Spectroscopic Methods in the Evaluation of Modified Vegetable Base Oils from Crambe abyssinica

Raw vegetable oil from Crambe abyssinica was subjected to oxidative treatment to enhance its viscosity. The oxidation processes were carried out in the presence of N-hydroxyphthalimide with or without supercritical CO2 as a solvent. Four spectroscopic techniques (Raman, UV-VIS, FT-IR, NMR) were applied to assess the chemical changes taking place during the oxidation. Raman and NMR spectroscopy proved best in the assessment of the chemical transformations leading to increased viscosity of the modified vegetable oil.


Introduction
Oils are a group of products which are characterized by a high variability of their chemical and physical properties. The chemical analysis of vegetable oils is arduous because they consist of complex mixtures of chemical compounds and the evaluation of the results can also be disturbed by a matrix effect [1]. However, characterization of these oils can be performed using a great variety of analytical methods and chemometric methodologies. Different techniques for fats and oils analysis have been developed so far and a large part of them is based on liquid or gas chromatography [2,3].
There has been a growing demand to replace traditional analytical methods with instrumental methods. The superiority of the latter is due to their high sensitivity, low limits of detection, speed of analysis and automated operation. Modern instrumental techniques include spectroscopic methods that are successfully used in qualitative and quantitative analysis in various industries, including the chemical and food industries, electronics or metallurgy. Furthermore, spectroscopic methods can be used in conjunction with advanced separation methods, which apart from comprehensive qualitative and quantitative analysis of isolated intermediates, enable the direct analysis of final products. Numerous scientific publications prove the usefulness of spectroscopic methods in the study of the properties of vegetable oils and, above all, the assessment of the quality of edible oils [4,5]. The analysis of vegetable oils can be carried out with the use of absorption spectroscopy in the area of the ultraviolet-visible spectral region (UV-Vis) [6][7][8], infrared spectroscopy (IR) [9][10][11][12][13][14], nuclear magnetic resonance spectroscopy (NMR) [15][16][17] and Raman spectroscopy [18][19][20][21].
The performance of base oils mostly depends on their physiochemical properties. The characteristics of base oils play a key role in the ultimate formulation of high-quality lubricants. Several publications concern the use of instrumental methods to determine the properties of various base oils [22][23][24][25][26] and lubricants [27][28][29][30][31][32]. However, these articles relate mainly to the analysis of mineral or synthetic base oils. Lubricants based on vegetable oils seem to be an attractive alternative to conventional lubricants. Structurally, vegetable oils are mixtures of natural fatty acid esters and are an alternative to replace the commonly used mineral oils due to their natural occurrence and their fast biodegradation rates which could reduce environmental risks.
The base oils obtained in the innovative process of vegetable oils modification, presented in our previous article [33], were subjected to various tests to determine the changes that have occurred in the chemical composition. The aim of this paper is to demonstrate the applicability of spectroscopic techniques, including UV-VIS spectroscopy, NMR, infrared and Raman spectroscopy, as the methods of analysis to evaluate the result of the modification reactions and determine their effect on the structural changes between unmodified vegetable oils and oil bases obtained as a result of oxidation reactions.

Oxidative Modification of Raw Crambe Oil
The raw vegetable oil from Crambe abyssinica (A) is characterized by a high content of both unsaturated erucic acid and natural antioxidants. Crambe abyssinica belongs to the cabbage family (Brassicaceae). It is an oilseed crop which originated from the Mediterranean region. It has also been grown in other regions e.g., tropical and subtropical parts of Africa, Central and West Asia, Europe, North and South America [34]. The oil is used as a lubricant, corrosion inhibitor and in the production of synthetic rubbers, nylons or plastic films [35]. The raw oil was modified to obtain higher viscosity class (VG 150) oils via an oxidative treatment. The oil, catalyst, radical initiator and solvent were vigorously stirred at elevated temperature and pressure under an oxygen atmosphere. The process was stopped after a specified amount of oxygen had been absorbed. N-Hydroxyphthtalimide (NHPI) was used as a catalyst, azobisisobutyronitrile (AIBN) as a radical initiator and supercritical carbon dioxide (scCO 2 ) as a solvent. Reactions without scCO 2 were performed for comparison. The modified oils were marked as A_CO 2 and A_O 2 accordingly. After the experiments all modified oils were tested to determine their kinematic viscosity, peroxide number (LN), iodine number (LJ), saponification number (LZ) and acid number (LK). The results are summarized in Table 1. In this Table, mean values of the measurements (three repetitions) and the values of standard deviations are presented.

