What Microwave Components Are Needed for Superconducting Quantum Computing Experiments?

A superconducting quantum computing setup is essentially a precision microwave measurement system operating at extremely low temperatures, often around 10 mK.

From room-temperature waveform generators to qubits mounted on the mixing chamber plate, microwave signals must travel through multiple temperature stages inside a dilution refrigerator. The components selected at each stage can directly affect signal integrity, thermal noise, qubit coherence, and readout fidelity.

This article walks through the complete microwave signal chain used in superconducting quantum computing experiments, from room temperature down to the millikelvin stage.

1. Two Fundamentally Different Signal Chains

A common misconception is that a superconducting quantum system has one bidirectional microwave path. In practice, the control line and the readout line are two very different signal chains.

Control Line

Room temperature → 10 mK

Typical signal path:

AWG → IQ Mixer → Attenuator Chain → Cryogenic Filters → Qubit

The main purpose of the control line is to deliver clean, precisely shaped microwave pulses to the qubit while suppressing thermal noise.

RF attenuators are therefore distributed across several temperature stages to progressively thermalize the signal.

Readout Line

10 mK → Room temperature

Typical signal path:

Readout Resonator → Circulator → JPA/TWPA → Circulator → HEMT Amplifier → Room-Temperature Amplifier → IQ Mixer → Digitizer

The readout line has a very different objective: amplify extremely weak microwave signals while adding as little noise as possible.

The first amplifier in the chain is especially critical because its noise performance strongly affects the total system noise, just as in a conventional RF receiver governed by the Friis noise formula.

Circulators and isolators are also essential because they prevent amplifier noise from propagating backward toward the qubit.

2. Room-Temperature Microwave Electronics

2.1 Arbitrary Waveform Generator

The AWG generates the baseband I/Q waveforms used to control qubit states.

A single-qubit gate is implemented by applying a microwave pulse with a precisely defined frequency, amplitude, phase, and duration.

Typical requirements include:

For multi-qubit systems, synchronization accuracy becomes increasingly important because two-qubit gates require tightly coordinated control pulses.

2.2 Microwave Source / Local Oscillator

Superconducting qubits commonly operate in the 4–8 GHz range, and a microwave source provides the LO signal required for upconversion and downconversion.

Phase noise is particularly important because excessive LO phase noise can contribute to qubit dephasing.

Typical design targets include:

• Phase noise of approximately −120 dBc/Hz at 10 kHz offset or better
• For high-performance systems, −130 dBc/Hz at 10 kHz offset or better
• Stable synchronization across multiple LO channels
• Careful frequency planning to prevent unwanted interaction between adjacent qubits and readout resonators

Frequency planning becomes increasingly important as the number of qubits grows.

2.3 IQ Mixer

The IQ mixer converts low-frequency I/Q control signals into microwave-frequency qubit control signals, or performs the reverse process on the readout path.

Three key imperfections must be controlled:

1. I/Q gain imbalance
2. I/Q phase imbalance
3. LO leakage

Poor IQ balance can generate unwanted image sidebands and degrade qubit gate fidelity.

IQ mixers therefore require careful calibration, and recalibration may be necessary when cable lengths, connectors, temperature, or system configuration change.

2.4 Digitizer / ADC

After amplification and downconversion, the qubit readout signal must be digitized for state discrimination.

Typical requirements are:

Fast digitization is particularly important for mid-circuit measurement and real-time feedback applications.

3. Cryogenic Microwave Components

A typical dilution refrigerator includes several temperature stages:

300 K → 50 K → 4 K → Still → Cold Plate → Mixing Chamber

The microwave components installed at these stages must balance RF performance, thermal conductivity, noise suppression, and mechanical reliability.

3.1 Cryogenic Attenuators

Attenuators are among the most important components on the control line.

Their role is not limited to reducing RF power. They also help thermalize the signal, progressively replacing high-temperature noise with noise corresponding to the local cryogenic stage.

A typical control-line attenuation budget may look like this:

Mechanical mounting is also important. Poor thermal contact between an attenuator and the cold plate can significantly reduce its thermalization effectiveness.

3.2 Cryogenic Filters

Cryogenic filters suppress unwanted microwave, harmonic, and infrared radiation entering the qubit environment.

Typical filter types include:

Insertion loss is particularly important on the readout chain because every fraction of a dB before the first amplifier directly reduces system SNR.

3.3 Circulators and Isolators

Circulators and isolators act as one-way RF elements in the readout chain.

