"Achieving 99% fidelity for a two-bit gate has been the most important experiment in the field over the past few years. The research groups around the world that are qualified to carry out this experiment are all engaged in research in this area, which can be said to be an open scientific research race. In other words, if it is consistently unable to achieve 99% two-bit gate fidelity, then there is no need for the system to continue research. When choosing the college entrance examination volunteer many years ago, Xue Xiao had wandered between finance, computers and natural sciences. Today, with many papers in his hands, he has truly lived up to his original choice. A few days ago, he achieved the above "99% target".
Figure | Xue Xiao (right) and the article's second author Maximilian Russ (Source: Mateusz Madzik)
On January 19, the Nature cover paper was published three times in a row, and it was all quantum themes. Xiao Xue is the author of one of the papers,[1] titled "Quantum logic with spin qubits crossing the surface code threshold."
Figure | Related papers (Source: Nature)
He said the work solves a long-standing bottleneck in a field: the spin qubit based on the spin of electrons in semiconductors, or spin qubits, whose two-qubit gates have always been relatively low fidelity.
Compared to the basic logical unit bits used by traditional computers, qubits are more susceptible to ambient noise, which leads to calculation errors. To develop a quantum computer with practical application value, a necessary technology is quantum error correction.
However, quantum error correction has extremely high requirements for fidelity, and the most promising error correction technology is called Surface code, which requires that in the process of performing quantum computing, each step of the operation includes bit initialization, single bit gate, two bit gate, and the fidelity of bit information reading needs to be higher than 99%.
Therefore, 99% fidelity is also called the fault tolerance threshold. Initialization of spin qubits, single-bit gates, and reading are relatively easy to achieve 99% fidelity, some of which are even close to 99.99%.
In the past, however, only two papers reported 92% and 98% of two-bit gate fidelity,[2] but in scientific research, 98% is a long way from 99%.
In the academic community, the research of semiconductor spin qubits has been carried out for more than a decade. The basic preparation method is to use micro-nano processing to make nanoscale metal electrodes on the surface of the semiconductor, and form a "potential well" inside the material by applying a voltage, binding individual electrons to it.
Such nanoscale structures are called quantum dots and bear a high degree of similarity to the transistor structures used in traditional computer chips.[3] This similarity has made semiconductor-based quantum computing a source of concern, especially its practical path.
At present, Intel, the Belgian Microelectronics Centre (IMEC), IBM and other industry giants are studying this direction.
Among them, Intel and IMEC have now used advanced integrated circuit processes to achieve large-scale device preparation on 300mm wafers. Once the system proves viable, then industry can apply the traditional integrated circuit process directly to quantum chips.
In contrast, other physical systems, such as ion traps and linear optics, are incompatible with existing advanced integrated circuit processes. As for superconducting qubits, while a similar process can be used, their qubits are in millimeter size, making it nearly impossible to integrate millions of qubits on a microscale chip.
The average size of spin qubits defined by quantum dots is around 100 nanometers, which gives them great advantages in scalability and future integration with chips with different functions.
As for the aforementioned "proof that a system is viable", a key indicator is the fidelity of the system. Of all operations, the fidelity of the two-bit gate is the most difficult to improve, which is true for all physical systems.
Xue Xiao pointed out that for a long time, the few propaganda points of semiconductor spin qubits are conducive to integration and expansion, but the fidelity of its two-bit gate has progressed slowly, so that the feasibility of this direction has been questioned.
This time, Xue Xiao's experiment proved that the fidelity of the single-bit gate and the two-bit gate were all stable above 99.5%, directly breaking through the previous limit.
The fidelity of the two-bit gate has been verified to reach 92%
Previously, Xue Xiao completed a series of silicon-based two-bit experiments in 2018-2019, the most representative of which was to verify that the fidelity of the two-bit gate reached 92%[2].
However, the qubit quality at that time was not good, mainly limited by the nature of the material. Silicon in nature contains three isotopes: Si-28; Si-29; and Si-30. The nucleus of Si-29 carries nuclear spins, which interfere with the spins of electrons used to encode qubits.
In early 2019, Xue Xiao's colleague Giordano Scappucci at Delft University of Technology in the Netherlands successfully improved the properties of the silicon substrate in the lab, including a silicon material that uses isotopic purification, that is, the removal of the vast majority of Si-29 atoms.
Colleague Nodar Samkharadze from the Netherlands Organisation for Applied Scientific Research (TNO) prepared the two-bit sample.
At this time, on the advice of his mentor Lieven Vandersypen, Xue Xiao began the study of two-bit gates again in 2020. Also joining is Maximilian Russ, a young theoretical physicist who is very good at semiconductor spin qubits, who is also the second author of the paper.
In the study, Xue Xiao et al. completely characterized and modeled these two spin qubits from both experimental and theoretical aspects, and fully grasped all the sources of errors that may cause errors in quantum computing in the experiment, such as environmental noise in the sample and external control systems. Finally, at the beginning of 2021, stable experimental results were obtained.
Simulating quantum chemistry and physics will be the "first mature application" of quantum computing
It is now widely believed that in recent years, simulating quantum chemistry and physics will be the "first mature application" of quantum computing.
The last experiment presented in this paper is that Xue Xiao and his team used high-fidelity quantum gate operations to perform quantum simulations of the ground-state energy spectrum of hydrogen molecules.
Figure | Quantum simulation of the ground-state energy spectrum of hydrogen molecules (Source: Nature)
In addition, another paper published earlier by Xue's colleagues in Nature [4] demonstrates the observation of Nagaoka's ferromagnetic state. This is a physical phenomenon proposed by Nagaoka, a Japanese theoretical physicist, but it is currently unobservable in nature.
