Characterizing Transport Properties of Surface Charged Nanofiltration Membranes via Model-Based Data Analytics

Area of Science:

• Membrane Science and Engineering
• Chemical Engineering
• Separation Science

Background:

• Accurate characterization of nanofiltration (NF) membrane transport properties is crucial for optimizing separation processes.
• Existing methods often lack precision in capturing dynamic transport phenomena, hindering efficient membrane development.
• Surface charge significantly influences solute transport in NF membranes, necessitating advanced characterization techniques.

Purpose of the Study:

• To present an integrated framework for characterizing surface-charged NF membrane transport properties.
• To compare the precision of different experimental startup modes (lag vs. overflow) for parameter estimation.
• To demonstrate the utility of dynamic diafiltration experiments and model-based data analytics in understanding transport mechanisms.

Main Methods:

• Dynamic diafiltration experiments were conducted to capture transient transport behaviors.
• An integrated framework incorporating startup dynamics, time corrections, and water flux effects was developed.
• Model-based data analytics, including A-, D-, and E-optimality, were employed for parameter estimation and model discrimination.

Main Results:

• The lag startup mode demonstrated superior precision over the overflow mode, with significant improvements in parameter estimates (8% A-optimality, 138% D-optimality, 83% E-optimality).
• Diafiltration experiments in the diluting regime effectively discriminated between dominant transport mechanisms.
• The developed framework accurately captures key transport properties governing NF membrane performance.

Conclusions:

• The proposed framework enhances the accuracy and efficiency of characterizing NF membrane transport properties.
• This work advances the development of self-driving laboratories (SDLs) and model-based design of experiments (MBDoE) for accelerated membrane development.
• The findings enable the inverse design of high-performance NF membranes tailored for specific separation applications.