ATR-FTIR Analysis for Oil Type Identification

Abstract

Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy (ATR-FTIR) analysis could greatly help in the identification of oil. By mere inspection of the IR spectra, saturated oils can be easily distinguished from coconut oils. Identification of oils can be enhance by the use of chemometrical techniques such Principal Component Analysis (PCA). Using this technique, subtle differences in the spectra can be used to classify vegetable oils. This study have shown that ATR-FTIR together with PCA can be used to differentiate unsaturated oils from saturated oil and discriminate virgin coconut oil (VCO) from ordinary refined, bleached and deodorized coconut oil (CCO).

Introduction

Virgin coconut oil (VCO) has a potential to be one of the Philippines’ high-value export products. In the first half of 2005 alone about 350 metric tons of VCO were shipped to foreign markets. Virgin coconut oil is a minimally process or a cold press oil obtained by mechanical means (Philippine National Standard for Virgin Coconut Oil 2005) such as crushing pressing and centrifugation. These procedures result in the retention of the natural flavor and minor components of the oil, which can otherwise be destroyed by chemical process such as refining, deodorizing and bleaching. Cold pressed oils commands a higher price compared to processed oils because these “natural” procedures deliver lower yield and uses higher quality raw materials to prevent spoilage (Ulberth and Buchgraber 2000).

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!

Essay Writing Service

To ensure the quality of VCO in the country, The Philippine National Standard provided a criterion (PNS/BAFPS 22:2004). This standard includes determination of fatty acid composition using gas-liquid chromatography (GLC), iodine value and sensory parameters such as taste, odor and color (Philippine National Standard for Virgin Coconut Oil 2005). However, it is a known fact that the fatty acid profile of oils is species dependent; therefore this classification method cannot distinguish VCO from ordinary refined, bleached and deodorized coconut oil (CCO). Iodine value in the other hand gives information on the degree of unsaturation of the oil and may be use only to distinguish coconut oils from unsaturated oil like corn oil (CO) and soybean oil (SO) but cannot discriminate VCO from CCO. Sensory parameters can easily be manipulated by bleaching and deodorizing the oil or by addition of flavoring agents.

In recent years, various spectroscopic methods have been employed to authenticate vegetable oils such as virgin olive oil (VOO) (Aparicio and Luna 2002; Bucci et al. 2002; Dupuy et al. 1996; Lankmayr et al. 2004; Ozen et al. 2003; Tapp et al. 2003; Tay et al. 2002; Vigli et al. 2003; Yang et al. 2005). These techniques have the advantages of minimal sample preparation and rapid analysis time (Ulberth and Buchgraber 2000). In the study conducted by Yang et al (2005), Fourier Transform Infrared (FTIR) spectroscopy was found to be the most efficient in classification of oils as compared to Fourier Transform Raman and Fourier Transform Near Infrared spectroscopy. An attenuated total reflectance (ATR) accessory is usually used with the FTIR spectrometer for oil and fat study due to its ability to handle liquid samples easily. The resulting infrared spectrum contains all vibrational information about the entire components of the sample. In order to take full advantage of the spectral information, including subtle features that may not be readily discerned, pattern recognition techniques such as Principal Component Analysis (PCA) are commonly used. These techniques can reveal relationships that were not previously suspected, which can lead to interpretations that are not readily apparent by mere visual inspection of IR spectra alone (Beebe et al. 1998; Davies and Fearn 2005). PCA is a mathematical manipulation used to reorganize the high-dimensionality of spectroscopic data into a smaller number of “Principal Components” or PCs that can account for the majority of the information in the data set.

Methodology

All samples of edible oils were obtained from reputable supermarkets in Iloilo City. Samples include four brands of VCO [a total of six samples, one brand have a total of three samples, two regular VCO with different manufacturing dates (labeled as VCO2 and VCO3) and one corn flavored VCO (labeled as VCO4) ] , three different brands of coconut oil ( CCO1, CCO2, and CCO3) and one brand each of corn oil (CO), soybean oil (SO), canola oil (CLO), virgin olive oil (VOO) and palm oil (PO). The oils were stored in their original containers until used. Minimal opening of bottles were observed to prevent oxidation of oils.

