Using experimental design to optimise precision of steam distillation for determining alcoholic strength in spirits

Abstract

An experimental design, in combination with oscillation-type densimetry is introduced as a novel procedure to optimise steam distillation for the determination of alcoholic strength in different types of alcoholic beverages. A central composite design was chosen to study the effects of variation in levels of receiver volume (25 and 50 ml), steam power (30–100%), distillation time (20–140 s) and sample volume (5–50 ml). Three commercial spirit drinks with different alcoholic strengths (4.9, 35.3, and 54.7% vol.) were compared on two automated steam distillation devices of different manufacturers by separately completing the experimental design for each. The models fitted for the prediction of the alcoholic strength, as indicated by an r 2 value of more than 0.93. Various surface plots were generated to describe the relationship between operating variables and predicted alcoholic strength. Optimum conditions were 50 ml receiver volume, 70% steam power, 100 s distillation time and 25 ml sample volume. Slight deviations from the optimum did not result in substantial decrease of alcoholic strength. Validation experiments carried out under predicted conditions showed excellent correspondence to the reference procedure (R=0.999).