Improved Cupressus sempervirens L. galls for methylene blue removal: adsorption kinetics optimisation using the DA-LS algorithm, characterisation, and machine learning modeling

ABSTRACT

In this study, an investigation was conducted on the effectiveness of three bioadsorbents – galbuli of Cupressus sempervirens L. almond shells, and Luffa – for removing methylene blue (MB) from water. Various activation techniques and operational conditions were employed to enhance the adsorption capacity. The study focused on adsorption kinetics and equilibrium capacity. The results showed that microwave-dried galbuli of Cupressus sempervirens L. exhibited the highest adsorption capacity of 38.05 mg/g with an adsorption efficiency of 17.4%. To improve this capacity, the Dragonfly Algorithm-integrated least squares (DA-LS) was proposed to optimise the fitting of 32 adsorption isotherm models, which included a proposed modified Langmuir model. After optimisation, the sixth chemical treatment of Cupressus sempervirens L. reached a Langmuir isotherm q_max value of 84.44 mg/g, suggesting the high efficiency of the bioadsorbent compared to previous studies. Characterisation using microscopy and ATR-FTIR techniques revealed that chemical modifications led to rougher, microporous cellulose fibres and significant changes in the chemical structure, potentially enhancing the surface area and adsorption efficiency. To predict the adsorption capacity, the Support Vector Machine for Regression with DA (SVMR-DA) was utilised, achieving a root mean square error (RMSE) of 1.357 and an R² of 0.9913. This high accuracy indicates a reliable model with a robust error distribution. These results suggest that Cupressus sempervirens L. can be an effective bioadsorbent for water treatment, with significant potential for industrial application.

KEYWORDS:

• Dragonfly algorithm
• isotherm models
• adsorption kinetics modelling
• Cupressus sempervirens L
• machine learning modelling

Acknowledgments

We would like to express our gratitude to the Quality Control department of SAIDAL Group, Medea, for providing the platform for our project.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/03067319.2024.2382374

Additional information

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.