Proteins in white wine can cause cloudiness or haze after bottling, which consumers may consider an indicator of poor quality. As a consequence, winemakers often use bentonite, a clay-based material that binds protein, to remove proteins and achieve protein stabilisation. However, removing bentonite from wine after treatment can result in a 3-10% loss of wine.
Membrane filtration technology is used in wine production for many purposes and ultrafiltration (UF) offers an easily-translatable process for protein removal. UF treatment of wine can produce heat-stable permeate and protein-enriched retentate, which enables targeted protein degradation. Heating the retentate, with or without protease significantly improved the heat stability of recombined wine in pilot scale trials. This study evaluated strategies for achieving protein stabilisation using membrane filtration.
Sauvignon Blanc wine (unfined) was fractionated by UF in triplicate, the resulting retentate subjected to protease and heat (62℃, 10 min) treatment, and the treated retentate recombined with the permeate. Traditional bentonite fining was performed as a positive control. Chemical and sensory analyses were carried out to evaluate the efficacy of treatment.
Heating retentate with protease reduced the concentration of haze-forming proteins by 54% compared with heating alone 40%. Chemical analyses and quality scores for recombined wine showed no significant difference with bentonite-fined wines. Sensory analysis indicated that UF/heat-treatment increased the green apple aroma, alcohol heat and overall flavour intensity of the wines compared to bentonite fined wines, suggesting UF-treated wines retained flavour without imparting oxidative characters.
Ultrafiltration combined with heat and protease treatment can reduce bentonite use without significantly affecting sensory properties. While results are promising, it is not yet a viable alternative to bentonite fining.
Poster presented at Macrowine virtual (June 23-30, 2021)
(1,267 Mb)
OPEN AND READ THE POSTER
