Theory

Cerenkov radiation has been generated by means similar to those proposed here, but not in the low-energy regime we will be working in. Nor have the calculations for a finite width, cylindrically symmetrical beam passing over a flat vacuum-dielectric boundary been done before now, even though Cerenkov radiation has been studied since the early 1900s [1].

In general, Cerenkov radiation (especially the emission angle) can best be explained by the Huygens construction. A better analogy can be made by considering the polarization induced in the dielectric as the particle passes through it. For a more quantitative treatment, an analysis of the radiated electromagnetic fields of the beam must be undertaken.

Huygens construction

The Huygens construction explains Cerenkov radiation as a phenomena very similar to the shockwave produced by supersonic jets. Suppose a particle emits a wave pulse every time t, and also that these pulses move out spherically at speed c'. Now allow the particle to move with speed v. If v < c', then the wavecrests merely bunch up in the direction of motion, but do not cross. When v > c' , however, the wavecrests do cross, adding up constructively at some angle θ, defined as

This wavefront moves out at the speed of light in the dielectric.

Dielectric Polarization

This explanation is closer to the actual matter at hand, as the wavecrest explanation is far less helpful when one considers the case where the particle is moving outside the dielectric (as is the case in the experiment at hand). It also does not address why “superluminous” motion would create a shockwave. Thinking about the polarization of the dielectric can help to explain this aspect of Cerenkov radiation a little better [1]. Consider that a charged particle, in passing by the atoms in the dielectric, momentarily polarizes them (pushing like charges in the atom away, and inducing a dipole state). Once the particle has passed, this polarized state collapses, causing each atom to emit Cerenkov radiation (if the velocity constraint is met, of course). For slow moving particles, the polarization is perfectly symmetrical, resulting in no electric field at long distances (and thus no radiation). When the particle is moving very quickly (beyond the constraint , hereafter written with the abbreviation as ), however, the polarization is no longer perfectly symmetrical.

Electron passing through a dielectric at .

Electron passing through a dielectric at .

The state is still symmetric in the azimuthal plane, but no longer along the axis of motion (a “cone” of dipoles develops behind the electron). There would now be distinct dipole field established in the dielectric, one that can only be collapsed with the emission an electromagnetic pulse (Cerenkov radiation). Radiation would be emitted perpendicular to the surface of this cone, with the angle of the cone found using the Huygens construction.

Theoretical Analysis

The following treatment of the problem makes several simplifying assumptions (from the Frank and Tamm theory detailed in [1]):

(i) The dielectric is considered as a continuum, in as much as the material is defined simply by its dielectric constant with all micro-structure ignored.

(ii) Dispersion is ignored.

(iii) Radiation reaction is ignored.

(iv) The material is considered to be a perfect isotropic dielectric, with no absorption of radiation (conductivity is zero, permeability is unity).

(v) The beam is considered to move at a constant velocity (no slowing due to ionization, of Coulomb scattering).

(vi) The material is unbounded and the beam path is infinite.

Electron sheet passing over a dielectric.

Even given these assumptions, there has been relatively little theoretical work done on the experimental geometry considered here. The closest work would be that of Danos [2, 3], who considered the motion of an infinitesimally thin sheet of charge moving at some distance from the vacuum-dielectric boundary. The following derivations follow the work presented in [2].

One begins by defining a charge density ρ

where

and, as before,

Also, is the Fourier coefficient of the charge density corresponding to the frequency , and is the average DC charge density. Of course, this is for a bunched sheet, moving in the z-axis, with zero extent in the x-axis (perpendicular to the surface of the dielectric).

The potentials in free space (above the vacuum-dielectric boundary) are

where is constant and is the z-axis unit vector.

Considering only the AC part of the charge density, one arrives at

where

To solve the homogeneous equation one makes the trial solution

By the Lorentz condition, , it follows that . Thus

The vacuum fields that go with the electron beam are

From above

Here signx is defined as

and

Above, the quantities and are parameters that can be adjusted to match the boundary conditions.

The vector potential inside the dielectric is just

where

From this the fields can be calculated

Now one can impose the boundary condition (the tangential field must match) at

From here one can find the equations of the outgoing wave

where

and

The electron beam radiates an average energy defined by the Poynting vector

where

Poynting vector decomposition.

The normal component of this vector corresponds to the energy radiated into the dielectric:

where

with N being the number of electrons in one coulomb (). Thus this component of the Poynting vector is in watts per centimeter squared if I is in amps per centimeter.

One should also note that the angle of the radiation () is given by

To simplify things, lets put the above derivation in terms of known quantities:

If the beam sheet in this example has width, this can be applied in the Poynting vector as such (in simplified notation)

where is measured from the surface of the dielectric.

A similar analysis was conducted by Linhart [4] for a single electron passing over an infinite plane interface between a vacuum and a dielectric (with similar constraints as above with the addition of and ). The total radiation output was found to be

where e is the charge of the electron, and the energy radiated per unit angular frequency is

Sketch of frequency dependence of power.

Consequently, there exists a frequency, and thus a wavelength, at which most of the radiation is emitted:

Summary

It has been shown that the general design used in this experiment, namely the passing of an electron beam over a dielectric, will produce Cerenkov radiation given the correct dielectric. This was done by Danos [3] when he generated microwave Cerenkov radiation by passing a 10 KeV electron beam over a plate of titanium dioxide . Since we are using a low-energy beam of ~3KeV, we need a dielectric constant of at least 85. This is found by

Of course, one can move easily from a potential difference like 3KV to a value for using

Solving one gets

 

Next Page

      | Website designed by Farzad Sadjadi

Mac and the Mac logo are trademarks of Apple Computer, Inc., registered in the U.S. and other countries. The Made on a Mac Badge is a trademark of Apple Computer, Inc., used with permission.

Google