Simulation and analysis of energy-efficient biasing networks for superconducting digital electronics using open-source SPICE models

Supervisor: Dr Joao Barbosa

School: Engineering

Description:

For decades, superconducting digital electronics has been investigated as a transformative beyond-CMOS technology, capable of delivering ultrafast data processing with unparalleled energy efficiency. A critical component of these circuits is the biasing network, which ensures reliable operation with exponentially low error rates. The simplest implementation of such a biasing network relies on small resistors connected to each circuit gate, which regulate the current flow to ensure proper gate operation. However, resistors continuously dissipate power, even when the circuit is idle, significantly limiting overall energy efficiency. 

To address this limitation, an alternative approach was proposed in the early 2010s, replacing resistive elements with superconducting inductors and Josephson junctions. This innovative design drastically reduces energy dissipation up to two orders of magnitude, presenting a highly promising solution for energy-efficient computing. However, these advanced biasing networks exhibit significantly more complex behaviour, characterized by multiple operation regimes, each with distinct biasing properties and dynamic responses. A thorough understanding and systematic optimization of these regimes are critical to unlocking the full potential of superconducting digital electronics and enabling their practical implementation in next-generation digital and quantum computing systems. 

This project aims to develop a simulation model using the three most popular open-source superconducting SPICE simulators—JOSIM, WRSPICE, and PSCAN2—to emulate the behaviour of energy efficient biasing networks. By leveraging the unique strengths of each simulator, we will perform a comprehensive analysis of energy efficiency across different operation regimes. The model will incorporate experimental data obtained in Glasgow (part of the Quantum Circuits group led by Professor Martin Weides) to account for noise sources such as flux and thermal noise, enabling its application to model practical circuits. Furthermore, the project will compare the performance and accuracy of these three simulators to identify their strengths and limitations in modelling superconducting circuits. 

In summary the project will be divided into the following stages: 

  • The student will first conduct a literature review to understand the theory of operation of superconducting digital circuits. 
  • The student will develop SPICE-compatible models for superconducting biasing networks using inductors, resistors and Josephson junctions, and implement them in JOSIM, WRSPICE, and PSCAN2 to perform comparative simulations. 
  • The student will simulate the behavior of the biasing network across different operation regimes and analyze energy efficiency. 
  • The student will validate the model by comparing simulation results across all three simulators and with experimental data.