Datasets for manuscript: Toward full autonomous laboratory instrumentation control with large language models

This dataset provides the supplementary materials, raw data, and analysis codes associated with the publication:

"Towards Full Autonomous Laboratory Instrumentation Control with Large Language Models"

The repository is intended to support the reproducibility and transparency of the experimental results presented in the paper.

Contents:

1. Supplementary document:   - Detailed descriptions of the experimental setup, AI-agent workflow, and instrument interfacing procedures.   - Includes implementation details for Standa XY stage and Keithley 2450 control.

2. Figure 3c:   - Raw scanning data (Excel format)   - Python script (Figure3cPlot.py) used for data visualization and reconstruction

3. Figure 4c:   - Raw scanning data for photocurrent mapping experiments

4. Figure 5:   - Compressed dataset including:     - AI agent generated code     - Iterative code evolution logs     - Automatically collected experimental data (I–V curves)     - Generated figures

Description of methodology:

The datasets were generated through automated or semi-automated experimental workflows using large language models (LLMs) for instrument control. The system integrates:

- Keithley 2450 SourceMeter for electrical measurements- Standa XY stage for spatial scanning- Python-based AI agent for autonomous experiment execution and error correction

Reproducibility:

All data and scripts provided here allow partial or full reproduction of the results presented in Figures 3–5 of the associated publication.

Notes:

- The Python scripts may require modification depending on local hardware configuration.- Instrument drivers (e.g., VISA, libximc) must be properly installed.- The datasets are provided in raw format without post-publication normalization.

Keywords:AI for Science; Autonomous Laboratory; Instrument Control; Large Language Models; Keithley 2450; Standa XY Stage; Photocurrent Mapping