Developing process simulation models for red wine fermentation: Anthocyanin mass transfer and extraction

Abstract

Phenolic compounds extracted from grape solids during maceration are critical components responsible for the overall quality of red wine. Among these compounds are anthocyanins, which account for the red and purple pigments in young red wine that are crucial to the long-term colour stability for aging red wine. As such, understanding the variables that impact the extraction of these compounds and the mechanisms in which this process takes place are vital to the informed manipulation of red wine phenolic composition for the purposes of improving the quality or consistency of the finished product. A review of the literature presented in Chapter 1 revealed that anthocyanin extraction rate and extractability is influenced greatly by liquid phase properties (composition and temperature) as well as the available solid-liquid contact area and liquid phase velocity. This review also identified the limitations of previously developed phenolic extraction models for predicting the outcome of future extraction scenarios and explored the potential of other well-established mathematical models that could be applied to anthocyanin mass transfer during red wine maceration. This thesis sought to establish new methods for modelling the extraction and mass transfer of anthocyanins during red wine maceration for the purposes of predicting and controlling the outcome of future winemaking scenarios. To this end, Chapter 2 describes the development of a mathematical model based on solutions to Fick’s second law of diffusion, utilising mass transfer variables in both the solid and liquid phases to model the extraction of malvidin-3-glucoside (M3G), the predominant anthocyanin across all red/black grape varieties, from pre-fermentative grape solids. This model was then applied to experimental extraction curves of M3G under forced convection at liquid-phase conditions simulating various stages of red wine maceration and fermentation in order to solve for relevant mass transfer parameters. Response surface methodology was applied to the internal diffusion coefficient and distribution constants solved using this method, allowing the estimation of these parameters at varying temperature, sugar and ethanol conditions within the design of the experiment. Analysis of variance showed that both temperature and ethanol as well as their combined interaction have a significant effect on the internal diffusion rate of M3G, while all individual factors (sugar, ethanol and temperature) as well as their combined interactions significantly influenced the systems distribution constant. Chapter 3 explores the impact of liquid-phase convection (forced and natural) on M3G extraction under simulated red wine fermentation conditions. This allowed for the calculation of the external (liquid-phase) mass transfer rate of M3G at various stages of red wine fermentation under convective conditions that more closely resemble liquid-phase conditions that occur during the majority of a red wine maceration regime, excluding mixing operations. Calculated rates of external mass transfer under natural convective conditions showed that both internal (solid-phase) and external (liquid-phase) mass transfer limited the overall extraction rate. By combining these calculated external mass transfer coefficients with the response surface models generated in Chapter 2, predictive simulations of M3G extraction in dynamic liquid phase conditions (fermenting must) were conducted at various rates of external mass transfer. These simulations yielded previously observed but as yet undescribed extraction patterns, whereby during active fermentation the extraction rate and maximum extractability is limited by the extent of fermentation. This new insight provides a rationale for previous studies monitoring the effect of pre-fermentative maceration techniques, whereby no significant increase in anthocyanin concentration could be observed despite having a prolonged maceration period. Chapter 4 further explores the impact of liquid phase conditions on total anthocyanin extraction from a different grape variety, Pinot noir (notorious for its difficulty in achieving a high level of extracted colour), using similar methods for modelling and simulation of fermentation scenarios. Predictive simulations of fermentation scenarios were explored, highlighting the impact of controllable process conditions that can be manipulated by winemakers, including nutrient concentration and fermentation temperature. Finally in Chapter 5, the mathematical model and mass transfer parameters determined throughout Chapters 2-4 were applied to real world fermentations under both laboratory and industry scale conditions in order to validate the models’ predictive capabilities under various conditions. Predictive simulations of anthocyanin extraction in this study showed good agreement with commercial red wine fermentations indicating the models robust ability to predict extraction profiles using a relatively small amount of fermentation data. These simulations represent a significant step forward towards the ultimate goal of process control of red wine phenolic extraction during fermentation. Collectively, this thesis enhances the understanding of the extractive behaviour of anthocyanins during red wine production, quantifying the impact of process variables that can be manipulated by winemakers and those that change as a result of fermentation. The mathematical models developed in this study could be used to predict the anthocyanin extraction potential in future red wine maceration scenarios based on fruit composition, providing new opportunities for real-time process control, equipment design, optimisation of product quality, and efficient use of fermentation infrastructure.