A photonic link for silicon donor-based quantum technologies

硅晶体管是大多数现代电子设备的基本组成部分，如果不能进一步缩小，量子力学就会使其无法操作。这个经典量子门槛实际上提供了一个巨大的机会：如果我们利用量子力学，而不是试图避免它，我们就可以建造一台量子计算机，它可以完成某些计算任务，否则将永远不切实际。药物模拟、大数据优化、线性代数、机器学习等许多基础计算将以指数级的速度加快，因此可以在现实的时间尺度上解决。

Silicon transistors, the essential building block of most modern electronic devices, cannot shrink much further without being rendered inoperable by quantum mechanics. This classical-quantum threshold in fact presents a tremendous opportunity: if we harness quantum mechanics, rather than attempt to avoid it, we could build a quantum computer, which could accomplish certain computational tasks that would otherwise be forever impractical. Numerous fundamental calculations for drug simulations, big-data optimization, linear algebra, machine learning, and more, would become exponentially faster and therefore possible to solve on a realistic timescale. One of the most promising candidate quantum bits (‘qubits’) are made from donor impurities in silicon: the very atomic defects preventing the smallest transistors from working properly. I have shown that these qubits have the longest solid-state lifetimes (>3 hrs) and the best solid-state quantum control properties (>99.9% accuracy) ever demonstrated. These excellent individual qubits also have key commercial advantages: they are atomically identical, and can be fabricated using the same techniques used to build modern silicon transistors. This is important because quantum computers will still need on-chip integrated “classical” computing power to operate effectively. What is urgently lacking is a reliable way to build connections, or ‘couplings’, between these atomic quantum bits — we need to invent something equivalent to a quantum transistor in silicon. My proposed solution to accomplish this is radically new. I plan to link silicon atomic qubits by mediating their interactions via photon qubits. This can be done in a number of ways: one way is to swap the quantum information between the atomic and photonic qubits. This strategy will make use of photonic structures in silicon which can, for example, direct light along predetermined paths within a silicon device. These structures, which are very similar to the ones developed for fibre-optics devices, could reliably link multiple atomic qubits. The designs are simple and the experimental tolerances are large. The biggest risk to my research plan is its urgency — the first research group to demonstrate a quantum transistor in silicon would gain an insurmountable first-mover advantage. The success of this research plan would launch silicon atomic qubits to the frontrunner position in the international race toward a large-scale quantum computer. The resulting revolution in computing power would have an enormous effect on the whole world, in ways we cannot yet predict. Imagine predicting the ubiquity of modern information technology based only on an evaluation of the first, huge, power-hungry, vacuum-tube-based digital computer in 1946. If quantum-information transistors can be developed for silicon, it will pave the way for silicon to revolutionize the information age once again.