This year’s Queen’s Lecture at TU Berlin will be given by Professor Winfried Hensinger of the University of Sussex. Professor Hensinger is working on a machine that could change all our lives in the future - the quantum computer. He has worked as an advisor to both the UK and German governments, founded startups in both countries, and regularly commutes between the UK and continental Europe. This capacity to be in two places at once, so to speak, is also a key element in his research.
Professor Hensinger, you completed the first stage of your undergraduate degree in physics in Heidelberg, before moving to Brisbane in Australia and doing your doctorate at the University of Queensland. After working in the USA at the University of Michigan for three years, you took up a post at the University of Sussex in England in 2005, where you still work today. Would you prefer to speak in German or English?
Oh, I can still manage both. But, let’s keep it in English, if only because of the technical terms. I am much more familiar with them in English.
You are working on constructing a quantum computer. In 2017, you made something of a name for yourself when you presented a blueprint for a chip-based quantum computer with millions of quantum bits that could be mass-produced on an industrial scale. What could such a computer do that even today's supercomputers can’t?
The honest answer? We still don’t know exactly. What we can say is that there are problems which today’s computers would need millions of years to solve that could be solved in a matter of hours by a quantum computer. Optimization problems, for example, like the “traveling salesman’s problem.” Perhaps you have heard of this? It’s the one about finding the shortest route to visit a number of cities. For a classical computer, the computing time increases exponentially with the number of places that need to be visited. A quantum computer, on the other hand, can calculate the optimal route in one fell swoop, so to speak.
So the logistics industry would benefit from such a computer, and optimization problems also exist in construction and process technology. What other applications are there for quantum computers?
A quantum computer will be able to recognize patterns in huge amounts of data much more easily; this is a prerequisite for machine learning and so helps advance artificial intelligence. In addition, a quantum computer is exactly what it says - something from the quantum world. So it’s predetermined to calculate and simulate precisely that. And quantum physics is part of everything that is truly revolutionary: complex biomolecules for new drugs and vaccines, superconductors that transport electricity without resistance, catalyst materials that in the future will remove CO2 from the atmosphere. But as I said: Our situation today is similar to that at the beginning of the computer age. If you had asked someone at Bletchley Park in England during the Second World War what a computer was useful for, they would probably have said “Well, for breaking the German Enigma encryption machine, of course, what else?”
Is it even possible for a layperson to grasp the concept of a quantum computer, that is to say its ability to calculate in one fell swoop, as you put it?
Yes, I think so. Join me briefly, if you will, on an excursion to an experiment that illustrates the peculiarities of the quantum world. It was even voted the most beautiful experiment of all time in a survey conducted by the British Physics Society.
I can’t wait.
Imagine you throw two stones into a pond. The circular waves collide and form a complex pattern as they attenuate and intensify each other. Now, in the famous experiment I am talking about, electrons are fired onto a wall with two holes positioned close to each other. If we then set up a camera for electrons behind this, the screen will show exactly the same pattern as with the waves on the pond. This demonstrates first of all that particles like electrons also have wave properties, because apparently circular "electron waves" emanate from the two holes. What’s really astonishing, however, is that this interference pattern also gradually emerges if one shoots the electrons one after the other at the holes. But for the pattern to emerge, you always require two circular waves. So what happened? Did each electron split, with one half flying through each hole? But this is physically impossible as the electron is an indivisible elemental particle.
OK, I see the problem. How is it solved?
Most physicists see the solution to this problem in so-called probability waves, which can be assigned to every quantum object - such as electrons. However, it is not clear whether they really exist or whether they are just a mathematical construct. However that may be, roughly speaking, the height of these waves indicates the probability of finding the particle at a given location. The probability that the electron will pass through a hole in our experiment is now 50 percent for each hole. So the probability wave of the electron splits into two partial waves in front of the holes. A partial wave with a height of "50 percent" goes through each hole, and behind it both overlap, like in the case of the two stones and the pond.
Isn’t that a case of physicists deluding themselves? I mean, saying that in some way the electron passes through both holes at the same time...
What we say is that the state of the electron is undetermined. Only when measured does it have to decide on a location. If we measure it, then in 50 percent of the cases we find the electron in the first hole, and in the other half of the cases in the second. Prior to this measurement, it exists in the superposition of two states. This is pretty crazy. In my doctoral thesis, I caused an atom to move forward and backward at the same time. Imagine this when trying to park a car! You only find out what the car has decided to do when it bumps in the Rolls Royce parked in front ...
So, what does all of this have to do with quantum computers?
Well, we can make use of this principle of superposition of states. For this purpose, we use ions, i.e. electrically charged atoms, which we can hold in place or move with electric fields. In a simplified picture of an atom, the electrons circle around the atomic nucleus on fixed paths. There are orbits with lower energy and orbits with a higher energy level. We are now looking at an electron that can be either in a low orbit or in a higher orbit. In one orbit the electron circles clockwise, in the other one counterclockwise. Now, the funny thing is that orbiting clockwise or counterclockwise affects the magnetism of the entire ion, so we can measure which orbit this one electron is in. Let's call the low orbit 0 and the higher orbit 1. So now we have made a quantum bit out of our ion with the digital values 0 and 1, which we can use to calculate.
