Edible oils were an important part of human diet, being consumed in a variety of forms. Numerous parameters
were used to assess the quality of edible oils including free fatty acids (FFA), iodine (IV), peroxide (PV) and
anisidine value (AV) and various other parameters. These parameters were separately analyzed with standardized
chemical and physical methods, approved by the German Society for Fat Science (Deutsche Gesellschaft für
Fettwissenschaft e.V.; DGF) and by the American Oil Chemist Society (AOCS). The assessment of quality and authenticity of
edible oils was traditionally performed based on several separate analytical methods such as chromatographic and volumetric
techniques. These methods have been developed and standardized several decades ago undergoing only minor changes since then,
which made oil analysis time-consuming and sometimes even erroneous.
of a new
A fast and non-destructive quality control was desired by edible oil producers and processors. Former researches described nuclear
magnetic resonance (NMR) as a technique extending the analysis of edible oils [1-6].
On the one hand, in this research a proton nuclear magnetic resonance (1 NMR) method was developed and validated for target
control of essential quality parameters within short time and in only one analytical run. Two completely new parameters were
established during the research project: the aldehyde and the double bond value. The aldehyde value described the content of
aldehydes independently of the aldehyde kind. The mean number of double bonds per fatty acid (FA) was described with the double bond value.
These values could replace the conventional AV and IV.
On the other hand an authenticity control was developed identifying the kind of oil. With the use of multivariate models like
Principle Component Analysis (PCA) the genuineness of edible oils was determined by visualizing the separation of the different oil types
A series of 130 samples of 40 different edible oil types, collected from local markets, Cologne, and numerous oil mills,
Germany, were measured using 1H NMR spectroscopy (spectrum in Fig. 1) by analyzing the following essential quality
parameters: free fatty acids (FFA), peroxides, aldehydes, α-Tocopherol, C18:3, C18:2, C18:1, Omega-3 fatty acids,
sterols, 1,3- and 1,2-digylcerides, unsaturated and saturated fatty acids, iodine value, aldehyde and double bond value.
Quality control of edible oils.
Fig. 1 1H NMR spectrum of olive oil, inclusively detailed schematic of quality parameters.
Statistical measurements were performed only on oils of which at least two samples were available.
Types and number of collected samples were:
Linseed Oil (8), Sunflower Oil (8), Walnut Oil (8), Rape Seed Oil (7), Almond Oil (5), Pumpkin Seed Oil (4),
Peanut Oil (4), Hemp Oil (4), Black Seed Oil (4), Gold-of-Pleasure Oil (4), Sesame Oil (4), Mustard Seed Oil (4),
Argan Oil (3), Chia Oil (3), Hazelnut Oil (3), Moringa Oil (3), Poppy Seed Oil (3), Evening Primrose Oil (2),
Olive Oil (24), Pistachio Oil (2), Macadamia Nut Oil (2), Grape Oil (2), Safflower Oil (2), Apricot Kernal Oil (2),
Avocado Oil (1), Borage Oil (1), Chili Oil (1), Dill Seed Oil (1), Germ Oil (1), Coconut Oil (1), Corn Oil (1),
Carrot Seed Oil (1), Nangai Oil (1), Rapeseed – Chili – Oil (1), Rice – Germ – Oil (1), Castor Oil (1), Soy Oil (1),
Tamanu Oil (1), Omega 3-6-9 Power Mix (1)
Fatty acid profile analyzed by 1H NMR.
The fatty acid (FA) profile was a very important quality parameter of edible oils. The relation of the different fatty a
cids could determine the kind of oil, the nutritional value and physical properties. The results from this study concerning
the saturated, monosaturated, C18:2 and C18:3 FAs [mol-%] of analyzed oils were presented in Fig. 2. The saturated FA of the
analyzed edible was determined in a range of 6.98 to 21.45 %, monosaturated FA of 0.00 to 80.66 %, C18:2 FA of 0.68 to
72.22 % and C18:3 FA of 0.24 to 57.50 %. In Fig. 6 the FA profiles were compared. Some oils showed similar profile which
could not be distinguished by eye. With the use of multivariate models like Principle Component Analysis (PCA) the genuineness
of edible oils was determined by visualizing the separation of the different oil types.
In Fig. 6 the FA profiles were chemometrically compared.
