Methods for Describing and Comparing TLS Impacts

The performance of superconducting qubits and two-qubit gates is often limited by two-level systems (TLS) losses. Characterizing TLS distributions and coupling strengths is critical for making informed decisions about chip design, material stack choices, and fabrication processes that affect these losses. These characteristics are complicated by the frequency-distributions of TLS, changes in these distributions over time, and variations between qubits, chips, and wafers. Here, we discuss analysis of data collected using swap spectroscopy to capture both coherent and incoherent TLS interactions. We outline several methods for characterizing TLS distributions over large sample sets, including frequency-density measurements (#/GHz) and a coupling-weighted frequency density metric, spectral pollution index (SPI). We also discuss various methods for collecting and analyzing TLS data, as well as extracting meaningful conclusions from limited datasets. Our findings highlight the importance of a systematic approach to TLS characterization.