Oxidative Stability Tests of the Modified Oils
The raw vegetable oil (A) with VG 46 viscosity class and VG 150 (A_CO 2 and A_O 2 ) oils obtained through modification were subjected to Rapid Small Scale Oxidation Test (RSSOT, PetroOxy) to evaluate their oxidation stability. This method consists in the acceleration of the oxidation process for a chosen sample at an elevated temperature and exposure to the excess of oxygen. The sample is placed in a small chamber and is exposed to the oxygen under pressure at a certain temperature. The temperature is kept constant and the pressure is constantly measured until a specified pressure drop is detected. During our research tests were performed in isothermal conditions, at 80 • C and 120 • C. Measurements were taken until pressure dropped by 10%. The time of oxidation induction was determined on that basis ( Table 2). In this Table, mean values of the measurements (three repetitions) and the values of standard deviations are presented.

Spectroscopic Methods
The obtained oils were analyzed using spectroscopic techniques. For the following materials, Raman, UV-VIS, IR and NMR spectra were recorded: raw vegetable oil (A) with VG 46 viscosity class before and after the PetroOxy test (A_80, A_120 for samples tested at 80 • C and 120 • C respectively); modified oils (A_O 2 and A_CO 2 ) with VG 150 viscosity class before and after the PetroOxy test (A_O 2 _80 and A_O 2 _120; A_CO 2 _80 and A_CO 2 _120, for samples tested at 80 • C and 120 • C respectively).

