Here we treat the classical oscillation of a toy Thomas-Fermi system.
Let n be the number density of spinless fermions of mass m and charge q. If we neglect relativistic eects (such as the Lorentz force, etc.) then we expect that u mn + (u )u = P nq nU, t where m is the mass of an electron, P is the degenerate pressure of the local electron gas, and is the eective local electrostatic potential caused by the charged particles and possibly the nucleus. u in the above is the bulk velocity of the particle cloud, which is traditionally assumed to be zero in the ThomasFermi theory. The pressure P is related to the local number density n through the quantum theory of free electron gas (details to be derived later). The U in the last term refers to any other external potential energy per particle that is still unaccounted for by , for instance, potential energy from some external conning eld or gravity. We also have charge conservation: n + (nu) = 0. t (1)
We next show how the static version of our formulation leads to the standard Thomas-Fermi theory. First, we derive the relation between P and n. From standard quantum physics, we know the kinetic energy of an individual particle inside a box of side L is PD 2 PD 2 h 2 h 2 j=1 lj j=1 lj = , = 2m L2 2m V 2/D where D (= 1, 2 or 3) is the dimension of the problem, and V is the size of a box in the D-dimensional space. The number of particles inside the box is N= X
lj
lmax
AD lD1 dl = AD
D lmax D
for some number AD depending on the dimension D, whereas the total kinetic energy is E = X
lj 2
lmax
0 ND AD
AD lD1 dl (D+2)/D
h 2 l2 h 2 1 lD+2 = AD max 2/D 2m V 2/D 2m D+2 V
h AD 2m
D+2
1 , V 2/D
and the Fermi energy is F =
2 h 2 lmax 2/D 2m V 2/D ND 2 AD h 2m V 2/D 2/D 2 h D n2/D . 2m AD
= = The pressure is
= =
E h 2 1 = AD V 2m D+2 D V (2/D)+1 (D+2)/D D 2 AD h 2 AD n(D+2)/D . 2m D+2 D
ND AD
(D+2)/D
(2)
We may view this as the equation of state which relates the pressure P of the gas to its density n, whether the system is in static equilibrium or in dynamical motion (at least in the classical sense). At static equilibrium, we have P qn nU = 0, 1 = P q U = 0, n (D+2)/D D AD h 2 2 D + 2 (2/D)1 n q U = 0, = AD n 2m D+2 D D (D+2)/D D AD h 2 D + 2 2/D = AD n q U = 0, 2m D+2 D
= (F ) q U = 0, = F + q + U = constant in space.
For the case U = 0, the last equation is seen to be exactly the assumption one imposes in the traditional Thomas-Fermi theory, provided q is set to e for electrons! Thus, our formulation does lead to the traditional theory when we consider the time-independent solution. Thus, it makes sense to investigate the dynamics of the system.
For the linearized problem we have mn0 u t P1 qn1 0 qn0 1 n1 U = P1 = = = = = = m u t =
n1 qn0 0 qn0 1 n1 U n0 n1 P1 + (P0 + n0 U ) qn0 1 n1 U n0 n1 P1 + P0 qn0 1 n0 P0 n1 n1 + P0 qn0 1 n0 n0 P0 n1 P0 n1 + n0 qn0 1 n0 n0 n0 P0 n1 P0 n1 P0 n1 + n0 n0 qn0 1 n0 n0 n0 n0 n0 n1 P0 qn0 1 n0 n0 n0 P0 n1 q1 n0 n0
where the subscript 0 means an unperturbed quantity, while the subscript 1 means it is a rst order perturbed quantity. For instance, P = P0 + P1 + , n = n0 + n1 + . Notice that we have a happy coincidence of ma = F = V here! In particular, we notice that this means the vorticity of the velocity eld u does not change with time, because ( u) u =m = 0. m t t I dont know yet if this may be of any use or signicance. The perturbation of the continuity equation (Eqn. 1) yields n1 + (n0 u) = 0, t which, together with Poissons equation and the equation of motion derived above, form the whole set of governing equations satised by the perturbed quantities u, n1 and 1 : 1 m u = n P0 n1 q1 , t 0 n0
n1 t
2 1 = q0 n1 , = (n0 u) . 3
