CARBON QUANTUM DOT-BASED PHOTODIODE INTRINSIC LAYER TUNED TO HUMAN CONE RESPONSES

Abstract

Background: Conventional retinal prosthetic systems rely on RGB photodiodes to approximate the color perception of human vision. However, such systems are computationally intensive, bulky, and poorly matched to the spectral sensitivities of natural photoreceptors. Quantum dots (QDs) offer tunable optical properties and high photostability, providing an alternative route for biologically coherent optical sensing. Carbon quantum dots (CQDs), in particular, offer low toxicity, cost-effectiveness, and biocompatibility, making them suitable for retinal biomimetic applications.

Objective: This study aimed to develop and characterize a CQD-based photodiode model that replicates the spectral response of human cone cells (L, M, and S) using computational modeling, optical analysis, and machine learning–assisted synthesis optimization.

Methods: CQDs were synthesized using a bottom-up hydrothermal method with citric acid as the carbon precursor. UV–Vis spectroscopy was employed to determine absorbance across 400–700 nm. The intrinsic layer was created by embedding CQDs in a 40% PMMA–chlorobenzene matrix, thermally cured at 80 °C. Absorption peaks were analyzed via Gaussian regression, while a neural network trained on synthesis variables (temperature, time, doping) predicted the relationship between processing conditions and spectral output. Reflectance was calculated using the Fresnel equation, and EQE profiles were modeled to align with human cone sensitivities.

Results: The L, M, and S CQDs demonstrated peak absorbances at 564 nm, 534 nm, and 420 nm respectively, matching the physiological cone response ranges. The intrinsic CQD–PMMA composite achieved over 85% optical transparency and a 2.5× increase in absorbance due to solvent shrinkage. Gaussian fitting yielded R² > 0.97, and neural network predictions reached 93% accuracy for target wavelength estimation.

Conclusion: The study confirms that CQD–PMMA composites can effectively replicate human cone spectral sensitivity with stable optical and mechanical characteristics. The integration of data-driven synthesis optimization establishes a pathway for developing compact, energy-efficient, and biocompatible optical prosthetic systems.