Quantitative Evaluation of the Lifshitz-Type Temperature Effect on the Casimir Force ()
1. Introduction
We consider the ideal situation of two parallel metallic plates presenting in the gap between them a reflection power of 100% and containing a vacuum at absolute zero temperature. In the spirit of Casimir’s initial work [1] the attractive force between the plates should give macroscopic access to the zero-point energy in the vacuum. Naturally these conditions are never met in nature, but as a minimum extension, the zero temperature can be removed. As is well-known, the Lifshitz theory of the force between dielectric bodies yields, with dielectric constants taken to infinity, in the limit
Casimir’s result
(1)
with
the plate distance, and the force that per unit area between the plates. In the case of finite temperatures we apply here a version of Lifshitz’s theory due to Dzyaloshinskii and Pitaevskii (1959) and discuss numerical results not found so far in the literature.
2. The Lifshitz Calculation
As reported in ref [2] , the Lifshitz force between parallel dielectric plates is derived by introducing Matsubara Green’s functions involving the Matsubara pseudo-frequency
(2)
Considering the limiting case of perfect metals, with dielectric constants taken to infinity and
the Casimir value of Equation (1), this expression leads to the result
(3)
with the value of the parameter
, and the prime indicating that for the term with
a factor 1/2 has to be inserted.
A derivation of the Equation (3) can be found in many places not listed here.
Notice that this expression depends on the combined parameter
.
Making the substitution
, valid for →0, the resulting double integral over p and
can be done analytically, yielding unity for the R value.
As a first step we have evaluated the quantity R numerically in ref [3] without the term
. The results, designated as
are shown on Figure 1(a), representing the variation of this quantity with temperature for several plate distances.
The term
has to be calculated separately. Taking the proper limit of the Equation (3) one finds as shown in the appendix
(4)
Adding these values to the results of Figure 1(a), we obtain for
the curves of Figure 1(b) showing an increase of the force with temperature. In the case, with the highest value i.e.
, the Equation (4) applies, since in this case
is negligibly small. This limit, corresponding to a linear temperature variation of the force, can be found in the literature [4] [5] . It is however nonrealistic since much smaller distances must be used to obtain a measurable force.
3. Discussion
In actual experiments one is far away from the ideal conditions presupposed in these calculations. In particular, in order to obtain small separation distances,
Figure 1. (a) Ratio
as function of temperature for several values of plate separations; (b) Ratio
for the same parameters as in Figure 1(a).
very special geometries have to be used. Despite these facts it seems that the temperature effect is still at the limit of observability, although progress has been made as shown e.g. in [6] .
According to our results, as long as very small distances are involved, temperatures up to room temperature don’t compromise the observation of the force. This could be different if measurements at larger distances could be made. In that case a force-increasing temperature effect would be more visible. Note that during the last 10 years, mostly experimental results have been presented for which a complete list can be found in reference [6] .
4. Conclusions
In this note we only want to stress the fact that a strict application of Lifshitz’s theory in its Dzyaloshinskii version confirms the older results obtained by more conventional methods.
But in addition we present exact numerical results, not found so far in the literature. However, the usual discussion of the temperature effect does not show clearly the fact that limiting expressions do not correspond to realistic measurable values of the force.
Appendix
We want the limit of the expression
(A1)
for
.
Changing integration variables by setting
,
,
, we have
(A2)
yielding for
(A3)
Using tables of integrals one finds
involving Riemann’s zeta function. Given the values
one finally obtains
(A4)
This numerical result remains valid if in Equation (1) s is replaced by xs with x any finite positive number.