Infrared (IR) spectra of samples were obtained at 4000 – 400 cm^-1, 1 cm^-1 resolution and 64 scans; using the Avatar 330 Thermo Nicolet FTIR equipped with DTGS (deuterated tri-glycine sulphate) detector and EZ OMNIC software at the University of the Philippines in the Visayas Chemistry Laboratory, Miag-ao, Iloilo. Samples were placed directly on a multi-bounce zinc selenide horizontal ATR crystal. The ATR crystal was totally cleaned using acetone after each run.

IR spectral data of different oils from EZ OMNIC software were converted to Microsoft Excel files. Data were arranged in tabular form, different oils in columns and transmittance at different wavelengths in rows. The excel file was then transferred to Unscrambler version 9.6 for Principal Components Analysis. The evaluation version of this software can be downloaded at www.camo.com.

Results and Discussion

The FTIR spectra of most fats and oils appeared almost visually similar; the slight differences in the spectral features reflect the variations in the fatty acid composition (Figure1 to 6). Every peak in the spectra represents structural and functional group, either of the oil or other minor components. At the high frequency end of the spectrum (4000-3500 cm^-1) only a few absorption bands can be observed. The most notable are the absorption bands due to O-H stretching vibration at about 3700 and 3400 cm^-1, which may be due to moisture, alcohol or fatty acids. Both alcohols and free fatty acids are degradation product of triglycerides which accumulate during prolong storage and in the presence of moisture. For a clean and dry oil, only the first overtone of the C=O vibration could be seen. The strong band dominating the 3000-2840 cm^-1 region is due to C-H stretches of methylene and terminal methyl groups. The =C-H stretch absorbed at slightly higher region, 3095 -3010 cm^-1. A strong absorption peak in the range of 1750-1735 cm^-1 is due to the C=O stretch of an aliphatic ester. If degradation has occurred a shoulder located on the lower frequency side of this will appear due to the C=O stretch of the resulting long chain carboxylic acid (1730-1700 cm^-1). A sharp peak at about 1660-1600 cm^-1 results from a C=C stretching vibration. The presence of this peak together with a C-H peak at wavenumber greater than 3000 cm^-1 is a good indication of unsaturation. The part of the spectrum lower than 1500 cm^-1 is called the fingerprint region. The absorption in this area of the spectrum is highly characteristic of the molecule as a whole, and can be used to identify the substance (Pavia et al. 2000; van de Voort et al. 2001).

As shown in Figure1-6 its is easy to distinguish unsaturated oils from saturated oils by observing the regions that indicate unsaturation, 3095-3010 cm^-1 and 1660-1600 cm^-1. For olive, corn, palm, canola and soybean oil sharp peaks are seen in both of these areas but no such peaks can be seen in the IR spectra of either CCO or VCO. Thus, visual examination of the IR spectrum can provide a way to distinguish unsaturated oils from coconut oils. However, visual inspection of the spectra alone is insufficient to differentiate VCO from CCO because their IR spectra are identical to the naked eye.

To compare such large data sets like IR spectra, which consist of more than a thousand wavelengths, Principal Component Analysis can be used. This chemometrical technique mathematically models the data into a more manageable data sets (consisting of a few principal components) that can then be interpreted more easily (Brereton 2003).

Principal Component Analyses were performed on the spectra of all oils using Unscrambler version 9.6 software. The software generated several important plots. The percent variance plot gives the total variation in the data set that is described by a particular PC. This plot can be used to determine the number of PC that can effectively describe the whole data set. The 2D scatter plot (PCy vs. PCx) represents how the samples are related to each other. The closeness of the samples in the plot is interpreted as chemical similarity(Beebe and others 1998).

The percent variance plot (Figure 7) of the oil samples spectra generated by the software shows that the first two principal components represent almost 100% of the variance in the data set. As a general rule the PCs that account for less than 5% of the data variation can be rejected (Beebe and others 1998), therefore the first two PCs are sufficient in making our investigation. The scatter plot of first principal component (PC1) versus the second principal component (PC2) shows that it possible to separate the oils into two separate groups, the unsaturated and coconut oil group (Figure 8). Unsaturated oil group consists of corn oil, soya oil, canola oil, palm oil while that of the coconut oil includes CVO and CCO.

Performing the same analysis on CCO and VCO spectra reveals that only the PC1 and the PC2 are highly significant (Figure 9). However, the scatter plot of the two PCs did not successfully separate VCO from CCO (Figure 10). Close inspection of spectra reveal that there are absorption bands at the high frequency region that might indicate the presence of moisture, fatty alcohol or fatty acids in some samples. This observation is confirmed by the Influence Plot (Figure 11) that shows that there is a high residual variance of the 3740 cm^-1, a wavenumber corresponding to H-O absorption. This indicates that the moisture content, fatty alcohol or the free fatty acid content of the oil may have greatly controlled the classification process. The other residual in the influence plot may be due to other components of VCO or CCO that are vital in the classification process.

