Computational-Chemistry-Based Prediction of Near-Infrared Rhodamine Fluorescence Peaks with Sub-12 nm Accuracy

Near-infrared (NIR) rhodamine dyes are pivotal for bioimaging due to the minimal tissue interference. Yet, their rational design is hindered by unreliable computational methods for excited-state property prediction. We benchmarked the time-dependent density functional theory (TDDFT) with the linear-...

Full description

Saved in:
Bibliographic Details
Main Authors: Qinlin Yuan, Hanwei Wang, Pingping Sun, Chaoyuan Zeng, Weijie Chi
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Photochem
Subjects:
Online Access:https://www.mdpi.com/2673-7256/5/2/15
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Near-infrared (NIR) rhodamine dyes are pivotal for bioimaging due to the minimal tissue interference. Yet, their rational design is hindered by unreliable computational methods for excited-state property prediction. We benchmarked the time-dependent density functional theory (TDDFT) with the linear-response (LR) and state-specific (SS) solvation models across five functionals (CAM-B3LYP, M06-2X, ωB97X-D, B3LYP, MN15) and optimized the ground/excited states for 42 rhodamine derivatives. A robust linear calibration framework was established by connecting the computed and experimental wavelengths, which was rigorously validated through six-fold cross-validation. The key metrics included the mean absolute error (MAE) and R<sup>2</sup> to assess the prediction robustness. CAM-B3LYP combined with LR solvation achieved the highest accuracy (absorption: MAE = 6 nm, R<sup>2</sup> = 0.94; emission: MAE = 12 nm, R<sup>2</sup> = 0.72). By integrating the TDDFT with a calibrated linear-response solvation model, we achieved sub-12 nm accuracy in predicting the NIR fluorescence peaks. This framework enabled the rational design of nine novel rhodamine derivatives with emissions beyond 700 nm, offering a paradigm shift in bioimaging probe development.
ISSN:2673-7256