FAU Physicist contributes to Google’s quantum computer

Prof. Dr. Michael J. Hartmann, Chair of Theoretical Physics at FAU played a part in developing Google’s quantum computer. Image: FAU/Georg Pöhlein)
Prof. Dr. Michael J. Hartmann, Chair of Theoretical Physics at FAU played a part in developing Google’s quantum computer. Image: FAU/Georg Pöhlein)

Universities and major IT companies have been investigating computers that store information in quantum states rather than as binary sequences for more than two decades. A research team at Google has now successfully used a quantum computer in a computing operation that would have taken thousands of years with the fastest supercomputer. The results of the quantum supremacy project were published in the latest edition of Nature. Prof. Dr. Michael J. Hartmann from the Chair of Theoretical Physics at FAU contributed to the development of the Google quantum computer and is co-author of the publication. We spoke to him about the project and future applications for quantum computing.

Prof. Hartmann, what is the difference between quantum computers and conventional computers?

Conventional computers store information as binary sequences of ones and zeros. A quantum computer stores information in quantum states. This means that the processing unit in quantum systems – known as a quantum bit or qubit – not only assumes the value of zero or one but for a specific period of time –known as the coherence time – it can maintain both states simultaneously. We call this property superposition. The values assumed by the qubits are measured after the computing operation. What makes quantum computers so powerful is the exponential increase in computing power that doubles with every qubit. A quantum chip with 300 qubits can assume more states than the number of atoms in the universe.

How many qubits does Google’s new quantum processor Sycamore use?

Sycamore comprises an array of 54 (6×9) qubits, however we were only able to work with 53 qubits in the quantum supremacy project due to an error. The processor can store 2⁵³ or approximately 10¹⁶ binary sequences at the same time which is a quantum leap beyond conventional computing. We conducted a random sequence of quantum logic operations using the chip. Using 53 qubits we were able to achieve quantum supremacy for at least 12 cycles of this quantum gate operation. This means that there is no longer a conventional algorithm which is capable of processing the binary sequences in this distribution in a reasonable time. The most complex calculation that we performed comprised 20 cycles and lasted for approximately three and a half minutes. A modern supercomputer would have needed approximately 10,000 years to perform this operation.

How can we verify calculations performed by quantum computers? This is not possible using a conventional computer.

However, we can verify smaller calculations. It is true that we cannot verify larger operations with conventional algorithms. In this case, we work with tendencies – calculated probabilities that predict which values the qubits will assume. It takes millions of calculations to generate statistical averages that we can use to verify our assumptions. Of course, we must bear in mind that the operations we performed using Sycamore do not have a practical application, they were merely designed to demonstrate the performance of the quantum computer. For most practical applications, we still need an effective method of error correction as the error rate is much higher for quantum computers in comparison with conventional computers.

How can we imagine a quantum computer – a supercomputer with a quantum chip?

Unfortunately, it’s not that simple. The Sycamore chip is designed to use superconducting circuits in which charges and currents exhibit quantum mechanical behaviour. The circuits are activated by magnetic and electric fields and thus change their energy state. Any influence of heat would produce noise and therefore errors, so we cool the chip to minus 273.13 °C which is close to absolute zero. For this reason, a quantum computer is similar to a freezer, however in the form of a cylinder that is suspended from the ceiling. This suggests that we can expect the practical application of quantum computers in data centres but it is unlikely that everyone will have a quantum computer in their home in ten to twenty years.

What tasks are quantum computers designed for?

One of the major challenges over the coming years is to find useful applications for the enormous computing power of quantum computers. Such theoretical research is also part of my role on the Google team. We are currently considering using quantum computers in fields which are related to the nature of the quantum computer such as simulating the quantum behaviour of electrons and atoms in chemical reactions and material properties. But quantum computers could also demonstrate their strengths in machine learning and artificial intelligence.

How did you become part of the Google team?

I have been working with quantum simulation and superconductors for over a decade and I was already part of a collaborative project with ETH Zürich involved in the development of a qubit–qubit coupling. I first got to know John Martinis, the lead hardware developer in the quantum supremacy project at Google while I was head of a research group in 2011 at TU Munich. A year ago, I applied to work with the Google team. At the end of 2019 I will return to the Chair of Theoretical Physics at FAU.

Further information

Prof. Dr. Michael J. Hartmann
Phone +49 131 8528461