Press release


A physical paradox:
how to cool an object by “heating” it

Paris, November 13, 2002


Physicists at Lyon’s Institut de Physique Nucléaire (CNRS/IN2P3(1) and the University of Claude-Bernard Lyon I), working in collaboration with the Institut für Lonenphysik of Innsbruck (Austria), have experimentally established, for the first time in a direct manner, the existence of a negative calorific capacity. The system, a finite set of 15 hydrogen molecules, suddenly lost heat during its liquid/gas transition, despite the contribution of external energy.

The phenomenon is commonplace. The temperature of water rises when it is heated. But as soon as the water begins to boil, its temperature ceases to rise until all the water has been turned into vapor. This behavior is typical of macroscopic mediums. Exposing a system to additional energy will increase the temperature of the system until it begins to change phase (i.e. evaporation or fusion); even though it is constantly heated, it will maintain a constant temperature until the evaporation (or fusion) is complete. In physics, this characteristic is referred to as the positive calorific capacity(2) of a system.

But, are there systems with a negative calorific capacity – that is, systems which lose temperature when heated? As early as 1970, the theoretical physicist Hans Thirring postulated such was the case with stars, whose core temperatures increase as they gradually lose energy via radiation. At the other end of the scale, the nuclei of atoms also drop in temperature just when they are vaporized.

A team from Lyon’s Nuclear Physics Institute, in collaboration with the Institut für Lonenphysik of the University of Innsbruck, has just demonstrated, for the first time in a direct manner, the existence of a negative calorific capacity in hydrogen aggregates, small units made up of about 15 molecules. Their experiments consisted of studying a number of collisions between a hydrogen aggregate and a fixed helium atom and detecting the resulting fragments of each collision. The collisions are of interest because they made it possible to deposit a significant amount of energy on the aggregate for a sufficiently short enough period to prevent the system from reacting. As the size of the resulting fragments changed with the amount of energy deposited on the aggregate, the researchers were able to verify a drop in temperature during the vaporization of the aggregates.

Between a gas and a liquid
In contrast to macroscopic systems, where the two phases (gaseous and liquid) can coexist, there can be no clear distinction between the two phases in a system made up of such a small number of particles: it is thus either gaseous or liquid. As soon as it is found in an intermediate state, which by nature is unstable, it seeks to return to a gaseous state “as soon as possible,” even if that means a temperature loss similar to that previously observed. In the macroscopic world, where a large number of molecules are involved, a sudden drop in temperature of all the system’s molecules is not possible; thus the two phases are observed coexisting at the same temperature while only a small fraction of molecules, caught between the liquid/gas interface, are affected by the transition phase.

Somewhere between macroscopic and microscopic

But one question remains unanswered. If such is the case, exactly where lies the fine line that divides the microscopic from the macroscopic world? The question is not merely academic. It has acquired importance in light of the current development of nanotechnologies, where the miniaturization of devices depends on electronic switches that may be composed of just a few atoms.

(1) Institut National de Physique Nucléaire et de Physique des Particules
(2) A system's calorific capacity is the energy it requires to increase its temperature by one degree Kelvin


Physical Review Letters, Vol. 89,183403, October 28, 2002

What is temperature?
Determining the temperature of microscopic systems is far from trivial. In this case, it is measured by the statistical distribution of the size of the aggregates detected after a high-speed collision. This measurement implies the conventional definition of the word “temperature” according to statistical thermodynamics: temperature is a scale that represents the average kinetic energy of a molecule with a system. Once again, the question of the fine line between the macro and the microscopic world is posed: to what extent can we still refer to a statistical distribution of kinetic energy?

Researcher contacts:
Bernadette Farizon
Tel : +33 4 72 44 83 89
e-mail :
Michel Farizon
Tel : +33 4 72 44 84 01
e-mail :

IN2P3 contact:
Dominique Armand
Tel : +33 1 44 96 47 51
e-mail :

Press contact :
Martine Hasler
Tel : +33 1 44 96 46 35
e-mail :