Uniformity of dielectrics

Hi guys,
I am interested in the optical properties of dielectrics at the moment. There are different types of dielectrics, which are often classified by their homogeneity or uniformity. From what I have read, uniformity is defined as follow: a dielectric is non uniform if the size of its particles are large compared to light's wavelength. I also read that homogeneity is not a good criterion, because a material with particles smaller than the wavelengths of visible light will has to be considered as inhomogeneous, although it behaves uniformly.
At first it sounded good to me. But then I thought about Rayleigh scattering. This occurs with particles that are much smaller than the wavelengths of light, doesn't it? THen surely that sould bread down the validity of the definition of uniformity, since particles smaller than light wavelengths generate a (weak) scatter.
Is that definition rubbish then? Any comments?
Thanks in advance
Enrique

Enrique Cruiz wrote:
  Quoted Text Here  
A dielectric "uniform" to kilohertz, megahertz or gigahertz electromagnetic waves may not be "uniform" to 10**14 hertz (1 with 14 zeros) visible light.
So, depending upon how you use it, a dielectric may or may not be homogeneous.
Are you talking about an optically used dielectric or a radio/TV used dielectric?
Then, you should realize that the above discussion is off the mark. In a polycrystalline material, there is almost no Absolute definition of homogeniety...... except functional.
It is easy to mix up apples and oranges and the result is rarely tangerines.
l

Thanks for the answer.
  Quoted Text Here  
Yes I am talking about an optically used dielectric.
  Quoted Text Here  
So, from what I understand, there is no absolute definiton of uniformity. On the other hand, you can talk about the concept of uniformity if the specifications of the studied applications are well defined. Is that right?

  Quoted Text Here  
I am interested in dielectric's optical properties. Is that ok then to define uniformity by comparing the size of the particles in the dielectric to the wavelengths of light?

Enrique Cruiz wrote:
  Quoted Text Here  
How do you propose to treat pores or voids of either the same size as the particles or smaller than the particles?
Are the pores randomly (uniformly) distributed, or are there gradients in the amount of porosity? That could make the material inhomogeneous, no matter what the ratio of the pore/particle size relative to the electromagnetic wavelength.
You need to introduce uniformity itself into the definition.
Isotropic ??
Then too you might require that the "performance" of the polycrystal should be identical to that of a pore free single crystal of the same substance (if it is polycrystalline and if the particles have isotropic optical properties, such as cubic materials symmetry). If the particles are anisotropic, then there is no trivial relationship between the properties of a pore free aggregate and the bulk.
The material may be statistically uniform with statistically predictable small scale deviations from the mean.
If the particles are distributed (and oriented) uniformly, then the response may be considered uniform.
You are trying to use the term "Uniform" for something, and you aren't clear on just what it is that you are seeking.
You should think and read other definitions of "uniform", rather than the snippets you have picked upon.
  Quoted Text Here  
invariable, invariant, regular, same, steady, unchanging, unvarying. See same/different/compare.
  Quoted Text Here  
analogous, comparable, corresponding, equivalent, like(2), parallel, similar. See same/different/compare.
  Quoted Text Here  
So, which of the above are you trying to use with the term?

  Quoted Text Here  
A good example of this is Rayleigh scattering off molecules of air. Although this is a bit controversial (and is not what any of the dumbed down explanations tell you), I was taught that this is due to density fluctuations (as predicted by statistical mechanics), which are more pronounced at high altitude/low pressures. Which would make it not a strictly molecular effect. I'm certainly no expert on scattering theory, however.

Enrique Cruiz wrote:
  Quoted Text Here  
http://www.lzos.ru/en/glass_nocolor_quality.htm
Optical Homogeneity Optical homogeneity of glasses and other optical materials is understood as homogeneity of the refractive index within a volume of material and can be defined by the maximum difference of values n for different areas of a piece . Breach of the homogeneity can be caused by chemical and physical inhomogeneities. Chemical inhomogeneities are striae which arise during glass melting and glass working and are conditioned through variations of glass composition because of incomplete melting or volatilization of some components, eating of glass melting furnace walls, and poor homogenization of melt. Striae are transparent, vitreous inclusions, which are thread-like, cord-like or knot-like formations with threads split off from them. Residual stress and structural non-uniformity account for variations in physical properties. Stress in glass initiates birefringence. Measurement of the refractive index value across a blank's field is a very difficult and labor-consuming process, requiring precision quality surfacing and application of complex device equipment. Therefore, in practice indirect methods of homogeneity estimation are used via setting of a strict tolerance system for striae and birefringence ensuring high level of optical homogeneity for supplied glass. Blanks for fabrication of optical components for high-accuracy collimators' objectives, for micro lithography, for measurement and astronomical devices , whose residual wave aberrations systems are comparable with distortions induced by glass inhomogeneity, are usually tested by interference methods judging by the wave-front distortion after a light beam passing a blank. For such components requirements for optical homogeneity are stipulated additionally when preparing the order.
http://amsglossary.allenpress.com/glossary/browse?s=o&p 
optically homogeneous—Homogeneous on the scale of the wavelength of the electromagnetic radiation of interest. Pure liquid water is optically homogeneous over the visible spectrum because one cubic wavelength of water contains many molecules, whereas a cloud of water droplets is not optically homogeneous. Optical homogeneity is more general than transparency, usually restricted to visible wavelengths. A body is said to be transparent if it transmits images. Optical homogeneity is necessary for transparency but not sufficient. A sufficiently thick sample of an absorbing optically homogeneous material would be described as opaque rather than transparent.
http://www.heraeusoptics.com/uploads/100005_datasheets/100033.pdf
The optical homogeneity, which is the main criteria for very low transmitted wavefront distortion, refers to three categories: