Information Extraction And Characterization Of Chemical Compositions Of Traditional Chinese Medicine By Vibrational Spectroscopic Imaging

Characterization of chemical compositions in complex mixtures has always been afundamental problem for analytical chemistry. Vibrational spectroscopic imaging (VSI),which can provide information about the molecular structure and spatial distribution ofchemical compositions in microscopic scale, has great potential in the characterizationof natural complex mixtures such as traditional Chinese medicine (TCM). Through thesystematical study of data collecting of infrared and Raman spectroscopic imaging andinformation extracting about major and minor compositions, the characterizationmethod, based on the combination of chemometrics and VSI techniques, for thechemical compositions in TCM was primarily established.Two new information extracting methods, i.e., PHA (principal component analysis–hierarchical clustering analysis–alterative least squares) and PTC (partial leastsquares–correlation coefficient), were established to analyze the VSI data. PHA wasdesigned for the non-target interpretation of the compositions without any priorinformation, while PTC was used to detect known compositions of interest. Basically,the limit of detection of PTC was better than PHA. In order to avoid the possible effectof chemical reagents on the original compositions, the direct VSI measurement ofdifferent types of TCM samples were studied. Attenuated total reflection infraredspectroscopic imaging was applied in the analysis of TCM samples for the first time,which provided mid-infrared spectroscopic images of various samples with clearspectral features by direct measurements.Using the previous sampling and interpreting techniques, the VSI methods werefirstly used to obtain the molecular structure information of chemical compositions inTCM materials. The active ingredient phillyrin in Forsythiae fructus, the adulterantingredient calcite in Margarita powder, and the characteristic ingredients calciumoxalate and starch in Aristolochia manshuriensis caulis were successfully detected,which proved the capability of VSI methods in the analysis of minor compositions andin the simultaneous detection of organic and inorganic compositions in TCM materials.Then, the VSI methods were used to characterize the spatial distribution of chemicalcompositions in TCM materials. The results showed the respective existence of-D-glucan and-D-glucan in the clumps and hypha of Poria. The testa of Arecae semen contained a lot of tannin, while the endosperm was composed of sugar, lipid, andprotein. The macroscopic and microscopic spatial distribution of eugenol and calciumoxalate in Caryophylli flos were also revealed.Based on the information extraction of chemical compositions in TCM materials,the VSI methods were applied in the recognition of chemical compositions andcompositional materials in finished TCM products. The objective identification methodsfor the powdered materials in Liuwei Dihuang Wan, Kouqiang Kuiyang San, andNiuhuang Jiedu Pian were established through the recognition of major and low-contentcharacteristic compositions such as starch, calcium oxalate, tannin, cutin, alisol,indigo, taurine, chrysophanol, As4S4, etc. by vibrational spectroscopic fingerprints.In conclusion, the methods based on VSI and chemometrics techniques for thecharacterization of molecular structure and spatial distribution of chemical compositionsin TCM samples is systematically studied in this research. The VSI methodssignificantly improved the analytical capability for the minor compositions in TCMsamples, and the inner relationship between the physical structures and chemicalcompositions of TCM samples can be studied in microscopic scales. The above resultsalso provide reference for the application of VSI methods in the analysis of other naturalcomplex mixtures.