Quantum Computing RF

Cryogen-Free System

Q: How cold can a cryogen-free dilution refrigerator get without liquid helium?

A modern cryogen-free dilution refrigerator reaches a base temperature of 7 to 10 mK at the mixing chamber, with continuous cooling power of roughly 12 to 20 microwatts at 20 mK and several hundred microwatts at 100 mK. The mechanical pulse tube cryocooler provides the first two stages at approximately 45 K and 4 K, replacing the liquid nitrogen and liquid helium baths of a wet system. From there a Joule-Thomson stage and the helium-3/helium-4 dilution unit carry the platform down to the millikelvin range. Cooldown from room temperature to base typically takes 18 to 36 hours, longer than a precooled wet system but fully automated and unattended.

Q: How much vibration does the pulse tube add and how does it affect RF measurements?

The pulse tube cryocooler runs at roughly 1.4 Hz and injects mechanical vibration plus its harmonics into the cold stages, with displacement amplitudes of tens of micrometers at the second stage if uncompensated. This couples into microwave cables and SQUID or qubit measurements as phase noise and microphonic sidebands. Mitigations include flexible copper thermal links, bellows or edge-welded vibration isolators between the pulse tube and the mixing chamber plate, remote motor heads, and gating the data acquisition to the quiet part of the pulse tube cycle. Well-isolated systems achieve residual vibration below 1 micrometer RMS at the sample stage.

Q: Why do quantum computing labs prefer cryogen-free systems over wet cryostats?

Wet systems consume liquid helium continuously, often 5 to 20 liters per day for a dilution refrigerator, which is costly, supply-constrained, and requires skilled cryogenic handling and transfers. A cryogen-free system recirculates a sealed charge of helium and only needs electrical power and cooling water for the compressor, so it runs for months unattended. This makes it the standard cold platform for superconducting qubit processors, where dozens of coax lines, attenuators, isolators, and cryogenic low-noise amplifiers must stay at fixed thermal anchors for long automated runs. The trade-offs are higher capital cost, added pulse tube vibration, and a longer initial cooldown.

/KRY-oh-jen free SIS-tuhm/

Closed-cycle refrigeration reaches deep cryogenic temperatures without any liquid helium or nitrogen bath, relying instead on a mechanical pulse tube cryocooler to precool the system to roughly 4 K and a recirculated helium-3/helium-4 dilution refrigerator to carry the mixing chamber down to 7 to 10 mK. Because the helium charge is sealed and recirculated, the platform runs for months on electrical power and cooling water alone, with no daily cryogen transfers. That stability makes the cryogen-free dewar the standard host for superconducting qubit processors and chains of attenuators, isolators, and cryogenic LNAs distributed across stages from 45 K to base.

Category: Quantum Computing RF

Base Temperature: 7 to 10 mK

Cooling Power: ~14 μW at 20 mK

Closed-Cycle Cooling from 300 K to Millikelvin

A cryogen-free system replaces the consumable liquid baths of a traditional wet cryostat with a sealed, mechanically driven cold chain. A two-stage Gifford-McMahon or pulse tube cooler, fed by a room-temperature helium compressor, establishes intercept stages near 45 K and 4 K. The 45 K plate intercepts conducted and radiated heat, while the 4 K plate condenses the dilution unit's helium-3/helium-4 mixture. Below 4 K, a Joule-Thomson or still stage and the dilution circuit take over, exploiting the finite solubility of helium-3 in superfluid helium-4 to produce continuous cooling into the millikelvin band. Nothing is vented and nothing is refilled; the same charge of helium cycles indefinitely.

The thermal architecture is hierarchical because cooling power shrinks dramatically as temperature falls. Tens of watts are available at the 45 K stage, a few watts at 4 K, but only microwatts at the mixing chamber. RF wiring exploits this by thermally anchoring (heat sinking) coaxial lines, attenuators, and DC bias at each successive plate so that very little heat reaches the coldest stage. Stainless steel or NbTi coax is used on the input lines to limit conducted heat load, and cryogenic isolators protect qubits from amplifier back-action while preserving the cold thermal budget.

The dominant penalty of going cryogen-free is mechanical vibration. The pulse tube cycles at roughly 1.4 Hz, and its harmonics couple into the cold mass as displacement and microphonic noise. Designers fight this with flexible thermal links, bellows isolation between the cooler and the experimental plates, remote rotary-valve motors, and measurement gating synchronized to the quiet portion of the cycle. A well-built platform holds residual sample-stage motion below 1 μm RMS, low enough that superconducting qubit coherence and SQUID readout are not degraded.

Governing Thermodynamics

Dilution Unit Cooling Power:
Q̇_mc ≈ 84 × ṅ₃ × (T_mc² − T_in²) watts

Stage Heat Load Budget:
Q̇_stage = Q̇_conduction + Q̇_radiation + Q̇_RF < Q̇_available(T)

Carnot Efficiency Limit:
COP_Carnot = T_cold / (T_hot − T_cold)

Where ṅ₃ = helium-3 molar circulation rate (mol/s), T_mc = mixing chamber temperature (K), T_in = inlet temperature to the mixing chamber (K). Example: ṅ₃ ≈ 400 μmol/s with T_mc = 0.02 K and T_in ≈ 0.012 K yields roughly 14 μW of cooling at base.

Cryogen-Free vs. Wet Cryostat

Attribute	Cryogen-Free (Dry)	Wet (Bath) Dilution	Pulse Tube Only
Base temperature	7 to 10 mK	5 to 8 mK	2.8 to 4 K
Liquid helium use	None (sealed charge)	5 to 20 L/day	None
Cooling power at 20 mK	12 to 20 μW	10 to 25 μW	n/a
Vibration at sample	< 1 μm RMS (isolated)	Very low	Tens of μm if uncompensated
Cooldown to base	18 to 36 hours	~8 hours (precooled)	1 to 3 hours
Unattended runtime	Months	Limited by refills	Months
Typical use	Qubit processors, cryo LNAs	Legacy / ultra-low vibration	4 K detectors, precooling

Common Questions

Frequently Asked Questions

How cold can a cryogen-free dilution refrigerator get without liquid helium?

Modern dry dilution units reach a 7 to 10 mK base at the mixing chamber, with about 12 to 20 μW of continuous cooling at 20 mK and several hundred μW at 100 mK. A pulse tube cooler supplies the 45 K and 4 K stages that a wet system would get from liquid nitrogen and liquid helium, then a Joule-Thomson stage and the helium-3/helium-4 circuit reach base. Cooldown from 300 K takes 18 to 36 hours, fully automated and unattended.

How much vibration does the pulse tube add and how does it affect RF measurements?

The pulse tube runs near 1.4 Hz and injects vibration plus harmonics, with tens of micrometers of motion at the second stage if uncompensated, appearing as phase noise and microphonic sidebands in qubit or SQUID readout. Mitigations include flexible copper links, bellows isolators between the cooler and the mixing chamber plate, remote motor heads, and gating acquisition to the quiet part of the cycle. Good systems hold residual motion below 1 μm RMS.

Why do quantum computing labs prefer cryogen-free systems over wet cryostats?

Wet dilution refrigerators burn 5 to 20 L of liquid helium per day, which is costly, supply-limited, and needs skilled transfers. A cryogen-free system recirculates a sealed helium charge and only needs power and cooling water, so it runs for months unattended. That suits superconducting qubit processors where dozens of coax lines, attenuators, isolators, and cryogenic LNAs sit at fixed thermal anchors. The costs are higher capital price, added pulse tube vibration, and a longer first cooldown.

Related Terms

← Cryocooler Cryogenic Amplifier Performance →