مطالعه آزمايشگاهي و مدلسازي رياضي عملكرد نانوفيلتراسيون داروي ضدتومور دوكسوروبيسين توسط غشا پلي‌سولفون در حضور عوامل حساس به pH شامل گروه‌هاي عاملي كربوكسيليك اسيد و پيريدين

Experimental study an‎d mathematical modeling of the nanofiltration performance of the antitumor drug Doxorubicin by polysulfone membrane in the presence of pH-sensitive agents including carboxylic acid an‎d pyridine functional groups

مهندسي شيمي

In the present study, Polysulfone pH-responsive membranes were fabricated an‎d eva‎luated for the separation of the antitumor drug doxorubicin from water. Three types of membranes were prepared: a membrane containing TiO₂@P4VP nanoparticles in the sublayer (PAA0/P4VPj), a membrane with TiO₂@PAA nanoparticles in the surface layer (PAAi/P4VP0), an‎d a hybrid membrane incorporating both types of nanoparticles (PAAi/P4VPj). The filtration performance of these membranes was investigated under various pH conditions. Incorporation of TiO₂@P4VP nanoparticles into the sublayer led to an 88% increase in porosity an‎d a fourfold enhancement in permeability compared to the pristine membrane. The presence of TiO₂@PAA nanoparticles in the surface layer improved hydro‎philicity an‎d surface charge, resulting in a 2.4-fold increase in permeability. In the hybrid membrane containing both nanoparticle types, a 22% increase in porosity led to a 2.2-fold enhancement in permeability. Increasing pH significantly improved the filtration performance; in PAA₀/P4VP₀ an‎d PAA0/P4VP4 membranes, permeability increased by 5.26 an‎d 1.14 times, respectively, with increasing pH. Regarding fouling behavior, TiO₂@P4VP nanoparticles reduced the total fouling ratio (TFR) by 19% through enhanced surface hydro‎philicity. TiO₂@PAA nanoparticles completely eliminated reversible fouling, an‎d the simultaneous incorporation of both nanoparticle types increased surface charge density fivefold, resulting in a 46% reduction in TFR. Moreover, increasing pH reduced the TFR of PAA0/P4VP0 an‎d PAA0/P4VP4 membranes by 58% an‎d 51%, respectively. Mathematical modeling using the DSPM-DE model revealed that Donnan exclusion was the dominant mechanism governing doxorubicin separation, whereas dielectric exclusion contributed minimally. Additionally, TiO₂@P4VP nanoparticles decreased separation efficiency by reducing effective charge density, while TiO₂@PAA nanoparticles improved rejection by decreasing MWCO an‎d increasing the effective surface charge. Finally, coupling the DSPM-DE model with the XGBoost algorithm enabled prediction of doxorubicin rejection with a sum of squared errors (SSE) below 0.001. These findings highlight the high potential of hybrid physics–machine learning models in analyzing an‎d predicting the performance of pH-responsive membranes.

5

151582