A typical configuration is:

Qubit → Readout Resonator → Circulator → JPA/TWPA → Circulator → HEMT

Their main purpose is to prevent amplifier noise and reflected power from propagating back toward the qubit.

This reverse isolation is critical because even small increases in the effective noise temperature seen by the qubit can degrade coherence and readout performance.

4. The Readout Amplification Chain

4.1 JPA and TWPA

At the millikelvin stage, Josephson parametric amplifiers are used to amplify extremely weak readout signals close to the quantum noise limit.

Two common architectures are:

JPA is often suitable for narrowband, smaller-scale systems, while TWPA is attractive for broadband and multiplexed readout architectures.

4.2 HEMT Amplifier

The next amplification stage is usually a cryogenic HEMT amplifier mounted near the 4 K stage.

HEMT amplifiers provide substantial gain while operating at temperatures where the refrigerator has enough cooling capacity to handle their electrical power dissipation.

Typical performance is approximately:

• Gain: 30–40 dB
• Noise temperature: 2–5 K

Because the parametric amplifier is placed first, the HEMT amplifier's noise contribution is significantly reduced by the gain of the first stage.

4.3 Room-Temperature Amplifier

After the signal exits the cryostat, an additional room-temperature LNA may be used.

At this point, the system already has substantial gain, so the noise figure requirement is less demanding. The main requirement is sufficient gain and bandwidth across the qubit readout band.

5. Cryogenic Cables and RF Interconnects

5.1 Microwave Coaxial Cables

Cryogenic microwave cables must provide a difficult combination of:

• Low insertion loss
• Low thermal conductivity
• Good shielding
• Mechanical stability
• Reliable connector performance

A common choice is 0.086-inch semi-rigid coaxial cable with stainless-steel outer conductors and CuNi or BeCu center conductors.

Pure copper cable is generally undesirable between room temperature and the millikelvin stage because its high thermal conductivity would introduce excessive heat into the refrigerator.

5.2 High-Density RF Interconnects

As quantum processors scale, the number of required RF and DC connections grows rapidly.

Even a small five-qubit setup may require approximately 20 microwave and bias connections. At 100 or 1,000 qubits, conventional point-to-point cabling becomes increasingly difficult to manage.

High-density microwave interconnects are therefore becoming a critical enabling technology for scalable superconducting quantum computing systems.

5.3 Connector Quality Matters

Every RF connector introduces some combination of insertion loss, mismatch, and thermal/mechanical risk.

At several gigahertz, a poorly tightened SMA connector can create reflections, ripple in S-parameter response, and instability across the signal path.

Best practices include:

• Minimize the total number of RF interfaces
• Use calibrated torque tools for SMA connectors
• Recheck connectors after thermal cycling
• Keep mating surfaces clean
• Monitor S11/S21 for periodic ripple caused by mismatch

Connector quality becomes especially important inside dilution refrigerators, where access is difficult and a single poor connection may require a complete warm-up and cooldown cycle to correct.

6. Typical Microwave BOM for a Small Quantum Test System

For a small multi-qubit superconducting quantum computing test setup, the microwave BOM may include:

Control Path

• Arbitrary waveform generator
• Microwave LO source
• IQ mixers
• Cryogenic attenuators
• Low-temperature filters
• Semi-rigid coaxial cables
• SMA connectors and adapters

Readout Path

• Cryogenic circulators or isolators
• JPA or TWPA
• 4 K HEMT amplifier
• Room-temperature amplifier
• IQ mixer
• High-speed digitizer
• Cryogenic coaxial cable assemblies

The original reference estimates that a five-qubit experimental system using predominantly imported equipment can easily reach a microwave electronics cost of approximately USD 105,000–275,000, excluding the dilution refrigerator.

7. Key Takeaways for RF Engineers

Superconducting quantum computing may appear highly specialized, but from an RF engineering perspective, many familiar concepts still apply:

• S-parameters
• Impedance matching
• Noise figure
• Gain budget
• Phase noise
• IQ modulation
• Isolation
• Insertion loss
• Cable phase stability
• Connector repeatability

The main difference is that all of these RF fundamentals must now operate under extreme constraints:

extremely low temperature, extremely low noise, and increasingly high interconnect density.

For this reason, microwave components used in quantum systems must be evaluated not only for RF performance, but also for cryogenic compatibility, thermal conductivity, repeatability, mechanical reliability, and low-temperature stability.

For RF and microwave engineers, superconducting quantum computing is therefore not an entirely new discipline. It is RF engineering pushed to one of its most demanding operating environments.