Liu, Yihan
(2025)
Microelectronic Devices for Optimizing Neuromorphic Computing Hardware Engineering.
Doctoral Dissertation, University of Pittsburgh.
(Unpublished)
Abstract
Neuromorphic computing, inspired by the structure and function of the human brain, represents a transformative approach to computing. By using artificial neuron and synapse hardware, biological principles can be applied to neural networks to perform tasks in a manner akin to brain function. Various emerging non-volatile memory (NVM) devices, such as resistive random-access memory (RRAM) and phase-change memory (PCM), have been proposed for neuromorphic applications. However, these devices face challenges due to the complex thermal-electrical processes and electrochemical reactions involved, which hinder their performance.
To address these challenges, we first present a study of the transient response of thermoelectric devices to optimize thermal management in spiking neuron networks (SNN) and spike-driven neuromorphic computing hardware. We fabricated a micro-scale silicon thermoelectric cooler (TEC) and conducted measurements to validate our model's predictions. Our results elucidate the transient temperature modulation capabilities of the TEC, its energy efficiency, and the effect of material properties and operation frequency. This work provides a foundation for the theoretical analysis of transient thermal management in emerging spiking-driven devices and neuromorphic hardware.
Additionally, we introduce a novel CMOS-compatible protonic memristor. In contrast to conventional oxygen-based or metal-cation-based memristors, protonic memristors exhibit a significantly smaller ion size, resulting in lower activation energy barriers and faster diffusion speeds. Our experiments demonstrate a notably faster pulse response in protonic memristors compared to their conventional counterparts. Moreover, due to the simple electrochemical processes governing its operation, the protonic memristor performs reliably in diverse environments, such as vacuum and high moisture conditions. This study underscores the potential of proton-based devices in high-speed neuromorphic computing applications, including spiking-rate dependent plasticity (SRDP).
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
7 January 2025 |
| Date Type: |
Publication |
| Defense Date: |
1 November 2024 |
| Approval Date: |
7 January 2025 |
| Submission Date: |
1 November 2024 |
| Access Restriction: |
No restriction; Release the ETD for access worldwide immediately. |
| Number of Pages: |
156 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
Swanson School of Engineering > Electrical and Computer Engineering |
| Degree: |
PhD - Doctor of Philosophy |
| Thesis Type: |
Doctoral Dissertation |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
Thermoelectric
In-memory computing |
| Date Deposited: |
07 Jan 2025 21:07 |
| Last Modified: |
07 Jan 2025 21:07 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/47054 |
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[feed]](/images/feed-icon-32x32.png){kind=icon}