Raman Spectroscopy
The raw vegetable oil from Crambe abyssinica (A) is composed mainly of unsaturated fatty acids with the following molecular structure: erucic acid (C22:1, 63.77%), oleic acid (C18:1, 15.07%), linoleic acid (C18:2, 13.16%) and linolenic acid (C18:3) [36,37]. Unsaturation is the main cause of the reactivity of the oil towards oxygen. Figure 1 shows the Raman spectra of oils obtained in the modification process (A_CO 2 and A_O 2 ) in comparison with the raw oil (A). The use of Raman spectroscopy allowed us to evaluate the influence of the modification process and thermal activation on changes in structure of the researched oils. The band assignments of the Raman spectra of initial A oil and modified A_CO 2 and A_O 2 oils are reported in Table 3 [38][39][40][41][42][43].  Based on the Raman spectra analysis of the studied A, A_CO2 and A_O2 oils, the modification influence on the fatty acids unsaturation degree change was evaluated. To do this, the ratio of bands intensity at 1656 and 1444 cm −1 (I1656/I1444) as well as bands at 3008 and 2854 cm −1 (I3008/I2854), resulting from the vibration of characteristic functional groups, was calculated. The ratio of the number of C=C double bonds to the number of C-C single bonds in the fatty acids of the evaluated oils was determined, and thus the influence of the modification process conditions on unsaturation degree change was estimated ( Figure 2).   Based on the Raman spectra analysis of the studied A, A_CO 2 and A_O 2 oils, the modification influence on the fatty acids unsaturation degree change was evaluated. To do this, the ratio of bands intensity at 1656 and 1444 cm −1 (I 1656 /I 1444 ) as well as bands at 3008 and 2854 cm −1 (I 3008 /I 2854 ), resulting from the vibration of characteristic functional groups, was calculated. The ratio of the number of C=C double bonds to the number of C-C single bonds in the fatty acids of the evaluated oils was determined, and thus the influence of the modification process conditions on unsaturation degree change was estimated ( Figure 2).  Based on the Raman spectra analysis of the studied A, A_CO2 and A_O2 oils, the modification influence on the fatty acids unsaturation degree change was evaluated. To do this, the ratio of bands intensity at 1656 and 1444 cm −1 (I1656/I1444) as well as bands at 3008 and 2854 cm −1 (I3008/I2854), resulting from the vibration of characteristic functional groups, was calculated. The ratio of the number of C=C double bonds to the number of C-C single bonds in the fatty acids of the evaluated oils was determined, and thus the influence of the modification process conditions on unsaturation degree change was estimated ( Figure 2).  The modification process of oil A, conducted for the purpose of obtaining base oils with a higher viscosity class, led to changes in the structure. It was found that the oil modification process carried out in the presence of CO 2 solvent was more beneficial because the established viscosity of A_CO 2 oil was obtained by smaller structure changes, evaluated on the basis of the I 1656 /I 1444 and I 3008 /I 2854 band intensity ratio. The basis for comparison with analytical parameters, especially iodine number LJ, was identified bands from the Raman spectra of the studied I 1656 /I 1444 oils, which characterized the unsaturation degree of the fatty acids ( Figure 3).
It was possible to compare the results of Raman spectroscopy with the iodine number because in both cases the number of double bonds in oil is determined. Raman spectroscopy allowed us to calculate the ratio of the number of C=C double bonds to the number of C-C single bonds in the oil, where a higher value of the ratio of those bands indicates a higher degree of oil unsaturation. The iodine number determines the number of grams of iodine necessary for the saturation of multiple bonds. The higher the iodine number value the greater the number of double bonds that the fatty acids in the oil contain. It was indicated that the Raman spectra allowed us to estimate the iodine number of oils. The modification process of oil A, conducted for the purpose of obtaining base oils with a higher viscosity class, led to changes in the structure. It was found that the oil modification process carried out in the presence of CO2 solvent was more beneficial because the established viscosity of A_CO2 oil was obtained by smaller structure changes, evaluated on the basis of the I1656/I1444 and I3008/I2854 band intensity ratio. The basis for comparison with analytical parameters, especially iodine number LJ, was identified bands from the Raman spectra of the studied I1656/I1444 oils, which characterized the unsaturation degree of the fatty acids ( Figure 3).
