Abstract

This study presents an inquiry case utilizing generative artificial intelligence (ChatGPT) to quantitatively analyze the non-ideal behavior of real gases observed in Boyle’s Law experiments, a common topic in secondary science education. Pressure-volume measurements of nitrogen () and carbon dioxide () revealed that both gases exhibited pressures higher than the theoretical values predicted by Boyle’s Law, demonstrating a clear deviation from ideal gas behavior. Although such deviations can be explained by the van der Waals equation—which accounts for intermolecular interactions and finite molecular volume—applying this model in secondary education is challenging due to the mathematical complexity of solving high-order equations. To overcome this educational barrier, this study employed ChatGPT to perform curve fitting on the experimental data, thereby deriving van der Waals constants and elucidating differences in intermolecular interactions based on molecular structure. The derived constants showed a significant correlation with literature values; specifically, it was confirmed that carbon dioxide, with its larger molecular size and greater number of electrons, exhibits stronger intermolecular attraction than nitrogen. These findings suggest that AI tools can effectively support complex data interpretation processes, thereby deepening students' scientific inquiry capabilities beyond the limitations of traditional calculation methods.