Rapid Analysis of Milk Using Low-Cost Pocket-Size NIR Spectrometers and Multivariate Analysis.

Jordi Riu,Barbara Giussani,Ricard Boqué,Giulia Gorla,Dib Chakif

doi:10.3390/foods9081090

Jordi Riu, Barbara Giussani + Show 3 more

Open Access

https://doi.org/10.3390/foods9081090

Copy DOI

Abstract

The miniaturisation of analytical devices, reduction of analytical data acquisition time, or the reduction of waste generation throughout the analytical process are important requirements of modern analytical chemistry, and in particular of green analytical chemistry. Green analytical chemistry has fostered the development of a new generation of miniaturized near-infrared spectroscopy (NIR) spectrometric systems. However, one of the drawbacks of these systems is the need for a compromise between the performance parameters (accuracy and sensitivity) and the aforementioned requirements of green analytical chemistry. In this paper, we evaluated the capabilities of two recently developed portable NIR instruments (SCiO and NeoSpectra) to achieve a rapid, simple and low-cost quantitative determination of commercial milk macronutrients. Commercial milk samples from Italy, Switzerland and Spain were chosen, covering the maximum range of variability in protein, carbohydrate and fat content, and multivariate calibration was used to correlate the recorded spectra with the macronutrient content of milk. Both SCiO and NeoSpectra can provide a fast and reliable analysis of fats in commercial milk, and they are able to correctly classify milk according to fat level. SCiO can also provide predictions of protein content and classification according to presence or absence of lactose.

Highlights

Cow’s milk is an important component of the human diet, with an estimated worldwide production of approximately 800 million tonnes in 2016 and growth expectations of 981 million tonnes in 2028 [1].From an economic point of view, around 150 million farms worldwide are involved in milk production.The composition of the milk largely determines its nutritional value and provides information on the health status of the cow
No thermostatic measurements or other pre-treatments were performed to better simulate rapid routine measurement conditions, some studies have reported that better results are obtained in multivariate calibration models after pre-treatments such as heating or sonication [33]
It is worthwhile to note that the number of samples is adequate according to ASTM E1655-17 [47], which indicates that the minimum number of samples for the calibration set must be equal to 6 (k + 1) for the mean-centred data and the prediction set must be equal to 4k

Summary

Introduction

Cow’s milk is an important component of the human diet, with an estimated worldwide production of approximately 800 million tonnes in 2016 and growth expectations of 981 million tonnes in 2028 [1].From an economic point of view, around 150 million farms worldwide are involved in milk production.The composition of the milk largely determines its nutritional value and provides information on the health status of the cow. Cow’s milk is an important component of the human diet, with an estimated worldwide production of approximately 800 million tonnes in 2016 and growth expectations of 981 million tonnes in 2028 [1]. From an economic point of view, around 150 million farms worldwide are involved in milk production. The composition of the milk largely determines its nutritional value and provides information on the health status of the cow. Milk has been analysed for decades and the standard methods of analysis in many cases are old and tedious. A common method for the determination of fat is the Gerber method (patented in 1891), which separates fats by adding sulphuric acid and reads the fat content using a calibrated butyrometer [3], and the reference method for the analysis of fats is a gravimetric method [4]. Instrumental techniques of analysis are widely used in milk analysis, and among them, chromatographic techniques are extensively used [5]

Objectives

Results

Conclusion