Lecture 13: Electron Beams and Applications

Important applications of electron beam technology includes + X-ray generation + High power microwave generation + Microscopy + Materials modification and processing X-ray generation The basic premise is that x-rays are generated when an electron beam strikes a target. The x-rays then pass through an object and form an image on a piece of film.

x-rays are attenuated, by absorption and scattering, in the target. The parameter describing this attentuation is the linear attentuation coefficient (mu).

The minimum feature size that can be imaged by an x-ray point projection source can be estimated by a simple analysis. Say an incident photon flux N_o [#/(s*cm^2)] falls on a feature of length l, with x-ray mean free path of lambda(E), which is a function of x-ray energy.

The smallest feature, Delta-l, corresponds to the minimum change in photon flux distinguishable from noise, Delta-N. From Poisson statistics, Delta-N ~ sqrt(N).

+ So the minimum resolvable feature size depends on the x-ray mean free path in the material (twice the mfp at a given energy) + the mean free path depends on the linear attenuation coefficient and the total mass attenuation coefficient for the x-ray of a given energy and the object material linear attenuation coeff

total mass attenuation coeff

+ the mean free path is related to the linear attenuation coefficient

+ There is no general expression for these coefficients + The mass attenuation coefficient is high at low energies, falls to a minimum and then rises again at higher energies. + For compounds, the mass attenuation coefficients can be used to combine by weight percentage

+ These values are tabulated for different materials, or can be found in the literature.

+ For high intensity radiography, relativistic electron beam diodes are usually used. + Four common varieties, paraxial, self-pinch, rod-pinch, magnetically immersed.

Paraxial diode

Self-pinched (left) and Rodpinched (right) diodes

Magnetically immersed diodes

All from Maenchen, Proc IEEE v92, n7, 2004.

High Power Microwave Generation The discussion starts with waveguides. + These are ducts that guide waves + Similar to optical fiber (which are also sometimes called waveguide) Maxwell's equations can be reduced to the wave equations. For a vacuum filled region, possibly bounded by conductors or dielectric, we have the following.

We could have three possible solutions, depending on boundary conditions. + Transverse Electromagnetic (TEM) -- E and B vectors are perpendicular to the direction of propagation, k. Get these in free space. Also hollow coax, or parallel plate waveguides, i.e. transmission lines. + Transverse magnetic (TM) -- B is perpendicular to k. Occur in hollow rectangular or circular waveguides. + Transverse electric (TE) -- E is perpendicular to k. Occur in hollow rectangular or circular waveguides. Using a Fourier-Laplace transform, where d/dt -> j(omega), we can write the following.

At a perfect conducting wall, however, the parallel E and perpendicular B must go to zero.

If we assume E and B have wave solutions (equivalent to the Fourier-Laplace transform idea), we will have fields like the following.

Using Maxwell's equations, we can show that + the perpendicular components of E and B (either the r and theta components, or the x and y compoents) can be expressed in terms of the axial field component (consult Humphries or Benford). + if E_z and B_z are both zero (TEM) the transverse components are also zero, and the field in waveguide is trivially zero. Hence no TEM waves in these. Rewriting the wave equations

Assuming separability

v = omega/k > c, always

cutoff: when k=0, or no propagation, omega = omega_co

+ Derived from transverse field boundary conditions + Must have parallel E and perpendicular B equal to zero at the wall + Apply these BCs to the TM or TE wave equations to find the k_perp rectangular waveguide, of x-dimension = a, and y-dimension = b.

for a cylindrical waveguide, of radius rho, B_r=E_theta=E_z=0 at radius rho. Solutions involve Bessel functions of the first kind.

In summary + Waveguides support TM or TE waves, with TM waves having an axial E + Waves come in specific modes, with specific field patterns + Each mode has a cutoff frequency, below which no EM energy can travel in that mode + Each mode travels at a phase velocity > speed of light Slow wave structures + Waveguides confine and conduct EM energy, but at faster than light speeds + If charged particles are to interact with these fields, need to slow down the waves + Slow wave structures slow down EM waves confined within them + Slow EM waves are produced within waveguides with a periodically varying boundary

waveguide

slow wave structure

Some periodic variations include a perturbed wall radius, a periodic array of resonating cavities, or a helical winding within the tube.

Consider a wall with radius that varies periodically

The frequency of normal modes is a periodic function of the wavenumber k_z

The field can be written as a sum, using Floquet's theorem

We can show then,

However, k_perp,m won't correspond to that derived for smooth waveguides, it will be a more complicated function. The end result on the dispersion relation is to make it periodic in k-space. Now sub-c phase velocities are possible.

Microwave generation + Microwaves are generated when space charge waves on a beam interact with normal cavity and waveguide modes + Fast wave interactions occur when the interaction involves a waveguide mode phase velocity faster than light + Slow wave interactions occur when the interaction involves a waveguide mode slower than the speed of light + O-type devices have an electron beam that travels parallel to any externally imposed magnetic field + M-type devices have electron beams traveling perpendicular to crossed electric and magnetic fields + Space charge devices have intense, space charge dominated electron beam interactions

+ We have dealt so far with electron beams that are uniform in the axial direction, or direction of travel. + Imagine a perturbation on the axial beam density

+ This perturbation gives rise to a perturbation in the potential and self-generated electric field along the beam + The perturbation forms a space charge wave that travels along the beam + If the space charge wave phase velocity matches the phase velocity of a normal mode, they can exchange energy -- the perturbation grows, and feeds more energy into the EM normal mode, etc. O-type interactions + The space charge waves on an electron beam also have a dispersion relation

+ The (+) in the dispersion relation gives the fast wave, and the (-) give the slow wave. + Slow space charge waves result in microwave interactions + How do we get a wave with phase velocity below c, to interact with a EM mode? Fill the waveguide with dielectric (dielectric Cerenkov maser)

Use a periodic slow wave structure (backward wave oscillator)

M-type interactions + Electrons move in an ExB drift + interact with an array of cavities + Magnetron

Space charge devices Vircator - an intense electron beam, exceeding the space charge limit enters a grounded gap. Oscillations ensue producing radiation