It was possible to compare the results of Raman spectroscopy with the iodine number because in both cases the number of double bonds in oil is determined. Raman spectroscopy allowed us to calculate the ratio of the number of C=C double bonds to the number of C-C single bonds in the oil, where a higher value of the ratio of those bands indicates a higher degree of oil unsaturation. The iodine number determines the number of grams of iodine necessary for the saturation of multiple bonds. The higher the iodine number value the greater the number of double bonds that the fatty acids in the oil contain. It was indicated that the Raman spectra allowed us to estimate the iodine number of oils. The use of Raman spectroscopy also allowed the evaluation of the thermal activation's influence on the unsaturation degree change of the researched oils. Samples of A, A_CO2 and A_O2 oils were subjects to thermal activation in a laboratory test under the conditions of continuous oxygen activity at 80 °C and 120 °C. Based on that, the oils' oxidation stability was evaluated as well as the oxidation induction time (Table 1). During the tests the Raman spectra of the oxidized oils were recorded. They were compared with the spectra obtained before the oxidation process. Exemplary spectra are shown below in Figures 4 and 5. The use of Raman spectroscopy also allowed the evaluation of the thermal activation's influence on the unsaturation degree change of the researched oils. Samples of A, A_CO 2 and A_O 2 oils were subjects to thermal activation in a laboratory test under the conditions of continuous oxygen activity at 80 • C and 120 • C. Based on that, the oils' oxidation stability was evaluated as well as the oxidation induction time (Table 1). During the tests the Raman spectra of the oxidized oils were recorded. They were compared with the spectra obtained before the oxidation process. Exemplary spectra are shown below in Figures 4 and 5.   The degree of unsaturation of the oxidized oils and the influence of oxidation on that change were evaluated on the basis of the obtained spectra. The band intensity ratio at 1656 and 1444 cm −1 (I1655/I1438) as well as at 3008 and 2854 cm −1 (I3008/I2854), resulting from the vibration of characteristic functional groups, was determined. Thus, the influence of the modification and oxidation process conditions on the unsaturation degree change was estimated. The band intensity ratio determined for individual oils before and after oxidation was compared-carried out at 80 °C and 120 °C for A oil before and A_80, A_120 after oxidation, A_CO2 oil before and A_CO2_80 A_CO2_120 after oxidation, and for A_O2 oil before and A_O2_80 A_O2_120 after oxidation ( Figure 6).   The degree of unsaturation of the oxidized oils and the influence of oxidation on that change were evaluated on the basis of the obtained spectra. The band intensity ratio at 1656 and 1444 cm −1 (I1655/I1438) as well as at 3008 and 2854 cm −1 (I3008/I2854), resulting from the vibration of characteristic functional groups, was determined. Thus, the influence of the modification and oxidation process conditions on the unsaturation degree change was estimated. The band intensity ratio determined for individual oils before and after oxidation was compared-carried out at 80 °C and 120 °C for A oil before and A_80, A_120 after oxidation, A_CO2 oil before and A_CO2_80 A_CO2_120 after oxidation, and for A_O2 oil before and A_O2_80 A_O2_120 after oxidation ( Figure 6). The degree of unsaturation of the oxidized oils and the influence of oxidation on that change were evaluated on the basis of the obtained spectra. The band intensity ratio at 1656 and 1444 cm −1 (I 1655 /I 1438 ) as well as at 3008 and 2854 cm −1 (I 3008 /I 2854 ), resulting from the vibration of characteristic functional groups, was determined. Thus, the influence of the modification and oxidation process conditions on the unsaturation degree change was estimated. The band intensity ratio determined for individual oils before and after oxidation was compared-carried out at 80 • C and 120 • C for A oil before and A_80, A_120 after oxidation, A_CO 2 oil before and A_CO 2 _80 A_CO 2 _120 after oxidation, and for A_O 2 oil before and A_O 2 _80 A_O 2 _120 after oxidation ( Figure 6).
In the oils' spectra after the oxidation test at 80 • C and 120 • C, bands at 2854 cm −1 and 1656 cm −1 were smaller than those for the initial oils. For all studied oils after the test, the change in structure was the smallest for initial A oil in comparison to modified A_CO 2 and A_O 2 oils.
The determined oxidation stability confirmed the higher resistance to oxidation of raw oil in comparison to modified oils, which is connected to the oil's molecular structure. The oil modification process initiated changes in the structure and the thermal activities augmented the changes. Oils in the VG 150 viscosity class result from the modification, but the modification process carried out in the presence of CO 2 had a smaller effect on the fatty acid unsaturation degree-diagnostic bands present in the Raman spectrum, which was also confirmed through analytical methods.  Figure 6. The influence of oxidation on the change in fatty acids unsaturation degree of the researched oils before and after oxidation, at 80 °C and 120 °C, expressed by I1656/I1444 and I3008/I2854 band intensity ratio.
In the oils' spectra after the oxidation test at 80 °C and 120 °C, bands at 2854 cm −1 and 1656 cm −1 were smaller than those for the initial oils. For all studied oils after the test, the change in structure was the smallest for initial A oil in comparison to modified A_CO2 and A_O2 oils.
The determined oxidation stability confirmed the higher resistance to oxidation of raw oil in comparison to modified oils, which is connected to the oil's molecular structure. The oil modification process initiated changes in the structure and the thermal activities augmented the changes. Oils in the VG 150 viscosity class result from the modification, but the modification process carried out in the presence of CO2 had a smaller effect on the fatty acid unsaturation degree-diagnostic bands present in the Raman spectrum, which was also confirmed through analytical methods.

