address this, we develop and deploy a human-AI collaborative (HAIC) workflow that integrates large language models for hypothesis generation and analysis, with collaborative policy updates driving autonomous pulsed laser deposition (PLD) experiments for remote epitaxy of BaTiO3/graphene. HAIC accelerated the hypothesis formation and experimental design and efficiently mapped the growth space to graphene-damage. In situ Raman spectroscopy reveals that chemistry drives degradation while the highest energy plume components seed defects, identifying a low-O2 pressure low-temperature synthesis window that preserves graphene but is incompatible with optimal BaTiO3 growth. Thus, we show a two-step Ar/O2 deposition is required to exfoliate ferroelectric BaTiO3 while maintaining a monolayer graphene interlayer. HAIC stages human insight with AI reasoning between autonomous batches to drive rapid scientific progress, providing an evolution to many existing human-in-the-loop autonomous workflows.
Cite
@article{arxiv.2511.11558,
title = {Human-AI collaborative autonomous synthesis with pulsed laser deposition for remote epitaxy},
author = {Asraful Haque and Daniel T. Yimam and Jawad Chowdhury and Ralph Bulanadi and Ivan Vlassiouk and John Lasseter and Sujoy Ghosh and Christopher M. Rouleau and Kai Xiao and Yongtao Liu and Eva Zarkadoula and Rama K. Vasudevan and Sumner B. Harris},
journal= {arXiv preprint arXiv:2511.11558},
year = {2025}
}