Exploring temperature-dependent photoluminescence dynamics of colloidal CdSe nanoplatelets using machine learning approach

Abstract

The study explore machine learning (ML) techniques to predict temperature-dependent photoluminescence (PL) spectra in colloidal CdSe nanoplatelets (NPLs), leveraging polynomial regression models trained on experimental data from 85 to 270 K spanning temperatures to forecast PL spectra backward to 0 K and forward to 300 K. 6th-degree polynomial models with Tweedie regression were optimal for band energy ([Formula: see text]) predictions up to 300 K, while 9th-degree models with LassoLars and Linear Regression regressors were suitable for backward predictions to 0 K. For exciton energy ([Formula: see text]), the Lasso model of degree 5 and the Ridge model of degree 4 performed well up to 300 K, while the Tweedie model of degree 2 and Theil-Sen model of degree 2 showed promise for predictions to 0 K. Furthermore, a GA-based approach was utilized to fit experimental data to theoretical model of Fan and Varshni equations, facilitating a comparative analysis with the ML-predicted curves.

Keywords: CdSe; Luminescence; Machine learning; Nanoplatelets; Temporal dynamics.

Fig. 1

Experimental setup.

Fig. 2

(a) TEM image, (b)–(d) histograms of the size distribution (, , ) and the fitting of the data to a Gaussian distribution, (e) XRD pattern, and (f) HR-TEM image for the synthesized colloidal CdSe NPLs.

Fig. 3

Optical density (black) and PL (red) spectra of CdSe colloidal NPLs at room temperature (300 K).

Fig. 4

Examples of double Gaussian approximation of CdSe NPL spectra at various temperatures.

Fig. 5

A representative series of PL spectra of CdSe NPLs as a function of temperature. The inset: PL spectra at temperatures of 85K and 270K.

Fig. 6

metrics for polynomial regression models for (a) predictions on range 270–300 K, and (b) predictions on range 0–85 K.

Fig. 7

Evolution of coefficeints and metrics during optimization of Varshni and Fan equations for fitting temperature-dependent shifts in and to experimental data.

Fig. 8

Comparison of experimental data with theoretical models for (a) the Fan equation and band energy, (b) the Fan equation and Varshni equation for exciton experimental data energy, (c) experimental data with ML predictions for integral intensity.