1：Experimental Design and Method Selection:

A blending profile was obtained by mixing different amounts of PLA granules and talc powder. The calibration model was built correlating wet chemical data (alkali digestion method) and NIR spectra. Using FT-NIR technique, a Partial Least Squares (PLS) regression model was set-up.

2：Sample Selection and Data Sources:

Thirteen different blends were created, increasing the amount of talc (0%–100%) and decreasing the amount of PLA granules (100%–0%), so as to maintain constant the final weight at 3.53 ± 0.05 g at room temperature.

3：53 ± 05 g at room temperature. List of Experimental Equipment and Materials:

3. List of Experimental Equipment and Materials: NIRFLex N-500 (Büchi, Flawil, Switzerland), equipped with the Solids Cell Module (Büchi, Switzerland), BT200 Batch Homogenizer (Dynaken, Silangor, Malaysia), 930 Compact IC Flex (Metrohm, Herisau, Switzerland).

4：Experimental Procedures and Operational Workflow:

Samples were well mixed using a lab-scale powder mixing system and then submitted to NIR analysis. FT-NIR diffuse reflectance spectra of samples were collected with a NIRFLex N-

5：Data Analysis Methods:

5 The raw optical data were pre-processed with a Standard Normal Variate (SNV) transformation. The Partial Least Squares Regression (PLSR) was used to establish the relationship between the off-line wet chemical data and NIR values.