Characterization of drive-induced unwanted state transitions in superconducting circuits

In the quest to build a quantum computer, scientists are racing to master the behavior of matter at its smallest scales. One of the most promising tools for this task is the superconducting circuit, a tiny electronic loop made from materials that conduct electricity without any resistance when chilled to temperatures near absolute zero. These circuits act as artificial atoms, capable of holding information in a state of quantum superposition, where they can exist in multiple configurations simultaneously. To make these circuits useful, researchers must be able to control them with extreme precision, using microwave pulses to flip their states, perform calculations, and read out the results. However, as these machines grow more complex, a persistent problem emerges: the very signals used to control the circuits can sometimes cause them to malfunction, jumping into unwanted states and losing the information they are meant to hold. This issue is a major bottleneck, limiting how fast and how accurately these quantum machines can operate.

A team of researchers at Yale University has now mapped out exactly how and why these malfunctions occur. By subjecting a specific type of superconducting circuit, known as a transmon, to a wide range of microwave frequencies and power levels, they discovered that unwanted state transitions are not caused by a single error, but by three distinct physical mechanisms. The first mechanism involves the microwave drive accidentally tuning the circuit into resonance with invisible, parasitic defects in the materials surrounding it. The second is an intrinsic property of the circuit itself, where the drive forces the system to absorb multiple photons at once, pushing it into high-energy states it was never meant to reach. The third mechanism is a form of scattering, where the drive causes the circuit to change state while dumping excess energy into stray electromagnetic waves trapped within the device's housing.

The researchers approached this problem like cartographers charting a dangerous landscape. They set up an experiment where they prepared their circuit in a known state and then bombarded it with a "pump" pulse of microwaves that varied in both frequency and power. They then measured where the circuit ended up. By sweeping the drive frequency across a massive range, from roughly 2.8 gigahertz to 11.8 gigahertz, and adjusting the power, they generated a detailed map of every possible way the circuit could be knocked off course. This map revealed a complex terrain of resonances, appearing as bright lines and fuzzy patches where the probability of an error spiked. The team's goal was to sort these features into their three categories, understanding the root cause of each so that future designs could avoid them.

The first category of trouble, which the researchers identified as resonant energy exchange, behaves differently from the others. These transitions do not depend on the specific frequency of the microwave drive, but rather on how much the drive shifts the energy levels of the circuit. When a strong microwave signal is applied, it pushes the circuit's natural frequency up or down, a phenomenon known as an ac-Stark shift. If this shift moves the circuit's frequency to match that of a tiny, hidden defect in the material—often called a two-level system—the circuit can suddenly dump its energy into that defect. The researchers observed this by watching how the circuit decayed over time. They found that these defects are not static; they drift and switch between different frequencies over periods of minutes or hours, creating a shifting target that is difficult to predict. This mechanism is a primary cause of energy loss, limiting how long the quantum information can survive.

The second category consists of transitions that are built into the circuit's own design. Even if the device were perfectly isolated from the outside world, the microwave drive could still force the circuit to absorb two or more photons at the same time. Because the circuit is non-linear, meaning its response changes depending on its state, it can jump directly from its ground state to a much higher, non-computational state by swallowing multiple drive photons in a single event. The researchers used advanced computer simulations based on the laws of quantum mechanics to predict exactly where these jumps would happen. The simulations matched the experimental data remarkably well, confirming that these transitions are an inherent feature of the circuit's physics. The researchers noted that these jumps become "fuzzy" or spread out because the circuit's internal electrical charge drifts slightly over time, blurring the exact conditions needed for the jump to occur.

The third category involves the circuit interacting with the electromagnetic environment created by the device's physical structure. When the microwave drive hits the circuit, it can scatter, changing the circuit's state while simultaneously exciting a stray electromagnetic wave trapped in the metal housing or the package surrounding the chip. These stray waves act like hidden resonators, and if their frequency matches the energy difference created by the scattering event, the transition becomes highly probable. The researchers found that these stray modes can exist at frequencies far higher than the circuit's own operating frequency, even reaching up to 28 gigahertz. By running electromagnetic simulations of the entire device package, they were able to identify the specific shapes and locations of these stray modes, showing that the physical layout of the device plays a critical role in creating these errors.

To distinguish between these three causes, the team developed a systematic method of elimination. If a transition appeared regardless of the drive frequency and depended only on the power-induced shift, it was attributed to the material defects. If the transition appeared in the computer simulations of the isolated circuit, it was classified as an intrinsic multi-photon event. Any remaining transitions that could not be explained by the circuit alone were identified as inelastic scattering events involving the external environment. This classification allowed them to pinpoint the origin of nearly every error observed in the experiment. They found that while some transitions were caused by the unpredictable behavior of material defects, the majority of the complex patterns they observed could be predicted by simulating the circuit's interaction with its electromagnetic surroundings.

The implications of this work extend beyond a single experiment. The researchers demonstrated that by understanding the specific mechanisms behind these errors, engineers can make informed choices to avoid them. For instance, they showed that certain drive frequencies are much safer than others, offering "windows" where the circuit can be controlled without triggering these unwanted transitions. They also highlighted that the design of the device package is just as important as the design of the circuit itself, as the metal housing can harbor resonant modes that sabotage the operation. While the microscopic defects in the materials remain a challenge that requires better fabrication techniques, the electromagnetic environment can be engineered and simulated to minimize its impact.

This study provides a comprehensive framework for diagnosing and mitigating the errors that plague superconducting quantum circuits. By separating the causes of failure into distinct, understandable categories, the researchers have offered a clear path forward for improving the fidelity of quantum operations. Their work suggests that the key to building more reliable quantum computers lies not just in making better circuits, but in understanding the complex dance between the circuit, the materials it is built from, and the electromagnetic waves that surround it. Through careful design and the strategic selection of control frequencies, it is possible to navigate around these pitfalls, paving the way for more robust and powerful quantum information processing.