But what about this superposition thing?
Well, that is the final trick, so to speak. With the help of microwaves, we can lift the electron from the lower energy level to the higher one, and also animate it to return to the lower level, meaning we can change from 0 to 1 and back again. However, to be completely sure that our quantum bit actually changes from 0 to 1, we would have to irradiate the microwaves for a certain time. But we don't do that. We only beam the microwaves in for half of that time. What do you think happens then?
A change has taken place in 50 percent of these cases, and in the other half there has not yet been a change.
Sounds logical. But welcome to quantum land! In fact, exactly the same happens as in the experiment with the electrons: A probability wave for 0 and 1 with a height of "50 percent" is formed. And since we use quite a few of these quantum bit ions, many such waves can overlap. By choosing the irradiation times of the microwaves differently, we can specify different proportions of 0 and 1 for each quantum bit, and thus different heights of the probability waves. If we then also use our microwaves to couple the ions in such a way that their energy states influence each other, then we can perform calculations and read off a solution from the resulting interference pattern. All we have to do at the end is measure the magnetic states of the ions.
Just like the sound waves of an orchestra, which, when everyone plays correctly, overlap to form a symphony from which something meaningful emerges.
Yes, except that each of our instruments is extremely delicate and sensitive. Every contact with the environment acts like a measurement and causes the superposition state of the quantum bits to collapse like a soufflé, so they spontaneously decide to be 0 or 1. That's why we need millions of them. Actually, a quantum computer with 300 quantum bits could compute 2 to the power of 300 numbers at once – that's more than the number of atoms in the visible universe. But how are we to determine whether one of the quantum bits has not just collapsed? Any measurement before the end of the calculation would lead to the undesired collapse of the superposition. Fortunately, there are cunning ways in which groups of quantum bits can watch out for each other, so to speak. We can then always measure one of them and thereby determine whether everything is still okay.
There are various methods of producing such quantum bits.
Tell us why yours with the ion chips is the best one!
I should perhaps add that I didn’t invent this method – that was Christof Wunderlich from the University of Siegen back in 2001. Since then, it has been applied by several research groups. We extended this idea and developed a method where you can simply apply voltages to the ion chip to perform calculations. And then we presented a blueprint on how to build such a quantum computer with millions of quantum bits that can also solve really important problems.
Now for the benefits of our method. There have long been experiments with ions in large vacuum chambers manipulated with lasers. But there you have to aim very precisely and adjust the lasers to micrometers; any shock is really disruptive. It would be difficult to build a quantum computer with millions of quantum bits this way. Then there's another chip technology that uses superconducting rings. In these, for example, you can have a current flowing clockwise and counterclockwise at the same time. But unlike our chips, you have to cool them down to temperatures of a few thousandths of a degree above absolute zero using liquid helium. This is very costly; moreover, due to thermodynamics, it only works for a relatively small number of quantum chips. We, on the other hand, could build a really large quantum computer with many chips in a modular fashion with moderate cooling.
What are the advantages of this modular approach?
It is the key to developing really big quantum computers. We achieved an important breakthrough just this year. We succeeded in assembling multiple microchips, rather like a jigsaw puzzle, and then transferring individual ions from one chip to the next – all with an extremely fast transfer rate and incredibly small error rates. This modular approach is also our unique feature.
What’s next? Major companies in the USA such as Microsoft, IBM and Google are already investing billions in different technologies for constructing a quantum computer.
I established a startup in 2018 with Sebastian Weidt because I really wanted to construct a quantum computer. This is something you cannot do at a university; you can’t get the right people for one thing and universities are there to conduct research and not develop products. Companies are much more suitable for constructing machines like this in large numbers and with the necessary performance. Incidentally, TU Berlin is a university that produces a lot of spin-offs, which is another reason why I am delighted to be giving this year’s Queen’s Lecture. We were able to raise venture capital with our company Universal Quantum and will now actually set about building two quantum computers in Hamburg for 67 million euros by 2026 on behalf of the German Aerospace Center. One of these will be a multi-chip quantum computer. We're only building a machine with about 100 quantum bits for now, but we will use this to integrate all our technological innovations to show that quantum computers with millions of quantum bits are no longer science fiction.
We are doing this through our German subsidiary, which we founded a few years ago, because we are convinced that Germany will become a leading country in quantum computer development. There is also a wealth of talent here, which will help with the rapid development of the technology.
67 million can’t of course be compared to investments of billions. Is this another case of Germany being too reluctant to take a risk?
No, quite the opposite! The German federal government will invest a total of three billion euros in funding for quantum computers and other quantum technologies between now and 2026. That really is a lot if you compare it with what the EU is investing. Here, we can say Germany definitely isn’t holding back. In the area of quantum technologies, we really have a chance to compete with the big players and not fall behind the USA or China.
Wow, do I detect some feelings of German patriotism? How do you see yourself today...as Anglo-Saxon, German, or cosmopolitan?
That’s a really good question. Perhaps you could best say I am somewhat quantum in terms of nationality and am many things at once (laughs).
Interviewer: Wolfgang Richter