**August 7, 2017**

Today we have a new paper out on the arXiv in which we describe a quantum thermal machine which generates maximal entanglement in any dimension. https://arxiv.org/abs/1708.01428

Thermal machines do a lot of useful tasks for us. Power plants turn heat into electricity. Refrigerators keep our beers cold. Steam locomotives pull trains (OK, maybe they mostly don’t any more – but a steam engine is the archetypical picture of a thermal machine). In general they are machines which move heat around, or transform it. Often by connecting different points of the machine to different temperatures and exploiting the heat flow between them.

Classical thermal machines are big. Think about power plants – and even a fridge is about the size of a person. At these scales, we don’t need to worry about quantum physics to understand what is going on. But what happens if make such a machine smaller and smaller, to the point where quantum effects become important? Say we take a locomotive and scale it down, down, down, until the boiler and gears and so on consist of just a few atoms? Of course, it won’t really be a locomotive any more – but maybe we can learn something interesting?

Indeed, we can. And the machine can still be useful.

In recent years, physicists (such as Paul Skrzypczyk and Marcus Huber) have learned a lot about thermodynamics on the quantum scale by studying such tiny thermal machines. Looking at how fundamental concepts from classical thermodynamics, such as the 2nd law or Carnot efficiency, behave in the quantum regime, we gain new insights into the differences between the classical and quantum worlds.

In the spirit of invention of steam engines (like at the time of the industrial revolution), we can also think about whether there are new tasks that such quantum thermal machines could do.

Entanglement is an essential quantum phenomena. Objects which are entangled behave as if they are a single entity even when separated and manipulated independently. This enables new, powerful applications such as quantum computing and quantum metrology, and is at the heart of the foundations of quantum physics. So creating and studying entanglement is very interesting from both fundamental and applied points of view. Might a thermal machine be used to generate entanglement then?

Entanglement is generally very fragile, and thermal noise tends to wash it out quickly. In fact, a lot of effort in quantum physics experiments goes into keeping the systems cold and isolated, so as to be able to observe the genuinely quantum effects. So it is not at all obvious that using a thermal machine for entanglement generation would work. However, it turns out that connecting with noisy environments can indeed help to create and keep entanglement stable, in certain systems.

This was already realised and studied by other researchers. Then, a couple of years ago, we described a minimal thermal machine generating entanglement (as I wrote about here). That setup was nice because it was really the simplest quantum thermal machine imaginable, using just two quantum bits and two different temperatures, and it turned out this was already sufficient to see entanglement.

However, the amount of entanglement which that machine could generate was rather limited. In our new paper, we present a new quantum thermal machine – not much more complicated – which generates maximal entanglement. And it does this, not only for two quantum bits, but also for two quantum trits, and in fact for two quantum systems of any dimension (which we prove analytically thanks to the hard work of Armin Tavakoli). The new machine again uses just two different temperatures and two quantum systems of some given dimension. One system is connected to a cold bath, one to a hot bath, and they interact with each other. When heat flows from hot to cold through the two systems, they become entangled.

So indeed, a quantum thermal machine can be useful. And maybe in the future we might see quantum computers, or sensitive quantum sensors, based on thermally generated entanglement.

** Published paper: **https://quantum-journal.org/papers/q-2018-06-13-73/