There’s a lot of talk about quantum computers and a coming revolution. Do you feel the same sort of optimism that the pioneers of personal computers felt as they began to put the first microcomputers together in their living-rooms and garages? Bloch: Yes and no. For one thing, they were a step ahead of us. They already had all the components – transistors, IC chips – needed to build their computers. We are still in the process of building the first integrated circuits, so to speak.
Really? How far has development progressed? Weinfurter: We now have machines that can handle 60 to 70 quantum bits, ‘qubits’ for short. Quantum computers are based on a new form of information processing, and qubits are the elementary units with which they work. Unlike a transistor, which is either OFF or ON, a qubit can store the values 0 and 1 simultaneously. This makes it possible to perform computational operations in parallel rather than in a linear fashion.
What can be done with these machines? Weinfurter: In terms of standard applications, like calculations in sprad sheets, nothing that a normal notebook couldn’t do just as well. But because qubits can be placed in states known as superpositions, operations can be carried out in a high-dimensional space. That allows one to simulate things that even today‘s supercomputers cannot manage – such as systems of more than 50 or 60 atoms in a solid. Bloch: With quantum simulators and quantum computers – for very special problems – we are now at a turning point. A universal computer that is capable of automatic error correction is still a very long way off. Such a system would require millions of qubits. Meanwhile, the current 60-qubit systems can be used to tackle a range of interesting scientific problems, such as phenomena related to magnetism, to mention just one example.
What does a 60-qubit computer look like? Something the size of a football field crammed with lenses, lasers and bulky cryostats? Bloch: The design depends on the specific platform. Several interesting platforms are under development. Some, such as those used in my group, are based on atoms or ion traps. Another system uses artificial, superconducting qubits. Our platform requires a relatively complex optical setup. The superconducting system doesn’t need that, but takes up more space per qubit, especially if the system is ever scaled up to millions of qubits. Google and IBM are investing heavily in superconducting-based systems.
Can these platforms be miniaturized? Bloch: Yes. It’s a little easier to do with the superconducting chip, because it has more to do with conventional electronics. Here again, the real problem lies in scaling this system up to handle millions of superconducting qubits. As things now stand, that would indeed require a chip as big as a football field. Weinfurter: – Which would need cryostats as big as football fields to reach the low temperatures required for superconductivity. The scalability issue is crucial, and it’s not yet clear which system will make the running. Bloch: Each platform has its own strengths and problems. Take the question of the quality of the qubits. Superconducting qubits are not supplied by Mother Nature. They have to be built up literally from the bottom up, and they must be indistinguishable from each other. The slightest imprecision reduces the fidelity of the entire system – and that’s very hard to control. Nature provides the atoms and ions that we use, and all particles of a given species have exactly the same properties. So I don’t need to worry about that. But there are other complications.
You work with atoms and ions. Do you regard the developers of rival platforms as competitors? Bloch: Of course. At conferences we have to rigorously demonstrate the sorts of calculations our platform is capable of performing and what level it can plausibly reach. Marketing hype is quickly cut down to size. Indeed, the discrepancy between the clamor on the market and what has been accomplished so far in the lab is often enormous. But I don’t want to play down what has been achieved. On the contrary, I find the progress that has made been made tremendously impressive. Weinfurter: One shouldn’t forget that only 40 years ago, each transistor in a commercial TV set had to be individually soldered into place. There were no microprocessors at that time. Systems undergo rapid development when a particular technology matures and becomes dominant. Quantum computers haven’t got that far yet. But the level of hype is already very high – it’s a curious situation.
Where does the new Cluster of Excellence stand in this competitive field? The Munich Center for Quantum Science and Technology (MCQST) aspires to a leading position among the protagonists of the second quantum revolution. At least that’s what the original proposal says. Bloch: As the only Center of its kind in Germany, we cover all the areas that form the central pillars of quantum technologies – simulation, computing, communication, quantum materials, metrology and sensor technology. Most of us have focused on basic research, but Harald Weinfurter’s work is a good example of how the development of quantum communication can reach the threshold of technological implementation. Apart from that, we see ourselves as pacesetters. Our aim is to extend the boundaries of the known – with regard to theoretical issues as well as hardware, software and algorithms. In Munich, we have experts in all these fields and who are among the best in the world. On the commercialization side, firms like Google or IBM have great advantages, because they have the necessary resources. That’s where Europe lags behind, at least as far as quantum computing is concerned.
Some of the leading members of the new Excellence Cluster recently published a paper on the subject of transmission protocols for superconducting qubits in the microwave region. Are these the sorts of basic elements that are needed? How important are such details in the larger scheme of things? Weinfurter: That’s indeed a good example of an essential building block, although the paper reports an important developmental step, not a complete protocol. It presents an initial concept that would allow two quantum computers to talk to each other. The method is designed for microwave frequencies but, in principle, it could be implemented at optical wavelengths.
