Longer coherence times and lower error rates for qubits will come not only from improved materials, but also through low-temperature operation, requiring dilution refrigeration for continuous use. Yet low-temperature operation is a more challenging problem than many realize: a colder refrigerator has lower cooling capacity, supporting fewer wires, optical fibers and control pulses, all of which contribute to the total thermal load.
We are determining the cooling capacity needed per qubit, accounting for the thermal cost of the interface and control, which will determine the number of qubits supported by dilution refrigeration technology and the achievable temperatures. Specifically, we are considering phosphorous donor electron spin qubits on enriched 28Si substrates by including the effect of finite temperature. By relating the qubit decoherence rate to the temperatures predicted by the qubit interface analysis, this project will determine precisely how the number of qubits trades off with temperature-dependent decoherence.