Still not enough information. For example, suppose I had a sample of copper plating bath solution -- basically a mixture or copper chloride, sulfuric acid, and proprietary organic modifiers. I could roast that to dryness, char the remains, and calculate the amount of organic by back calculating the mass of the copper oxide. (Note: I probably wouldn't do it that way, but this is just an example) This method would be worthless for an accurate account of say, a food stuff. But there may be an assay like you described that is adequate for some food stuffs, maybe, like pet food often has a listing on its label % ash, as a general indicator of quality. You seem to be doing something like the second one, but want much better results, but you're not specific -- what elements do you need to quantitate, how accurately do you need to know them, and what is the matrix these elements are in , so we can help you with problems and pitfalls.