Upcoming Event:

New Insights into Ionic Functionality in Angstrom-Scale Confinement: From Separations to Neuromorphic Computing

Mohan Teja Dronadula,

Abstract

At the angstrom scale, fluid and ion transport enters a regime where classical continuum descriptions fail. Under extreme confinement, ion transport is heavily influenced by dehydration energetics, surface-ion interactions, and solvent structure changes. Biological channels operate in this regime, leveraging these non-classical effects to achieve highly selective transport, nonlinear conduction, and memristive behavior. This dissertation investigates how synthetic angstrom-scale confinement can be engineered to emulate such biological functionalities for applications in critical mineral separation and neuromorphic computing. 

 

First, we address the challenge of separating tantalum and niobium, which are chemically similar critical elements that are difficult to isolate due to their comparable physical and chemical properties. Using catalytically active sub-nanometer MXene pores, we demonstrate selective transport among their fluoro-complexes. Machine learned molecular dynamics simulations reveal a catalytic dissociation mechanism that induces differential size modulation: tantalum complexes undergo enhanced reduction compared to niobium complexes. Leveraging this emergent size asymmetry, we achieve approximately fourfold higher tantalum flux, establishing a novel strategy for separating these chemically similar metal complexes. Next, we investigate nonlinear ion transport behavior in angstrom-scale confinement, akin to biological pores. Within this regime, we observed conductance switching and memristive ion transport, in two different systems: graphene Angstrom channel and Janus MoSSe Angstrom pore. 

 

In both cases, the solvent structure changes with electric field, leading to nonlinear transport behavior. Further analysis reveals that the delayed relaxation of this perturbed solvent structure after field removal results in a pinched loop current electric field hysteresis, characteristic of memristors. This intrinsic memory enables neuronal functionality such as plasticity, which we demonstrate using the Janus system.  Finally, leveraging these nonlinear ion dynamics, we implement a reservoir computing architecture and demonstrate learning capability in handwritten digit classification (MNIST) and DNA base recognition tasks. Together, these results establish engineered angstrom-scale confinement as a platform for achieving both selective separation and bioinspired information processing.

Biography

I am a 5th-year PhD candidate in the Mechanical Engineering Department and I am from Hyderabad, India. Before coming to UT Austin, I did Master’s in Mechanical Engineering from the University of Illinois Urbana-Champaign (UIUC) and a Bachelor of Technology in Mechanical Engineering from VNIT, Nagpur. I currently work as a Graduate Research Assistant in the Oden Institute with Dr. Narayana R. Aluru, where my research focuses on designing nanofluidic memristive systems with applications in neuromorphic computing.

 

Contact: dronadula@utexas.edu