In Xue Xiao's laboratory, this phenomenon was successfully observed in an artificially prepared array of quantum dots.
With the increase in the number of qubits, the results are expected to be achieved in the future to simulate the climate, optimize urban transportation, and quickly decipher codes. However, these expectations are difficult to achieve in the next 5-10 years.
Speaking of this, he added: "As a digression, I am personally curious about whether quantum computing can be applied to mining and improving the cloud gaming experience in the short term. ”
He added: "This work can eventually be completed, thanks in particular to the second author of the article, Maximilian Ross, and the third author, Nordar Samharadze. The former joined our lab three years ago as a postdoctoral fellow in theoretical physics, primarily responsible for theoretically supporting experiments. There is actually a big generation gap between theoretical researchers and experimental researchers, which also caused some difficulties for the cooperation between the two sides at the beginning. But we've always been proactive in learning from each other. In the later stages, it was full of tacit understanding. SamHaradze is currently a researcher at TNO. He is primarily responsible for Quantum Inspire's projects. The project aims to put spin qubits on the cloud, allowing ordinary users to operate directly through the Internet. ”
At first, Samharaze prepared the sample for use in Quantum Inspire, but their team didn't have enough experience to debug the experiment. So, he and Xue Xiao began to cooperate and provided samples to the latter, and Xue Xiao gave them feedback on the debugging results and experience. In the end, in addition to cooperating to complete the paper, another almost identical quantum chip was also put on the cloud as scheduled.
In addition, the chips used in this paper study were used in an earlier paper titled "CMOS-based cryogenic control of silicon quantum circuits",[6] and published in May 2021. In that paper, Xue Xiao et al. initially demonstrated the possibility of integrating traditional control/reading instruments and qubits.
For the follow-up plan, he said that 99% fidelity is required by quantum error correction, so the next step is naturally to conduct quantum error correction experiments. First of all, a sufficient number of qubits is required; second, high-fidelity initialization and reading needs to be done while completing high-fidelity single/double qubit gates; and finally, a fast feedback system is needed to correct in real time according to the errors in the experiment.
Deeply influenced by the study of a large number of sub-information in China
According to reports, in 2014, Xue Xiao graduated from the Department of Physics of the University of Science and Technology of China. He said: "HKUST has always been the best quantum information research 'treasure land' in China, and it is also one of the strongest scientific research units in the world. As early as the end of the college entrance examination, I saw the discussion of quantum computing in the online forum of HKUST. At that time, although I almost didn't understand it at all, I subconsciously felt that I might be interested in it, but in fact, I originally discussed with my parents to learn a more profitable professional direction such as economics or computers. ”
He recalls that when he was studying at HKUST, he was affected a lot. At that time, HKUST had already made breakthroughs in multiphoton entanglement and 100-kilometer quantum key distribution. The Quantum Satellite (Mozi) project has also been launched.
Many of the original teachers were scientists directly engaged in quantum information research. In his spare time, he listened to the academic reports of the three academicians of "GDP" (Guo Guangcan, Du Jiangfeng, and Pan Jianwei) many times, and also visited their laboratories.
In his sophomore year, he decided to pursue quantum research, and later "played soy sauce" in the laboratory of Professor Lu Chaoyang, one of the leaders of the "Nine Chapters" quantum computer experiment.
After graduating from undergraduate, Xue Xiao studied at Tsinghua University at the postgraduate level until she decided to go abroad in 2016. In 2017, he joined the laboratory of Professor Livin van der Sippen of Delft University of Technology in the Netherlands as a PhD candidate.
In 2020, his team worked with Intel Corporation to implement the operation of spin qubits using a cryogenic control chip. This is a landmark experiment leading to integrated quantum chips. The so-called integrated quantum chip is to integrate the qubit with the traditional electrical control and reading system on the same chip. The article was finally published in Nature in 2021.[6]
Now, Xiao Xue has completed his doctoral thesis and continues to pursue post-doctoral research in the Livin van der Sippen Laboratory. For his current mentor, he admired and introduced: "During his doctoral studies at Stanford, Levine van der Sippen completed the experimental verification of the world's first Shure algorithm using nuclear magnetic resonance experiments on liquid molecules [7]. "The algorithm quickly decomposes an integer into the product of two prime numbers, which is the mathematical basis of many modern ciphers." Later, Levine van der Seepen came to Delft University of Technology, established the laboratory of semiconductor spin qubits, and completed most of the work that laid the foundation for spin qubit experiments in the early days, and he also won the spinoza Prize in the Dutch natural science in 2021 for his contributions to semiconductor spin quantum computing.
To have the opportunity to learn from quantum giants at home and abroad is Xue Xiao's luck and the result of his efforts. The post-90s youth from Zibo, Shandong Province, who is also considering a teaching position in China, said: "Returning to China has always been at the forefront of my many options. Of course, this also depends on the opportunity. ”
-End-
reference:
1、X.Xueet al., Nature601, 343 - 347 (2022)
2、X.Xueet al.,PRX9,021011 (2019)
3、Lieven Vandersypen, Mark Eriksson,Physics Today72, 8, 38 (2019)
4、J.P.Dehollain et al.,Nature579, 528–533 (2020)
5、https://www.quantum-inspire.com/
6、X.Xue et al.,Nature593, 205–210 (2021)
7、L.M.K.Vandersypen et al.,Nature414, 883–887 (2001)