Physics of X-ray Imaging
The design of an X-ray imaging system is difficult because of the
constraints imposed by the interaction of X-rays with matter. First, X-rays
impinging at normal incidence on any material are largely absorbed rather than
reflected. (It is this property that makes possible medical X-rays. X-rays
are absorbed by dense tissue like bone, and transmitted through less dense
tissue like skin and muscle. Thus, a medical X-ray shows the
"shadow" of the bone.) Normal incidence mirrors, like those used
for optical telescopes, are ruled out. Second, the index of refraction, n, is
~1 at X-ray wavelengths for all materials. Any refracting system
(i.e., lens) sufficiently thin to transmit X-rays must therefore
possess a long focal length; such a system would be highly impractical for use
on a rocket or satellite.
For most materials, however, the index of refraction is slightly less than
unity at X-ray wavelengths. This property offers the possibility of using
"total external reflection" of X-rays incident on a surface near
grazing incidence. The index of refraction at X-ray wavelengths may be
written:
n = 1 - d - ib
where d and b depend on the material and the wavelength of the incident
X-rays. If d > 0 and b ~ 0, and the incident X-rays are propagating in a
vacuum (for which n = 1), then by Snell's law X-rays will undergo total
external reflection for angles theta < thetac, where
cos(thetac) = 1 - d. Thus thetac ~ (2d)1/2.
The visible light
analogy to this phenomenon is the "total internal reflection" which,
among other things produces the glistening of a diamond. In that case, the
index of refraction of the diamond is higher than air, so light within the
diamond reflects efficiently off the various facets.
Generally, the dependence of d, and thus thetac of a material is
proportional to its atomic number, Z. Thus high Z materials reflect X-rays
more efficiently than low Z materials. The most commonly used reflecting
materials are gold and nickel, for which the critical angle at 1 keV is about
1 degree.
Contributed by Rob Petre of the Laboratory for High-Energy Astrophysics, GSFC
|