On-chip quantum memories based on erbium dopants in silicon waveguides

通过量子中继器实现全球量子网络是当前量子科学中最热门的课题之一。它不仅有助于量子理论的新的基础和精度测试，还将使量子信息处理中的许多应用成为可能。要实现一个大型量子网络，需要满足严格标准的量子存储器。首先，它们需要提供足够的存储时间，大约是S的1倍。其次，它们必须促进高效率和保真度的量子比特的存储和按需检索。第三，它们必须在低损耗的电信频段中充当单光子的源(或与源兼容)。最后，它们应该提供一种简单的有效多路复用的策略，这通常需要多模容量、低成本材料、健壮性和可扩展的制造技术。尽管在不同的物理平台上进行了大量的努力，但这种量子存储器的实现仍然是一个突出的挑战。特别是，后两个标准在以前的实验中被证明是困难的。因此，克服现有物理系统瓶颈的新技术似乎势在必行。为此，我们计划研究由晶体硅制成的纳米光子波导中的掺铒。一般来说，基于掺杂剂的存储器特别有希望，因为它们提供了任何量子系统中最长的相干时间-长达6小时-并且它们提供了一条通向可伸缩性的明确道路。在迄今研究的所有掺杂剂中，铒的发射波长落在1530 nm到1565 nm之间的光通信主波长范围内，因此脱颖而出。这有两个优点：第一，硅在这个波长范围内的透明性确保了与硅纳米光子学成熟平台的兼容性，为多路复用提供了一条清晰的路径。其次，光纤在该波长下的最小损耗是跨越全球距离的量子网络的关键要求。我们的新实验平台建立在半导体行业的标准工艺基础上，与所有其他正在研究的平台相比，这大大降低了实验开销。通过利用硅纳米加工的潜力来实现集成的纳米光子量子存储器，我们希望建立一种关键的能力，用于可证明安全的通信和基于网络的量子信息处理。

The realization of global quantum networks via quantum repeaters is one of the most intensely pursued topics in current quantum science. It would not only facilitate novel fundamental and precision tests of quantum theory, but also enable numerous applications in quantum information processing.To implement a large quantum network, quantum memories that fulfil stringent criteria are required. First, they need to offer sufficient storage time, on the order of 1 s. Second, they have to facilitate storage and on-demand retrieval of qubits with high efficiency and fidelity. Third, they must act as a source (or be compatible with sources) of single photons in the low-loss telecom bands. Finally, they should provide a simple strategy for efficient multiplexing, which typically requires multimode capacity, low-cost material, robustness and scalable fabrication techniques.In spite of numerous efforts in different physical platforms, the realization of such quantum memory is still an outstanding challenge. In particular, the latter two criteria have proven difficult in previous experiments. Therefore, a new technology that overcomes the bottlenecks of existing physical systems seems mandatory. To this end, we are planning to investigate erbium dopants in nanophotonic waveguides made of crystalline silicon.In general, memories based on dopants are particularly promising as they offer the longest coherence times of any quantum system – up to six hours – and they provide a clear path towards scalability. Among all dopants studied to date, erbium stands out because its emission wavelength falls within the main wavelength band of optical telecommunication between 1530 nm and 1565 nm. This has two advantages: first, the transparency of silicon in this wavelength regime ensures compatibility with the mature platform of silicon nano-photonics, providing a clear path for multiplexing. Second, the minimal loss of optical fibers at this wavelength is a key requirement for quantum networks that span global distances.Our new experimental platform builds on standard processes of the semiconductor industry, which dramatically reduces the experimental overhead compared to all other platforms under investigation. By harnessing the potential of silicon nanofabrication towards the realization of integrated nanophotonic quantum memories, we thus expect to establish a critical capability for provably-secure communication and for network-based quantum information processing.