11 April 2019

Microsoft invests millions of Danish kroner in a state-of-the-art quantum physics lab and a team led by a guitar playing school dropout who became a professor in physics. All as part of Microsoft’s plan to develop quantum computers.

When Peter Krogstrup dropped out of school to follow his dream of becoming a guitar-playing rock star, it seemed impossible that a few decades later he would become a professor in Physics at Copenhagen University and be working with Microsoft to develop the ultimate quantum computer.

As impossible as having a bit which holds the value 0 and 1 at the same time.

As impossible as the building material for that bit is an elementary particle which is its own antiparticle.

Which, in the mind-boggling world of quantum physics, is quite possible.

Today, Peter Krogstrup is Director of the Microsoft Quantum Materials Lab as well as a professor in material physics at the renowned Niels Bohr Institute, named after the Danish physicist who made groundbreaking contributions to quantum theory.

I had the pleasure recently to talk with Peter Krogstrup and thought the subject would interest the readers of this blog.

When you read, please bear in mind the following quote from the philosophical inclined Niels Bohr: “If you can fathom quantum theory without getting dizzy, you cannot possibly have understood it.”

A quantum computer works with qubits (quantum bits) that can hold the value 0 and 1 simultaneously.

This is, of course, different from the traditional bits we know which are either 0 or 1. The superposition of a qubit in both a 0 and 1 state means that adding more qubits will give an exponential rise in computational power, so a quantum computer should quite easily outclass the world’s fastest supercomputers.

We already have quantum computers with qubits, but the existing quantum computers haven’t fulfilled the potential that is ascribed to proper quantum computers, because the qubits are quite fragile.

Microsoft and Peter Krogstrup’s team of researchers want to change that by developing qubits that are much more stable than the conventional ones.

Improving quantum computers

“Quantum states are very easily affected by their surroundings. We operate at the elementary particle plane and the smallest interaction with the surroundings means that your quantum state will lose information,” Peter Krogstrup says.

The short life span of quantum states in existing quantum computers inevitably leads to errors in the computations. Thus, it is necessary to correct the errors, and this error correction means that existing quantum computers do not exceed the capabilities of traditional computers.

”Quantum computers today have a certain amount of errors which is measured by qubit-fidelity. If you have a fidelity at 99%, you will have an error rate at 1%. This is a problem as you need to have a lot of algorithms to adjust your computations for errors. So basically you don’t gain anything by using those kinds of quantum computers. If you want a quantum computer that exceeds classical computers, then you need a qubit-fidelity at 99.999%, so you get rid of the need for most of the error correction,” Peter Krogstrup explains.

Conventional qubits are made from superconducting materials such as niobium and aluminium, patterned on a silicon substrate. That is the kind of quantum computer with 20 and 50 qubits you may know.

IBM, Google and Intel are trying to increase the life span of conventional qubits and subsequently the quality of the quantum computations.

Crystals as qubits

Microsoft, Peter Krogstrup and his fellow researchers have chosen a different path. Instead of trying to improve conventional qubits, they want to build some new, more stable qubits called topological qubits.

“We have chosen a more challenging direction,” Peter Krogstrup says with a small grin and continues:

“Theoretically, we know that there are some exotic quantum particles, fractional elementary particles called topologically bound state, which has bounded states on the edge of your material with characteristics that mean they do not as easily interact with their surroundings. They do not react with normal electrons and protons as they have a different structure.”

Please note the word theoretically.

The exotic quantum particles with the even more exotic name Majorana fermions does not exist naturally.

“To obtain these topologically protected states is difficult. They do not exist in nature, so we have to build them,” Peter Krogstrup explains before he starts describing how you build Majorana fermions:

“Semiconducting crystals, superconducting crystals, ferromagnetic insulators” and some other stuff which I do not really pay attention to as my brain is busy trying to process the information that a Majorana fermion is its own antiparticle.

“Wait … what?” I think, desperately trying to understand the implications of this apparent impossibility. I do not succeed, but despite my mental short circuit I managed to note that crystals are the building blocks for topological qubits and there are stringent requirements to the purity of the materials and the perfection of the crystals.