That will require specialized hardware. At the moment, the qubits can be transmitted through 60-cm superconducting cable, and the next goal is to develop a 10-m cable for this purpose. That doesn’t sound like much, but it’s being hailed as a milestone. Bloch: We already touched on the problems involved in scaling up systems based on superconducting qubits. Cryostats the size of football fields, which would be required in principle, do not exist. But the problem can be attacked from another direction. One could load several dilution refrigerators with, let’s say, 200 qubits each, and connect them via a superconducting cable to create a single chip with a much higher performance. Such interfaces and communication protocols are essential, they are core components. They might make it possible to build computers that can work with 10,000 qubits. Weinfurter: And such a system would also be compatible with the classical approach to computer architecture. Bloch: It’s not just a matter of building a working computer. The real question concerns the construction of quantum networks, and here too we need to develop feasible and efficient approaches to the problem. Weinfurter: For communication over longer distances, we will end up working with optical wavelengths. The quantum world will make use of a variety of technologies, just as conventional computers utilize bits in the form of electrical pulses as well as pulses of laser light.
So the development of quantum computers will entail the construction of the quantum internet? Weinfurter: Yes, we need to re-create every element of our current modes of communication. That‘s the idea. Maybe the quantum devices we use ourselves will be a comparatively small device, while the real work is done in the background by much larger units. That would be no different from what we have now. All we use of the Internet are our smartphones. Yet, a simple Google query is now processed in computer centers the size of football fields; the result appears on our smartphones again. Bloch: The whole complexity of our classical system, including its imperfections, is essentially hidden from the user. And lot of software protocols simply protect us from these imperfections. The world of quantum computing requires the same kinds of resources, but designing them is far more challenging.
The capacities of today’s quantum computers are quite limited, as you say. Yet fears are already being expressed that they will be able to crack our current encryption systems. Weinfurter: Quantum computers could, at some time in the future, crack our present public-key crypto systems. That’s why cryptography experts are now trying to develop methods that are immune to attacks by quantum computers also in the future. The field is now known as post-cryptography! Bloch: In my view, code-breaking is one of the most boring of all conceivable applications for quantum computers, though intelligence services might take a different view. As of now, attacks carried out by quantum computers are the least of my worries. Research is nowhere near the stage needed to build such systems. This estimate could change in the next 5 or 10 years. But it’s much more likely that advances will lead to breakthroughs in fields like pharmacology or quantum chemistry, or more economical ways to solve the travelling-salesman problem. The TSP is the classical optimization problem, and concerns a commercial traveler who needs to work out the most economical route for his next sales trip to customers in so-and-so many different cities. It sounds trivial, but the problem often defeats classical computers.
The first performance test for Europe‘s quantum computer OpenSuperQ will be to compute the structures of polymers of hydrogen cyanide. Bloch: That’s a very hard nut to crack. We can now solve simple problems in quantum chemistry, such as calculating the energy levels of bi-atomic molecules. But we are miles away from doing the same for larger molecules, such as pharmaceuticals. The goals are ambitious, but we are still at a stage where we focus on exemplary demonstrations. However, simulation-based approaches to problems that are just beyond the reach of classical algorithms look very promising. Indeed, the prospect of competition from simulations has boosted efforts to develop new classical algorithms. This sort of effect is often overlooked.
Can quantum cryptography provide us with secure keys based on quantum operations? Weinfurter: Yes, absolutely. As always, the tricky question relates to how much extra effort one wants to invest to obtain a secure key. Currently, quantum cryptography is not as easy to handle as public-key systems. But I believe that, once it becomes fully integrated optic fibers or free space communication systems, the technological effort involved won’t be very much more. Quantum cryptography allows one, for the first time, to assess the level of security achieved by a key as we can measure the amount of information that a potential eavesdropper could possibly have.
In other words, the technique is not absolutely secure, but one can determine whether an eavesdropper has tapped the line and react appropriately? Weinfurter: It is absolutely secure. There is a signal-to-noise parameter that tells us how much information an eavesdropper could have maximally picked up. With that knowledge, it’s possible to reduce the size of the key such that the eavesdropper knowledge of the key vanishes – he can’t intercept any communication at all. That‘s the trick.
You have worked on secure quantum keys for years. Your early experiments were quite spectacular. In 2007, you securely transmitted a quantum key between Tenerife and La Palma, a distance of almost 150 km and you later sent a key to a ground station from an aircraft at an altitude of 20 km. Weinfurter: In our earliest experiments, we transmitted keys between the Zugspitze and the Karwendelspitz, Alpine peaks that are 20 km apart. That’s about the distance through the atmosphere you need to bridge in order to communicate with a near-Earth satellite. The higher you go, the better the weather conditions. That’s true of the volcanic islands of Tenerife and La Palma as well, over a distance of 144 km. There we were above the clouds, where the air is dry and the visibility was optimal. The technique now works quite well here in the city.
