A novel picture of thermal conductivity20/10/17Science & Technology
Dr Marc de Boissieu, senior scientist at CNRS, tells PEN about new research on thermal conductivity in complex materials
The engineering of thermal conductivity in semiconducting materials is a central issue in the development of modern nano- and microtechnologies, and low thermal conductivity is important in materials used in technology products as it provides thermal insulation and thus the reduction of heat transfer, ensuring the products do not overheat. In new neutron experiments conducted at the Institut Laue-Langevin (ILL) and the French National Centre for Scientific Research (CNRS), researchers have provided a direct quantitative measurement of phonon lifetimes in a clathrate, offering a novel picture of thermal conductivity in complex materials. The study has also highlighted the importance of neutron techniques in overcoming the challenging task of accessing, and therefore successfully measuring, phonon lifetimes.
Pan European Networks spoke to Dr Marc de Boissieu, a senior scientist at CNRS, about this research and some of the challenges involved.
What would you say are the biggest experimental and theoretical challenges involved in measuring the lifetime of individual phonon states?
Phonon lifetimes are most easily measured using inelastic neutron or X-ray scattering. However, the current instrument resolution allows the measurement of a relatively short lifetime (5ps) and mean free path (of the order 5-10nm at most).
How did you overcome the barriers to measuring phonon lifetime/quantifying thermal conductivity in your recent research?
Using a high-resolution inelastic neutron scattering instrument and comparing the data with pure Germanium, we could first evidence a clear signal indicating a mean free path over 20nm. A level of accuracy almost ten times better was then obtained by using the neutron resonant spin echo technics (NRSE) that allow the measuring of phonon lifetimes up to 50ps, and mean free path up to 100nm. This technique requires large single grains of extremely high quality, grown within the research collaborative team. Due to the very long measuring time required by the NRSE techniques, a careful selection of the ‘interesting’ points was carried out using large overviews of phonons measured on other instruments.
Could you describe the key results of the project? What impact do you hope this research will have on improving modern nano- and micro-technologies moving forwards?
Understanding how heat propagates in a material is a key issue in many different fields. For applications such as nano- and micro-technology, heat dissipation is a key element. For thermoelectric materials, which can transform heat into electricity, researchers design new materials and try to reduce as much as possible their thermal conductivity, a key parameter in the efficiency equation in these materials. However, an understanding of the detailed mechanism of heat propagation in complex systems is still lacking. The heat is carried out by quasiparticles named phonons, related to atomic vibration, which travel in the material at the speed of sound. While travelling in a material, a phonon collides with defects or with other phonons, like billiard balls, in turn reducing the thermal conductivity. The more collisions there are, the smaller the thermal conductivity. This is measured by the mean distance a phonon travels without collision (phonon mean free path) or through the mean time between two phonon collisions (phonon lifetime – the phonon mean free path is the product of the phonon lifetime by the phonon velocity).
A multi-partner study has measured, experimentally, for the first time the phonon lifetime in a clathrate, a thermoelectric material, renowned for its very low ‘glass-like’ thermal conductivity (i.e. where phonons are expected to experience a large number of random collisions). This material contains Ge, Ba and Au atoms arranged periodically with a unit cell containing 54 atoms. The Ge and Au network form cages that enclose the Ba atoms.
The team found surprisingly long mean free paths ranging from tens to hundreds of nanometres (1nm=10-9m). This is much larger than the short phonon mean free path of the order 0.5nm that is commonly associated with such glass-like thermal conductivity. The study also demonstrates a large reduction of the number of phonons effectively carrying heat. Those results challenge current theoretical calculations and open the way for a new understanding of heat transport in complex materials.
The result came as a real surprise, although previous measurements had already given some hint toward this result. To really confirm that all measurements were correct, we did a careful comparison with pure Ge single crystal (the clathrates already contain 75% Ge), for which lifetimes were already known to be very long. The input of the high resolution experiment, as accessible with the NRSE technique, was also a key in these studies.
We have also carried out a detailed structural analysis to fully understand the atomic structure and the disorder. Combined with atomic scale simulations we could also understand the distribution of phonons velocities.
Our second important result is a strong reduction of the number of effective phonons that carry heat. This allowed us to derive a simple model to extract lifetime from the measured thermal conductivity that gives results in very good agreement with our experimental findings.
All these crosschecks gave us confidence about the validity of our result, and our work was made possible through an impressive collaboration within the European C-MAC network across 24 laboratories in several European countries.
How important was this? How difficult was it to achieve?
The collaboration was set up within the European C-MAC network (www.eucmac.eu) that gathers about 24 laboratories across Europe. The network is dedicated to the study of complex intermetallic compounds. The collaboration was essential as no laboratories had the required skills to conduct the experiment successfully alone.
Cutting-edge experimental and theoretical facilities were required for single crystal growth in addition to atomic structure determination including the disorder, physical property measurements, inelastic neutron scattering measurement, and atomic scale simulations; expertise that belongs to different teams. It is only because all of those ‘ingredients’ were gathered together that the result could be obtained. This result demonstrates the importance of collaboration at the European level, which certainly goes against the current research policies that largely promote competition between teams.
Are there plans to now build on this research? Where will future priorities lie?
While we have evidenced a very long phonon lifetime, the atomic scale mechanisms responsible for this phonon lifetime are not yet understood: whether it is a result of collisions with defects, with the structural complexity, or with other phonons remains to be determined. This requires both further experiments on the same sample but also on other samples, and new atomic scale simulations. New theoretical inputs certainly need to be incorporated in current atomic scale simulations and into theory. Our experimental results, however, put severe constraints on any modelling.
We hope to derive general principles explaining thermal conductivity in structurally complex materials.
Dr Marc de Boissieu
This article will appear in Pan European Networks: Science & Technology issue 25, which will be published in December, 2017