Extraction and Simulation of the Impact of Flux Trapping in Moats of AC-Biased Shift Registers

TL;DR

This study simulates how flux trapping in moats affects the operation margins of large-scale ac-biased superconducting shift registers, finding minimal impact when moats are properly designed, with junction variability being the dominant factor.

Contribution

It introduces a simulation framework combining inductance extraction and circuit simulation to assess flux trapping effects on large superconducting shift registers.

Findings

Flux trapping has minimal impact on operation margins with proper moat design.

Junction critical current variation dominates threshold variability.

Simulation results align with experimental threshold distributions.

Abstract

Moats in superconducting ground planes are used to trap magnetic flux away from sensitive parts of superconductor integrated circuits. We simulate the effect of magnetic flux trapped in moats on the operating margins of ac-biased SFQ shift registers (ShReg) with two ground planes for various congruent moat geometries, moat sizes, locations in the ShReg cells, number and polarity of the trapped of fluxons. Using inductance extractor InductEx, we extract mutual couplings between the moats and the ShReg inductors and include them in the refined netlist. Then, we use JoSim to simulate the circuit operation and find the threshold ac clock amplitude above which the register starts to operate correctly. The relative change in this threshold is used to characterize the influence of flux trapping on the circuit operation. Monte-Carlo simulations are used to investigate the effect of the circuit parameter variations on the statistics of the threshold, using the standard deviation of the Josephson junction critical current and of the circuit inductance extracted from the fabrication process control monitors. The obtained distributions are compared with the independently measured distributions of the threshold amplitude for the individual cells of ac-biased shift registers with 108,500 Josephson junctions. The results show that flux trapping in properly designed moats have a very small effect on the operation margins of the large-scale shift registers whereas the main contribution to the cell-to-cell variation of the threshold comes from the variation of the junctions' critical currents.