The French inventor Jacques de Vaucanson is remembered for, among other things, producing three curious automata in the 18th century. A poster from the time advertised them all side by side: a figure that played a real flute, another that banged a tambourine and a duck that gobbled up corn and seemingly turned it into pellets of… well, use your imagination.
For physicist , based at the National Institute of Standards and Technology in Maryland, these antiquated automata have a resonance with some of today’s most cutting-edge technology. Vaucanson’s inventions prefigured the industrial revolution, during which mechanisation went from being a quirky curiosity to a force that reshaped the globe. We may be at an analogous turning point today when it comes to quantum technology, says Yunger Halpern.
Advertisement
The steam-powered world of the industrial revolution may seem far removed from the quantum realm. But this period of dramatic change was bolstered by thermodynamics, which deals with heat, work and energy. And, recently, physicists have been applying its ideas to the subatomic realm to devise the new field of quantum thermodynamics. This has seen the development of machines like quantum fridges, batteries and clocks.
However, these are just the start of the quantum technology revolution, says Yunger Halpern. She was a co-author of a recent manifesto aiming to chart a path towards greater things. She spoke to New Scientist about what advanced quantum machines might look like, the astounding benefits they could bring and how we can work towards making them a reality.
Thomas Lewton: You have been thinking a lot about automata recently. But what exactly are they?
Nicole Yunger Halpern: An automaton is a machine that performs some function on its own, without external control that changes over time. When I was little, for instance, I was enchanted by a toy dog that would bark and do a backflip at the push of a button. Today, automata vacuum floors, and self-driving cars are becoming increasingly common.
People have built automata for many centuries. For example, in 1739, the inventor de Vaucanson built a “digesting duck” that appeared to eat grains and then expelled pellets afterwards. It’s a good example of how automata achieved an impressive level of sophistication and yet were used primarily for entertainment. On the other hand, though, Vaucanson also contributed to the rise of the industrial revolution by inventing an automatic loom.
So those centuries-old automata spurred wider technological progress?
Indeed, there was a shift during the industrial revolution: automata and machines that needed less control than machines used throughout most of human history were harnessed on a large scale for practical benefit, transforming the face of civilisation. The toys served as a stepping stone to industry.
How did these machines lead to a new understanding of physics?
The industrial revolution inspired thermodynamics, the study of energy in forms such as heat and work, and the relationships between them. Engines were powering factories so, naturally, people wanted to understand how efficiently engines could operate. For example, a heat engine is a very simple type of engine; steam engines are heat engines, as are the engines that operate in many cars. A heat engine interacts with a hot environment and a cold environment. Heat tends to flow from hot to cold. The engine siphons off some of this heat and transforms it into work. Where does thermodynamics come in? Well, in 1824, the French engineer Nicolas Léonard Sadi Carnot identified an upper limit on the efficiency of heat engines – and he designed an engine cycle that could achieve this ideal efficiency. So thermodynamics, especially at first, was inspired by concerns about performing useful work.
There’s been a surge of interest in what happens when concepts in thermodynamics collide with the quantum realm. Why?
The laws of thermodynamics were developed by people who had in mind steam engines and other classical systems of everyday size. In quantum thermodynamics, we are extending and reformulating these laws to see how they govern quantum systems such as atoms, single particles of light and other minuscule systems. We have long known that energy is intertwined with information. Thermodynamics, for example, famously features entropy, a measure of uncertainty, which also features in information theory. Naturally, quantum systems can have their own version of this: quantum information. It features in phenomena like entanglement, where two particles can behave as if they are connected even when separated by vast distances.
We can look at a system that is processing information and energy and ask: what behaviours does it exhibit only if it’s quantum? Thermodynamics can thereby help us understand better where the border is between the classical world and the quantum world. Pinpointing this border has both fundamental and practical implications. First, it elucidates the nature of our universe, which is ultimately quantum but which – a bit mysteriously – appears classical in our everyday experiences. Second, pinpointing the border can guide us towards building new quantum technologies, akin to quantum computers, and identifying their limitations.
Why have you sometimes referred to all this as “quantum steampunk”?
Steampunk is of course that genre of literature, art, and film that features 19th-century settings, such as smoky Victorian London and the Wild West and embeds futuristic technologies in those throwback settings. It conveys a sense of adventure and exploration. To me, quantum thermodynamics is the real-world version of steampunk. Thermodynamics developed during the Victorian era, inspired by steam-powered technologies. In contrast, quantum information science is partially cutting-edge and partially futuristic, as quantum technologies remain under development. Plus, quantum thermodynamics is a fast-moving field full of opportunities, so it shares the steampunk spirit of adventure.
