Multivariate determination of cloud point in palm oil using partial least squares and principal component regression based on FTIR spectroscopy

A rapid FTIR spectroscopic method was developed for quantitative determination of the cloud point (CP) in palm oil samples. Calibration samples were prepared by blending randomized amounts of palm olein and palm stearin to produce a wide range of CP values ranging between 8.3 and 47.9°C. Both partial least squares (PLS) and principal component regression (PCR) calibration models for predicting CP were developed by using the FTIR spectral regions from 3000 to 2800 and 1800 to 1600 cm⁻¹. The prediction capabilities of these calibration models were evaluated by comparing their standard errors of prediction (SEP) in an independent prediction set consisting of 14 palm oil samples. The optimal model based on PLS in the spectral range 1800-1600 cm⁻¹ produced lower SEP values (2.03°C) than those found with the PCR (2.31°C) method. FTIR in conjunction with PLS and PCR models was found to be a useful analytical tool for simple and rapid quantitative determination of CP in palm oil.     

Rice bran FFA determination by diffuse reflectance IR spectroscopy

Rice bran with FFA levels above 0.1% cannot be used as a food ingredient due to oxidative off-flavor formation. However, extracting high FFA oil from bran by in situ methanolic esterification of rice bran oil to produce methyl ester biodiesel produces greater yields relative to low-FFA rice bran oil. Therefore, high-FFA bran could be exploited for biodiesel production. This study describes an FTIR spectroscopic method to measure rice bran FFA rapidly. Commercial rice bran was incubated at 37°C and 70% humidity for a 13-d incubation period. Diffuse reflectance IR Fourier transform spectra of the bran were obtained and the percentage of FFA was determined by extraction and acid/base titration throughout this period. Partial least squares (PLS) regression and a calibration/validation analysis were done using the IR spectral regions 4000-400 cm⁻¹ and 1731-1631 cm⁻¹. The diffuse reflectance IR Fourier transform spectra indicated an increasing FFA carbonyl response at the expense of the ester peak during incubation, and the regression coefficients obtained by PLS analysis also demonstrated that these functional groups and the carboxyl ion were important in predicting FFA levels. FFA rice bran changes also could be observed qualitatively by visual examination of the spectra. Calibration models obtained using the spectral regions 4000-400 cm⁻¹ and 1731-1631 cm⁻¹ produced correlation coefficients R and root mean square error (RMSE) of cross-validation of R=0.99, RMSE=1.78, and R=0.92, RMSE=4.67, respectively. Validation model statistics using the 4000-400 cm⁻¹ and 1731-1631 cm⁻¹ ranges were R=0.96, RMSE=3.64, and R=0.88, RMSE=5.80, respectively.     

Multivariate calibration of fourier transform infrared spectra for determining thiobarbituric acid-reactive substance content in palm oil

Fourier transform infrared (FTIR) spectra of palm oil samples between 2900 and 2800 cm⁻¹ and 1800 and 1600 cm⁻¹ were used to compare different multivariate calibration techniques for quantitative determination of their thiobarbituric acid-reactive substance (TBARS) content. Fifty spectra (in duplicate) of palm oil with TBARS values between 0 and 0.25 were used to calibrate models based on partial least squares (PLS) and principal components regression (PCR) analyses with different baselines. The methods were compared for the number of factors, coefficients of determination (R ²), and accuracy of estimation. The standard errors of prediction (SEP) were calculated to compare their predictive ability. The calibrated models generated three to eight factors, R ² of 0.9414 to 0.9803, standard error of estimation (SEE) of 0.0063 to 0.0680, and SEP of 1.20 to 6.67.     

Fourier transform infrared spectroscopy evaluation of low density lipoprotein oxidation in the presence of quercetin,catechin, and α-tocopherol

We investigated the changes in human LDL primary and secondary lipid oxidation products and modification of the apolipoprotein B-100 (apoB-100) secondary structures during Cu²⁺-mediated oxidation by FTIR spectroscopy in the presence of catechin, quercetin, and α-tocopherol at physiological concentrations. Catechin- and quercetin-containing samples had slower rates and longer lag phases for conjugated diene hydroperoxide (CD) formation than α-tocopherol-containing samples; however, all antioxidant-treated LDL samples generated similar CD levels (P<0.05). A lower maximum (98.4 nmol/mg LDL protein) of carbonyl compounds was produced in the quercetin- and catechin-treated samples than in α-tocopherol samples. Modification of the apoB-100 secondary structures corresponded closely to the formation of carbonyls and was hampered by the presence of antioxidants. Physiological concentrations of catechin and quercetin offered similar levels of protection against modification by carbonyls of the apoB-100 at advanced stages (carbonyls ∼96.0 nmol/mg LDL protein) but not at the intermediate stages (carbonyls ∼58.0 nmol/mg LDL protein) of LDL oxidation probably owing to differences in the protein-binding mechanisms of catechin and quercetin. Relationships between peroxide formation, carbonyl products, and LDL protein denaturation were shown by the FTIR approach. The FTIR technique provided a simple new tool for a comprehensive evaluation of antioxidant performance in protecting LDL during in vitro oxidation.     

Determination of slip melting point in palm oil blends by partial least squares and principal component regression modeling of FTIR spectroscopic data

The determination of slip melting point (SMP) in refined, bleached, and deodorized palm oil is a common and regular requirement in the manufacture of this oil. An FTIR spectrometer in the mid-IR wavelength region (4000-750 cm⁻¹) can provide the basis for a suitable on-line and off-line quantitative analytical methodology if used with a suitable calibration model. In this study, by using data from selected regions, the calibration models, including principal component regression and partial least squares regression, have been examined and evaluated for efficacy in determining SMP directly in palm oil blends. Results indicate that orthogonal models using selected wavelength data offer superior predictive performance.     

Quantitative analysis of palm carotene using fourier transform infrared and near infrared spectroscopy

β-Carotene content is usually determined by using ultraviolet (UV)-visible spectrophotometry at 446 nm. In this study, two spectroscopic techniques, namely, Fourier transform infrared (FTIR) and near infrared (NIR) spectroscopy, have been investigated and compared to UV-visible spectrophotometry to measure the β-carotene content of crude palm oil (CPO). Calibration curves ranging from 200 to 800 ppm were prepared by extracting β-carotene from original CPO using open-column chromatography. Separate partial least squares calibration models were developed for predicting β-carotene based on the spectral region from 976 to 926 cm⁻¹ for FTIR spectroscopy and 546 to 819 nm for NIR spectroscopy. The correlation coefficient (R ²) and standard error of calibration obtained were 0.972 and 25.2 for FTIR and 0.952 and 23.6 for NIR techniques, respectively. The validation set gave R ² of 0.951 with standard error of performance (SEP) of 25.78 for FTIR technique and R ² of 0.979 with SEP of 19.96 for NIR technique. The overall reproducibility and accuracy did not give comparable results to that of spectrophotometric method; however, the standard deviation of prediction was still within ±5% β-carotene content over the range tested. Because of their rapidness and simplicity, both FTIR and NIR techniques provide alternative means of measuring β-carotene content in CPO. In addition, these two spectroscopic techniques are environmentally friendly since no solvent is involved.     

Quantitation of estolides by Fourier transform infrared spectroscopy

Estolides were produced from meadowfoam oil fatty acids, oleic, linoleic, petroselinic, andcis-5,cis-13 docosadienoic acids. Estolide reaction mixtures were quantitated by Fourier transform infrared spectroscopy and compared to the area percentages determined by high-performance liquid chromatography. The absorbance frequency of estolide carbonyl (1737 cm⁻¹) is different than the lactone carbonyl (1790 cm⁻¹) and the acid carbonyl (1712 cm⁻¹). Estolide standards were obtained by wiped-film molecular-still distillations and column chromatography.     

Differentiation of Lard From Other Edible Fats and Oils by Means of Fourier Transform Infrared Spectroscopy and Chemometrics

Fourier transform infrared (FTIR) spectra at mid infrared regions (4,000–650 cm⁻¹) of lard and 16 edible fats and oils were compared and differentiated. The chemometrics of principal component analysis and cluster analysis (CA) was used for such differentiation using FTIR spectra intensities of evaluated fats and oils. With PCA, an “eigenvalue” of about 90% was achieved using four principal components (PCs) of variables (FTIR spectra absorbances at the selected frequency regions). PC1 accounted for 44.1% of the variation, while PC2 described 30.2% of the variation. The main frequency regions that influence the separation of lard from other evaluated fats and oils based on PC1 are 2,852.8 followed by 2,922 and 1,464.7 cm⁻¹. Furthermore, CA can classify lard into its group based on Euclidean distance.     

Characterization of edible oils,butters and margarines by Fourier transform infrared spectroscopy with attenuated total reflectance

The combination of attenuated total reflectance (ATR) and mid-infrared spectroscopy (MIRS) with statistical multidimensional techniques made it possible to extract relevant information from MIR spectra of lipid-rich food products. Wavenumber assignments for typical functional groups in fatty acids were made for standard fatty acids: Absorption bands around 1745 cm⁻¹, 2853 cm⁻¹, 2954 cm⁻¹, 3005 cm⁻¹, 966 cm⁻¹, 3450 cm⁻¹ and 1640 cm⁻¹ are due to absorption of the carbonyl group, C−H stretch, =CH double bonds of lipids and O−H of lipids, respectively. In lipid-rich food products, some bands are modified. Water strongly absorbs in the region of 3600–3000 cm⁻¹ and at 1650 cm⁻¹ in butters and margarines, allowing one to rapidly differentiate the foods as function of their water content. Principal component analysis was used to emphasize the differences between spectra and to rapidly classify 27 commercial samples of oils, butters and margarines. As the MIR spectra contain information about carbonyl groups and double bonds, the foods were classified with ATR-MIR, in agreement with their degree of esterification and their degree of unsaturation as determined from gas-liquid chromatography analysis. However, it was difficult to differentiate the studied food products in terms of their average chainlength.     

Determination of free fatty acids in crude palm oil and refined-bleached-deodorized palm olein using fourier transform infrared spectroscopy

A rapid direct Fourier transform infrared (FTIR) spectroscopic method using a 100 μ BaF₂ transmission cell was developed for the determination of free fatty acid (FFA) in crude palm oil (CPO) and refined-bleached-deodorized (RBD) palm olein, covering an analytical range of 3.0–6.5% and 0.07–0.6% FFA, respectively. The samples were prepared by hydrolyzing oil with enzyme in an incubator. The optimal calibration models were constructed based on partial least squares (PLS) analysis using the FTIR carboxyl region (C=O) from 1722 to 1690 cm⁻¹. The resulting PLS calibrations were linear over the range tested. The standard errors of calibration (SEC) obtained were 0.08% FFA for CPO with correlation coefficient (R ²) of 0.992 and 0.01% FFA for RBD palm olein with R ² of 0.994. The standard errors of performance (SEP) were 0.04% FFA for CPO with R ² of 0.998 and 0.006% FFA for RBD palm olein with R ² of 0.998, respectively. In terms of reproducibility (r) and accuracy (a), both FTIR and chemical methods showed comparable results. Because of its simpler and more rapid analysis, which is less than 2 min per sample, as well as the minimum use of solvents and labor, FTIR has an advantage over the wet chemical method.     

Free fatty acid formation and lipid oxidation on milled rice

Milled rice was stored at 37°C and 70% humidity and sampled regularly for 50 d. Rice surface lipid was extracted with isopropanol and analyzed for free fatty acids (FFA) and conjugated diene (CD) contents. Diffuse reflectance Fourier transform infrared (DRIFTS) spectra of the rice samples were also obtained. FFA and CD levels increased together during rice storage and exhibited three distinct phases. DRIFTS identified a decrease in intensity at 1746 cm⁻¹ (ester, −C=O) and increases in intensity at 1731 cm⁻¹ (aldehyde, −CO) and 1714 cm⁻¹ (fatty acid, −C=O) during storage, which correlated well with the chemical analysis data. DRIFTS spectral data were analyzed by a partial least squares regression method to identify spectral regions that correlate strongly with measured FFA and construct prediction models. Overall, the mid-infrared region (4000–400 cm¹) gave the best model (R=0.98, root mean square error of cross-validation=0.05) and also predcted the FFA content of milled rice well. The DRIFTS technique has potential for use in studying qualitative chemical changes on the milled rice surface lipids and for predicting FFA on milled rice.     

Milled rice surface lipid measurement by diffuse reflectance fourier transform infrared spectroscopy (DRIFTS)

Diffuse reflectance Fourier transform infrared spectroscopy was investigated as a method for rice surface lipid determination. Long- and medium-grain rice was milled at four degrees of milling to obtain samples with various levels of residual bran, and total lipids were determined by solvent extraction. Fourier transform infrared spectra were collected between 4000 and 400 cm⁻¹. Weighted regression analysis identified changes in surface chemical functional groups with bran removal. Groups typical of lipids increased with bran content whereas those typical of carbohydrates and protein decreased. Partial least squares (PLS) regression analysis showed a high degree of correlation between the spectra in the 4000–400 cm⁻¹ range and extracted lipids of long-grain rice (R ²=0.96) and medium-grain rice (R ²=0.96); a high degree of correlation was also observed when long- and medium-grain rice data were combined (R ²=0.96). There was a high positive correlation between the spectra and extracted lipids in the 1300–1000 cm⁻¹ range for the long-grain rice (R ²=0.98), medium-grain rice (R ²=0.98), and combined long-/medium-grain rice data (R ²=0.94). PLS selected spectral regions that correlated positively with functional groups of lipid/lipid oxidation products and negatively with functional groups of protein and carbohydrates.     

A new method for determining gossypol in cottonseed oil by FTIR spectroscopy

A new method was developed to determine the gossypol content in cottonseed oil using FTIR spectroscopy with a NaCl transmission cell. The wavelengths used were selected by spiking clean cottonseed oil to gossypol concentrations of 0–5% and noting the regions of maximal absorbance. Transmittance values from the wavelength regions 3600–2520 and 1900–800 cm⁻¹ and a partial least squares (PLS) method were used to derive FTIR spectroscopic calibration models for crude cottonseed, semirefined cottonseed, and gossypol-spiked cottonseed oils. The coefficients of determination (R ²) for the models were computed by comparing the results from the FTIR spectroscopy against those obtained by AOCS method Ba 8-78. The R ² were 0.9511, 0.9116, and 0.9363 for crude cottonseed, semirefined cottonseed, and gossypol-spiked cottonseed oils, respectively. The SE of calibration were 0.042, 0.009, and 0.060, respectively. The calibration models were cross-validated within the same set of oil samples. The SD of the difference for repeatability and accuracy of the FTIR method were better than those for the chemical method. With its speed (ca. 2 min) and ease of data manipulation, FTIR spectroscopy is a useful alternative to standard wet chemical methods for rapid and routine determination of gossypol in process and/or quality control for cottonseed oil.     

Application of a Portable Handheld Infrared Spectrometer for Quantitation of <Emphasis Type="Italic">trans</Emphasis> Fat in Edible Oils

Trans fat poses serious health risks to consumers. In order to meet the FDA labeling requirements for trans fatty acids, development of fast, accurate, easy-to-use analytical methods for oils, fats and related products is desirable. Fourier transform infrared spectroscopy (FTIR) is a well-established analytical technique for quantifying trans fats, and the development of handheld FTIR units over the past decade presents new application opportunities. Our objective was to evaluate the performance of a handheld FTIR sensor for measuring trans fat content between 0.1 and 20% trans (w/w) in edible saturated and unsaturated oils. Calibration models were built by measuring height of the band at 966 cm⁻¹ and by partial least squares regression (PLSR) using benchtop FTIR as a reference method. Predictive accuracy of the models was validated with an independent test set of commercial edible oils. Calibration models developed using PLSR and linear regression of band heights gave correlation coefficients R ² > 0.98. Multivariate analysis for the handheld unit gave standard error of prediction (SEP) of approximately 1%, comparable to values obtained with benchtop systems. This study demonstrates that handheld FTIR spectroscopy coupled with chemometrics is a suitable method for quantitation of trans fat content.     

Application of FTIR Spectroscopy for the Determination of Virgin Coconut Oil in Binary Mixtures with Olive Oil and Palm Oil

Rapid Fourier transform infrared (FTIR) spectroscopy combined with attenuated total reflectance (ATR) was applied for quantitative analysis of virgin coconut oil (VCO) in binary mixtures with olive oil (OO) and palm oil (PO). The spectral bands correlated with VCO, OO, PO; blends of VCO and OO; VCO and PO were scanned, interpreted, and identified. Two multivariate calibration methods, partial least square (PLS) and principal component regression (PCR), were used to construct the calibration models that correlate between actual and FTIR-predicted values of VCO contents in the mixtures at the FTIR spectral frequencies of 1,120–1,105 and 965–960 cm⁻¹. The calibration models obtained were cross validated using the “leave one out” method. PLS at these frequencies showed the best calibration model, in terms of the highest coefficient of determination (R ²) and the lowest of root mean standard error of calibration (RMSEC) with R ² = 0.9992 and RMSEC = 0.756, respectively, for VCO in mixture with OO. Meanwhile, the R ² and RMSEC values obtained for VCO in mixture with PO were 0.9996 and 0.494, respectively. In general, FTIR spectroscopy serves as a suitable technique for determination of VCO in mixture with the other oils.     

Effects of Dietary Cholesterol on Tissue Ceramides and Oxidation Products of Apolipoprotein B-100 in ApoE-Deficient Mice

Oxidized LDL (oxLDL) has been shown to activate the sphingomyelinase pathway producing ceramide in vascular smooth muscle cells. Therefore ceramide, which is a biologically active lipid causing apoptosis in a variety of cells, may be involved in the apoptotic action of oxLDL. In this study, we examined whether cholesterol enriched diets affected ceramide metabolism and oxidation product of LDL, represented by degradation of apolipoprotein B-100 (apoB) in apoE-deficient (apoE^−/−) mice. ApoE^−/− and wild type mice were fed a standard (AIN-76) diet or 1% cholesterol-enriched diet for 8 weeks. Tissue ceramide levels were analyzed using electrospray tandem mass spectrometry (LC-MS/MS). Ceramide levels in the plasma and the liver of apoE^−/− mice were intrinsically higher than those of the wild type. In apoE^−/− mice, dietary cholesterol significantly increased several ceramides and degradation products of apoB in plasma compared to those fed the control diet. Dietary cholesterol did not affect tissue ceramide levels in the wild type mice. Based on these results, plasma ceramides possibly correlate with the increase in LDL oxidation and are a risk factor for atherosclerosis.     

Influence of triacylglycerol characterics on the determination of free fatty acids in vegetable oils by fourier transform infrared spectroscopy

A rapid and direct Fourier transform infrared (FTIR) spectroscopic method using a 25-μm NaCl transmission cell was developed for the determination of free fatty acids (FFA) in six important vegetable oils (corn, soybean, sunflower, palm, palm kernel, and coconut oils) that differ in fatty acid profile. The calibrations were established by adding either standard FFA (oleic, lauric acids) or a representative mixture of FFA obtained after saponification of the refined oils. For all oils, up to a FFA level of 6.5% for coconut oil, the best correlation coefficient was obtained by linear regression of the free carboxyl absorption at 1711 cm⁻¹. All correlation coefficients were greater than 0.993, and no significant difference between the calibration methods could be detected. Upon validation of the calibration, no significant difference (α=0.05) between the “actual” and the “FTIR predicted” FFA values could be observed. The calibration models developed for the six oils differed significantly and indicate the need to develop a calibration that is specific for each oil. In terms of repeatability and accuracy, the FTIR method developed was excellent. Because of its simplicity, quick analysis time of less than 2 min, and minimal use of solvents and labor, the introduction of FTIR spectroscopy into laboratory routine for FFA determination should be considered.     

Rapid Method for the Determination of Moisture Content in Biodiesel Using FTIR Spectroscopy

A new, rapid, and direct method was developed for the determination of moisture content in biodiesel produced from various types of oils using Fourier transform infrared (FTIR) spectroscopy with an attenuated total reflectance (ATR) element. Samples of biodiesels used in this study were produced using sludge palm oil (SPO). The calibration set was prepared by spiking double-distilled water into dried biodiesel samples in ratios (w/w) between 0 and 10% moisture. Absorbance values from the wavelength regions 3,700–3,075 and 1,700–1,500 cm⁻¹, and the partial least square (PLS) regression method were used to derive a FTIR spectroscopic calibration model for moisture content in biodiesel samples. The coefficient of determinations (R ²) for the models was computed by comparing the results obtained from FTIR spectroscopy against the values of the moisture concentrations (%) determined using the American Oil Chemists’ Society (AOCS) oven method Ca 2d-25. Same comparison was done using International Union of Pure and Applied Chemistry (IUPAC) distillation method 2.602. R ² was 0.9793 and 0.9700 using AOCS and IUPAC methods, respectively. The standard error (SE) of calibration was 1.84. The calibration model was cross validated within the same set of samples, and the standard deviation (SD) of the difference for repeatability (SDDr) and accuracy (SDDa) of the FTIR method was determined. With its speed and ease of data manipulation, FTIR spectroscopy is a useful alternative method to other methods for rapid and routine determination of moisture content in biodiesel for quality control.     

Determination of olive oil free fatty acid by fourier transform infrared spectroscopy

A new procedure for determining free fatty acids (FFA) in olive oil based on spectroscopic Fourier transform infrared-attenuated total reflectance spectroscopy measurements is proposed. The range of FFA contents of samples was extended by adding oleic acid to several virgin and pure olive oils, from 0.1 to 2.1%. Calibration models were constructed using partial least-squares regression (PLSR). Two wavenumber ranges (1775–1689 cm⁻¹ and 1480–1050 cm⁻¹) and several pretreatments [first and second derivative; standard normal variate (SNV)] were tested. To obtain good results, splitting of the calibration range into two concentration intervals (0.1 to 0.5% and 0.5 to 2.1%) was needed. The use of SNV as a pretreatment allows one to analyze samples of different origins. The best results were those obtained in the 1775–1689 cm⁻¹ range, using 3 PLSR components. In both concentration ranges, at a confidence interval of α = 0.05, no significant differences between the reference values and the calculated values were observed. Reliability of the calibration vs. stressed oil samples was tested, obtaining satisfactory results. The developed method was rapid, with a total analysis time of 5 min; it is environment-friendly, and it is applicable to samples of different categories (extra virgin, virgin, pure, and pomace oil).     

An infrared absorption study of autoxidized methyl linoleate

Summary A study has been made of the infrared absorption spectra of autoxidized methyl linoleate in samples ranging from PV 1 to PV 940 m.e./kg. Principal changes occur in the frequency range 3400–3550 cm⁻¹ where bonded−OH groups absorb and at 1650–1775 cm⁻¹ where >c=0 groups absorb. Two maxima were observed in oxidized samples in the −OH absorption range: one sharp and distinct at 3467–70 cm⁻¹ which increased in intensity with increase in PV and a broad band which increased with increasing PV until it resolved into a true maximum at 3430 cm⁻¹. Reduction of typical oxidized samples with KI reagent resulted in disappearance of the 3430 cm⁻¹ band and appearance of a new band above 3500 cm⁻¹. The band at 3430 cm⁻¹ was attributed to −OOH groups associated by hydrogen bridging. The band at 3467 cm⁻¹ and the band appearing above 3500 cm⁻¹ were attributed to −OH groups, the band at the higher frequency resulting directly from reduction of a hydroperoxide. Absorption due to ketone and aldehyde carbonyl groups appeared only as an indefinite shoulder on the band due to the ester carbonyl. These were resolved by using the intensity of the sample with PV 1 as I_o and that for the oxidized samples as I. A plot of Log I_o/I then revealed three maxima. These indicate the presence of two and possibly three carbonyl containing substances other than the ester carbonyl in autoxidized methyl linoleate. Absorption in the two critical frequency ranges of fractions of autoxidized methyl linoleate eluted from an adsorption column correlate with interpretations made from ultraviolet absorption studies of the same substances. This paper reports research undertaken in cooperation with the Quartermaster Food and Container Institute for the Armed Forces and has been assigned number 211 in the series of papers approved for publication. The views or conclusions contained in this report are those of the authors. They are not to be construed as necessarily reflecting the views or indorsement of the Department of the Army. Presented at the Fall Meeting of the American Oil Chemists’ Society, New York, Nov. 17, 1948.