Determination of α-HMX Phase Impurity Content in HMX by Near Infrared Spectroscopy

For rapid detection of α‐HMX phase impurity content in HMX ,based on the preparation of samples for building model ,the mathematical model between HMX spectra and α‐HMX impurity phase content was established with partial least squares (PLS) method using near infrared spectroscopy technique .The representativeness of sam‐ples for building model ,the selection of the spectral range and the optimization process of the model were discussed . The results show that the model has a broadly representativeness ,the prefered spectral ranges are 6476-6446 and 4602-4424cm -1 ,the correlation coefficient of cross‐validation is 0 .996 ,the root mean square error of cross‐vali‐dation(RMSECV) is 0 .20% ,and the root mean square error of prediction (RMSEP) of external‐validation is 0 .27% .The error of the method is less than 0 .13% and the standard deviation is 0 .1% .The method has the ad‐vantages of simple operation ,rapid ,non‐destructive and green enviromental protection ,and can be used to detect the α‐HMX phase impurity content in HMX .