(4fr) Theory-Guided Design of Membrane and Processes for Efficient Separations

The sustainable production of critical metals (CMs) will play a pivotal role in the energy transition, with demand forecast to rise six-fold by 2050.¹ Batteries, magnets, photovoltaic cells, electrolyzers, fuel cells, and catalysts, rely on a diverse array of CMs. Rapidly increasing production is essential to mitigate the risk of CM demand outstripping supply and stalling the net-zero transition.² Efficient separations processes can extract CMs from natural (salt lakes, magmatic brines) and anthropogenic (geothermal, oilfield) brines, increasing CM supplies.

Existing separations technologies, including solvent extraction and ion exchange, are energy, water, and chemical intensive, generating large quantities of waste requiring extensive post-treatment, which limits their use in many locations.⁴ By contrast, nanofiltration (NF) uses hydraulic pressure to drive water and selected ions through selective membranes (Figure 1). Electrodialysis (ED) uses an electric field to drive the permeation of cations and anions through cation (CX) and anion exchange (AX) membranes. Using nanopores with tailored size and functionalization to achieve ion-ion selectivity, NF, CX, and AX membranes can separate CMs from complex aqueous mixtures without requiring large volumes of organic solvents or acids.

Membranes have been widely studied for water treatment applications, which comprise dilute (≤10 g_salt/L), binary and ternary (≤3 salts) solutions, while the non-selective transport of concentrated binary electrolytes has been studied for battery design. However, the extraction and purification of CMs from brines requires selective ion separations in concentrated (>100 g/L), multi-ion (>3 salts) aqueous solutions (Figure 1). Current thermodynamic and mass transfer models rely on semi-empirical adjustable parameters with minimal physical significance and predictive capability. Furthermore, experimental measurements for every ion composition are infeasible due to combinatorial explosion. Without a rigorous understanding of selective multi-ion permeation, membrane and process design for CM extraction and purification is limited to a slow and expensive trial-and-error approach with new experiments required to study each new brine composition.⁵

Research Approach

My research group will build a robust, generalizable models, overcoming combinatorial explosion, for selective transport in thermodynamically nonideal, multi-ion mixtures combining thermodynamics, mass transfer, fluid mechanics, and process design. The models built and physical understanding developed will enable the rational design of membrane materials and processes for a new generation of CM separations.

Thrust 1: Building Robust Models for Selective Mass Transfer in Nonideal, Multicomponent Mixtures

Developing membranes and processes that leverage selective transport to separate CMs requires accurate chemical potential calculations in concentrated multi-ion solutions (Figure 2). Industrially relevant CM-containing brines comprise a wide range of cations, from Li to REEs and other heavy metals. Current excess Gibbs free energy models rely on adjustable semi-empirical ion-specific parameters to capture ion-solvent and ion-ion interactions. In Thrust 1, we will build generalizable models to predict unknown binary and ternary ion-solvent and ion-ion interaction parameters in excess Gibbs free energy models using tensor completion methods and by encoding stereoelectronic ion and solvent information. Developing a robust understanding selective ion transport is crucial for the design of efficient, membrane-based CM separations (Figure 2). Fickian models, in which diffusion coefficients are highly composition dependent, can lead to significant errors in concentrated, multi-ion mixtures. My group will develop a physics-based continuum mass transport framework. Initially, formulations based on the Poisson-Nernst-Planck equations will be used to model convective and diffusive transport of ions through porous media with the Stokes-Maxwell-Stefan equations for multi-ion transport replacing the Fickian diffusion assumption. These frameworks will be coupled with cheap high-throughput experimental measurements to estimate unknown diffusion coefficients.

Thrust 2: Developing Multiscale Computational Frameworks for Ion Fractionation

Multi-ion separations using membranes are inherently multiscale with ion permeation governed by selective pores (nanometers), transmembrane transport (microns), concentration boundary layers (millimeters), and module-scale composition variations (meters) (Figure 3). In Thrust 2, my group will build multiscale computational frameworks for NF and ED, starting from the nonideal multi-ion transport models developed in Thrust 1. Transmembrane transport models will be extended to account for the effect of concentration polarization, the formation of concentration gradients in boundary layers adjacent to selective membranes. PDE solvers will be developed to integrate the coupled flux models over a finite membrane area, building an understanding of how composition variations during selective extraction impact ion permeation in NF and ED. Surrogate models will be built to form a low-dimensional approximation of the full module-scale models. The surrogates will be used to build system-scale models for ion fractionation using multiple NF and ED modules in a membrane cascade. Finally, computationally efficient membrane cascade models will be used to quantify how membrane properties, including ion permeance and ion-ion selectivity, impact optimal CM separation performance.

