1530-437X (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/JSEN.2016.2530873, IEEE Sensors Journal IEEE SENSORS JOURNAL, PRE-PUBLICATION, 2016 1 A Low Cost NIR Digital Photometer Based on InGaAs Sensors for the Detection of Milk Adulterations With Water Maur´ıcio Moreira, Jos´ e Alexandre de Franc¸a, Dari de O. Toginho Filho, Vanerli Beloti, Alberto K. Yamada, Maria Bernadete de M. Franc¸a, Lucas de Souza Ribeiro Abstract—The adulteration of milk with water is a very com- mon problem, which reduces its nutritional value and industrial yield, and can also represent a major risk of contamination. The standard method for detecting this kind of fraud is the cryoscopy, which fails when the adulteration is done along with other substances. Therefore, there is a need for new and more robust techniques for this purpose. With this goal, we propose a prototype of a digital photometer to quantify water added to milk. This is a microcontrolled, portable device, which uses three LEDs with emission in the NIR region and was developed without the use of lenses, filters or moving parts. This equipment measures the transmittance of IR radiation through milk samples to assess the addition of water. In this paper, we present results of experiments that were conducted with diluted milk samples containing 0 to 25 % of added water, using the proposed equipment and the standard method of cryoscopy. In the measurements of the percentage of added water, the mean absolute error was less than 1 %. In terms of repeatability, in two sets of 10 measurements we obtained absolute deviations from the average corresponding to less than 0.7 % of added water. Results show that the prototype response is similar to the one of a commercial cryoscope, but faster. Index Terms—Near Infrared; NIR Spectroscopy; Photometry. I. I NTRODUCTION Milk, along with its derivatives, is a food widely consumed worldwide. Only dairy beverages accounted for about 18 % of the world’s consumption of beverages in 2012, and exceeded the consumption of water. The world milk production exceeds 600 million tons per year. Brazil, for example, is the second largest milk producer in the Americas, after the United States, and one of the largest producers in the world, with 30,715 million liters produced in 2010. Much of this amount comes from farmers who produce up to 50 liters per day, what char- acterizes it as a family activity. Most of these producers is part M. Moreira, J. de Franc¸a, M. B. de M. Franc¸a & L. de S. Ribeiro are with the Laborat´ orio de Automac¸˜ ao e Instrumentac¸˜ ao Inteligente, Departamento de Engenharia El´ etrica, Universidade Esta- dual de Londrina, Caixa Postal 10039, Londrina, PR, 86057-970, Brazil (e-mail: mauromauriciomoreira@hotmail.com, franca@la2i.com, mb- morais@uel.br,lucassribeiro12@gmail.com) D. de O. Toginho Filho is with the Departamento de F´ısica, Universidade Estadual de Londrina, Caixa Postal 10.011, Londrina, PR, 86057-970, Brazil (e-mail: darit@uel.br) V. Beloti & A. K. Yamada are with the Departamento de Medicina Vete- rin´ aria Preventiva, Universidade Estadual de Londrina, Caixa Postal 10.011, Londrina, PR, 86057-970, Brazil (e-mail: lipoa.uel@gmail.com) The authors would like to thank CAPES and Fundac¸˜ ao Arauc´ aria for the financial support for the research. of associations and cooperatives that distribute the product to dairy industries throughout the country. In such cooperatives, adulteration of milk with water is the most frequently observed fraud. Among other consequences, the addition of water to milk has a negative effect on the quality of derivatives, and increases the production costs. Besides the reduction in the nutritional value of milk and in its industrial yield, the addition of water can be a major risk of contamination, according to the conditions in which the adulteration is done. Unfortunately, since these are community associations and milk from different producers are mixed together, it is very difficult to detect which of them is responsible for the fraud. Thus, the losses could only be avoided with the use of a rapid, inexpensive and accurate method for quantification of the water added to milk. Currently, the most accepted method to detect water addition is the cryoscopy, which consists in estimating the amount of added water present in a sample of milk by measuring its freezing point. Although it is quite accurate, the cryoscopy is likely to fail due to the addition of other substances, such as urine, salt and other solutes which also change the freezing point and mask the addition of water. Accordingly, we intend to develop a new technique to measure the water added to milk based on near-infrared (NIR) spectroscopy, which consists in observing the absorption of electromagnetic radiation in this spectral region by samples of this product. The objective is that this system be fast, accurate, and more robust in fraud detection, since the interactions between electromagnetic radiation and the analyte depends on the types of molecules existing in the analysis medium. So, the purpose of this project is to develop a portable NIR digital photometer, based on LEDs, which could be used on a large scale by most of producer cooperatives. The molecular absorption bands in the NIR region have been used for qualitative and quantitative analysis several applications. For example, Li et al. [1] presented a quantitative measurement of the chemical compositions of gas mixtures, such as natural gas, based on near-infrared spectroscopy and hollow-core photonic bandgap fiber cell. The presented technique was considered viable for measurements of low gas concentrations [2]. Even more recently, Al Hosani et al. [3] used three methods based on near-infrared (NIR), mid- infrared (MIR), and Raman spectroscopy to acquire various spectra for online determination of very fine particles in a gas pipeline. In addition, Giusto et al. [4] used a white-light time- resolved spectroscopy system to monitor absorption changes