Virial expansion of molecular Brownian motion versus tales of "statistical independency"
arXiv:0802.0288
Abstract
Basing on main principles of statistical mechanics only, an exact virial expansion for path probability distribution of molecular Brownian particle in a fluid is derived which connects response of the distribution to perturbations of the fluid and statistical correlations of its molecules with Brownian particle. The expansion implies that (i) spatial spread of these correlations is finite, (ii) this is inconsistent with Gaussian distribution involved by the ``molecular chaos'' hypothesis, and (iii) real path distribution possesses power-law long tails. This means that actual Brownian path never can be disjointed into statistically independent fragments, even in the Boltzmann-Grad gas, but behaves as if Brownian particle's diffusivity undergoes scaleless low-frequency fluctuations.
21 pages, 1 figure, IOPART, to be submitted to JSTAT
Cited by in corpus (3)
- Thermodynamic restrictions on statistics of molecular random walks
- Short remarks on the so-called fluctuation theorems and related statements
- Hamiltonian Brownian motion in Gaussian thermally fluctuating potential. I. Exact Langevin equations, invalidity of Marcovian approximation, common bottleneck of dynamic noise theories, and diffusivity/mobility 1/f noise