UV-VIS Spectroscopy
The raw oil (A) does not exhibit any absorption above 300 nm. In the 200-400 nm region, a weak absorption band is present λ = 234 nm, ε = 0.046. In the spectra of both modified oils, two additional absorption maxima are present: λ = 218 nm, ε = 0.18 and λ = 268 nm, ε = 0.032 for A_O2 and λ = 218 nm, ε = 0.168 and λ = 268 nm, ε = 0.036 for A_CO2 (Figure 7). Figure 6. The influence of oxidation on the change in fatty acids unsaturation degree of the researched oils before and after oxidation, at 80 • C and 120 • C, expressed by I 1656 /I 1444 and I 3008 /I 2854 band intensity ratio.

FT-IR Spectroscopy
By comparing the spectra of raw and modified oils in 3800-2400 cm −1 region, the change in the absorbance at 3465 cm −1 (γ O-H in OH groups) is noticeable, as the height of this band increased after modification. The band at 3005 cm −1 decreased slightly after treatment and the increase in absorbance at 970 cm −1 indicated the formation of trans C=C bonds.
The examination of ATR-FTIR spectra of raw and modified oils revealed the changes in the following regions: the OH stretching region of hydroxyl groups between 3600 and 3100 cm −1 , hydrogen stretching vibrations between 3050 and 2800 cm −1 and carbonyl stretching vibrations between 1800 and 1680 cm −1 . These changes reflect the modification process leading to the alteration of structure and properties of the raw C. abyssinica oil.
Comparing the spectra of raw and modified oils, absorbance changes at 3465 cm −1 (γ O-H in OH groups) are observed. The height of this band increased after modification most likely due to the formation of -COOH. Likewise the absorbance between 1800 and 1680 cm −1 after modification decreased and the broadening of the band is observed, which is associated with appearance of new carbonyl groups originated from ketones, aldehydes and free fatty acids. The resulting decrease of band at 1744 cm −1 is caused by the loss of the carbonyl groups of fatty esters. The band at 3005 cm −1 and 1650 cm −1 slightly decreased after treatment, which is associated with the decrease of cis C=C bonds. At the same time absorbance at 970 cm −1 increased indicating the formation of trans C=C bonds.
Deconvolution of the spectra of raw oil A in the carbonyl range between 1770 and 1680 cm −1 reveals two carbonyl bands at 1745 cm −1 and at 1733 cm −1 while deconvolution of the modified oils A_O 2 and A_CO 2 spectra shows additional two carbonyl bands at about 1720 cm −1 and at 1700 cm −1 which in comparison does not show in case of raw oil (A) (Figure 9a-c). The band at 1720 cm −1 results from the creation of aldehydes, ketones, esters and band at 1700 cm −1 represents the appearance of free fatty acids [45]. Gaussian/Lorenzian deconvolution model with middle sensitivity and wave width of 4 cm −1 gave good compatibility of original spectrum and the total of all deconvolution bands.
ATR-FTIR spectroscopy thus seems to be an appropriate and comprehensive tool for monitoring the chemical changes appearing during modifications of the Crambe abissinica oil.

