Quantum Computer Flaws Pinpointed Using Novel Energy Decay Spectroscopy Technique

Scientists are increasingly focused on understanding the microscopic origins of energy relaxation in superconducting qubits, a critical challenge for advancing quantum computing. Tanay Roy, Xinyuan You, and David van Zanten, working with colleagues at the Superconducting Quantum Materials and Systems (SQMS) Center and Fermi National Accelerator Laboratory, present a new spectroscopy method to characterise these defects without requiring frequency tuning. Their research details how analysing correlated multilevel relaxation, specifically, repeatedly preparing the second excited state of a qubit and extracting information from both first and second excited state decays, allows identification of dominant two-level systems (TLSs) and reconstruction of their frequency drift over time. This innovative approach, detailed in their paper, significantly expands the scope of TLS spectroscopy and offers a powerful tool for mitigating decoherence in superconducting quantum devices.

These defects, known as two-level systems (TLSs), are a primary source of noise that limits the coherence and performance of these devices. The study centres on transmon qubits, a widely used type of superconducting circuit where temporal fluctuations in energy relaxation are often attributed to TLSs present in the device materials. Traditionally, isolating these individual defects required precise tuning of either the qubit or the TLS into resonance, adding complexity to scalable quantum architectures and introducing sensitivity to external noise. This work introduces a novel spectroscopy method based on analysing multilevel relaxation, repeatedly preparing the qubit in its second excited state and simultaneously measuring the decay rates of transitions between all three energy levels. By carefully examining the correlations in these decay rates, researchers can identify dominant TLSs and reconstruct their frequency drift over time. This technique circumvents the need for frequency tuning, offering a pathway to stabilise coherence and improve the reliability of quantum computations. Furthermore, the ability to map TLS frequency fluctuations provides crucial insights into the microscopic mechanisms driving coherence loss, potentially enabling strategies to engineer and mitigate these detrimental effects. The proposed method offers a powerful diagnostic tool for TLS spectroscopy, particularly valuable for fixed-frequency transmons favoured in scalable architectures due to their simplicity and reduced sensitivity to low-frequency noise. Remarkably, the team found that TLSs detuned by more than 100MHz can still exert a significant influence on qubit relaxation, challenging previous assumptions about the range of TLS influence and highlighting the importance of addressing even distant defects. Initial measurements with device A reveal a median T1e of 155 microseconds and a median T1f of 64 microseconds, establishing a baseline for temporal relaxation studies. Subsequent monitoring over a 60-hour period demonstrates a pronounced anti-correlation between T1e and T1f, highlighted by a region where fluctuations are most significant. Analysis of device B yielded a median T1e of 111 microseconds and a median T1f of 90 microseconds, but exhibited minimal correlation between the two relaxation times, necessitating a two-TLS model for accurate fitting. The observed anti-correlation in device A suggests the influence of a single, fluctuating TLS near the qubit transition frequencies. Fitting the data to a single-TLS model allows reconstruction of the TLS frequency drift over time, revealing fluctuations superimposed on a baseline frequency. The extracted TLS linewidth is approximately 100MHz, indicating a relatively well-defined defect. Notably, TLSs detuned by as much as 200MHz from the qubit transition can still significantly influence relaxation. For device B, the lack of strong correlation between T1e and T1f necessitates a more complex model incorporating two TLSs, each contributing to the observed relaxation dynamics. The frequency of the identified TLS in device A fluctuates between 4542MHz and 4822MHz, while the qubit transition frequencies ω01 and ω12 serve as reference points for comparison. This methodology leverages the fact that even TLSs detuned by greater than 100MHz from the qubit transition can still significantly influence relaxation processes. The experimental setup incorporates a transmon circuit coupled to a readout resonator and probed via a transmission line, allowing for precise measurement of population dynamics. By analysing the fluctuations of decay rates over several days, researchers reconstruct the frequency drift of the identified TLSs over time, providing insights into their dynamic behaviour and contribution to coherence fluctuations. Scientists recognise that maintaining the delicate quantum states of superconducting qubits is as important as building them. Decoherence, the loss of quantum information, is often traced back to microscopic defects within the materials constituting these qubits, specifically TLSs acting as unwanted noise sources. Pinpointing these individual defects has proven difficult, typically requiring laborious tuning of the qubit itself to bring the defect into resonance for detection. This new work sidesteps that limitation with a fixed-frequency spectroscopy method, allowing researchers to characterise TLSs even when they are significantly detuned from the qubit’s operating frequency, revealing that distant defects can still exert a considerable influence on qubit performance. While the reconstruction of TLS properties relies on statistical analysis and assumptions about their distribution, and the method currently focuses on identifying dominant TLSs, this fixed-frequency technique promises to accelerate materials characterisation and inform the design of quieter, more robust qubits. The next step will likely involve integrating this spectroscopy with advanced materials engineering, aiming to suppress or neutralise these troublesome TLSs and deliver on the promise of scalable quantum computation.