Testing modularity and scaling in electromagnetic simulations of superconducting quantum devices

Superconducting circuits have quantum properties determined by the circuit design. Electromagnetic simulation is an essential step of the design process. Generally, simulations for devices such as transmon qubits use a finite-element solver to simulate electromagnetic field distributions across a device and extract circuit parameters. These simulations are computationally intensive, and full device simulation can take days, even on high-performance computers.

We compare the results of a full simulation of superconducting devices with that of a lumped-element model that discreetly evaluates each component individually. We determine how much our results vary across these two methods, and compare their computational intensity. We then analyze each method’s accuracy by contrasting our simulated results to experimental data retrieved from superconducting devices fabricated and measured in-house in a dilution fridge. We propose suggestions for future optimization of the simulation process.