To correct for the effect of the O-H absorption band, it was decided to eliminate wavelength above 1800 cm^-1 from the spectra of the oil and use only the area encompassing the carbonyl carbon (C=O) and fingerprint region, 1800-400 cm^-1. Principal component analysis of this part of the spectra shows that the first two PC can be used to correctly classify VCO and CCO into separate groups (Figure 12 and 13).

In conclusion, FTIR identification of oils can be enhance by the use of chemometrical techniques such Principal Component Analysis. Using this technique, subtle differences in the spectra can be used to classify oil. FTIR together with PCA can differentiate unsaturated oils from saturated oil and can discriminate virgin coconut oil from ordinary refined, bleached and deodorized coconut oil (CCO).

Acknowledgement

The authors are very grateful to the UP System for upgrading the UPV Chemistry Laboratory and for the purchase of the FTIR used this study possible.

Literature Cited

Aparicio R, Luna G. 2002. Charaterization of Monovarietal Virgin Olive Oils. European Journal of Lipid Science Technology 104:614-627.

Beebe KR, Pell RR, Mary BS. 1998. Chemometrics: A Practical Guide. New York: John Wiley & Sons, Inc. 348 p.

Brereton RG. 2003. Chemometrics Data Analysis for the Laboratory and Chemical Plant. West Sussex: John Wiley & Sons Ltd. 489 p.

Bucci R, Magri AD, Magri AL, Marini D, Marini F. 2002. Chemical Authentication of extra Virgin Olive Oil Varieties by Supervise Chemometric Procedure. Journal Agricultural Food Chemistry 50:413-418.

Davies AMC, Fearn T. 2005 Back to Basics: The Principles of Principal Component Analysis. Spectroscopy Asia:35-38.

Dupuy N, Duponchel L, Huvenne J, Sombret B, Legrand P. 1996. Classification of Edible Fats and Oils by Fourier Transform Infrared Spectra. Food Chemistry 57(2):245-251.

Lankmayr E, Mocak J, Serdt K, Balla B, Wenzl T, Bandoniene D, Gfrerer M, Wagner S. 2004. Chemometrical Classification of Pumpkin Seed OIl using Uv-Vis, NIR and FTIR Spectra. Journal of Biochemical and Biophysical Methods 61(1-2):95-106.

Ozen BF, Weiss I, Mauer L. 2003. Dietary Supplement Oil Classification and Detection of Adulteration Using Fourier Transform Infrared Spectroscopy. Journal Agricultural Food Chemistry 51:5871-5876.

Pavia DL, Lampman GM, Kriz GS. 2000. Introduction to Spectroscopy. New York: Brooks Cole.

Philippine National Standard for Virgin Coconut Oil. 2005. PNS/BAFPS 22:2004.

Tapp HS, Defernez M, Kemsley K. 2003. FTIR spectroscopy and Multivariate Analysis can Distinguish the Geographic Origin of Extra Virgin Olive Oil. Journal Agricultural Food Chemistry 51:6110-6115.

Tay A, Singh RK, Krishnan SS, Gore JP. 2002. Authentication of Olive Oil Adulterated with Vegetable Oils Using Fourier Transform Infrared Spectroscopy. LWT-Food Science and Technology 35:99-103.

Ulberth F, Buchgraber M. 2000. Authenticity of Fats and OIls. European Journal of Lipid Science Technology 102:687-694.

van de Voort F, Sedman J, Russin T. 2001. Lipid Analysis by Vibrational Spectroscopy. European Journal of Lipid Science Technology 103:815-840.

Vigli G, Philippidis A, Spyros A, Dais P. 2003. Classification of Edible Oils by Employing 31P and 1H NMR Spectroscopy in Combination with Multivariate Statistical Analysis. A Proposal for the Detection of Seed Oil Adulteration in Virgin Olive Oils. Journal Agricultural Food Chemistry 51:5715-5722.

Yang H, Irudayaraj J, Paradkar M. 2005. Discriminant Analysis of Edible Oils and Fats by FTIR, FT-NIR and FT-Raman Spectroscopy. Food Chemistry 93:25-32.