Cryogenic Chip Calibrates Microwave Signals in Milliseconds

Researchers have developed a novel method for the accurate calibration of microwave attenuation and gain, crucial for the performance of sensitive superconducting circuits. Thomas Descamps and Linus Andersson, leading the work at Chalmers University of Technology alongside colleagues Vittorio Buccheri, Simon Sundelin, Mohammed Ali Aamir, and Simone Gasparinetti, have demonstrated a compact, self-calibrating cryogenic noise source. This device, integrating an on-chip chromium attenuator directly into a coaxial microwave line, allows for in situ determination of attenuation and gain without prior knowledge of the attenuator temperature. The significance of this research lies in its ability to provide a simple and accurate characterisation technique for near quantum-limited parametric amplifiers, vital components in superconducting-qubit readout systems.

Until recently, verifying the delicate balance of signals in quantum processors proved exceptionally difficult. Now, a miniature, self-checking device allows precise tuning of the components that read information from these systems, promising more reliable and powerful quantum computers. Scientists developing advanced superconducting quantum circuits require precise calibration of microwave signals at extremely low temperatures.

Accurate measurement of both attenuation and amplification chain noise is essential for interpreting experimental results and characterising amplifier performance. Scientists have created a compact, self-calibrating cryogenic noise source integrated directly into the microwave line at the mixing-chamber stage of a dilution refrigerator. This new device utilizes an on-chip chromium attenuator. This can be heated with remarkably low power levels, on the order of nanowatts.

By comparing the Johnson-Nyquist noise generated through both direct current (Joule) heating and microwave power dissipation within the attenuator. The attenuation of the input microwave line is determined. Crucially, this method does not require prior knowledge of the attenuator’s temperature, simplifying the calibration process — once established, the input line attenuation allows for precise quantification of the gain and added noise of a cryogenic amplification chain. Typically used to boost the weak signals from superconducting qubits.

Here, this offers a simple and accurate method to characterise near-quantum-limited parametric amplifiers, vital components in superconducting-qubit readout. Precise microwave measurements at cryogenic temperatures are fundamental to diverse areas including quantum optics, quantum sensing, and mesoscopic transport experiments, though achieving accurate calibration within these low-temperature environments presents challenges.

Various techniques have been explored, including those leveraging superconducting qubits or bolometers. These approaches often have limitations in bandwidth or introduce experimental complexity. Unlike previous methods, this newly developed noise source is designed for direct integration into existing experimental setups. Minimising potential errors from impedance mismatches and insertion losses.

The system features a T-network attenuator fabricated from a 60nm chromium film deposited on a sapphire substrate. Incorporating niobium leads for excellent conductivity and thermal isolation. In turn, scientists found the device exhibits millisecond-scale response times and generates negligible heat that could affect the overall cryostat temperature. Meanwhile, the attenuation was determined by comparing Johnson-Nyquist noise generated by Joule heating with that generated by microwave power.

Cryogenic microwave gain and attenuation calibration using an integrated noise source

Precise knowledge of microwave line attenuation and amplification chain gain is essential for accurate data interpretation when performing measurements at cryogenic temperatures. At the same time, this effort details a compact, self-calibrating cryogenic noise source integrated directly into a coaxial microwave line at the mixing-chamber stage. By comparing Johnson-Nyquist noise generated through Joule heating and microwave dissipation within an on-chip chromium attenuator, the attenuation of the input line. Therefore the gain of the amplification chain, is determined without needing to know the attenuator’s temperature.

Initial tests revealed millisecond-scale response times for the device, with negligible heating observed on the cryostat base plate. At the core of this project is the determination of gain and added noise within a cryogenic amplification chain across the 4-8GHz band. Here, the on-chip attenuator, designed with resistors of 26Ω and 70Ω, demonstrated a clear relationship between applied power and temperature.

Specifically, the temperature of the 5.5GHz microwave field followed a power law relationship with the Joule power applied to the attenuator, allowing for precise calibration. Thermal response times were also characterised, showing that the attenuator thermalizes after a microwave heating pulse with a time constant of 680 microseconds. In turn, the effort quantified the attenuation of the input line by comparing noise generated by Joule heating with that from microwave power.

Meanwhile, the difference in noise power spectral density between these two methods directly yields the attenuation value. At the same time, the project team successfully demonstrated the ability to characterise near quantum-limited parametric amplifiers, vital components in superconducting-qubit readout systems, providing a simple and accurate method for in situ calibration.

In-situ cryogenic noise calibration utilising on-chip attenuation

This effort is underpinned by a compact, self-calibrating cryogenic noise source constructed around an on-chip chromium attenuator directly integrated into a coaxial microwave line at the mixing-chamber stage. Once fabricated, the attenuator was heated with nanowatt-level power, allowing for in situ calibration of microwave attenuation and amplification-chain noise.

By comparing Johnson-Nyquist noise arising from both Joule heating and microwave dissipation within the attenuator, the attenuation of the input line and, as a result, the gain of the amplification chain were determined without needing to know the attenuator’s temperature. This method circumvents issues associated with temperature dependence and potential errors from impedance mismatches.

The experimental setup involved careful consideration of signal routing and filtering. A vector network analyzer (VNA) was used for microwave measurements, connected through bandpass and low-pass filters to isolate the signal and minimise unwanted noise. A bias tee supplied direct current to the attenuator. Meanwhile, an infrared blocking filter prevented stray infrared radiation from affecting the sensitive cryogenic components.

Amplifiers operating within the 4-8GHz band further boosted the signal before it reached a spectrum analyzer for power spectral density measurements. The attenuator comprised two resistors, with values of 26Ω and 70Ω, arranged to create a controlled attenuation. The project directly measured the generated noise power using a room-temperature spectrum analyzer, a departure from previous RF reflectometry techniques.

The placement of this noise source upstream of the device under test minimised experimental complexity, unlike bolometer-based calibration methods. The choice of using an on-chip attenuator was deliberate, exhibiting millisecond-scale response times and negligible heating of the cryostat base plate, unlike variable-temperature attenuators assembled from commercial parts.

Unlike shot-noise tunnel junctions, the method did not require fitting procedures for temperature extraction and proved less susceptible to electrostatic discharge. By directly integrating the noise source at the mixing-chamber stage, The project team achieved a simple and accurate method for characterising near-quantum-limited parametric amplifiers used in superconducting-qubit readout.

Integrated noise source improves qubit measurement fidelity

Scientists building the quantum computers of tomorrow face a persistent challenge: accurately measuring the faint signals from individual qubits. For years, calibrating the extremely sensitive microwave components needed to read these signals has relied on methods prone to error and requiring painstaking temperature control. Now, a team has devised a self-calibrating noise source integrated directly onto the cryogenic chip. Offering a simpler and more precise approach to characterising these delicate circuits.

Instead of attempting to measure absolute temperatures, the device cleverly compares different types of noise to determine signal loss and amplification gain. The significance of this effort extends beyond mere technical refinement; accurate calibration is not simply about achieving higher numbers. But about reducing uncertainty in qubit measurements. Once error rates fall below critical thresholds, the potential for meaningful computation increases dramatically.

By streamlining the calibration process and improving its accuracy, this development brings practical quantum devices a step closer to reality. Limitations remain, as the current setup is tailored to a specific frequency range and requires integration with existing cryogenic infrastructure. The real power of this approach lies in its potential for automation.

Currently, calibration is a manual, time-consuming process. A fully integrated, self-calibrating system could continuously monitor and adjust the amplification chain, compensating for drifts and fluctuations. This technique doesn’t depend on knowing the precise temperature of the attenuator itself, a simplification that reduces complexity and potential error.

Beyond the immediate application to superconducting qubits, the principles behind this noise source could be adapted to calibrate sensitive detectors in other fields. Such as astronomy and materials science. The field has sought ways to move calibration from an external process to an intrinsic one — this development represents a genuine advance, not because it achieves record-breaking performance. But because it addresses a fundamental bottleneck in quantum computing.

Further research could focus on broadening the frequency range and miniaturising the device. We might anticipate seeing these self-calibrating noise sources become standard components in dilution refrigerators worldwide, paving the way for more reliable and scalable quantum systems.