Design and Validation of an LLM-Based Digital Twin Educational Ecosystem: Convergence of Quantum-Thermodynamic Simulation of Smart Drug Delivery Nanocomposites and Futures Studies in Basic Science Education
Design and Validation of an LLM-Based Digital Twin Educational Ecosystem: Convergence of Quantum-Thermodynamic Simulation of Smart Drug Delivery Nanocomposites and Futures Studies in Basic Science Education
Fereshteh Hemmat Yar1 Seyedeh Mina Shayeganmehr2 Poriyas Rahimi Sam3
1) M.Sc. in Organic Chemistry, Islamic Azad University (Karaj Branch, Iran); Chemistry Teacher, Ministry of Education –
2) B.Sc. in Physics and M.Sc. in Meteorology, Islamic Azad University (Tehran North Branch), Iran; Physics Teacher, Ministry of Education –
3) M.Sc. in Medicinal Chemistry, Islamic Azad University (Science and Research Branch, Iran); Chemistry Teacher, Ministry of Education –
IOI Code
:
AAHD-ABHEZD
Publication :
National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
Background and Objective: Teaching complex and unobservable phenomena at the nanoscale presents profound cognitive challenges for learners. This study aims to investigate the effectiveness of a novel educational ecosystem integrating a Digital Twin with Generative Artificial Intelligence (Large Language Models) to teach the synthesis and hydrodynamic behavior of targeted nanocarriers.
Methodology: A quasi-experimental design with a pretest-posttest and a non-equivalent control group was employed, comprising 160 twelfth-grade students. The intervention featured an interactive platform simulating quantum physics equations and fluid dynamics (utilizing an atmospheric advection-diffusion analogy) in real-time, translated through Socratic AI dialogues. To mitigate extraneous cognitive load, the platform was developed utilizing a professional, Bright User Interface (Bright UI) design, strictly avoiding distracting elements.
Results: Analysis of Covariance (ANCOVA) demonstrated that the experimental group achieved statistically significant improvements in comprehending quantum physics, reaction kinetics, and fluid dynamics compared to the traditional control group (p < 0.001). The partial eta squared
(eta^2= 0.80) indicated that 80% of the variance in deep conceptual learning was directly attributable to the intelligent ecosystem intervention.
Conclusion: By elucidating the black box of nanoparticle synthesis via digital twins and AI facilitation, this approach successfully transforms abstract theoretical concepts into a tangible, high-level cognitive learning experience.
Keywords : digital twin artificial intelligence nanoparticle synthesis fluid dynamics cognitive load bright ui