Agricultural Research Management Information System

Home
Research Summary
Search
About Us
- ARMIS
- Brochure
Contact Us
Report
User Request
Data Input
Help
- Operation Manual
  - PDF
  - Video
- Program Area & Commodity
We have reached 37600 number of research entries at this moment.

- Logout

Research Detail

Home
Research
Detail

Chemometric Model for Rapid Detection of Urea and Hydrogen Peroxide in Milk

M. Basak
Department of Applied Chemistry and Chemical Engineering, University of Dhaka, Bangladesh

M. N. Uddin
Bangladesh Council of Scientific and industrial Research (BCSIR), Dhaka-1205, Bangladesh

R. A. Jahan
Centre for Advanced Research in Sciences, University of Dhaka, Dhaka-1000, Bangladesh

A. A. Rana
Department of Applied Chemistry and Chemical Engineering, University of Dhaka, Bangladesh

M. M. Karim
Department of Applied Chemistry and Chemical Engineering, University of Dhaka, Bangladesh

Abstract/ Executive Summary:

A simple and cost effective method has been developed for determination of adulteration in milk with urea and hydrogen peroxide by using chemometric modeling with Fourier Transform Infrared (FTIR) spectroscopic data. Milk samples were purchased from a dairy farm (South Banasree, Dhaka, Bangladesh) and spiked at different concentrations of urea and hydrogen peroxide. Spectral data of all samples were collected using ATR-FTIR spectrophotometer. After acquisition of spectral data, they were preprocessed with transformation techniques such as multiplicative scatter correction (MSC) and savitzky-golay derivative. The predictive performance of principal component regression (PCR) and partial least-square regression (PLSR) methods were assessed by relative prediction errors and recoveries (%) were compared . PLSR shows better prediction efficiencies over PCR with R2 value 99% for urea and 95% for hydrogen peroxide. Six brands of commercial milk samples have been evaluated by this method and the samples contain 21.66-44.73 mg urea and 1.62-2.86 mg hydrogen peroxide in 100 ml milk. This method can be easily used in the quality assessment of milk.

Keywords: Milk, Fourier-Fransform Infrared spectroscopy; Multiplicative Scatter Correction; Savitzky-Golay derivative; Partial Least-Square Regression (PLSR)

Location: South Banasree, Dhaka, Bangladesh.

Start Date:

End Date:

Program Area: Quality and Nutrition

Commodity: Milk

Objectives:

To develop a method to determine urea combined with hydrogen peroxide adulterants in milk.

Methodology:

Raw cow milk was supplied by a dairy farm near South Banasree, Dhaka, Bangladesh. Six different local commercial samples were purchased from the local shops. Acetone (HPLC-Grade), urea and hydrogen peroxide with highest purity were used for spiking. Total 35 preparations (calibration set) of standard solutions (Table I)containing mixture of milk, urea and hydrogen peroxide were prepared for the development of a chemometric calibration model. Raw milk was mixed with urea and hydrogen peroxide in different concentration by following Orthogonal Experimental Design (OED). This experimental design was performed by using software SPSS (version 22.0). Transmittance measurement of the milk samples was carried out by using an ATR-FTIR spectrophotometer (Model- IR Prestige 21, Shimadzu, JAPAN). Spectral data of raw milk sample and standard samples were collected using Fourier transform infrared (FTIR) spectrophotometers in the wavenumber range of 4000-700 cm^-1. The instrument was standardized before each day’s analysis. Weighing of the samples was carried out by using a standard calibrated balance (Model-H0001, A & D Company Ltd., JAPAN). A vortex mixture of Benchmark Scientific Inc. USA (Model-BV 1000) was also used to mix the sample solutions. Obtained data were processed by a licensed copy of CAMO the Unscrambler (Ver. 10.5). Infra-Red spectra were recorded by transferring each sample (approximately 0.5ml) to the ATR plate attached to the FTIR spectrophotometer. Multiple spectra (three times) were collected for each sample. All spectra were acquired in the range of 4000 to 700 cm^-l at a resolution of 1cm^-l. Acetone was used to clean the ATR stage l and dried after each experiment to ensure a clean surface of the crystal so as to obtain the best possible spectra. The acquired IR data was preprocessed with some transformation to extract the meaningful information and to enhance the predictive capability for determining the analytical parameters out of the spectral data(Ahmad et al., 2016). This preprocessing is done for introducing changes in the values of the variables so as to make them better suited for analysis. A wide range of transformations can be applied to data before they are analyzed. In this study, transformation techniques such as Multiplicative Scatter Correction (MSC), Savitzky-Golay derivative and their combinations were applied and their efficiency was assessed. Multiplicative Scatter Correction (MSC) is a transformation technique that is used to compensate for additive and/or multiplicative effects in spectral data. MSC was initially intended to manage multiplicative scattering alone. However, a number of similar effects can be successfully treated with MSC, such as path length problems, interference, offset shifts, etc. Another transformation technique Savitzky-Golay algorithm depends on performing a least-squares linear regression fit of a polynomial around each point in the spectrum so as to obtain smooth data. It is showed by Savitzky and Golay that a set of integers could be derived and utilized as weighting coefficients to do the smoothing operation. In the study, Savitzky-Golay smoothing with 1^st and 2^nd derivatives and their combined treatments have been applied. Chemometrics that emerged from chemistry has presented new techniques equipped for managing a lot of compound information by means of multivariate data analysis. In recent years, multivariate chemometric strategies appear to be the techniques showing the best execution in terms of complex compounds. In each spectrum, there is an enormous number of absorbance values for each wave point. Each wave point or data point is considered a spectroscopic variable. These variables are tremendous in number and are mutually correlated. In this situation, Ordinary Least Square (OLS) method cannot be used as there is a problem of singularity. So we are going to use Principal Component Analysis (PCA), Principal Component Regression (PCR), and Partial Least Square Regression (PLSR) in our study and compared their prediction efficiency. After pre-processing the spectral data, PLSR has been used to predict different parameters of milk. A recently developed generalization technique of multiple linear regressions (MLR) is partial least squares (PLS) regression (Martens and Naes, 1996; Wold et al., 2001; Brereton, 2000; Naes et al., 2002). PLS regression is specifically noteworthy on the grounds that, unlike MLR, it can examine data with strongly collinear (correlated), noisy and redundant variables (X variables or wavelengths) and furthermore model several characteristics (Y values) at the same time (Lindgren et al., 1993; Tenenhaus et al., 2015; Fornell and Cha, 1994; Geladi, 1988).

Source of Information: Bangladesh J. Sci. Ind. Res.56(1), 1-8, 2021

Article Link (Online Journal): DOI: https://doi.org/10.3329/bjsir.v56i1.52689

Funding Source:

Budget:

Total Budget (Tk.) :

Achievement/Findings:

From this study, we can conclude that FTIR spectroscopic data and PLS regression can be used to determine urea (R²=99%) and hydrogen peroxide (R²=95%) as adulterant in milk when the spectroscopic data are pretreated with MSC and MSC+S-G filtering.The proposed technique is rapid, non-dangerous, straightforward, and easy to utilize. Compared to traditional methods, multivariate data analysis combined with FTIR instrumental techniques gives a new and better insight into complex problems by measuring a great number of chemical compounds at once. The chemometric methods developed in the present study can be used to analyze a large number of samples within a very short time with great precision.

Knowledge Management: Journal