Oxford scientists unlock the properties of turbulence

15 April 2019

Our world and the whole Universe is full of turbulent fluids, usually in the plasma state. Most people are familiar with the notion of turbulence. Whether it is the chaotic swirls that appear as you add milk to your coffee or tea or the unpredictable motions of the atmosphere all too familiar to frequent fliers. However, despite this ubiquity, it is exceptionally hard to pin down in precise mathematical terms, with current theories either derived empirically or through dimensional analysis.

Moreover, it is not easy to study astrophysical plasmas. They are very distant and evolve on timescales of millions of years. Luckily, a central tenet of science is that the physical laws we discover on Earth work everywhere (and everywhen). We can use laboratory experiments to improve our understanding of physical processes and apply these results to processes throughout the Universe. Just like putting a car in a wind tunnel to predict how it will behave out on the road, we can perform experiments in the laboratory to predict how the turbulence will evolve. This emerging field is known as “laboratory astrophysics”.

In space, the chaotic motions that govern the dynamics of the interstellar medium, molecular clouds, stars, supernovae and accretion disks is the driver of turbulence. However, unlike most terrestrial turbulence, in these astrophysical systems, the motions are often supersonic. i.e., faster than the speed of sound, making it even harder to understand. An international collaboration led by the University of Oxford - Prof Gregori (Atomic and Laser Physics), Prof Schekochihin and Prof Sarkar (Theoretical Physics) - was able to study the behaviour of turbulence generated from the collision of two supersonic plasma jets created by high-energy lasers and use it to better understand turbulence in the cosmos (see Figure 1). Prof Thomas White (now at the University of Nevada, Reno) says: “by utilizing a suite of space and time-resolved diagnostics, we were able to observe the transition from subsonic to supersonic turbulence, with the formation of small shockwaves, in the latter case. In our experiment, we were able to accurately characterize the distribution of the density clumps (resulting from shock compression) and velocity motions, and compare our results with several numerical simulations and found that only some models are able to reproduce what we see in the experiments”.


Why is this so? A prominent characteristic of turbulence is that larger scale motions drive smaller ones. Physicists call this the “turbulent cascade”, and in subsonic fluids is it well understood in terms of a theory developed by Russian physicist Andrei Kolmogorov in the 1940s. Kolmogorov stated that large vortices in a turbulent flow break down into increasingly smaller ones, eventually ‘forgetting’ the cause of the large-scale motion. This is why much of subsonic turbulence looks so similar. As put poetically by English scientist Lewis Richardson:

Big whorls have little whorls,
which feed on their velocity,
And little whorls have lesser whorls,
and so on to viscosity.

However, supersonic turbulence behaves quite differently. It is altered by the formation of those small shock waves. That it is very hard to simulate on the computer, where numerical approximations are required in order to capture the sharp discontinuities in the properties of the plasma caused by the shocks. Knowing how the mass is distributed in clumps around those shocks is very important in the astrophysical context as it triggers gravitational instabilities and e.g., the rate of star formation in molecular clouds.

What comes next? Well, we know that magnetic fields are also present in molecular clouds, as well as in many other astrophysical systems. The magnetic field is predicted to modify the behaviour of the turbulence, so we’re currently looking into ways we can add a magnetic field in our laboratory experiments. Several decades ago Richard Feynman called turbulent flow one of “the last unsolved problems of classical physics”. This is still the case now.

The paper has been published online on 15 April 2019 in Nature Communications. The lead author is Prof Thomas White, a former graduate student and postdoc in the Department of Physics at the University of Oxford.


This project was carried out with the support of the European Research Council (ERC) and the UK Engineering and Physical Sciences Research Council (EPSRC).

A copy of the paper is available at: doi:10.1038/s41467-019-09498-y

For more information or an interview, please contact:
Prof Gianluca Gregori, Tel: 01865 282639, Email: Gianluca.Gregori@physics.ox.ac.uk