In addition to the suggestion by Dr. Towe you may also want to check out the NASA JPL ASTER spectral library which has 400+ rock spectra and 60+ soil spectra in the 1.4µm to 25µm range. You can try to compare the lab measured spectra with the known mineral spectra from the library. Usually there is a description associated with most of the mineral spectra found in the library which is quite helpful to understand the characteristics of the minerals based on the spectral signatures. You may also want to check out the TSG software. 

What is the aim of your rock and soil spectral measurements? Are you looking for some specific minerals?

http://speclib.jpl.nasa.gov/search-1

http://www.thespectralgeologist.com/

http://speclab.cr.usgs.gov/PAPERS.refl-mrs/refl4.html

Dear Sir,

Thank you for your reply . My aim behind DRIFT or FTIR measurements is to predict minerals present in sample possibly without aid of Xray diffraction. But as we know there occours overlap of mineral peaks in natural sample . Is there any other method appart from refering some spectra and interpreting?

Nitesh, I am not aware of any automatic technique using Infrared Spectroscopy. If you find out one, please let me know. If I were you, I would first begin with identifying spectral signatures of pure minerals. Only then I can understand mixed spectra from natural samples which is often the case in nature. Depending on the wavelength range I'm looking at, it will determine whether I can add linearly the pure spectra to derive the mixed spectra or vice versa. Take a look at the TIR spectra in Fig. 8-1 from the attached book chapter by Prof. Ramsey. I guess you need to do a lot of reading anyway. I must say identifying minerals in a mixed spectra is not easy, but if you want to learn, it is better to follow the long and tedious way rather than following some 'black box' method. I hope you are not calling me 'Sir' because I am also a student like you. 

http://www.pitt.edu/~mramsey/papers/GAC1.pdf

Nitesh,

First you'll need to have a good spectral library that is dedicated to the type of measurements you plan to make (i.e. reflectance, transmission, etc.). Once you have that I recommend looking for a software that does non negative least squares on full pattern. The first software that comes through my mind is EssentialFTIR, which unfortunately is not free. Otherwise you have to look for free packages written in Matlab or IDL languages. Search for NNLS Matlab or NNLS IDL for some free codes. It doesn't matter if they were tailored for other purposes (e.g. Fraunhofer spectra in astronomy etc.) it can work for your purpose once you transform your spectra (e.g. %T--> absorbance).  Then the problem reduces to mixing pure mineral end members from spectral library till you match your matrix spectrum. The same thing is done using full pattern technique in XRD, XRF etc. 

For pure DRIFT measurements I recommend CSIRO's Spectral Geologist which again is not free. A free software to do this is MGM (modified gaussian model) written in various languages see the description here http://www.planetary.brown.edu/mgm/LPSC99MGM.pdf and do some searches online to find the code, but works best with broad absorptions of olivines and pyroxenes. Good luck!