Quantum Computing with Cs Atom Qubits

纠缠是量子力学的一个基本特征。例如，如果两个相同的粒子可以分别处于状态A或B，它们可以处于纠缠态AA+BB，这意味着粒子处于同时处于A或同时处于B的叠加状态，但永远不会一个在A，另一个在B。这些高度非经典的状态是量子计算机工作的核心。到目前为止，原型量子计算机已经由多达14个量子比特（量子位）制成，但它们的输出可以很容易地用经典计算机再现。随着纠缠态变得越来越复杂，它们不再能够在经典计算机上建模。拥有超过50个量子比特的量子计算机可以解决某些无法解决的问题。量子计算正在使用几种不同类型的量子比特，包括离子、超导约瑟夫森结、量子点、光子、钻石中的氮空位中心和中性原子。每个候选量子比特都有其优缺点。被困在光学晶格中的中性原子可以很好地与环境隔离，因此它们具有相对较长的相干时间，这是量子比特的一个基本特征。用光捕获它们提供了一条相对简单的途径，可以实现远远超过50个量子位的可扩展性。尽管如此，对捕获中性原子的研究仍少于对大多数其他量子比特候选者的研究。这里提出的工作是针对开发用于量子计算的中性原子。将开发中性原子量子计算机所需的实验技术。以前，原子在5微米间隔的三维光学晶格中被捕获并冷却，其中一半的位置上有一个原子。使用精确的站点占用图和在5×5×5站点卷中寻址单个站点的能力，将执行一个对该卷中的原子进行任意排序的过程。例如，将创建完全占用的3×3×3立方体和5×5平面。由于原子可以在排序后冷却到接近其振动基态，因此可以检查排序过程并在需要时纠正小错误，从而为量子计算提供了理想的起点。本文将展示一种测量中性原子量子比特内部状态的新技术，该技术通过基于原子内部状态的相干分裂，然后用较短长度尺度的光学晶格将它们锁定在适当的位置，从而实现无原子损失的中性原子量子比特的内部状态测量。然后，它们就可以在这个新的晶格中被可靠地检测到，在这个晶格中，它们的位置编码了它们的初始内部状态。将展示一种新型的单量子比特微波门，其中原子不需要离开它们的存储基，这保证了极高的保真度。此外，还将继续展示双量子比特里德伯门，利用原子的低温和相关的优秀定位。在所有这些技术开发之后，该系统将允许在任何原子损失之前在25个原子上实现~3000个栅极。这将构成中性原子系统可扩展性原理的充分证明，以刺激这些系统的纠错和扩展的进一步工作。

Entanglement is an essential feature of quantum mechanics. For instance, if two identical particles can each be in either state A or B, they can be in an entangled state AA+BB, which means that the particles are in a superposition of both being in A or both being in B, but never one in A and the other in B. These highly non-classical states are central to the working of quantum computers. To date, proto-quantum computers have been made with up to 14 quantum bits (qubits), but their outputs can be readily reproduced with classical computers. As entangled states become increasingly complex they can no longer be modeled on classical computers. A quantum computer with more than 50 qubits could solve certain kinds of problems that are otherwise unsolvable.Quantum computing is being pursued using several different types of qubits, including ions, superconducting Josephson junctions, quantum dots, photons, nitrogen vacancy centers in diamonds, and neutral atoms. Each candidate qubit has its strengths and weakness. Neutral atoms trapped in optical lattices can be well-isolated from their environment, so they have relatively long coherence times, an essential qubit feature. Trapping them with light presents a relatively straightforward path to scalability well beyond 50 qubits. Still, there has been less work on trapped neutral atoms than on most other qubit candidates. The work proposed here is directed toward developing neutral atoms for quantum computation.Experimental techniques needed for a neutral atom quantum computer will be developed. Previously atoms in a 5 micron spaced 3D optical lattice have been trapped and cooled, with an atom at half the sites. Using accurate site occupancy maps and the ability to address individual sites within a 5×5×5 site volume, a procedure to arbitrarily sort the atoms within that volume will be executed. For instance, perfectly occupied 3×3×3 cubes and 5×5 planes will be created. Since the atoms can be cooled to near their vibrational ground state after sorting, the sorting procedure can be checked and small errors corrected if need be, giving an ideal starting point for a quantum computation.A new technique for measuring the internal states of a neutral atom qubit without atom loss by coherently splitting atoms based on their internal states, and then locking them in place with a shorter length scale optical lattice will be demonstrated. They can then be reliably detected in this new lattice, where their location encodes their initial internal state. A new type of single qubit microwave gate where atoms do not need to leave their storage basis will be demonstrated, which promises exceptionally high fidelity. Also work will continue to demonstrate two-qubit Rydberg gates, taking advantage of the low temperature of the atoms and the associated excellent localization. After all these techniques are developed, the system will allow for the implementation of ~3000 gates on 25 atoms before any atom loss is expected. This would constitute a sufficient proof of principle of scalability in neutral atom systems to stimulate further work in error correction and scaling in these systems.