So you’re still chasing the next record? Weinfurter: We hope to extend the current limits, to demonstrate that the method is close to practical application. Our devices are now hand-held, and the optical unit for the transmitter is about the size of three matches. and can be integrated easily into conventional communication systems
And the next step? Weinfurter: We want to go beyond the exchange of keys to build a proper network. That’s a much bigger challenge, because the idea is to transmit information of quantum states by means of quantum teleportation, which necessitates to distribute entangled particles over long distances. That requires new systems, called quantum repeaters that act as relay stations. That’s what we are now working on.
How does one go about building such systems? Weinfurter: You need two quantum systems and a channel for communication of quantum states of light between them, and then you have to entangle the states of the two systems. We use two atoms, one located here in Physics Building, the other in the Faculty of Economics, which are coupled by 700 m of optic fiber. These two atoms then become entangled with each other.
In what sense? Bloch: Entanglement is a specific quantum mechanical state, which is distributed or shared between two spatially separated particles, and forms a link between them. It’s a difficult concept to grasp, I must admit. Even Einstein had trouble with it, and once referred to it as ‘spooky action at a distance’. But if we measure the properties of each of the entangled particles, we find correlations between them that cannot be explained by classical statistics.
And entangling particles that are far apart is obviously not a trivial task. Weinfurter: Each of the two atoms is stimulated to emit a photon, and these two photons are then brought together and measured like in quantum teleportation. In this way, the entanglement of the two atom-photon pairs can be transferred to the two atoms. At present, the efficiency of the process is not very high. We often have to try it a million times before we get the atoms entangled.
And they then remain entangled? Weinfurter: That’s a good question. Quantum states are highly sensitive to environmental factors such as fluctuations in temperature and magnetic fields, which we do our best to minimize.
Your next trick will be to entangle particles located in Munich’s city center and Garching and separated by a distance of 20 km. How can you manage that? Weinfurter: The emitted photons have a wavelength of 780 nanometers. Unfortunately, conventional optic fibers strongly absorb at this wavelength. But the absorption rate is much lower in the region around 1500 nanometers, which is used of course for communication over glass fibers. So we added a frequency converter to our system, which enabled us to use that type of fiber, and allowed us to transmit the entanglement between atom and photon over a stretch of 20 km. So far we used a 20-km length of fiber coiled on a big spool in our laboratory. It would be impossible to do this overland at present – but we’re working on it.
Have other fields of application made more rapid progress? Bloch: I would cite sensor technology here. In quantum communication and quantum computers, systems must be shielded from unwanted disturbances from the environment. Sensor technology exploits the sensitivity of quantum states to the environment – their fragility – to make extremely precise measurements.
Can you give me an example? Bloch: I’m thinking here of sensors that can rapidly measure magnetic or electric fields, and rates of acceleration with levels of precision and spatial resolution that are unattainable by any other means. This is an area that is very promising for applications in many different contexts. Weinfurter: We have already learned to control particular quantum systems very well. These systems are tiny, but their responses are easily read out. There are basically two approaches. One is to use atomic systems like in the quantum simulators. The other is based on what are called nitrogen vacancies in diamond. Bosch is using it to develop components for magnetometry.
How does this work? Weinfurter: An atom has a magnetic moment, just like a compass needle. So do defects in crystals, which give crystals their colors. In diamonds, the replacement of a carbon atom by a nitrogen atom together with a vacancy in the lattice nearby, and this nitrogen vacancy center behaves like an artificial atom. It has similar properties to real atoms and can be used to make extremely sensitive measurements. Bloch: Tiny diamond crystals can be relatively easily attached to virtually anything, even biological samples, which are otherwise not easy to make measurements on. Weinfurter: We’re talking here of nanocrystals with dimensions of around 200 nanometers. One can create such defects in a pattern just below the surface and use it like an MRI scanner to measure the magnetic field strengths of molecules. That could turn out be a tremendously useful tool for chemists.
Are advances like this part of Quantum Revolution 2.0? Bloch: That term actually refers to the application of the entanglement concept, the ability to link elementary building blocks to form a single unit that can be manipulated in accordance with the laws of quantum mechanics, in various ways and for many different purposes. Weinfurter: Up to now, the main reason for learning about quantum mechanics was that it enabled one to understand the operation of transistors, and how to improve their efficiency. Now we are in a position to manipulate individual quantum systems. Bloch: The next step is the transition from single systems to entangled systems. But working with entangled systems is technologically far more difficult. Entangled states permit one to make measurements with significantly higher sensitivities than those attainable with non-entangled systems. The challenge lies in generating these entangled states on demand, or rather, keeping them in existence for longer periods of time.