NMR Spectroscopy
Based on the 1 H-NMR spectra of crude Crambe abyssinica oil (Figure 10), it was determined that signals at a chemical shift of 0.9 ppm originate from the -CH 3 groups. The integral calculations (Table 4) were made by assuming that there are three such groups in the triglyceride molecule, equal to nine protons. On this basis, it was determined that the multiplet at 5.35 ppm corresponds to eight hydrogen atoms derived from vinyl groups (-CH=CH-). Therefore, the distribution of double bonds in the molecular chains of triglyceride is possible in two ways: two moieties consisting of two double bonds connected by a methylene group (the so called bis-allylic group =CH-CH 2 -CH=) or one such moiety and two isolated double bonds ( Figure 11). hydrogen atoms derived from vinyl groups (-CH=CH-). Therefore, the distribution of double bonds in the molecular chains of triglyceride is possible in two ways: two moieties consisting of two double bonds connected by a methylene group (the so called bis-allylic group =CH-CH2-CH=) or one such moiety and two isolated double bonds ( Figure 11).     hydrogen atoms derived from vinyl groups (-CH=CH-). Therefore, the distribution of double bonds in the molecular chains of triglyceride is possible in two ways: two moieties consisting of two double bonds connected by a methylene group (the so called bis-allylic group =CH-CH2-CH=) or one such moiety and two isolated double bonds ( Figure 11).     Based on the integrals of peaks originating from the allylic methylene groups (2.0 and 2.8 ppm), it was determined that there are two hydrogen atoms originating from the bis-allylic methylene group according to the structure =CH-CH 2 -CH= and 12 hydrogen atoms that are derived from the allylic protons -CH 2 -CH=CH-CH 2 -. This information allowed us to figure out the general triglyceride structure eliminating one of the above possibilities. In only one chain are there two double bonds ( Figure 12).
Based on the integrals of peaks originating from the allylic methylene groups (2.0 and 2.8 ppm), it was determined that there are two hydrogen atoms originating from the bis-allylic methylene group according to the structure =CH-CH2-CH= and 12 hydrogen atoms that are derived from the allylic protons -CH2-CH=CH-CH2-. This information allowed us to figure out the general triglyceride structure eliminating one of the above possibilities. In only one chain are there two double bonds ( Figure 12).  Figure 12. The general structure of the triglyceride.
In addition, the number of hydrogen atoms forming the -CH2-chains was calculated (70 H) and this allowed us to estimate the triglyceride structure. The fatty acid chains that create the molecule probably originate from erucic acid 22:1 and linoleic acid 18:2. Therefore: The structural changes that occurred as a result of the oxidation of Crambe abyssinica oil were determined with the use of 1 H-NMR spectra (Figures 13 and 14). These changes are mainly demonstrated by the decreased intensity of three signals (2.0, 2.8 and 5.35 ppm). Comparing the spectrum of the modified oil with the crude oil, a decrease in the value of the integral of the signal deriving from the vinylic protons -CH=CH-is noticeable (5.35 ppm). Using these values (Table 4), it is possible to calculate the degree of conversion of double bonds: Another issue worth considering is the signal with 2.8 ppm chemical shift coming from the hydrogen atoms between the two double bonds =CH-CH2-CH=. This signal also decreases considerably which indicates the disappearance of double bonds.
In addition, the signal with 2.0 ppm chemical shift derived from the hydrogen atoms in the CH2-CH=CH-CH2 structure is reduced which is another proof of a change in the structure confirming the efficiency of the oxidation of double bonds. In addition, the number of hydrogen atoms forming the -CH 2 -chains was calculated (70 H) and this allowed us to estimate the triglyceride structure. The fatty acid chains that create the molecule probably originate from erucic acid 22:1 and linoleic acid 18:2. Therefore: The structural changes that occurred as a result of the oxidation of Crambe abyssinica oil were determined with the use of 1 H-NMR spectra (Figures 13 and 14). These changes are mainly demonstrated by the decreased intensity of three signals (2.0, 2.8 and 5.35 ppm). Comparing the spectrum of the modified oil with the crude oil, a decrease in the value of the integral of the signal deriving from the vinylic protons -CH=CH-is noticeable (5.35 ppm). Using these values (Table 4), it is possible to calculate the degree of conversion of double bonds: Another issue worth considering is the signal with 2.8 ppm chemical shift coming from the hydrogen atoms between the two double bonds =CH-CH 2 -CH=. This signal also decreases considerably which indicates the disappearance of double bonds. In addition, the signal with 2.0 ppm chemical shift derived from the hydrogen atoms in the CH 2 -CH=CH-CH 2 structure is reduced which is another proof of a change in the structure confirming the efficiency of the oxidation of double bonds.  The analysis described above also applies to Crambe abyssinica oil oxidized in the presence of carbon dioxide. The differences between the spectra of modified oils are slight. Decreased intensity of signals from vinylic protons (5.35 ppm chemical shift) allowed us to calculate the degree of conversion of double bonds:  The analysis described above also applies to Crambe abyssinica oil oxidized in the presence of carbon dioxide. The differences between the spectra of modified oils are slight. Decreased intensity of signals from vinylic protons (5.35 ppm chemical shift) allowed us to calculate the degree of conversion of double bonds: The analysis described above also applies to Crambe abyssinica oil oxidized in the presence of carbon dioxide. The differences between the spectra of modified oils are slight. Decreased intensity of signals from vinylic protons (5.35 ppm chemical shift) allowed us to calculate the degree of conversion of double bonds: In the presence of AIBN (a radical initiator), the oxygen reacts initially with the alkyl radicals of fatty acid chains in the triglyceride to form primary oxidation products-the hydroperoxides.
Those highly reactive hydroperoxides follow several pathways, leading to oxygenated and cross-linked products. For example, they react preferably with the allylic hydrogens to form relatively stable allyl radicals. Those radicals in turn, can react with oxygen finally yielding some oxygenated products (e.g., epoxides, cyclic peroxides, aldehydes, alcohols) or they can recombine with other alkyl radicals yielding cross-linked glycerides. The ratio of the products attributable to propagation and termination steps is mostly influenced by the relative rates of branching and termination pathways [46].
In the IR and Raman spectra of the oxidized oil bands due to hydroxy group are absent. This supports the assumption that the reaction, does not stop at the hydroperoxide or epoxide stage.
On the other hand, during such oxidative treatment radical polymerization of the vegetable oil occurs, resulting in the increase of viscosity. Examination of the 1 H-NMR spectra reveals lack of peaks that can be attributed to epoxides [47]: 3.00 ppm for protons on the epoxy ring is missing and the peak at 1.50 ppm for the methylene protons adjacent to these epoxy groups shows only a negligible intensity. Signals that can be attributed to aldehydes and free acids are barely above the noise level.
Combined evaluation of IR, Raman and NMR data supports the assumption that the double bonds were mainly converted to C-C bonds and ether (C-O-C) and/or peroxy (C-O-O-C) bridges.

