†Corresponding author. E-mail: huangyongs@gmail.com
*Project supported by the National Basic Research Program of China (Grant No. 2011CB808104) and the National Natural Science Foundation of China (Grant No. 11105233).
As an important QED effect to detect the vacuum polarization, birefringence in the presence of a strong electric and magnetic field, E0 ⊥ B0, E0 ≤ cB0, is considered. The directional dependence of birefringence is obtained. In two special cases: E0 = 0 and E0 = cB0, our results are consistent with the previous ones. The refractive index of the probe wave propagating in the E0 × B0 direction decreases with E0/ cB0, while that in the − E0 × B0 direction increases with E0/ cB0. The physics of the direction dependence of birefringence maybe the E0 × B0 drift velocity of the virtual electrons and positrons.
In the strong field,
In general, in the field (E, B) of a tightly focused low-frequency laser, | E| < c| B| . If the laser frequency is smaller than 1 × 1020 rad/s, the scaling length of the field is much larger than the Compton wavelength of electron. In that case, the field can be equivalent to the homogenous field and the effective Lagrangian 𝔏 eff[5] is available. Therefore, birefringence in a crossed field (E0 ⊥ B0 and E0 ≤ cB0) should be reconsidered. In order to check birefringence in a more general regime, we consider the interaction between a high-frequency electromagnetic wave and a strong electromagnetic field:
In this paper, we show that the refractive index of a forward wave with the wave vector k ‖ E0 × B0 is different from that of a counter propagating wave in a vacuum with a strong homogeneous static electric and magnetic field, solving the Maxwell equations corrected by the effective Lagrangian[5]
where E and B is the total electric and magnetic field respectively, using
and
which are valid for | E| , c| B| ≪ Ecrit ≈ 1018 V/m and the frequency of the field ω ≪ ω e ≈ 1020 rad/s. The quantum parameter,
First, the external field is assumed to be
Let E = E0 + Eq and B = B0 + Bq be the total electromagnetic field, and Eq and Bq are the nonclassical parts which are expressed as
Let | E1| , | E2| , | cB1| , | cB2| ≪ | E0| , and | cB0| , where c is the light velocity. The wave vector
where α 0 = 1 + 2ξ L̂ 0, c0 = E0/B0, β 0 = c0/c, and
With the wave assumptions that E1, 2, B1, 2 ∝ exp[i(kz − ω t)], the linearization of the corrected source-free Maxwell equations
gives
With Eq. (8), the dispersion relationship for parallel polarization is obtained after some straightforward algebra
With Eq. (7), the dispersion relationship for perpendicular polarization is given by
Let n = ck/ω be the refractive index. With Eqs. (9) and (10), the refractive index of a probe wave propagating in the positive z direction, i.e., in the E0 × B0 direction for parallel polarization and perpendicular polarization, is respectively given by
and
Similarly, the refractive index of a probe wave propagating in the negative z direction, i.e., in the − E0 × B0 direction for parallel polarization and perpendicular polarization, is respectively given by
and
With Eqs. (11) and (14), the difference of the refractive indices for the parallel polarization and the perpendicular polarization in the E0 × B0 direction is obtained as
In the − E0 × B0 direction, it is obtained as
Equation (15) shows that the difference of the refractive index for the parallel polarization wave and the perpendicular polarization wave is proportional to ξ . Here the wave propagates in the positive z direction, i.e., E0 × B0. Equation (16) shows the refractive index for the parallel polarization wave and the perpendicular wave propagating in the negative z direction, i.e., − E0 × B0. The results predict the difference of the refractive indices is proportional to ξ . Figure 1 shows the drift effect on vacuum birefringence in a homogenous strong crossed field. The birefringence is amplified by the drift effect for the counter-propagating probe wave and is reduced for the probe wave propagating in the direction of E0 × B0.
To compare our results with that of the literatures, two special cases are considered: (a) E0 = 0, i.e., vacuum birefringence in the presence of a homogenous magnetic field; (b) E0 = cB0 in a crossed field.
For E0 = 0, c0 = β 0 = 0, L̂ 0 = − 1, and the refractive indices are
for parallel polarization, and
and the difference of the refractive indices is
All the results are the same as those of Refs. [10]– [13]. Therefore, in a homogenous magnetic field, there is no drift velocity of virtual particles and there is no direction dependence of the refraction.
Another well-known case is E0 = cB0, β 0 = 1, and c0 = c. Therefore, the refractive indices in the E0 × B0 direction are trivial and n+ , par = n+ , per = 1. There is no interaction between the probe wave and the crossed field. It is also consistent with Eqs. (16) and (17) in Ref. [19] and Eq. (15) in Ref. [20] in the E0 × B0 direction.
However, for the counter-propagating probe wave, the refractive index is given by
for parallel polarization, and
which is also consistent with Eqs. (16) and (17) in Ref. [19] and Eq. (15) in Ref. [20] in the − E0 × B0 direction.
The refractive indices of the probe wave in the E0 × B0 direction and in the − E0 × B0 direction are different from each other. The difference of the refractive indices for parallel polarization and perpendicular polarization are also different in the two directions. Define
which tends to be zero as E0 → 0.
The physics of the direction dependence of the refractive index maybe the E0 × B0 drift of the virtual electrons and positrons in the strong field. The virtual electrons and positrons have the same E0 × B0 drift velocity,
In the experiment to measure the birefringence, the light is reflected and propagates through the field forward and backward repeatedly for N times. Set the longitudinal length of the field as L, the elipticity value[14] is given by
which is
In conclusion, vacuum birefringence is different in the E0 × B0 direction and in the − E0 × B0 direction. The E0 × B0 drift of the virtual electrons and positrons may be the key point. The influence of the drift effect on the ellipticity in the experiments has been given explicitly.
We are grateful to Felix Karbstein for the interesting and enlightening discussion.
1 |
|
2 |
|
3 |
|
4 |
|
5 |
|
6 |
|
7 |
|
8 |
|
9 |
|
10 |
|
11 |
|
12 |
|
13 |
|
14 |
|
15 |
|
16 |
|
17 |
|
18 |
|
19 |
|
20 |
|