Tsinghua University implements quantum memory enhanced non-local graph state preparation

Recently, Luming Duan research group of Institute for the Interdisciplinary Information Sciences, Tsinghua University has made important progress in the field of quantum information. For the first time, it has realized the efficient preparation of non-local graph states by using quantum memory in experiments, demonstrating the application prospect of quantum memory for quantum information processing and quantum measurement in large-scale distributed systems. The research paper《Quantum-memory-Enhanced Preparation of Nonlocal Graph States》was recently published in the international journal Physical Review Letters.

Graph states are an important class of multi-body entangled states, which have attracted the attention of researchers due to their wide application prospects, including testing the basic concepts of quantum mechanics, quantum computing, quantum cryptography, quantum metrology, and so on. The previous experimental preparation of graph states, especially the Greenberger-Horne-Zeilinger (GHZ) state in linear optical systems, presents an exponential decay law with the size of the system, which limits its application in large-scale quantum networks. In order to overcome the scalability problem of nonlocal graph state preparation, Professor Luming Duan and others proposed a theoretical scheme for efficient graph state preparation (Phys. Rev. Lett. 97143601 (2006)), and the preparation efficiency decreases polynomially with the scale of the system. The core idea of this scheme is that photons that do not participate in the later operation can be measured in advance and selected later; Then, large-scale graph states are generated by divide and conquer method using long-lived quantum memory. However, due to the difficulties in experimental technology, the scheme had not been realized before.

In this work, the researchers successfully increased the coherence time of the cold atom quantum memory to tens of milliseconds by trapping the ultra-low temperature rubidium atomic gas in the optical lattice and using a pair of clock level transitions in the atomic ground state energy levels. Researchers first use the first quantum memory (QM1) to generate and store a pair of entangled states between photons and atoms, then use the second quantum memory (QM2) to generate a second pair of entangled states between photons and atoms, and finally read the quantum states inside the two quantum memories at the same time and project them onto the four-photon GHZ state of the target. The results show that the preparation efficiency of the four-photon GHZ state generated in this way is linearly proportional to the preparation efficiency of a single EPR entangled pair. Compared with the quadratic relationship between the preparation efficiency caused by the absence of a quantum storage scheme and the preparation efficiency of a single EPR entangled pair, it changes the complexity of the preparation efficiency in scale. When n pairs of entangled pairs need to be connected in the future, the preparation efficiency will rise from the exponential attenuation level to the polynomial level, which shows the superiority of the quantum storage scheme in preparing large-scale graph states. In addition, the researchers further verified the mabk inequality by using the prepared four-photon GHZ state, and demonstrated the quantum secret key distribution protocol in quantum cryptography. This work has realized an efficient prototype for preparing large-scale graph States, which has taken an important step for its various applications in quantum information science and quantum metrology.

The first co-author of this paper is Sheng Zhang, a postdoctoral, and Yukai Wu, an assistant professor of the Institute for Interdisciplinary Information Sciences, Tsinghua University. The corresponding author is Professor Luming Duan. Other authors include Chang Li, a doctoral graduate of the school of cross information (now a postdoctoral fellow of the University of Strasbourg, France), Nan Jiang, a lecturer at Beijing Normal University, and Yunfei Pu, an assistant professor.

Link:

Institute for Interdisciplinary Information Sciences, Tsinghua University