Abyssinian Oil Modification
The oxidation processes of Abyssinian oil were conducted using equipment composed of three devices: a "limbo" pressure reactor (Büchi, Uster, Switzerland), a metering pump (Blue Shadow, Berlin, Germany) and a thermostat (Julabo, Seelbach, Germany). Starting materials crambe oil, 0.05% (w/w) NHPI and 0.05% (w/w) AIBN were fed to the pressure reactor. Then, the reactor was pressurized with 0.25 MPa with oxygen, the temperature was raised to 100 • C and the pressure was increased to 10 MPa with carbon dioxide. For reactions in pure oxygen, the reactor was pressurized with 0.4 MPa. During the process, the temperature and pressure were controlled. If necessary, pressure in the reactor was corrected by the addition of oxygen (for the process without solvent) or CO 2 (for the process with solvent).

Raman Spectroscopy
Raman spectra were obtained using an NRS 5100 confocal grating Raman microspectrometer (Jasco Corporation, Tokyo, Japan) which was equipped with a pumped laser with wave length of 532 nm and a Charged Coupled Device (CCD) detector. The operating conditions of the spectrometer were as follows: diffraction grating 1800 lines/mm, laser power 5.1 mW, numerical aperture 4000 µm, resolution 8.4 cm −1 , lens magnification 20×, exposure time 15 s.
The method of research was developed. The procedure was the same for each sample. Samples were applied to glass slides having a recess in order to maintain the same thickness of the sample layer. The measurement conditions were established (focusing on the sample, selection of integration time, number of scans, laser power). The Raman spectra of the oils were measured under selected conditions in a way to obtain the appropriate signal-to-noise ratio in the shortest possible time in the entire spectral range. The number of scans of the sample was 5, the exposure time was 15 s for a single accumulation, a laser with an excitation line of 532 nm was used and the measurements were taken at room temperature.
Obtained spectra were the basis for assessing the degree of unsaturation of the tested oils. An intensity ratio of bands at 1656 and 1444 cm −1 was taken into consideration for the analysis. The bands at 1266 and 1300 cm −1 in the analyzed oils were not very characteristic and they were not analyzed.

UV-VIS Spectroscopy
Spectra were recorded on a JASCO V-650 spectrometer (Jasco Corporation, Tokyo, Japan) using a 1 cm cuvette. The absorption spectra of the oils diluted with n-hexane (C = 5 × 10 −5 g/mL) were measured in the region 200-400 nm.

NMR Spectroscopy
1 H-NMR spectra were recorded on Varian spectrometer (Palo Alto, CA, USA) at a frequency of 600 MHz using NMR solvent (CDCl 3 ) which was purchased from ACROS Organics (Geel, Belgium). Coupling constants (J) are in Hertz (Hz). All chemical shifts (δ) are expressed in ppm downfield from TMS as an internal standard.

Determination of Modified Oils Properties
The viscosity-temperature oil properties were evaluated based on the kinematic viscosity

Conclusions
The aim of the research was to determine the changes in the chemical structure of oil components caused by its oxidation. For this purpose Raman, UV-Vis, IR and NMR spectroscopy were used for the sample analysis. In a way, all methods confirmed the structure changes occurring during the oxidative modification process. The analysis of the obtained Raman and NMR spectra led to the determination of the degree of fatty acids' unsaturation. Furthermore, the change in the absorption bands and hence structural changes can be confirmed with the use of the IR and UV-Vis spectra. Spectroscopic methods are effective in observing the changes caused by the oxidation of Crambe abyssinica oil and determining the degree of unsaturation of fatty acids.