The potential of quantum technologies is regarded as huge. Governments are providing research funding, industry is investing, especially in quantum computers. IBM wants to collaborate with the Fraunhofer Association, Google with the Jülich Research Center. The EU has initiated Flagship Programs and the Federal Government intends to put 650 million euros into research on quantum computers. The investment involved is massive. Bloch: Please forget the figure of 650 million euros. In fact, the BMBF will provide up to 100 million for quantum technologies over the next few years. – With that amount of money, you can’t expect any quantum leaps. With regard to the involvement of American companies in Germany, as far as we know, Google has signed an agreement on the exchange of information with the Jülich Research Center. Weinfurter: And we’ll have to wait and see whether IBM actually installs one of its machines at a Fraunhofer Institute.
Does Munich have the potential to become a Qubit Valley? Bloch: I tend to avoid comparisons like that. But the new Excellence Cluster is already of global significance. Munich is an important player on the international scene, and hosts many companies, such as Microsoft and Google, that are very interested in the field. Furthermore, Munich is home to lots of companies that play an important role in the photonics sector. – So many of necessary conditions are in place. Basic research remains the driving force, but many other forces are at play – including commercially oriented businesses that have great power and influence. In the future, we will need authoritative entities to assess the claims made by these firms in relation to the performance levels of their products. Last year Google announced a decisive breakthrough …
… which was published in Nature. Bloch: Indeed. In that paper, the company claimed to have shown that, for one specific class of problem, its quantum processor had successfully exceeded an efficiency threshold which conventional supercomputers cannot reach. IBM doubts the validity of the claim. Here, universities can provide useful reference points and develop standards. In addition, we should concern ourselves more with concrete problems in various areas of the science, test how good the different quantum machines really are, and determine whether and where they can be of use to us. It’s an exciting environment for researchers like us, but also for the industrial sector.
Earlier on, you mentioned in passing that you hoped to use quantum simulations to gain a better understanding of physical phenomena such as magnetism. Bloch: Certainly. One of the central questions in condensed-matter physics concerns what I call the social behavior of large ensembles of particles. How do they act as a collective, what are they actually doing? Although we know the rules that determine how particles interact with each other, it is often difficult to predict what swarms of particles will actually do when they are allowed to interact in accordance with these rules. This is what accounts for the properties of materials and phenomena like magnetism and superconductivity.
How do you intend to tackle the problem? Bloch: We want to understand how these interactions work, using model systems that allow use to localize atoms in precise positions on a lattice, to visualize them individually and to follow what each of them is doing. This microscopic view gives access to new analytical methods, which answer the question of why different materials behave in the ways that do.
So is the real revolution to be expected in basic research rather than in the field of computing? Bloch: When I compare what we as researchers could do 30 years ago with how we explore these systems today, then I have to say that we are now in the middle of a revolution. When it was first demonstrated 30 years ago that atoms could be cooled by lasers, no-one could have imagined that the technique would one day allows us to analyze ensembles of particles in diverse fields such as solid-state physics and high-energy physics. The rate of progress has increased enormously. We now make use of discoveries that once received Nobel Prizes as basic elements in our experiments, and we can now test theories with the aid of quantum simulators. It may well turn out that quantum computers and quantum simulators will have their greatest impact on the sciences, in biology and chemistry too, and that the average person will never actually interact with a quantum computer. In my opinion, that is very likely to be the case.Moderation: Hubert Filser and Martin Thurau
Prof. Dr. Immanuel Bloch holds a Chair in Experimental Physics at LMU and is a Director of the Max Planck Institute for Quantum Optics (MPQ) in Garching. Bloch (b. 1972) studied Physics at Bonn University. After obtaining his PhD from LMU, he carried out postdoctoral research at the MPQ and LMU. He was appointed to a full professorship at Mainz University in 2003, and returned to Munich in 2008. Immanuel Bloch has won many prizes for his research, including the Leibniz Prize awarded by the Deutsche Forschungsgemeinschaft (DFG), and an Advanced Investigator Grant from the European Research Council. Bloch is one of three Coordinators of the Munich Center for Quantum Science and Technology (MCQST).
Prof. Dr. Harald Weinfurter is Professor of Experimental Physics at LMU. Weinfurter (b. 1960), studied Technical Physics at the Technical University of Vienna, where he obtained his PhD. He has worked as postdoc at the Hahn-Meitner Institute in Berlin and at Innsbruck University, where he completed his Habilitation. Weinfurter joined LMU in 1999. He has received the Descartes Prize awarded by the European Commission and the DFG‘s Copernicus-Prize. In 2010 he became a Fellow of the Max Planck Institute for Quantum Optics in Garching.