UC Davis

This dissertation focuses on gaining fundamental understanding of the extraction dynamics and sensory quality of full immersion brewed coffee. The overall goal is to use a diffusion-limited mass transfer model following a “universal brewing curve” to capture and predict the strength, measured in total dissolved solids (TDS), and extraction yield, denoted E, of full immersion brews under various brewing conditions, and to relate these physical properties (TDS and E) to sensory properties using sensory descriptive analysis to help achieve desired cup quality of full immersion brewed coffee.The dissertation is composed of four main stages. First, the equilibrium values of TDS and E for full immersion brews were investigated over a range of brewing parameters and interpreted using an equilibrium desorption model. The results indicate that the equilibrium TDS is approximately inversely proportional to the water/coffee mass brew ratio, while E is independent of the brew ratio. Surprisingly, the species-averaged equilibrium constant K was insensitive to grind size, roast level, and hot brew temperature over the range 80-99°C. Furthermore, an analysis of the standard oven-drying method for measuring E demonstrated that it yields significant underestimates of the true value at equilibrium, due to retained brew within the spent moist grounds. The second main stage involves the dynamics of extraction. A diffusion-limited mass transfer model was updated and extended to predict the TDS and E of full immersion brewed coffee in terms of three parameters: an initial “rinse” concentration (C_0), an equilibrium concentration (C_eq), and an effective mass transfer coefficient (k). The model was validated with a series of systematic brewing experiments using various coffee preparation factors, including varied coffee types, roast levels, brewing temperatures, grind sizes, and agitation rates. The experimental results indicated that (i) increasing brew temperature strongly increases the extraction rate, (ii) light roasted beans extract more quickly than dark roast, and (iii) coffee types from different origins using different post-harvest processing methods yielded no statistically significant differences in extraction kinetics. Furthermore, our agitation experiments demonstrated that occasional and slow manual stirring was as effective as continuous high-speed agitation in terms of increasing the extraction rate, and that the overall extraction rate was primarily limited by diffusion through the interior of the coffee ground. The third stage involved the impact of coffee ground sedimentation on the dissolved solids concentration in ostensibly quiescent (non-externally agitated) brews. Time-lapse photography observations indicated that dark roasted coffee grounds sediment more slowly than light roasted. Simultaneous measurements of the TDS versus both time and vertical position in the brew revealed that full immersion brews undergo a concentration gradient inversion midway through the brewing process. The time scale of the inversion depends sensitively on the roast level of the coffee grounds, which occurs up to a factor of three slower for darker roasts. A key implication is that variations in TDS due to different sedimentation rates could possibly confound sensory evaluations of coffees with different roast levels. Finally, the fourth stage explored the relationship between coffee sensory quality and extraction dynamics of full immersion brewed coffee. A descriptive analysis method was used to explore how the sensory profile of full-immersion brews evolved versus time for different roast levels and brewing temperatures. Our results indicate roast level had the greatest impact on the sensory profile of the coffees, followed by brewing temperature. Surprisingly the effect of brew time, particularly the longer brew time as the TDS approaches equilibrium, had subtler impact than expected at least for the range of brew temperatures and brew time points studied here. The intensity of sweetness was negatively correlated with TDS, while 19 other attribute intensities were positively correlated with TDS. To increase efficiency, coffee brewing practitioners can reduce brew time beyond equilibrium without changing their products. Interestingly, certain long time cold brews were found to have similar sensory profiles to those of some short time hot brews in light roast, suggesting the potential of achieving similar sensory profiles of long-time cold brews using short-time chilled hot brews through controlled preparation method. Overall, this dissertation aims to advance the understanding of how coffee preparation factors impact the extraction dynamics, and corresponding TDS and E, of full immersion brewed coffee, with the ultimate goal of helping coffee practitioners best obtain desired sensory qualities.