Statistical Physics
- From: Christophe <chatelai@xxxxxxxxxxxxxxxx>
- Date: Mon, 30 May 2005 16:49:56 +0000 (UTC)
In the microcanonical ensemble, one assumes that all microstates with the
same energy have the same probability. Following Jaynes, this hypothesis is
the best choice when nothing is known about the system apart from energy
conservation. Defining two subsystems, one can then show that the
probability of a microstate of the first subsystem is the Boltzmann weight
exp(-\beta E_1)/Z provided the second subsystem is much bigger.
Thermodynamic averages calculated with this probability distribution are in
amazing agreement with experiments if there is no additional conservation
law.
Why does it work so well ? Let us consider for example a system at
equilibrium with a thermostat. If one could measure experimentally in which
microstate the super-system system+thermostat is and reconstruct the
probability distribution from the time series of these measurements, I do
not see any reason why one should always obtain a flat distribution.
Despite of this, one recovers experimentally the thermodynamic averages
predicted by theory.
Does anybody know experiments where there is no conservation law but where
thermodynamic averages are not in agreement with canonical ensemble
predictions ?
Thank you by advance,
Chris.
.
- Prev by Date: Re: Why physicists should pay attention to the mind
- Next by Date: Re: Wave Function of the Universe?
- Previous by thread: upper/lower *theoretical* limits for photon wavelength
- Next by thread: Boundary conditions in lattice QFT
- Index(es):
Relevant Pages
|
