基于半导体量子器件的量子信息处理的理论研究

The quantum information technology, as a young and multidisciplinary subject, has prospects of a variety of applications. The problems in this field are twofold: the decoherence and scalability. In this project, we focus on these problems, which arise in controlling semiconductor quantum dot devices, including dynamically corrected gates, gate crosstalk, and their relationship to fault-tolerant quantum computing. The principal investigator proposed the SUPCODE control protocol in 2012 and have made contributions in related areas. Based on existing results, in this project we are going to start with benchmarking, optimizing and extending existing dynamically corrected gates, and will gradually proceed to multi-qubit, multi-electron systems and other spin-based qubit systems. The research will focus on the merit of dynamically corrected gates in realistic noisy environments, the influence of the gate-crosstalk on gate control and its remedies, and the optimization of dynamically corrected gates, including shortening the gate time duration and optimize them for particular types of noise. The research will also focus on the generalization of dynamically corrected gates to multi-qubit and multi-electron system. This project will place a particular emphasis on improving the applicability and effectiveness of dynamically corrected gates in experiments. We expect 5-7 scientific journal papers as part of the outcome of this project.

量子信息技术作为一门新兴的综合性学科有着广泛的应用前景。目前该领域存在两个关键问题：退相干和可扩展性。本项目专注研究涉及操控半导体量子点量子器件的这些问题，包括动力学修正量子门，门串扰，及其和容错量子计算的关系等。申请人于2012年提出了SUPCODE控制理论，并且在相关领域里也积累了一些动力学修正量子门的工作。基于已有成果，本项目将由标定、优化和拓展现有的动力学修正量子门开始，逐步深化推广到多比特、多电子体系和其他基于电子自旋的量子比特系统。重点研究动力学修正量子门在实际噪声体系中的效能；在多电子体系中门串扰对操控的干扰以及解决办法；对动力学修正量子门的优化，包括缩短其运行时间以及对特殊的噪声谱进行针对性的优化；以及将动力学修正量子门在多量子比特、多电子体系中的推广。本项目尤其注重提高动力学修正量子门在实验上的可行性和有效性。成果预计将在国内外核心期刊发表论文5～7篇。

在该项目中，我们研究了动力学修正量子门和环境噪声的相互关系，并且引入机器学习中的监督学习方法找到了一种测量噪声谱的新方法。我们进一步研究了双量子点体系中势垒操控和失谐操控的区别，使用随机标定方法发现了现有实验的不足，并提出一套新的方法来进行抗噪操作。我们对动力学修正量子门技术进行了进一步优化和推广，研究了在不同的状况下量子门的最简分解，以及一维自旋链上的量子传输。我们还对多量子点体系的能谱以及门串扰进行了仔细的微观计算。最后，我们将机器学习中的强化学习方法应用到量子控制和量子测量中，取得了一些新的结果。本项目共发表论文17篇，包括npj Quantum Information 2篇，Physical Review Letters 2篇，Physical Review A 10篇，Physical Review B 1篇，《中国科学:物理学、力学和天文学》1篇，Advanced Quantum Technologies 1篇。总体来说，本项目圆满完成了研究目标。