[15a-K405-1~11] 12.5 Organic and hybrid solar cells

11:45 AM - 12:00 PM JST(2:45 AM - 3:00 AM UTC)

[15a-K405-11] Machine Learning Approach Using LOOCV for Designing High J_SC Donor Molecules in Fullerene-Based Organic Thin-Film Solar Cells

Yumi Morishita¹, Misato Yarimizu¹, Masanori Kaneko², 〇Azusa Muraoka¹ (1.Japan Women's Univ., 2.Yokohama City Univ.)

Keywords:

Fullerene-based organic solar cells,Short-circuit current density,Machine learning

The Donor molecule materials for fullerene-based organic solar cells were categorized into training and test datasets, and machine learning was conducted using leave-one-out cross-validation (LOOCV) with donor J_SC as the objective variable to generate donor molecules with high J_SC values. The J_SC was calculated using a machine learning method. The donor molecule structure that exhibited the highest contribution to J_SC comprised three five-membered ring structures with thiophenes at both termini of the unit structure and a partially fluorinated molecular backbone.