# Optimizing the Flavor Profile of Brazilian Spirits: Torrefaction Modeling of Native Woods for Cachaça Maturation

## TL;DR

This study explores how heating different Brazilian woods affects the taste of cachaça, a Brazilian spirit, and uses AI to optimize the aging process for better flavor.

## Contribution

The study introduces a novel AI-driven model to optimize wood selection and toasting for cachaça maturation, enabling customized sensory profiles.

## Key findings

- Each wood species develops unique sensory characteristics based on toasting parameters like time and temperature.
- A five-wood cachaça was produced with enhanced complexity and tailored sensory attributes using the predictive model.
- Current Brazilian practices of uniform toasting are challenged by the study's data-driven approach.

## Abstract

Cachaça, a traditional Brazilian spirit, undergoes significant sensory refinement through barrel aging. In this study, we investigated how heat treatment of Brazilian woods (Balsam, Jaqueira, Jequitibá, Amburana, and Ipê) affects the sensory profile of cachaça, using Oak as a benchmark. Physicochemical characterization, toasting assessments, sensory analysis, and artificial intelligence (AI) were integrated to develop a predictive model for optimizing wood selection and heat-treatment conditions to achieve targeted sensory profiles. Applying this model, we produced a five-wood cachaça, a novel spirit distinguished by its complexity and customized sensory attributes. This approach reveals that each wood species develops distinct characteristics depending on toasting parameters such as time and temperature, challenging the current Brazilian practice where a single toasting condition is applied to all woods without prior physicochemical analysis. Linking wood composition with sensory outcomes through AI, this work introduces an unprecedented product innovation and demonstrates the potential of multi-criteria analysis to guide spirit maturation, enhance product design, and reshape the beverage industry.