You choose a good quality standard powder sample (available from NIST I guess) with relative large paricle size and defect free. This should give delta shaped diffraction peak with practically no width but due to the instrumental broading this will really give reflections with finite width whch is due only to he instrumental broadening of the diffractometer used. You will thus get from the measured diffraction profiles of the standard samples 2thta dependent instrumental broadening and line sphases experimentally.
I agree with Prof. Tapan Chatterjee. However, The sample needs to be scanned slowly. The instrumental broadening is generally done while determining particle sizes of nano crystalline powder samples. But one has be very careful while determining such particle sizes of a film sample. Not only instrumental broadening but stress could also contribute. You may refer to our paper on nanocrystalline CaSO4:Dy.
You can also calculate the instrumental broadening from the instrumental geometry of your diffractometer. The relevant formulae is available in standard books of X-ray diffraction. If you have the B.E. Warren's book X-ray diffraction then go to p. 257 section 13.2 (Dover edition) and read it carefully. If you use neutron diffraction then there are old paper by Caglioti which give formulae for calculating instrumental broading or resolution function. These are all calculations based on sume assumptions but work pretty well. But I prefer to measure it experimentally by using standard samples which you can buy.
Please try to find paper by Johnson and Hindeleh from 1975 - 1980 times for the original description of dealing with this problem. It was published in Polymer in those years. The precise year may be 1976
Collect data from NIST LaB6. Use the smallest step size you can manage reasonably and collect for long enough to get reasonable signal to noise. You assume that LaB6 is 100% crystalline and that all broadening is instrumental. Samples with reflection widths larger than this include some sample broadening.
Here is an example of using Bruker LEPTOS to generate an "instrument profile" or "strain free state profile" for Si (113)-. This "standard" profile is then used to compare Bragg profiles for a LiF2 mono-crystal topographically. The resuts are interesting at the least :-)
In our case we were using a Bruker HRXRD system with the appropriate optics. The software requires minute details of the optics and diffractometer dimensions in order to simulate the most meaningful information.
The challenge generally with powders or polycrystalline specimen/standards is the ubiqutous presence of "preferred orientation" and non-spherical coherrent diffracting domain "shape". Never-the-less, any material may be used as "standard" as long as its Nano structure doesn't change with time, environment etc. The NIST standards are universally demonstrated to provide superior homogeniety and precision. Good idea to run the standard and sample contemporaneously if possible.
BTW silicon would probably work better than silica :-) LaB6 works great due to its high diffraction efficiency. Moisture sensitivity is important when considering powder "standards"