Getting back to practical matters, can an improved understanding of quantum behaviour help us build better machines?
Yes, we already know that quantum phenomena such as entanglement can enhance information-processing tasks such as computing and cryptography. Quantum computers have already been claimed to outperform classical computers in certain tasks. But it’s not all about information. There are energy-processing tasks, such as powering factories and charging batteries and refrigerating, where we can use quantum phenomena to get an enhancement. People have designed – and to some extent experimentally realised – quantum engines, quantum batteries, quantum refrigerators and more.
Most of the quantum machines built so far aren’t very useful, though. For example, in one , researchers intentionally wasted the work outputted by an engine because otherwise they would have had to devise a whole other plan for capturing it. Also, quantum phenomena such as entanglement usually come into play only at temperatures close to absolute zero, so experimentalists have to spend a lot of energy cooling down the system. Because of things like this, much more work is often spent on operating quantum machines than you actually get out of them.
It sounds like we are still at the “digesting duck” stage when it comes to quantum machines…
Yes, very much so. Most of the machines I just mentioned don’t even qualify as automata because they are driven by external control that changes over time. But I’m advocating for leaving the duck stage! Since we have experienced such great successes with the fundamental research aspect of quantum thermodynamics, we can start to look beyond it. Now is the time to shift our attention to realising useful autonomous quantum machines.
What would a really transformative quantum machine be like?
A good example would be around quantum computers. Large-scale quantum computers will be able to solve certain problems far, far more quickly than ordinary computers, even supercomputers. They could be used for things like advanced cryptography, as well as for discovering new materials and drugs. But our current models are too small to do these things. So we would like to free quantum computers, as much as possible, to perform each step of their computations on their own. If we are to build an autonomous quantum computer, we need to include an autonomous quantum clock, which would tell the computer when to begin and end each step. These autonomous quantum computers would be scalable and use fewer resources.
Tell me about your recent call to arms…
In November, my collaborators and I published for how to achieve useful autonomous quantum machines. It’s a call to arms that echoes the DiVincenzo criteria for quantum computing. In 2000, David DiVincenzo published seven guidelines for building a quantum computer or material. These criteria have steered the construction of quantum computers for decades.
Inspired by the DiVincenzo criteria, we looked at a lot of examples of existing or proposed autonomous quantum machines, such as autonomous quantum refrigerators, and also looked at natural, autonomous quantum machines found in biochemistry. For example, our bodies contain molecular switches and enzymes that speed up chemical reactions. Then we abstracted eight criteria that we thought more or less united them.
What are these criteria?
The first four core criteria concern the components you need to build quantum automata, and how to engineer the right interactions amongst them. The next four concern questions such as: Is the output of the machine worth the input? How will the machine shut itself down after completing its task? How can autonomous quantum machines communicate with each other and navigate a landscape?
We would like to free up quantum computers to perform calculations on their own
You and your collaborators recently developed a useful quantum machine, which could help reset quantum computers between calculations. How did that come about?
I described my frustrations about how quantum thermodynamics should be more useful to my colleague the Chalmers University of Technology in Sweden. I explained that I was looking for an environment that was already cold, so that a machine in it could behave quantum mechanically. This environment should include a hot sub-environment and a cold sub-environment, so that the machine could essentially extract its own work. Moreover, there should already be a need for an autonomous quantum machine here so that we could just slot it in. Simone said: “Oh, I already have this environment in the quantum computer being built at my university!”
Simone is developing a quantum computer made of superconducting qubits. These already have to be kept extremely cold with a large refrigerator, so if we put an autonomous quantum refrigerator in there, it won’t require much extra energy to stay cold. There was also a need in this setting: if you have finished a quantum computation and you want to perform another, you need to reset or erase qubits by cooling them down a great deal. My colleagues recently tried doing this with an autonomous quantum refrigerator we designed – and I was surprised at how well it performed. Now I am on the lookout for more of these opportunities.
In your wildest dreams, what do you imagine autonomous quantum machines could look like someday?
One could draw inspiration from the small autonomous machines that operate in biological systems, though these are not necessarily quantum. Molecular motors carry cargo across cells, and RNA uses DNA construction manuals to build proteins, for instance. We can wonder whether adding quantum coherence or entanglement could enhance these processes. Or you could imagine quantum drones delivering atoms here and there, or quantum sensors that might be able to move around and report on what they encounter.
