Fats, sugars, theobromine, and caffeine are important compounds present in chocolate, capable of influencing the chemical, physical, and sensory characteristics of products. These compounds are generally determined using conventional analytical methods that require lengthy analysis times and the use of chemicals with significant environmental impact.

In this context, the aim of a recent study conducted by a group of Hungarian researchers (Benes et al., 2025) was to evaluate the use of near-infrared spectroscopy (NIRS) in combination with partial least squares regression (PLSR) to predict the compositional characteristics of chocolate.

The results demonstrate that the proposed method can accurately determine fat (coefficient of determination, R2, of 0.98 and concentration range of 30.79–63.55 g/100 g), sucrose (R2 of 0.92 and concentration range of 0.55–46.55 g/100 g), and theobromine (R2 of 0.94 and concentration range of 3.32–9.66 mg/g) contents. Predicting caffeine, however, was difficult, likely due to the low concentration of this compound (0.37–1.11 mg/g). However, the model’s performance could potentially be improved by incorporating additional samples or using alternative data preprocessing and variable selection methods.

In conclusion, the authors argue that the developed technique can be usefully used as a rapid and environmentally friendly quality control method for chocolate products.

Bibliographic references: E. Benes et al., Food Chemistry, 469, 2025, 142562