Why do flux qubits have to be micrometer-sized?

Flux qubits are made using micrometer sized Josephson junctions. They exploit superconducting properties to create and interfere with the magnetic flux between them.

My question is that I've seen superconducting quantum interference devices that use JJs that are a few inches big, so why do flux qubits have to be ludicrously small? Do they have to be small, or is that a question outside of physics altogether, more so for a question of mass production instead?

The reason behind this is that I'm looking to create a pulse tube based flux qubit machine with Pb as my material for the JJs. Since the prototype SQUIDs are rather big, I figured I could build them too, without expensive equipment. Considering a pulse tube refrigerator can be made with copper, PVC pipes, and reciprocating saws, I'm sure cooling wouldn't be an issue. The noise, however...is another question. I realize this is extremely ambitious, but I appreciate any help you can give me.

• First guess: I don't see the merit in a big flux qubit. We need to combine and perform multiple particle gates on several qubits. The problem in experimental quantum computing are not so much control of one qubit but scaling up to multiple qubits. For all of this, I'd say that the qubits should be small. Whether this is the reason that they are, I don't know. – Martin Apr 22 '15 at 15:36
• The merit is being able to build one at home :) I don't want to perform ultra-amazing calculations, rather, I'd like to experiment with two, four, maybe eight qubits. The people who currently own a D-Wave (which this is based heavily on) don't take kindly to this kind of experimentation :) – OrangeCalx01 Apr 22 '15 at 15:40
• When you say a pulse tube fridge can be made with copper, PVC pipes, and reciprocating saws, I think you may be missing the important fact that superconducting qubits must be operated a temperatures below 1 Kelvin. I don't think there is any known way to do that without either an adiabatic demagnetization system or a helium dilution refrigerator. – DanielSank Apr 22 '15 at 16:12
• I'm pretty sure this is going to come down to getting the junction energy scale to match the $\hbar \omega$ energy scale where $\omega$ is constrained by the thermal environment: $\omega \gg k_b T / \hbar$. I have to really go through the numbers to give a good answer though. Will do later today, I hope. – DanielSank Apr 22 '15 at 16:49
• The superconducting temperature isn't the issue. The problem is that for a circuit to behave quantum-mechanically, the resonance frequency of the circuit must satisfy $\omega \gg k_b T$. For 11 Kelvin that's a frequency of 200 GHz! That frequency range is essentially impossible to work with because of a complete lack of any commercial hardware. Skin loss in metals is very high, parasitic capacitance/inductance will screw up everything you try to build, etc. For this reason superconducting qubits work in the ~5 GHz range, and for that you need milli-Kelvin temperatures. – DanielSank Apr 22 '15 at 18:54