Good Vibrations

John Bechhoefer was the classic 1960s child. "I wanted to be an astronaut," he says. Many of those who saw Neil Armstrong walk on the Moon during their formative years longed to join him exploring the vast ocean of space. For those whose space travel dreams didn’t come true, they were left with a lasting love of the cosmos. "I did a lot of stargazing as a kid," Bechhoefer says. "Then I found out that to understand much about astronomy you needed to learn some physics." The night sky became Bechhoefer’s gateway drug to a career in physics.

Fast forward and Bechhoefer is now a quantum physicist at Simon Fraser University (SFU) in Burnaby, British Columbia, studying the atomic instead of the astrophysical. Along with his SFU colleague David Sivak and Susanne Still from the University of Hawaii at Manoa, Bechhoefer was recently awarded over US$630,000 from FQXi to optimise information processing in the quantum world.

Information processing sucks up a lot of energy in the real-world applications. Bechhoefer points to the fact that 5 per cent of the energy usage in North America can be linked to data processing. "It’s even more stark in biological systems like ourselves," he says. Your brain uses up about 20 per cent of your calorie intake, despite the fact it represents just two per cent of the body by weight.

Such levels are hundreds to thousands of times less efficient that the thermodynamic limits calculated by quantum information pioneers Rolf Landauer and Charles Bennett in the 1960s and 1970s. They showed that it’s most efficient to do the processing slowly. But in the real world we want answers fast. You don’t want to wait a week for your PC to save a file and a mental calculation over whether to run away from a tiger can hardly wait more than a split second. "We’ve been looking at how you can optimise an operation, even if you want to do it quickly," Bechhoefer says.

Natural Fluctuations

One of Bechhoefer’s experiments involves using an optical tweezer—a sharply focused laser beam that can keep a glass particle suspended in a fluid (see diagram, below). The particles are a micron in size, similar to a bacterium. The laser ("the support") is horizontal and the particle is heavy and so it sags a little under the weight. Thermal fluctuations in the liquid buffet the particle so that it jiggles up and down at random. Now comes the trick that allows the team to exploit these natural fluctuations to extract useful energy, or ’work,’ from the system: "Imagine that every time the particle moves up you move the support up by the same amount," says Bechhoefer. You haven’t had to expend work lifting the particle, just moving the support. The particle now has more potential energy, something you can extract by lowering it at some later stage. We do something similar when we raise water in a hydroelectric power station.

"This is a new kind of engine," Bechhoefer says. Combustion engines consume fuels such as gasoline and their biological equivalents are powered by the molecule ATP (adenosine triphosphate). By contrast, Bechhoefer explains, "Our engines are powered by the information they gather—seeing and then acting upon favourable fluctuations." As such, the engine is a realization of "Maxwell’s demon"—a thought experiment dreamt up by nineteenth-century physicist James Clerk Maxwell in which an all-knowing demon could use knowledge of a system to extract work from it.

"It was a brilliant idea to use an optically trapped, micron-scale bead in water to realise a way to extract work from thermal fluctuations," says Gerard Milburn, a quantum physicist from the University of Queensland, Australia. "The key technical innovation that makes this experiment possible is the ability to measure the position of the particle quickly and accurately."

Milburn believes that the mini engine could have a huge number of practical uses. "The principle they have demonstrated can be applied to a very wide class of machine, especially clocks and…to design and build molecular scale machines that do not need a battery," he says.

Marcus Huber, from the Atomic Institute in Vienna, Austria, is a touch more cautious. "The main question I would have is how generalisable the results will be beyond this particular setup," he says. Huber goes on to say that what ultimately comes out of this research "depends on the achievable efficiencies, especially taking into account the work costs of running the experimental setup. If these even come close to reasonable efficiencies that would be quite a major milestone."

Both Milburn and Huber point to the fact that this isn’t actually a quantum experiment, but a classical one. "The system is large enough to avoid many of the quantum conundrums of information processing, yet apparently small enough to reach very high efficiencies and detailed controllability," Huber says. "This is an interesting zone to work in for sure." Quantum effects such as the uncertainty principle would make it harder to make precise measurements of the particle’s position.

So what has Bechhoefer discovered so far? Early results are in and Bechhoefer was surprised by the power the system could generate. "It’s the same as the power involved in the motors that power machines within living cells," he says. So it’s already possible to see how this may be applied to understanding the efficiency of information processing in biology.

Erasing Information

Bechhoefer and his colleagues are also finally grappling with another experiment, one that had to be delayed due to the impact of the COVID-19 pandemic (see "Lockdown Lab Life"). It involves the erasure of information. While this may sound like simply deleting information stored in a computer’s memory, it is a little more nuanced than that. Computers work by storing information as a series of 0s and 1s. These are "binary digits," which is why a unit of storage is called a bit. Imagine that all the bits in a computer memory start off as 0s and then you perform some operation that turns them into a mixture of 0s and 1s. Then you want to erase what you’ve done to return every bit to zero again. "It’s that reset operation that’s costly because each bit goes from being able to hold two possible states to being one possible state," Bechhoefer explains. "That requires some energy to perform and some heat to be generated as a consequence."

The team has mocked up an idealised version of this scenario in the lab. The same small, micron-sized glass particle is suspended in water once more, but this time the equipment is setup to produce two regions that the glass particle can be confined in. These twin "potential wells" mimic the binary digits of 0 and 1. It takes energy to move the particle between one region and the other, just as it takes energy to reset a bit in the memory of a computer. The team can then look at which reset procedures are the most efficient. "It’s a system that we can control precisely and do precise measurements on," Bechhoefer says. "It’s a test case—if we can understand this better then we can start to understand more complex systems more completely."