Located in a quiet corner of the main Science building of the University's Hongo campus, the laboratory is home to ten graduate students, three postdocs, Assistant Professor Kazumasa A. Takeuchi and, of course, to Professor Masaki Sano. There is space for experiments on turbulent liquid crystals, two-phase thermal convection, Janus particles and cell locomotion in the basement of the same building.

Kazumasa Takeuchi shares his office with five students, various shelves full of books and reports, a myriad of softly buzzing desktop computers and a makeshift kitchen with an espresso machine. Himself a graduate of the laboratory, he has worked on experiments on electroconvection in liquid crystals and interface dynamics. Recently, he started a new project on covariant Lyapunov vectors in an ongoing collaboration with Hugues Chaté of CEA-Saclay, France, and Fransesco Ginelli of the Institute for Complex Systems and Mathematical Biology, Aberdeen, UK. Covariant Lyapunov vectors are perturbation vectors of dynamical systems that grow exactly at the rates given by the Lyapunov exponents. They convey more information about the system than the Gram-Schmidt vectors that one obtains with most conventional algorithms for the computation of the Lyapunov spectrum that use repeated orthogonalisation, he explains. In two recent papers, Kazumasa and his co-authors used covariant Lyapunov vectors to measure hyperbolicity in chaotic dynamical systems and to estimate the dimension of the inertial manifold of dissipative PDEs. Impressive results, but, as I soon find out, the computation of these vectors is no stroll in the park. As Kazumasa explains the details of the algorithm on a minuscule white board, he lists many subtle considerations about transient times, convergence and memory requirements for forward and backward time-stepping.

It is time for lunch, and we are joined by Yohei Nakayama, a first year PhD-student, who has little trouble convincing us to go for a bowl of ramen. This ubiquitous noodle soup defies many clichés of Japanese food: it is heavy, salty and comes with a few slices of roast pork and toppings like chilly oil and raw garlic. As usual in company like this, the conversation settles on what type of ramen is best, depending on the time of day or night and ones state of inebriation. On the way back Yohei explains that his work is more theoretical. Together with Kyogo Kawaguchi, also working towards a PhD, he is moving into non-equilibrium thermodynamics.

Directed percolation criticality in the electrically driven convection of nematic liquid crystal.

Back in the lab, Kazumasa introduces me to Yutaro Matsui, a PhD-student about to submit his thesis on thermal convection in water. The catch is, that in his experiment the gaseous and liquid phase of water coexist. In this setup, the upward heat transfer is enhanced by the flux of latent heat, as water evaporates at the liquid surface and condenses on the lid of the vessel. Moreover, the water intermittently starts to boil, resulting in sudden changes in the dynamics. Yutaro shows me time series that look tantalizingly like relaxation oscillations with a fair amount of noise on top. For now, he only has experimental results. The modelling of this experiment is rather involved, given the thermodynamics, the presence of a free surface, the formation of droplets and so forth. However, there is a keen interest for this system from engineers, who wish to exploit the potentially very high rate of heat transfer that can be achieved. "Just looking at the time series, I started to see an analogy with neural systems and bursting", says Yutaro. "It is just an analogy, but I hope to find some inspiration there for the analysis."

Later in the afternoon, I present some of my work in the Laboratory seminar. A lively discussion ensues, kick-started by regular visitor Helmut Brandt of Bayreuth. "I met Masaki for the first time in Sendai, twenty-nine years ago," he tells me later, "and I have been visiting whenever I could since then." Masaki Sano graduated from the Engineering Department of Tohoku University in 1979 and spent his early years there in the 80s and 90s. "I was reading books by Chandrasekhar and Prigogine. The former became the base of my study on Rayleigh-Bénard convection and instabilities in fluid dynamics, the latter drew me into nonequilibrium systems. That's why I am still working on nonequilibrium thermodynamics and a wide class of nonlinear systems, aiming to find some general laws and principles."

