Back

@roomey it's really a technology-dependent question. Superconducting and quantum dot systems need their actual chips cooled to millikelvin temps, AND some control circuits cold-but-not-that-cold temps. Optical systems using photons as qubits, however, are fine at room temp, but some single-photon detectors need to be very cold.

@roomey ion traps are kind of complicated. The outside of the vacuum chamber is room temperature, but the ions themselves are individually near ground state (lowest possible) energy, so what temperature are they at? And some implementations are cooling a few parts of the system to cryogenic temps.

@roomey and then NV diamond technically can support a coherent quantum state at room temp, or near room temp, but when you need to read it out or get photons out they get lost in the other noise unless the device is cooled.

@roomey so really "Is this a cryogenic system?" is less the right question than, "How much cooling do you use, and what do you use it for?" And "how much" means a lot of things -- dollars, watts, exotic materials (He3, for example), how much expertise it takes to implement and operate.
As systems engineers build better packages, more self-contained and reliable, this question will matter less.

@rdviii Interesting about the NV diamond needing low temp to read.

My thesis is that heat will need to be removed to avoid premature "measurements" ie. collapse, and removing heat required large infrastructure, so that QPUs will tend to be expensive and large.

Sounds like I am not 100% correct here for some forms of qubit.

Either way, thank you for the valuable feedback!

@roomey @rdviii

Qubits have two states with different energies. Higher temperature means larger energy fluctuations, so the relevant comparison is between (1) the energy difference between the two states and (2) the energy scale of the qubit's thermal environment. If (1) < (2) or (1) ~ (2), then the environment will be constantly screwing up your qubit; if (1) >> (2), then your qubit will be fine. That's the most basic and fundamental reason that some kinds of qubits require cryo temperatures.

@roomey @rdviii Then as Rod nicely lays out, there are lots of other more detailed technical considerations. Researchers can (and do) cool atoms and ions down to extremely low temperatures with lasers, but from an experimental infrastructure POV this isn't anything like cooling a chunk of material down in a fridge as in the superconducting/silicon cases.

@benbenbrubaker @rdviii Would I be correct in thinking that even with laser cooled atoms the detector/measurement widget would need to also have a cryo temp?

@roomey @benbenbrubaker there are several technologies for detecting single photons. Photomultiplier tubes operate at hot temps, but are noisy and low efficiency. The trndy thing these days is SNSPDs, superconducting nanowire single photon detectors. They are kept right near the temperature where the transition to/from superconducting. They are incredibly efficient, and not cheap.

@roomey @benbenbrubaker Ben's correct that most atom/solid-state implementations have slightly different energy levels, and that tiny difference is related to the amount of energy it takes to corrupt a qubit.

@roomey @benbenbrubaker When you're working with photons, though, the energy levels are generally much, much higher, so that's not the reason the detectors need to be low temp.

@roomey @benbenbrubaker (In fact, the two photonic states for |0> and |1> are often the same energy, being the prominent exception to Ben's earlier statement about energy levels.)

@rdviii @roomey

All very important qualifications! I was just trying to give high-level context to Roomey's intuition about "higher temperature --> more noise" This is absolutely true in a general sense. But what counts as "high temperature" is very different for different systems. It's like asking "is a meter a long or a short distance?" Depends on what you're talking about!

@roomey @benbenbrubaker it's a critical question, and there are some basic principles worth knowing, but the details matter, and what will drive that in the end will be how good the engineers get at packaging any given technology.
Alpine Quantum Tech (Rainer Blatt's company) claims to be able to deliver 19" rack mount ion traps.
For superconducting, improvements in cooling tech are in a symbiotic balance with the race to build bigger systems.