Abstract ​Stephan SommerJutta KeiserSeth Cohen

Rapid and Matrix-Independent Method to Analyze Anthocyanins in Red Grape Juice and Wine by FT-MIR

Stephan Sommer,* Jutta Keiser, and Seth Cohen
*Appalachian State University, 730 Rivers St., Boone, NC 28608 (sommers@appstate.edu)

Qualitative and quantitative analysis of monomeric anthocyanins can be used to evaluate enological treatments or identify the grape variety or age of a wine. Analysis of food and beverages by Fourier Transform Mid-Infrared Spectroscopy (FT-MIR) has increased in recent years. This indirect fingerprinting method allows easy and rapid determination of multiple attributes at a very low cost per sample. In practical industry applications, the main disadvantage of FT-MIR is the necessary differentiation between juice and wine matrix because high sugar or ethanol levels have distinct effects on the accurate detection of most analytes. This often leads to defective data, especially in young and still-fermenting wines. Anthocyanins, in contrast, have a variety of properties that allow analysis independent of the matrix because they are often bound to sugars and acids, which makes the whole molecule more accessible to MIR spectroscopy. Over 300 grape and wine samples were selected, processed, and analyzed for their anthocyanin composition. Reference analysis was performed by HPLC-DAD, and the basic calibration for the FOSS FT 120 WineScan was set up with the FOSS CalibrationMaster using Partial Least Square Regression modeling. The validation and quality control tools show that several parameters like total anthocyanins, malvidin-3-glucoside, and the sum of acylated anthocyanins can be determined with sufficient accuracy as they are present in high enough concentrations. The coefficient of determination (R2) in relation to the reference method ranges from 0.73 to 0.85 for these parameters. Other minor components, for example single anthocyanins, mostly lack that precision, although current literature suggests otherwise. However, the quality of the correlation can be improved by including more samples into the calibration. The extremely low cost of this method compensates for the possible lack of precision for industry applications.

Funding Support: none