Novel smoothing routine for the data processing in thermogravimetric analysis

摘要

For a certain short interval of a TG scan, the correlation between mass and time is linear. A smoothing and filtering routine based on use of linear regression and error analysis was developed and successfully applied in the TG data processing. This method provides a filter to smooth the noise fluctuations and, at the same time, to introduce no distortions into the TGA experimental data. The computer program required is quite simple and effective. The method used in the program promises auto-convergence.