That’s why Microsoft built a whole new lab for Peter Krogstrup and his team of fellow researchers at Microsoft’s Danish headquarters in Lyngby, a few miles north of Copenhagen.

”If we’re going to build the quantum computer we want, we have to build a whole new platform of crystals, that haven’t been made before. That requires material research in a super controlled environment. Vibrations, acoustic noise, electromagnetic noise has to be kept at a minimum.”

Experiments near absolute zero

The material research is a mix of experimental quantum physics and hardcore quantum theory. The experiments in the Lyngby lab are conducted at temperatures near absolute zero; -273.15C or 0 Kelvin.

“Temperature is vibrations, energy and we want to avoid that as it can destroy the quantum states, so we keep it as cold as possible, so the vibrations are minimised as much as possible. It’s impossible to reach absolute zero, but we are close to a temperature at 20 milliKelvin.”

It is groundbreaking research taking place in the lab, so it is very difficult to establish a traditional project plan with milestones. Peter Krogstrup can’t say when the researchers will have the Majorana fermions ready.

“We have done a lot of experiments which indicates that we have the Majorana fermions. Our equipment suggests that we have realised the states we want, but we still have to do quite a lot before we are able to isolate them. That’s what we are working on at the moment.”

Quantum computing in Azure

Even if the researchers can create stable qubits in the near future, you should not expect to buy your own quantum computer any time soon.

Besides the lower layer of qubits with a temperature near absolute zero, there are other layers that do not consist of standard hardware. On top of the qubit layer, there is a layer with control electronics consisting of gates and switches used to control and direct impulses to the quantum layer. This layer is kept at a temperature at 4 Kelvin (- 269.15 C) with the help of liquid Helium.

Next, there’s another layer where the temperature is kept at 70 Kelvin (- 203.15 Celsius) with liquid Nitrogen. This layer is what Peter Krogstrup describes as “a lot of classic control electronics to adjust for error”.

If the researchers succeed in creating a paradigm shift for quantum computers and obtain stable topological qubits, the result of the researchers’ work will be available for the masses via cloud computing, for instance, Azure.

The end of an unstructured life

Peter Krogstrup’s journey from guitar playing school-dropout to professor in quantum physics seems almost as incomprehensible as elementary particles that are their own antiparticles and qubits that can be 1 and 0 at the same time, but not for Peter Krogstrup.

“I was obsessed with music and played in different bands and studios, but economically it was very hard. There were months when I didn’t earn any money. When my girlfriend became pregnant, I decided that I could not continue my unstructured life any longer.”

Consequently, he started studying in the evening and “fell in love with science.” During his school time he had always liked physics, but “I had too much energy to sit down and listen.”

He was 30 years old when he began to study Physics at Copenhagen University and since that he has been so busy that he hasn’t got the time to play the guitar anymore.

You can try quantum coding!

If you are in Copenhagen beginning of May, you can hear Peter Krogstrup explain more about his work with Microsoft and quantum computers at the Infosecurity-conference.

Meanwhile, you can have a look at Microsoft’s Q# programming language for quantum computers.

Even though you do not have your own quantum computer with several layers near absolute zero, you can run quantum programs on your traditional computer as the development kit simulates a quantum computer. Of course, the simulation will not be as powerful as a real quantum computer, so don’t expect to break RSA encryption with Shor’s algorithm or accelerate breakthroughs in medical research.

We need Peter Krogstrup to get hold of those Majorana fermions so that Microsoft can build a proper quantum computer.

What are your thoughts?


Author: Dan Mygind

Dan is a Journalist and Computer Scientist with a strong interest in technology, technology-related businesses, and the transforming effect source code can have on society.

He has worked for startups, SMEs and global IT-organisations such as IBM as a developer, consultant, and IT-architect. With a solid technology background, he has written extensively for a wide variety of publications such as Computerworld as well as writing technical white papers for Microsoft and other companies.

He is also a published author, ‘World Storytellers

Contact Dan Mygind: mygind{at}writeit{dot}dk

The views expressed are those of the author and do not necessarily reflect the view and opinion of Curo Services.

Back to Curo News