Thrust 3: Understanding Membrane Formation for Novel Architectures

The transition from water-salt separations to ion-ion selectivity has vastly improvement the design space for new membrane materials and structures (Figure 4). A limited understanding of how membrane synthesis impacts membrane design and performance restricts the development of new membranes currently requiring an expensive trial-and-error approach. In Thrust 3, my group will build analytical and computational models for the reaction-diffusion-convection processes that govern membrane fabrication through interfacial polymerization, solution casting, and phase inversion. These models will be used to develop a fundamental understanding of how monomer chemistry, solvent choice, support layer design, and other innovations can impact membrane structure. These models will combine heat and mass transfer with fluid dynamics to understand how reaction rate control and other synthesis conditions such as temperature can affect resulting membrane chemistries and architectures. Combining a theory-based continuum approach with molecular dynamics simulations will allow fundamental insight to be developed to identify rate-dominating phenomena, ultimately enabling the physics-based design of membrane synthesis.

Long-Term Research Vision

Going forward, I hope to lead a diverse and collaborative independent research group that designs new membranes and processes to tackle tough separations from ion and chemical fractionation to biomolecular separations for sustainable drug manufacturing. Thrusts 1-3 will form the building blocks of a high-throughput membrane development platform that combines online mixture characterization with data-driven experimental design. By building collaborations with other research groups and industrial partners, the platform will be used to design (1) new membrane materials and (2) membrane cascades for complex mixture separations. In the long-term, the computational platform will be extended to explore the data-driven design of membranes and processes for partially specified mixtures, where composition information might be incomplete.

Teaching Interests

My core teaching interests center on separations, heat and mass transfer, thermodynamics, numerical methods, process design, and process control for a new generation of chemical engineers equipped to tackle the new challenges at the heart of the clean energy transition. Through each of these subject areas, I hope to provide students with 1) analytical and theoretical insights to enable rapid, low-resource, first-order estimation and 2) an understanding of open-source computational tools for higher-order calculations.

In separations, I am passionate about incorporating new modules building beyond distillation-based separations to use the thermodynamics, heat and mass transfer, and fluid dynamics to study membranes and adsorption. In thermodynamics, I will incorporate new case studies on refrigeration cycles and energy conversion processes to illustrate practical applications for emerging green technologies. I will also endeavor to help students build an understanding of electrolyte thermodynamics, moving beyond neighboring-molecule based interaction models. In process design and control, I will focus on holistic project-based learning, helping students designing complete chemical processes, such as biofuel production or desalination plants, incorporating process simulation and optimization techniques.

Throughout my teaching, I will strive to use open-source computational tools and packages, including those developed by my group and others in Python and Julia, to help students build a versatile skillset for diverse future career paths. By combining strong analytical intuition for core chemical engineering concepts with broadly applicable computational skills and project experience, I will aim to equip students with the practical skills and theoretical knowledge they need successful careers in the chemical engineering industries of the future and beyond.

Education and Training

Postdoctoral Associate (2019-2023) & Research Scientist (2023-), Mechanical Engineering, Advisor: Prof. John H. Lienhard, Massachusetts Institute of Technology

PhD (2019) & MSc (2015), Chemical and Environmental Engineering, Advisor: Prof. Menachem Elimelech, Yale University

BA & MEng (2013), Chemical Engineering, University of Cambridge

References

¹IEA, The Role of Critical Minerals in Clean Energy Transitions (2021). ²Olivetti et al., Joule (2017). ³DuChanois et al., Nat. Water (2023). ⁴Vera et al., Nat. Rev. Earth Environ. (2023). ⁵Lair et al., Annu. Rev. Chem. Biomol. Eng.(2024).