Route to high-precision positioning of single ion-implanted impurities in silicon

唯一的量子技术（QT）制造技术可以很容易地利用微电子制造工艺和现有的大规模扩展能力，实现足够大的量子比特阵列进行纠错，或者可以潜在地重复制造大量相同的设备，这是通过植入结合单个杂质量子比特。然而，除非完全确定的单离子注入（ISI）被开发出来，否则基于杂质的QT的优势将无法实现。基于离子阱、超导电路和使用少量量子比特的半导体量子点的量子计算是非常先进的，但非常大规模的复制对它们来说都是一个主要挑战。硅中的少量杂质量子位也可以使用基于扫描探针技术的氢光刻技术以高质量制造出来，这种技术已经实现了原子尺度的精度，导致了单原子晶体管等突破性成就(然而，它很慢，并且不能提供一个容易扩展的途径，以达到可制造量子计算机所需的数百万个量子位。在硅中植入单个杂质量子比特原子提供了一种解决方案，但该领域的大多数研究都集中在随机加入杂质的样品上，并且通过掩模或聚焦光束对放置进行了一些有限的控制。因此，与离子阱等相比，这里的挑战正好相反——大规模重复很容易，但每个量子位的定位（以及随之而来的错误率）较差，必须加以改进。放置精度受到注入离子聚焦和离子进入目标材料后的运动的限制，称为冲击散列。注入也会对晶体宿主造成意想不到的损害，因为高能离子会通过晶体中的通道反弹。这是我们试图解决的挑战，我们使用一种推测性的想法，不仅可以修复这种冲击损伤云，而且最重要的是，可以对植入的杂质进行更高精度的定位。我们提出了一个基于横向固相外延再生（L-SPER）的解决方案。简而言之，通过聚焦离子束或广域光刻和离子注入，对目标区域进行预非晶化（将硅离子注入硅中，断开键，但不引入杂质，甚至可以提高同位素纯度）。在注入单个离子后，低温退火通过外延再生恢复晶体，外延再生由周围的晶体材料播种。与完全由植入过程引起的部分非晶化相比，完全预非晶化在退火后会导致更高的结晶度。这个提议的本质是考虑L-SPER对单个植入原子的影响。我们有充分的理由认为，在再生长过程中，当非晶化区域缩小时，杂质原子在晶体重组时被缓慢地推向中心。如果我们能够证明这一点，那么原子最终位置的精度可能会受到预非晶化区域的中心位置的更强烈的影响，而不是受到注入离子聚焦不确定性和散乱的限制，其中前者可以达到纳米的数量级，从而在最终位置上得到数量级的提高。

The only quantum technology (QT) fabrication technology that can readily leverage microelectronic fabrication processes with the existing ability of large scale-up, enabling big enough qubit arrays for error correction, or that can potentially repeatably manufacture large numbers of identical devices, is the incorporation of single impurity qubits through implantation. However, unless fully deterministic implantation of single ions (ISI) is developed, the advantages of impurity-based QT for scale-up will not be realized. Quantum computing based on ion traps, superconducting circuits and semiconductor quantum dots using a small number of qubits are well advanced, but very large-scale reproduction constitutes a major challenge for each. Small numbers of impurity qubits in silicon can also be made with high quality using hydrogen lithography, which is based on scanning probe techniques, that have enabled atomic-scale precision leading to such ground-breaking achievements as the single-atom transistor (However, it is slow and does not provide an easily scalable route to the millions of qubits needed for manufacturable quantum computers. Implantation in silicon of single impurity qubit atoms offers a solution, but most of the research in this area centres on samples with stochastic incorporation of impurities with some limited control over the placement through masks or with focussed beams. The challenge here is therefore the opposite compared with ion traps etc - large scale repetition is easy, but the positioning (and consequent error rate) of each qubit is poorer and must be improved. The placement precision is limited by the focusing of the implanted ion and the movement of the ion after it enters the target material, known as the impact straggle. Implantation also causes undesirable damage to the crystal host, as the energetic ion ricochets through channels in the crystal. This is the challenge we seek to address, using a speculative idea that will not only repair this impact damage cloud but also, and most importantly, allow much higher precision positioning of the implanted impurity. We propose a solution based on lateral solid phase epitaxial regrowth (L-SPER). Simply put, the target area is pre-amorphised (implanting silicon ions into silicon breaks bonds but does not introduce impurities and can even improve isotopic purity) by a focussed ion beam or through broad area lithography and ion implantation. Following implantation of a single ion, a low-temperature anneal restores the crystal through epitaxial regrowth, which is seeded by the surrounding crystalline material. Full pre-amorphisation is well known to result in higher crystallinity following annealing, compared to the partial amorphisation caused solely by the implantation process. The nature of this proposal is to consider what effect L-SPER has on an individual implanted atom. There is every reason to expect that, as the amorphised region shrinks during regrowth, the impurity atom is slowly pushed to the centre as the crystal reforms. If we can demonstrate this, then the precision of the final placement of the atom may be affected more strongly by the central positioning of the pre-amorphised regions rather than limited by the focusing uncertainty and straggle of the implanted ion, where the former can be of the order of a nanometer giving an order of magnitude improvement in the final positioning.