Low Power Electronics and Thermal Management for the Cryogenic Temperature Regime

Ridgard, George and Prance, Jonathan and Tsepelin, Viktor (2025) Low Power Electronics and Thermal Management for the Cryogenic Temperature Regime. PhD thesis, Physics.

Text (2025ridgardphd)
2025ridgardphd.pdf - Published Version
Available under License Creative Commons Attribution-NonCommercial.
Download (6MB)

Abstract

Understanding how power dissipation affects the local temperature distribution is key when conducting sensitive experiments at low temperature. This thesis develops strategies for reducing the power consumption of electronics at cryogenic temperatures and for managing the thermal gradients created by the dissipated power. First, the low temperature characterisation of low threshold Field Effect Transistors devices is shown, to assess their suitability for low supply voltage cryogenic circuits. Secondly, the \textit{in situ} testing of resistors is used to optimise two differential amplifiers for gain, power consumption, and bandwidth. Finally, results from both experiments lead to the demonstration of a novel way of reducing supply voltages in low temperature electronics, `cryogenic threshold engineering'. This is the use of FETs with negative room temperature thresholds, that move close to zero once cooled to low temperatures. This technique was demonstrated in a common source amplifier. Both theory, simulation and experiment showed a factor of 8 improvement in efficiency from 295K to 4K. In this thesis a setup designed to mimic a quantum computing was thermally investigated at milli-kelvin temperatures. This involved monitoring the copper mount, interposer, PCB, and internal IC temperature as a function of power dissipated in the IC. A experiment was performed to show aluminium wirebonds can act as superconducting heat switches between hot and cold devices in the same setup. Voltage noise thermometry with the cross-correlation technique was demonstrated inside the IC. This was demonstrated to work in the range 300~mK to 8~K. At 500~mK the benchmark error was 5\%. Also, expression for the relative error of the cross-correlation measurement is derived. Finally, a technique to speed up cross-correlation, frequency averaging, is demonstrated in room temperature resistors. Then it is applied to a literature example to demonstrate its application in reducing the uncertainty in noise thermometry.