in EDX the depth of the penetration (and therefore the region from which characteristic X-rays are generated = 'interaction volume') is related to the energy of the incident beam and the density of the material. It is possible to estimate the beam paths within your material by using a Monte-Carlo Simulation program. Have a look at the link I have provided for some suitable software.

Generally for any given material the interaction volume will increase with increased incident energy and for any given beam energy the interaction volume will decrease with increasing mean atomic number. So by using the Monte-Carlo simulation you can see the effect of varying the beam energy in your material on the size of the interaction volume. 

I have attached a couple of images ( I have guessed at your composition - but this is just for an example) of simuations using 10kV and 5kV, both at the same scale (2um scale bars). I hope this is useful for you.

http://www2.arnes.si/~sgszmera1/html/monte_carlo.html

Dear Abul F. Huq,

Energy Dispersive X-ray Spectroscopy (EDS) can providing chemical analysis in areas as small as 1 µm in diameter, and can be used to map regions of your sample. However, because of the depth penetration of x-ray is quite high, this techique is not capable to provide you with depth analyses. The following link provides a small article with radiation penetration.

Hi Abdul, in addition to previous answer,  Energy-dispersive X-ray spectroscopy (EDS/EDX)  (in polymer science) is composition characterization technique that uses x-ray excitation of electrons to discrete energy levels to quantify chemical composition. This technique provides characterization at surface depths of 1–10 nanometers, approximately the range of oxidation in plasma and corona treatments. In addition, this process offers  the benefit of characterizing microscopic variations in surface composition.

If you need depth resolution and you are not interested in lateral spatial resolution, you could try to use Total Reflection XRF.

Or else PIXE - particle induced x ray emission. It will probe through matter much deeper than EDS or XPS.  

in EDX the depth of the penetration (and therefore the region from which characteristic X-rays are generated = 'interaction volume') is related to the energy of the incident beam and the density of the material. It is possible to estimate the beam paths within your material by using a Monte-Carlo Simulation program. Have a look at the link I have provided for some suitable software.

Generally for any given material the interaction volume will increase with increased incident energy and for any given beam energy the interaction volume will decrease with increasing mean atomic number. So by using the Monte-Carlo simulation you can see the effect of varying the beam energy in your material on the size of the interaction volume. 

I have attached a couple of images ( I have guessed at your composition - but this is just for an example) of simuations using 10kV and 5kV, both at the same scale (2um scale bars). I hope this is useful for you.

http://www2.arnes.si/~sgszmera1/html/monte_carlo.html

According to Ian's pictures you can have generation depth from 0.2 to 2 microns, depending on accelerating voltage. Accelerating voltage should be chosen depending on elements you have in your specimen. If it is just plane regular polymer, I doubt very much EDS will be of use to you. 

By using confocal micro-XRF you are able to perform planar/depth/3D scanning. The spatial characteristics is than depending on the primary beam diameter/shape and focussing optics characteristics (focus diameter). Combination of the two will define your spatial resolution. However by this analysis you need to account for absorption of both, the exciting beam and produced secondary radiation. Since the secondary radiation will be of lower energy than the primary, it will determine the effective depth from which you will be able to collect the data. Moreover, this depth will be depending on the element (since different elements will produce characteristic radiation of different energy).

What elements do you want to detect from the regular polymers? X-rays usually penetrate deeper than electrons (0.2-2 um may be as provided by Ian depending on energy of electron beam and effective atomic number of sample matrix). If you want better "information depth" you can try confocal X-ray fluorescence with higher energy photons or even confocal PIXE (generally available at accelerator facilities).

Thanks all for their contributions. It appears from the answers that we can expect depth resolution of the order of microns.  Dhrubajyoti, I would like to get percent compositions. Filipe, I don't see a link that you talked about. Thanks

So, can the EDS analyzer perform elemental analysis to 1-2 microns depth? We did EDS analysis for the composite coating, but we were not sure how deep the measurement was.