Dear Mr. Sekar,

you refer to what was done in the past, as the so-called norm minerals were calculated from the chemical analyses. It was no proper mineral assemblage but an artificial one. There is no correct mineral analyses without a mineralogical analyses, involving a set of basic tools and methods. It starts off with a thin section or a polished section when the opaque components prevail over the transparent ones. The next step based upon this qualitative examination is to use methods like Rietveld XRD (which needs good crystallographic experience and a precise knowledge of the X structure). The MLA (mineral liberation analyses) under the SEM-EDX is a good tool but even in this case you have to know precisely the composition of your reference minerals. Rietveld XRD in combination with MLA most closely complies with your desire. A simple calculation of minerals from the XRF analyses is only possible in case of simple two- or three-component systems, e.g., an impure dolomitic limestone. I know that people want to obtain more and more the quantity of minerals of a certain rock or ore type straight from the XRF or ICPMS data. We have used software of different types during deep drilling to create down-hole mineral logs. It is sufficient for an overview but you cannot avoid a proper sequential mineral analyses if you want to get precise information. If your ore is of simple natural design with two or three minerals and your study is planned as part of a routine analysis it may work. But keep in mind even in an Fe ore the valence state is crucial in your calculation and may fraught your calculation with difficulties. The approach taken in such a case depends upon the accuracy of your results you want to accomplish and the complexity of the rock and ore type under study.

Best regards

Harald G. Dill

Dear Pragadeesh,

The most common instrument used to analyze minerals is the Electron Probe Micro-Analyser (EPMA), where one uses a carbon coated polished thin-section of a rock/ore containing the minerals. For common rocks the analytical results from the instrument are obtained in the form of oxide percentages of relevant elements (eg. MgO, CaO, Al2O3 etc.). If the sample is a sulphide ore, you can get the analytical result in the form of sulphide percentage of the relevant elements. It takes only about 3 minutes to get a complete analysis of a sample, even if you analyze 8-9 elements. As Kenneth  and Harald mentioned above, it is always important to know, at least qualitatively, the minerals present in your sample and the elements you are expecting, so that you can chose correct standards before you embark on the analysis. (Before analysis you will be first analyzing the standard sample to make sure that the instrument is giving you accurate results.)  If you know the elemental analysis, you should calculate the weight of the oxides that each element will produce when combined with oxygen. Then from the total oxide weights you can calculate oxide weight percentages. The same with sulphides.

Dear Pragadeesh:

Minerals are chemical compounds, just a few are native elements.The ONLY instrumental technique which gives the NAMES of the minerals analyzed in a rock sample, or in crystals separated from it, is X-Ray Diffraction. Electron microprobe, X-Ray Fluorescence, and many other analyitical methods only give a chemical composition of minerals, either expressed as weight percent of elements, or oxides, but they DON'T identify minerals in a direct way. In particular, the electron microprobe is not capable of analyzing water contents as H, and some other light atomic weight elements, such as B, F, Li, therefore the results obtained, for example, analyzing Mg-silicates, such as forsterite, enstatite, chondrodite or, serpentine, are quite similar: only peaks of Si, O, Mg, and some small peaks of Fe, are present, but no peaks of H or F. Instead the peaks in a diffractogram can lead, after very simple steps are taken, and with the use of appropiate tables, to the very name of the mineral, not just the chemical composition. If you have a list of oxide %´s of a single mineral species, it is quite easy to reconstruct an empyrical formula of that species. Numerical methods are given in Hurlbut & Kein (2002) Manual of Mineralogy, if you have a mixture of minerals, unless you now precisely the amount of each, if will be a quite difficult task. You said you had an iron ore (impure bauxite, I assume) containing 30% FeO and 40% Al2O3, and 30% of other oxides  (probably SiO2, TiO2, or MnO). The only way to know which mineral is present is the one I said before: X-Ray Diffraction powder method. The main costituents, probably goethite, hematite, gibbsite, and quartz, will produce identifiable peaks in the diffractorgram, othe minor species probably will be hard to identify. However, to ascertain the proportion of each species in the ore, the only way is to construct a calibration curve using known measured mixtures, of two main constituents, as the intensity of the peaks (or the counts in the detector)  is proportional, more or less, to the amount of each species present.  

The only instrumental technique which gives the bulk chemical analysis of  a rock  sample and best to use in your case is X-Ray Fluorescence (XRF).

XRD first step then you can use EPMA or EDX technique 

As per the suggestions given earlier, I got an XRD for a rock sample for identifying mineral composition.But I could not find the database for identifying the mineral composition from the internet. Please could you suggest how I can move further and complete my identification process?

Any data base to crosscheck the mineralogical composition?

Is quantitative analysis possible?

Yes, Kenneth M Towe and thank u so much, in this case the answer is XRD ;)

Dear Pragadeesh: a colleague of your, Hillol Chacravarty, said that he wasn't able to find an XRD database in the Internet to which compare the results obtained by him performing XRD of a rock sample. Well I know an excepcionally complete mineral database, which includes all the XRD peaks (listed and as a difractograms) and intraplanar distances of thousands of minerals, among other characteristics such as Raman spectra, IR spectra, and optical properties: it is called RRUFF. Simply write RRUFF in the Google finder and it will appear. This database is easy to use and extraordinarily efficient. There are many others, but RRUFF is my favorite. With regards, Sebastián Grande. Central Univ. of Venezuela, Caracas.

Developed electron micro probe facilities available in various organisations like national Geophysical Research Institute, Hyderabad ad.