Chaos and fractals were hot topics in the late seventies and early eighties, and Prof. Sano worked both on theoretical and experimental questions. "But after I proposed an algorithm for estimating the whole Lyapunov spectrum from time series and a thermodynamic formalism of chaos in '85 and '86, I felt that a large part of the study of low dimensional chaos was finished and decided to move to problems of turbulence and spatially extended nonequilbrium systems." This decision led to a series of experiments on various forms of convection using liquid crystals, low-temperature Helium gas and mercury. The experiments with Helium gas took place at the University of Chicago, during a two-year visit. "The group led by Kadanoff and Libchaber had reached the highest point in the world in the late 80's in the field of nonlinear dynamics and chaos. They worked on pattern formation, the Saffmann-Taylor instability, viscous fingering, directional solidification, front instabilities and so on. The last universal route to chaos, from torus to chaos, had been investigated experimentally by Libchaber's group and the multi-fractal formalism to characterize such a strange attractor had been just developed before I arrived." The main focus in the experiments was the transition between different turbulent states, and they led to the discovery of Large Scale Circulation (LSC), a mean flow within turbulence that could be reconstructed from measurements with two small, nearby thermometers.

A group portrait of Sano Laboratory. Back row: Masafumi Kuroda, Yutaro Matsui, Takaki Yamamoto and Jean-Baptiste Delfau. Second row: Kyogo Kawaguchi and Koutarou Otomura. Third row: Daiki Nishiguchi, Yohei Nakayama and Yuta Hirayama. Front row: Kazumasa A. Takeuchi, Masaki Sano and Ken H. Nagai.

Sometimes, the search for universal behaviour in the convection experiments was successful. For instance, in 2007, a team consisting of Prof. Sano, his then graduate student Kazumasa, Masafumi Kuroda and Hugues Chaté found the directed percolation criticality in liquid crystal convection. This behaviour had been conjectured by Pomeau many years prior. Other manifestations of universality and criticality proved more evasive. In particular, Prof. Sano has been trying to find the ultimate turbulent state, characterised by a scaling of the Nusselt number as the square root of the Rayleigh number. Still in Chicago, he set the Helium experiment up so that Rayleigh numbers of $10^{15}$ could be reached - a world record that stood until Sreenivasan and Donnely beat it by two orders of magnitude in 2000. "In the summer of 1989 I saw a precursor to the ultimate turbulence state, but it was not reproduced after I left for Japan in August. This became a nightmare for me. Since then, the struggle for finding the new transition has been continued by many good researchers such as Castaing, Sreenivasan and Donnely, Ahlers and Bodenshatz and so on." The discussion is still going, as is apparent from recent papers by He et al. and <a target="external" href="http://prl.aps.org/abstract/PRL/v109/i 15/e154301"> Urban et al. in Physical Review Letters.

Apart from convection, Prof. Sano worked on a range of other projects, such as the oscillating instability in crack propagation and flocking phenomena from the view point of nonlinear dynamcs and fluid dynamics - both through model studies and field observations. "I shifted interests to nonequilibrium and dynamical aspects of biological systems in the late nineties, because the questions raised by Prigogine have not been answered yet for me. Are there general principles governing nonequilibrium systems? Can we understand biological systems through the study of nonequilibrium systems?" Prof. Sano launched a number of investigations, ranging from single DNA molecules to cell motility, always taking a physicist's point of view. The focus is on phase transitions, nonlinear dynamics and fluctuations. "Recent finding on locomotion of amoeba cells suggests that they are making use of instabilities and macroscopic fluctuations originating from chaotic dynamics of the actin-myosin system in the cytoskeleton. This idea is now becoming a hot subject, sometimes called active matter or active soft matter. Active fluids, active gels, active colloids, and their collective behavior is a very rich field for researchers in statistical mechanics, soft matter, fluid mechanics and biophysics." These studies are made possible by recent technological developments that allow us to observe dynamics of single molecules. Also, new theory for nonequilibrium systems, such as the fluctuation theorem, the Jarzynsky equality and stochastic energetics, was developed in the late nineties. "Now I feel this is a really exciting moment. I am hoping to find out novel roles of fluctuations in structural transitions of mesoscopic systems. There might be a fundamental law of thermodynamics connecting fluctuation and structural formation. Our recent experiment on Maxwell's demon might trigger research in this direction."