Hot Spin Qubits Pushing for Fault-Tolerance
Cooling is one of the most cited critical issues for quantum technologies in general. In fact, especially when talking about scaling up to useful quantum computers possibly made of tens of thousands or millions of qubits, I do wonder how much cryogenics systems would need to scale to support such advancements. Just sending and receiving control signals from and to the chip would generate a very high heat load to counterbalance.
Some people work towards technologies that don’t rely as much on cryogenics, such as photonics or atom/ion-based systems. Others work to make cryogenics ready to support tens of thousands of qubits at millikelvin temperatures. A few try alternative routes to increase the operating temperature, a very challenging task indeed for delicate systems such as qubits.
The Hot Topic of Hot Qubits
This is the reason why, at Quantum Machines, we are so excited to see our customers’ results reported in a Nature article by Huang et al., 627, 772–777 (2024). The team sees prominent names in the semiconductor spin community, such as Prof. Andrea Morello, Prof. Andrew Dzurak, and Prof. Natalia Ares, whose teams joint work produced some incredible results: operation of spin qubits above one Kelvin showing fidelities approaching the fault-tolerance threshold. This is clear proof that (1) thermal energies don’t necessarily need to be well below qubit energies, determined by the external magnetic field, and (2) spin qubits might just be able to offer fault-tolerant quantum computation above 1K, making cooling power requirements much more reasonable. This is impressive news for a platform with already a lot going on, from the long coherence to the industrial compatibility of mass manufacturing, which promises a bright future for scaling.
To some 1 Kelvin might seem an unimpressive change from e.g. 10mK. But mind you, there is a huge difference between cooling to a few Kelvins and cooling to a few millikelvins, both in terms of cryogenic technologies used and heat load to be handled. A closed cycle Helium cryostat can get to 1K and still cost almost an order of magnitude less than a dilution refrigerator while providing an order of magnitude more cooling power at those higher temperatures. In the low-temperature physics communities anything above 0.5 K is considered Hot.
Novel Initialization Technique for High-Temperature Spins
In their work, Huang et al. utilize a gate-defined double quantum dot system, using two electron spins as qubits and a way to control both the qubits individually (for single-qubit gates) and their interaction (for two-qubit gates). As shown in the figure below the OPX control platform is used to generate the baseband control for the qubits, the I and Q quadrature signals for the microwave control, and to drive and readout the RF-SET sensor. Nowadays, the QDAC-II and the new QDAC-II-compact offer ultra-stable DC voltages that seamlessly integrate with the OPX, making the controls for such experimental setups straightforward.
Figure – Full experimental setup, taken from [Huang et al. Nature 627, 772–777 (2024)] Extended Data Fig. 1.
One of the great novelties shown in this work is an algorithmic initialization protocol that works even when the energy of the qubit is comparable to (or less than) the thermal energy of its environment. The authors show fidelities exceeding 99.5% at a temperature of 1K, a truly remarkable result. The protocol works in purification steps that bring the spins from a mixture of |↓↓>, |↑↓>, |↓↑>, and |↑↑> (stage 1) to a mixture of |↑↑> and |↓↓> with odd-parity states filtered out (stage 2). Then, a zero–controlled NOT gate converts |↑↑> into |↑↓> (stage 3), which is again filtered out, ensuring initialization in the |↓↓> state. It is worth noting that implementing such a comprehensive control flow is unique to OPX and intuitively programmable using QUA.
Figure – The Diraq team next to the experimental setup used for their work published in Nature [Huang et al. Nature 627, 772–777 (2024)].
They evaluate the temperature dependence of the system in detail, shedding some light on the underlying physics that tends to kill qubits at higher temperatures and still shows high fidelities well beyond 1K. The paper offers insights on the temperature dependence of initialization, readout, single-qubit (Clifford Fidelity, T1, and T2), and two-qubit performances (2-qubit Randomized Benchmarking and more), making it truly a well-rounded report and proof of concept to high-fidelity operation of semiconductor spin qubits at temperatures exceeding what was thought possible given the qubit inherent energy levels.
The Race is Not Over
If the SiMOS spin qubit platform was a great contender for fault-tolerant quantum computation, demonstrating algorithmic initialization and high-fidelity operation at temperatures above 1K might be a game changer. This technological platform becomes more affordable and less complex to use, offering a much sharper view of how the scaled-up system might look on the way to fault-tolerance.
As the authors themselves comment, all the techniques employed in this work can be efficiently implemented in FPGA, allowing advanced calibrations done in real-time (this is not the first time we have discussed embedded calibrations when talking about the OPX). We are ecstatic to see the OPX platform once again used as enabling technology for such breakthroughs, and we congratulate the team for this incredible work.
For all of you inclined towards superconducting and atomic systems for scaling up and fault-tolerant machines, the spin community says the race is not over.
