Skip to content

carlosdprados/phd-thesis

Repository files navigation

Polymer-Electrolyte Organic Memristors for Temporal Computing

PhD thesis repository of Carlos David Prado-Socorro at the Institute of Molecular Science (ICMol), Universitat de Valencia.

Read the current thesis draft | Introduction | Proof-of-concept chapter | Comparative chapter | Temporal-computing chapter

Thesis In Brief

This thesis studies solution-processed organic memristive devices whose electrical response is shaped by mobile ions inside a polymer composite. These devices do not merely store a static resistance value: their conductance carries a decaying record of recent electrical activity.

The central idea is to use that fading memory as a computational resource. The current experimental result is strongest for polymer-electrolyte composition: in the replicated PEO/LiOTf grid, composition tunes the switching window, pulse integration, and fading-memory time. Host, anion, and cation chemistry are retained as an illustrative, sample-limited tuning landscape rather than as a powered Li > Na > K law.

Two-terminal polymer-electrolyte composite memristor and its chemically tunable fading-memory timescale

The active layer combines a semiconducting conjugated polymer, an oxygen-rich ion-transport host, and a dissolved salt. An applied electric field redistributes the ions. The residual ionic profile changes the device conductance and acts as its internal memory state.

Why This Matters

Conventional computers repeatedly move data between memory and processing units. For data-intensive workloads, this transfer can cost more energy and time than the computation itself. Memristive hardware offers two complementary routes around that bottleneck:

  1. In-memory computing, where programmed conductances perform operations inside a memory array.
  2. Event-driven temporal computing, where the natural relaxation of a physical device processes time-dependent signals.

This thesis focuses on the second route. A device that gradually forgets is not treated as a failed non-volatile memory. Its relaxation time becomes a useful feature for processing streams, spikes, transients, and other time-resolved signals.

Classical, in-memory, and event-driven temporal computing paradigms

Research Finding

The working physical hypothesis was that memory timescale is linked to ion transport inside the polymer electrolyte. Ether oxygens in the host coordinate alkali-metal cations, so cation identity was expected to change the relaxation dynamics:

  • Li+: stronger coordination and longer fading-memory timescales
  • Na+: intermediate behaviour
  • K+: weaker coordination and shorter fading-memory timescales

The current Chapter 3 draft keeps that HSAB argument as qualitative framing, but the archive does not support a robust host- and anion-independent Li > Na > K ordering. The defensible quantitative result is the composition dependence in the PEO/LiOTf grid; chemistry beyond that is explicitly sample-limited.

Polyether coordination chemistry, ion-transport modes, and the cation-timescale hypothesis tested in Chapter 3

Thesis Roadmap

Chapter Focus Status
1. Introduction Computing bottlenecks, synaptic inspiration, memristors, organic materials, and the polymer-electrolyte strategy Draft available
2. Proof of concept A fully characterised SY/Hybrane/LiOTf two-terminal device with analogue switching, short- and long-term retention, EPSC-like response, and STDP Draft available
3. Comparative study How composition (the PEO/LiOTf grid) tunes volatile dynamics — switching, potentiation, and fading memory — and how electrolyte chemistry shifts them further as sample-limited side evidence Draft available
4. Temporal computing Data-driven reservoir simulations from Chapter 3 parameter cards: memory-capacity/NARMA benchmarks, WESAD physiological temporal-context reconstruction, and WESAD affective-computing demonstrations; coincidence detection is cut and filter-bank logic is folded into the reservoir framing Draft available
5. Conclusions Contributions, limitations, and future directions Planned

The proof-of-concept chapter expands the published work:

Carlos David Prado-Socorro et al., "Polymer-Based Composites for Engineering Organic Memristive Devices", Advanced Electronic Materials 8, 2101192 (2022).

Repository Guide

Path Contents
thesis.tex Top-level LaTeX entry point
chapters/ Standalone chapter sources and shared thesis formatting
figures/ Figures used throughout the thesis
bibliography/ Shared BibLaTeX database
exports/ Committed PDF snapshots
handouts/ Planning documents, outlines, and working notes
docs/ Reference docs about the experimental archive and analysis pipeline (data and code live in the sibling Nanomem_Devices_Library/)

Build The Thesis

The chapter files compile both independently and as part of the complete thesis. A local LaTeX installation with latexmk and biber is required.

make chapter1       # build/chapters/chapter1_introduction.pdf
make chapter2       # build/chapters/chapter2_proof_of_concept.pdf
make thesis         # build/thesis.pdf
make all            # build chapters and thesis
make exports        # refresh committed PDF snapshots
make clean          # remove generated LaTeX artefacts

Run these commands from the repository root. Generated LaTeX files are written to build/ and excluded from version control.

About

PhD thesis on chemically tunable polymer-electrolyte organic memristors for neuromorphic and temporal computing.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors