Access Official Technical PDF

Share Summary:LinkedInEmail

Cryogenic quantum processor testing must address more than the behavior of individual qubits. A processor-on-chip can combine quantum structures with analog and mixed-signal interfaces, microwave or millimeter-wave circuits, and digital control. Testing these elements together at low temperature is challenging because measurement access, cryostat power limits, and interactions between the circuitry and qubits all shape what can be validated.

Measurements on 22nm fully depleted silicon-on-insulator (FDSOI) CMOS illustrate why test planning needs to cover devices, circuits, and processor integration. They also show the value of establishing a relationship between measurements at cryogenic temperatures and at 300 K, rather than relying exclusively on slow, resource-intensive low-temperature tests.

Why quantum processors create a system-level test challenge

Semiconductor quantum-dot qubits confine electrons or holes within energy potential wells formed in structures such as silicon, Si/SiGe, FDSOI MOSFETs, and FinFETs. Nearby quantum dots can be coupled through a barrier. In a charge qubit, the electron’s position across a double quantum dot provides the relevant degree of freedom; gate voltages control tunnel coupling, while an applied voltage can create energy detuning between the dots.

Other approaches use hole-spin qubits. In either case, the quantum structure is only one part of the processor. Control and readout electronics must operate alongside it, and the test setup must make it possible to assess device behavior without overlooking the influence of the surrounding circuitry.

This integration complicates measurement. A cryogenic package may include a chip or daughterboard assembly, cabling that must be thermally managed, and circuitry whose impact on the qubit can be difficult to separate from the qubit’s own response. On-die probing offers another route, but the padframe, probe complexity, and available cryostat capacity constrain what can be measured.

Key requirements for cryogenic quantum processor testing

A useful test strategy has to accommodate several kinds of devices and operating conditions. The measurement environment described for this work covered temperatures from 2 K to 300 K, DC through 67 GHz and selected millimeter-wave bands of 140–220 GHz, magnetic fields below 2.5 T, and DC power below 30 mW. These limits underline the need to prioritize measurements and design test structures that make efficient use of cryostat time and power.

Test planning should include common transistors and passive components as well as single-electron transistors, double quantum dots, circuits, and processor-level padframes. Characterizing only a transistor cannot establish how a complete quantum processor behaves; conversely, processor measurements are difficult to interpret without device-level data.

For quantum-dot arrays, relevant measurements include the position and shape of current peaks, the effects of temperature, and how operating points respond to bias. For control and readout blocks, engineers also need to understand the cryogenic behavior of transistors, transimpedance amplifiers, resistors, capacitors, transmission lines, inductors, and transformers.

Device behavior and process variation at low temperature

Measurements of minimum-size MOSFETs at 2 K showed quantum behavior under low drain-source voltage. In an 18 nm by 70 nm p-MOSFET, current peaks were observed at temperatures beyond 20 K. The reported peaks were separated by approximately 50–60 mV and reached currents of 5–10 pA; they flattened as temperature increased. The work identifies biasing at the first valley as a condition relevant to qubit manipulation.

Device characteristics are sensitive to manufacturing variation. Changes in channel width and length, silicon or silicon-germanium thickness, oxide thickness, surface roughness, and spacer or source/drain-to-channel barriers can affect energy-level spacing and operating parameters such as Rabi frequency. Threshold voltage and other qubit-related quantities may therefore vary between devices or dies.

Back-gate adjustment can correct or tune some of these variations. Measurements also indicated that the gate voltage of a peak could be tuned using the back gate, and that the first MOSFET and cascode peaks occurred at similar gate voltages in the reported comparison. This makes back-gate control an important consideration in both device characterization and circuit design, while not eliminating the need to account for process spread.

What 22nm FDSOI circuit measurements indicate

The measurements found that 22nm FDSOI active devices and passive components operated across the tested cryogenic range, with several circuit characteristics improving at 2–3 K compared with 300 K. Transistor and transimpedance-amplifier performance improved, while transmission-line, inductor, and transformer losses also improved. Poly resistors and metal-oxide-metal capacitors varied by less than 5% in the reported results.

Low temperature also changes transistor operating points. Threshold voltage increased, but could be readjusted through the back gate. The undoped channel was reported to avoid a kink effect. These findings matter for integrated control and readout: cryogenic circuit design must account for shifts in device bias as well as the performance of passive elements, rather than assuming room-temperature conditions carry over unchanged.

Building an efficient validation strategy

Cryogenic measurement is slow and expensive, so a production-oriented program needs to connect low-temperature test results with more accessible measurements. The practical goal is to establish correlation between 4 K and 300 K testing. Once characterized, that relationship can help reduce the amount of repeated cryogenic testing required, while preserving low-temperature validation for behaviors that cannot be inferred reliably from room-temperature data.

For engineering teams, a structured approach can include:

  • Define test coverage across the stack: Include qubit structures, transistors, passive components, control and readout circuits, and processor interfaces.
  • Plan around measurement limits: Account for temperature, frequency, magnetic-field, DC-power, probe, and packaging constraints before finalizing test structures.
  • Track tunable operating points: Measure how back-gate bias affects device peaks and other characteristics that may shift with process variation.
  • Correlate temperature conditions: Compare cryogenic and room-temperature measurements to identify which tests can support a more efficient production flow.

Conclusion

Quantum processor validation is a system-level problem: quantum dots, control electronics, readout circuits, and the test environment all influence what engineers can observe. The 22nm FDSOI results show that active and passive components can function at cryogenic temperatures, while also highlighting temperature-dependent behavior and process variation that require careful characterization. By planning measurements around cryostat constraints and building a dependable link between 4 K and 300 K results, teams can make cryogenic testing more targeted without losing sight of the processor’s integrated behavior.

Access Official Technical PDF

About The Author