Feasibility Research on Non-Invasive Analysis of Tanreqing Injection with Near Infrared Spectroscopy
Wenlong Li,Zhiwei Cheng,Yuefei Wang,Haibin Qu
DOI: https://doi.org/10.1255/jnirs.1022
2012-01-01
Journal of Near Infrared Spectroscopy
Abstract:A non-invasive analysis method for bottled Tanreqing injection was developed based on near infrared (NIR) spectroscopy. The NIR trans-mission spectra were collected using a home-made accessory. The quality indicators of Tanreqing injection, which include baicalin, ursodesoxycholic acid (UDCA), chenodeoxycholic acid (CDCA), chlorogenic acid (CA), forsythin, total amino acids (TAA) and soluble solid content (SSC), were determined with standard methods. Several outlier samples were eliminated using the Dixon test, leverage and studentised residual test and analogous samples were amalgamated. Partial least squares regression (PLSR) was used for establishing the calibration models with the pre-treated spectra and the reference values. The root mean square errors of prediction ( RMSEP) values of external validation samples were 0.098 g L−1, 0.059 g L−1, 0.140 g L−1, 0.038g L−1, 0.007 g L−1, 0.053 g L−1 and 0.538 gL−1 for baicalin, CA, UDCA, CDCA, forsythin, TAA and SSC, respectively, which indicated that the established models possessed satisfactory predictive abilities. The presented method can reduce the workload of product inspection to a large extent. This will allow the sampling proportion to be increased and enable improved quality control during Tanreqing injection production.