Evaluation of macro and micronutrient elements content from soft drinks using principal component analysis and...

Area of Science:

• Analytical Chemistry
• Food Science
• Environmental Science

Background:

• Carbonated soft drinks are widely consumed globally.
• Understanding the elemental composition of beverages is crucial for quality control and consumer safety.
• Brazilian soft drink market presents diverse flavors and manufacturing practices.

Purpose of the Study:

• To determine the concentration of nine essential and trace elements in various Brazilian carbonated soft drinks.
• To investigate the potential of multivariate statistical methods, specifically Principal Component Analysis (PCA) and Kohonen maps, for classifying soft drinks based on their elemental profiles and flavors.
• To assess the efficacy of different machine learning approaches in differentiating beverage types.

Main Methods:

• Inductively Coupled Plasma Optical Emission Spectrometry (ICP OES) was employed for the quantitative analysis of nine elements (Cu, Fe, Mn, Zn, Ca, K, Na, S, P).
• Multivariate statistical analyses, including Principal Component Analysis (PCA) and Kohonen maps (self-organizing maps), were utilized to explore relationships between elemental concentrations and drink flavors.
• Samples comprised carbonated soft drinks of various flavors and from different manufacturers within Brazil.

Main Results:

• Concentrations of the analyzed elements varied significantly across different soft drink samples.
• PCA indicated a tendency for grouping based on flavor (orange and cola), with cola showing distinct separation.
• Kohonen maps demonstrated a higher accuracy in classifying drinks into three flavor groups: cola, orange/lemon, and guarana.

Conclusions:

• ICP OES is a suitable technique for elemental analysis in carbonated soft drinks.
• Multivariate statistical methods, particularly Kohonen maps, can effectively classify soft drinks based on their elemental composition and flavor profiles.
• The findings suggest that elemental fingerprinting combined with machine learning can be a valuable tool for quality control and authentication in the beverage industry.