consider balloon of volume V held within thermally-isolated vacuum chamber of volume 2V, although the gas expands in volume,
no work is done on the surroundings
entropy in a joule-free expansion: balloon bursts and
the gas is free to expand
entropy in a joule free expansion: for adiabatic walls
δQ is also zero so Uf=Ui
entropy in a joule free expansion - for an ideal as, U=
U(T) so there is no temperature change
entropy in a joule free expansion - in practice most gases
cool slightly and we can measure the joule coefficient
uj=(dT/dV)u
joule free expansion - initial and final states are
equilibria so delta s is well defined
joule free expansion - for an ideal gas with s(T,V), delta s=
cvlnT2/T1 + Rln V2/V1 = Rln2
so can have a change of entropy without heat input or thermal changes
in the context of the kinetic theory of gases, entropy is often described as
the disorder of the particles
hard to quantify
consider two dice, each with six degenerate states or microstates.
the macrostate in this scenario is
the sum of the dice
eg for macrostates =3 can have microstates (1,2) or (2,1)
in a fluctuating system which macrostate is the most probable
the one with the largest number of microstates
the joule free expansion suggests that entropy depends on
how a gas is distributed
boltzmann then formulated the link as S=S(omega)
where omega = number of microstates
as entropy is extensive, the entropy of the total system should be
the sum of the two entropies SAB=SA+SB
if the two systems are independent then the number of microstates should be
the product of the individual numbers
omegaAB=omegaAomegaB
for the composite system, Boltzmann’s hypothesis should obey
S(ΩAB) = S(ΩAΩB) = S(ΩA)+S(ΩB)
S=
KB ln Ω
microstates for monatomic gas
for gas particles in a box, each particle has position r and momentum p
the likeliness of a microstate i is proportional to its
boltzmann factor e^-Bei
B is the thermodynamic beta used to simplify notation
lower energy states are more likely since
the exponent is negative
the actual probability of each state is also correlated to
how many states the system has
this means the probability is given by Pi=e^-Bei /Z
Z norm const
to quantify the normalsiation constant, use
the property that all probabilities have to sum to unity
normalisation constant Z is called
the partition function
the expectation value for a property A can be found by
averaging over the allowed microstates
if you interpret different microstates as fluctuations, then you could consider this average as a
time-average over the fluctuations
link between thermodynamic potential and the partition function
F=-KBT ln Z
from F we can derive most other thermodynamic properties
link between statistical and classical approaches