Evaluating the Effects of Sterile Membrane and Other Filtration on the Sensory and Chemical Properties of Wine

The goal of our project is to evaluate the effects of sterile membrane and other filtration on the sensory and chemical characteristics of wine. To do this, we have filtered two red wines, a Cabernet Sauvignon and a Merlot, and one white wine blend through 0.45 µm PVDF and PES membrane filters and compared the sensory and chemical characteristics of these wines to unfiltered control wines. Treatments were expanded with the Merlot and white blend to also examine the effects of a pad filter and cartridge depth filter used as prefilters. We have examined changes in dissolved oxygen content, tannin content, and color during the course of filtration and found only minor changes. Sensory panels were trained for each of these wines and each of the treatments evaluated immediately after filtering and then on a regular basis for 9 weeks (for the Cabernet), 24 weeks (for the Merlot), and 20 weeks (for the white blend). While all three wines changed significantly over time in the bottle, very few significant differences were observed in aroma or mouthfeel between filtration treatments. In other words, our results thus far indicate limited impact of sterile filtration on the sensory or chemical properties of the wine, regardless of the type of filter material used. We did observe a small decrease in tannin and astringency with the pad filter for the Merlot wine, but this may not be due to the filtration treatments. In addition, we have evaluated the effects of filtration of both a red and white wine using a cross-flow filter. No effects on sensory characteristics were found for the white wine. Cross-flow filtration of the red wine seemed to stabilize the flavor profile of the wine, compared to the control that took on earthy characteristics, due possibly to growth of another organism. We also studied the effects of the three types of pumps typically used in wineries and found no significant effects of pumping either for short or extended periods of time. Finally, we evaluated the use of various gases for rinsing bottles prior to filling and their effect on the bottle shock phenomena observed. No differences were seen between bottles rinsed with argon, nitrogen, or carbon dioxide. Differences were observed when oxygen was introduced into the bottles.