I have a problem of analytical chemistry. We are working on a spectroscopy technique (FTIR) for the quantification of a mineral phase in complex matrices. We intended to use chemometric techniques to train models for the identification of different samples and proper estimation. The problem is to find the "true" value for our mineral phase to train the models. External and internal calibration are not good options because there is for sure matrix effects and the other mineral phases are generally not pure. Standard addition seems the logical choice. The problem is that we don't have a way to have a blank matrix without our target. Two reasons 1) the matrix might contain traces of the target; 2) the combination/possible matrices are just too many. Do we have other options?