Unravelling structure–property–performance relationships in nanocomposite membranes using principal component analysis
Abstract
This study investigates the structure–property–performance relationships of advanced membrane materials for crystal violet (CV) dye removal using Principal Component Analysis (PCA). A dataset of electrospun nanocomposite membranes, including pristine, PES/PVDF-blended, and nanofiller-modified variants incorporating ZIF-67, graphene oxide (GO), and graphene nanoplatelets (GNPs) was analysed. PCA reduced dataset complexity and revealed two dominant components explaining 76.50% of the total variance. The first component (PC1, 54.66%) represented a performance axis driven by viscosity, porosity, pore size, flux, and rejection, while the second component (PC2, 21.84%) captured a wettability–processing axis governed by contact angle and flow rate. Based on the dominant PCA loadings, two reduced PCA-derived descriptors were formulated: the Rheology–Structure–Transport Index (RSTI) and the Hydrophilicity–Processing Index (HPI). These descriptors provide compact, physically interpretable summaries of the main PC1 and PC2 tendencies within the analysed dataset, rather than externally validated predictive models. The results show that high membrane performance can be achieved through distinct physicochemical pathways depending on nanofiller type. The proposed PCA-based framework offers a robust and interpretable tool for analysing complex membrane datasets and provides valuable guidance for the rational design and optimisation of advanced membranes for wastewater treatment.
Related
- Enhanced antifouling performance of polyvinyl alcohol membranes via magnetite-graphene oxide nanocomposite modification
- Why does PTFE still matter? Revisiting its structure–property–performance-applications relationships beyond the PFAS debate
- Polymer-modified MXene thin-film nanocomposite membranes for forward osmosis
- Multifunctional polycaprolactone–lignin–silver nanocomposite membranes for simultaneous removal of heavy metal, nutrient and microbial contaminants from wastewater
- Response-surface optimization of high-performance GO-PET-PEG forward osmosis membranes for bentazone removal
BibTeX
@article{younes2026unravelling,
title = {{Unravelling structure–property–performance relationships in nanocomposite membranes using principal component analysis}},
author = {Khaled Younes and Samer El-Zahab and Eddie Gazo-Hanna and Mohamed A.O. Abdelfattah and Omar Mouhtady and Baghdad Ouddane and Jalal Halwani and Jack Arayro},
journal = {Journal of Membrane Science Letters},
year = {2026},
month = oct,
doi = {10.1016/j.memlet.2026.100119},
url = {https://hal.science/hal-05823643v1},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Unravelling structure–property–performance relationships in nanocomposite membranes using principal component analysis
AU - Khaled Younes
AU - Samer El-Zahab
AU - Eddie Gazo-Hanna
AU - Mohamed A.O. Abdelfattah
AU - Omar Mouhtady
AU - Baghdad Ouddane
AU - Jalal Halwani
AU - Jack Arayro
PY - 2026
DA - 2026/10/20
JO - Journal of Membrane Science Letters
DO - 10.1016/j.memlet.2026.100119
UR - https://hal.science/hal-05823643v1
AB - This study investigates the structure–property–performance relationships of advanced membrane materials for crystal violet (CV) dye removal using Principal Component Analysis (PCA). A dataset of electrospun nanocomposite membranes, including pristine, PES/PVDF-blended, and nanofiller-modified variants incorporating ZIF-67, graphene oxide (GO), and graphene nanoplatelets (GNPs) was analysed. PCA reduced dataset complexity and revealed two dominant components explaining 76.50% of the total variance. The first component (PC1, 54.66%) represented a performance axis driven by viscosity, porosity, pore size, flux, and rejection, while the second component (PC2, 21.84%) captured a wettability–processing axis governed by contact angle and flow rate. Based on the dominant PCA loadings, two reduced PCA-derived descriptors were formulated: the Rheology–Structure–Transport Index (RSTI) and the Hydrophilicity–Processing Index (HPI). These descriptors provide compact, physically interpretable summaries of the main PC1 and PC2 tendencies within the analysed dataset, rather than externally validated predictive models. The results show that high membrane performance can be achieved through distinct physicochemical pathways depending on nanofiller type. The proposed PCA-based framework offers a robust and interpretable tool for analysing complex membrane datasets and provides valuable guidance for the rational design and optimisation of advanced membranes for wastewater treatment.
SN - 2772-4212
ER -