UNSW Sydney’s research has opened the door to large-scale silicon-based quantum processors that can be used in real-world manufacturing.
Australian researchers have demonstrated that quantum computing is almost error-free. This opens the door to building silicon-based quantum devices compatible with all aspects of current semiconductor manufacturing technology.
“Today’s publication Nature proves our operations were 99 percent error-free,” said Professor Andrea Morello from UNSW, who led the work.
“When errors are rare, it is possible to detect them and correct them as they happen.” This proves that quantum computers can be built with enough power and scale to perform meaningful computations.
Quantum computing in silicon reaches the 99% threshold.
Morello’s paper was one of three published in Nature today, proving that reliable, robust quantum computing in silicon has become a reality. This groundbreaking feature is on the journal’s front cover.
- Morello et al. achieved 1-qubit operation fidelities of 99.95% and 2-qubit fidelity at 99.37% with a 3-qubit system consisting of an electron and two phosphorous atoms. These were introduced into silicon via ion implantation.
- Lieven Vandersypen, a Delft team from the Netherlands, achieved 99.87% 1-qubit and 99.65% 2-qubit fidelities using electron spins in quantum dots created in stack silicon and silicon-germanium alloy (Si/SiGe).
- Seigo Tarucha, a RIKEN Japan team, achieved similar 99.84 percent 1-qubit results and 99.51 percent 2-qubit fidelities using a two-electron system with Si/SiGe quantum dots.
The UNSW and Delft teams verified the performance of the quantum processors by using a complex method called gate set tomography. This was developed at Sandia National Laboratories in America and made available to the public.
Morello had demonstrated before that he could keep quantum information in silicon for 35 seconds due to the extreme isolation of nuclear spins and their environment.
Professor Morello says that 35 seconds in the quantum world is an eternity. For example, the superconducting quantum computers IBM and Google life expectancy is only about 100 microseconds. This is almost a million times shorter than the average lifetime.
The tradeoff was that the qubits were isolated, making it almost impossible for them to interact with one another as is required to perform computations.
The nuclear spins can interact with accuracy.
His team solved this problem using an electron that encompasses two nuclei from phosphorus atoms.
Dr. Mateusz Madzik is one of the leading experimental authors. He says that if you have two nuclei connected to the same electron, you can make them do quantum operations.
While you can’t control the electron, these nuclei securely store their quantum information. You can make them talk to one another via electron to achieve universal quantum operations that can adapt to any computational problem.
Dr. Serwan Asaad is another experimental author. He says, “This is an unlocking technology.” The core quantum processor is the nuclear spins. Once you have entangled them with the electron, the electron can be moved to another location and further afield. This opens the door to large arrays of qubits capable of performing robust and valuable computations.
David Jamieson, University of Melbourne research leader, says: “The phosphorous Atoms were introduced into the silicon chip via ion implant, the same process used in all other silicon computer chips. This allows us to ensure that our quantum breakthrough is compatible with all aspects of the broader semiconductor industry.
Every computer has some data redundancy and error correction, but quantum physics places severe limitations on how this correction can be done in quantum computers. Professor Morello explained that quantum error correction protocols require less than 1% error rates. Now that we have achieved this goal, it is possible to design silicon quantum processors which scale up and work reliably for valuable calculations.
The three papers
The semiconductor spin qubits have the potential to be the platform of choice when it comes to reliable quantum computers. They can store quantum information for extended periods and can be scaled using the same techniques as modern semiconductor manufacturing technologies.
Professor Morello states, “up to now, however, the challenge was performing quantum logic operations at sufficiently high accuracy.”
“Each of today’s three papers shows that this challenge can be overcome so that errors can be corrected quicker than they appear.”
- Andrea Morello, a UNSW team member, created a two-qubit universal quantum logic operation between two nuclear spins formed by a phosphorous donor. These spins were introduced into silicon using the industry-standard method for ion implants. Quantum operations require an electron whose probability waves are spread across both nuclei. According to gate set tomography, individual cores could operate with 99.95% and two-qubit functions with 99.37% accuracy. The electron spin is a qubit, which can be entangled with the two nuclei to create a three-qubit quantum entangled state with a fidelity of 92.5%.Paper:
- Lieven Vandersypen, a Delft team member, created a two-qubit system from a material consisting of a carefully grown stack of silicon and silicon-germanium alloy (Si/SiGe). Quantum information is encoded by spins of electrons contained in quantum dots. Gate set tomography was used to quantify and improve the precision of quantum operations. They reached 99.5 percent accuracy on the two-qubit logic gates. “Pushing the two-qubit gate fidelity beyond 99 percent required improved materials, specially designed qubit control, and calibration methods,” Xiao Xue (lead author of the publication in Nature) said.
Collaborations and exchanges
Although the papers present independent results, they demonstrate the benefits of free academic research and the free flow of ideas, people, and materials. The silicon and silicon-germanium used by the RIKEN and Delft groups were grown in Delft, and the material was shared between them. Kohei Itoh from Keio University, Japan, provided the isotopically pure silicon material for UNSW.
The gate set tomography (GST), crucial in quantifying and improving quantum gate fidelities in UNSW and Delft papers, was developed by Sandia National Laboratories and made public. While the Sandia team developed methods for the UNSW nuclear spin system using direct collaboration, the Delft group also adopted the method for its own research.
The movement of people among the teams has allowed for significant ideas sharing.
- Mateusz Madzik was an author of the UNSW paper. He is now a postdoctoral researcher for the Delft team.
- Serwan Asaad was once a student at Delft.
- Lieven Vandersypen was the Delft leader and took a five-month sabbatical at UNSW in 2016, hosting Andrea Morello.
- Giordano Sappucci is the leader of the material-growth team and a former UNSW researcher.
UNSW’s paper was the result of an extensive collaboration that included researchers from UNSW, the University of Melbourne (for ion implant), the University of Technology Sydney, Sandia National Laboratories (Invention and refinement of GST method), and Keio University (supply of isotopically pure silicon material).
Organizations
- School of Electrical Engineering and Telecommunications UNSW Sydney, Australia
- Centre for Quantum Computation and Communication Technology (Australia)
- University of Technology Sydney
- Ain Shams University Cairo, Egypt
- Sandia National Laboratories Albuquerque & Livermore, USA
- Center for Computing Research at Sandia National Laboratories Albuquerque (NM 87185), USA
- School of Fundamental Science and Technology at Keio University in Kohoku-Ku Yokohama (Japan).
Acknowledgment of funding
The UNSW-UTS consortium was established as part of the AUSMURI Project. The Next Generation Technologies Fund, Defence Department, funded this multi-university Australia-US initiative. Multi-qubit systems are being used to reduce quantum gate errors. This is a significant breakthrough in the direction of high-fidelity silicon quantum processors.
The US Army Research Office provided additional funding for the silicon quantum computing initiative, which supports UNSW, Melbourne, and Sandia National Labs.
The ARC Centre of Excellence supported the UNSW and Melbourne work for Quantum Computation and Communication Technology. The facilities at the Australian National Fabrication Facility (ANFF) UNSW node were used to fabricate the quantum device.
The Dutch Government supported Delft’s work through the same scheme that supports UNSW work.