Title:Does the second law of thermodynamics really hold good without exception?

Abstract:A major part of the many thermally driven processes in our natural environment as well as in engineering solutions of Carnot-type machinery is based on the second law of thermodynamics (or principle of entropy increase). An interesting link between macroscopically observable quantities of an ensemble (state variables) and the thermal velocity of its individual constituents such as molecules in a liquid is provided by the Brownian motion of suspended larger particles. We postulate a "frustrated Brownian motion" that occurs if these particles get partially trapped in an environment of suitable geometrical conditions. This dissipates a small fraction of the kinetic energy attended with the Brownian motion and deposits it inside the trap, and so this constitutes a mechanism that by itself transfers thermal energy from cold to warm. We note that this is just a marginally admitted, slowly evolving effect driven by a thermodynamic quasi-equilibrium, thus being of limited efficiency in terms of energy density attainable per unit of time. However, a simple experiment suggests that this process indeed is allowed to take place, and the envisaged applications then are straightforward.