Flash presentation: Advancing butter quality analysis: a comparative study of four near-infrared spectrometers across diverse spectral ranges.

This study evaluates the performance of four near-infrared (NIR) spectrometers—ISC, JDSU, FOSS, and Aurora—for analyzing key nutritional components (Moisture and fat content) in butter. Each spectrometer operates within a distinct spectral range: ISC (1100–1600 nm), JDSU (900–1650 nm), FOSS (850–2500 nm), and Aurora (950–1650 nm). These differences in spectral range provide a unique opportunity to compare the effectiveness and accuracy of each device for routine food quality assessments. Ninety-six butter samples were collected over a period of 12 months, at different plants located from the low land of the Po Valley, up to the higher altitude of the Alps. Each sample was scan in borosilicate glass and then in disposable plastic petri dishes. Spectra collected in plastic petri dishes showed distinct absorption peaks, overlapping those typical of fat showed The results for glass petri dishes showed variation in prediction accuracy across the instruments, particularly in relative prediction errors (RMSE/mean). Among the devices, FOSS and Aurora demonstrated the highest accuracy with RMSE around 0.5-0.6% and RSQ of 0.93. ISC and JDSU had lower performances that was similar between these two instruments, with RMSE around 1-1.15% and RSQ of 0.7-0.8. Despite the fact that spectra collected with plastic petri dishes had marked signal in the fat spectral region, calibration performances for Fat and Moisture had an increase in prediction error of 10-20% compared to the glass petri dishes. While the compact spectrometers (ISC and JDSU) exhibited reduced analytical sensitivity compared to the benchtop FOSS model, they performed sufficiently well for routine analysis of butter composition. These portable devices are cost-effective and practical for field-based or small-scale operations, where ease of use and portability are priorities. Aurora, being versatile in its spectral range, offers a middle ground between accuracy and practicality. On the sample presentation, the use of disposable plastic dishes greatly simplify routine work compared having to wash and reuse glass dishes, which, even with the slight increase of error, makes it easier to implement in routine analysis.