Fig. 2 Saturated FA (blue), C18:1 FA (red), C18:2 FA (green) and C18:3 FA (purple) of different kinds of edible oils (mean ± deviation)
The 13C NMR spectroscopy is a versatile tool to identify and quantify the distribution of triacylglycerides (TAGs) with the respective
isomers sn-1/3 and sn-2 (Fig. 7 to Fig. 9). TAGs like saturated (sat.) and unsaturated (unsat.) fatty acids can be distinguished.
Between C16:0 and C18:0 a distinction is not possible. Unsaturated fatty acids can be distinguished into C18:1, C18:2, C18:3, γ-Linolenic
acid (ALA), γ-Linolenic acid (GLA), C20:1, C22:1 and C22:1.
Gold-of-Pleasure oil is a representative oil, rich in γ-Linolenic acid. Borage oil is rich in γ-Linolenic acid.
Fatty acid profile analyzed by 13C NMR
Fig. 3 13C NMR spectra from Gold-of-Pleasure Oil (blue) and Borage Oil (red)
Iodine Value – an obsolete value?
The iodine value (IV) in oils described the amount of iodine which is absorbed by a fat or oil determining the
amount of unsaturation in fatty acids. Unsaturated fatty acids contained double bonds which reacted with iodine compounds.
The higher the iodine value, the more double bonds were present in the fatty acid.
Problematically was that other olefinic double bonds in by-products like sterols were also quantified leading to an
increased, falsified IV. Fig. 4 described the comparison of the conventional IV values and the 1H NMR values of double bond integrals.
Conventional Quality Parameter
Fig. 4 Comparison of specific NMR integrals with the conventional iodine value.
Due to the falsified values by simultaneous quantification of other double bond containing substances,
the double bond value (DBV) was provided. This value described the mean amount of double bonds per fatty
acid in the analyzed oil.
Conventionally, molecular weights of TAGs were determined with liquid or gas chromatography. But, the molecular
weight could be determined also with 1H NMR spectroscopy. The three TAGs Glyceryl trioleate, Glyceryl tristearate
and Glyceryl trinonadecanoate with known molecular weights were analyzed and quantified with 1H NMR (Tab. 1).
The from literature known molecular weights were compared with the quantified molecular weights (Fig. 5). The
comparison proved that the NMR spectroscopy is a versatile tool for the determination of molecular weights of edible
oils. Using NMR spectroscopy for molecular weight analysis was that all parameters except of the fatty acid distribution
were determined in only ONE analysis run leading to VERY FAST RESULTS.
Tab. 1 Molecular weights of TAGs.
Fig. 5 Comparison of quantified and known molecular weights of TAGs.
which is which?
Edible oils were analyzed for their FA profile being measured without an internal standard.
The FA profile of 10 different edible oil types (peanut, hazelnut, pumpkin seed, gold-of-pleasure,
linseed, almond, olive, rapeseed, sunflower, walnut oil), mostly used in households, were measured
with 1H NMR and evaluated with Principle Component Analysis (PCA) due to the high complexity and
signal overlapping. For the statistical evaluation the signals at d 0.85; 2.00; 2.75 and 2.80 ppm
were used. The samples could be separated into clusters identifying and confirming the authenticity of
edible oils. The first three components described 99.89% of the total variance (Fig. 6).
Fig. 6 PCA scores plot on the first and the second principle components for 1H NMR integrals of
the fatty acid profile of edible oils, most frequently used in the household.
Another PCA was performed for oils with identical or very similar FA profile, divided in Tab. 2. The first two
components described 88.28% of the total variance Fig. 7 illustrated the PCA score plot of edible oils with
the four similar FA profile groups. The clusters showed a wide internal dispersion. Group 1 (argan and peanut oil)
and group 3 (pumpkin and black seed oil) respectively could not be distinguished by the PCA. In group 2 (apricot,
hazelnut and sunflower oil), the sunflower oil could be separated by the clusters of apricot and hazelnut oil. Almond and
rape seed oil, group 4, could be distinguished by their separated clusters.
The 1H NMR results supported the capability of NMR combined with chemometric techniques to be used in quality
assessment and authenticity control of edible oils and in detection procedures of adulteration and incorrect labelling.
Fig. 7 PCA scores plot on the first and the second principle components for 1H NMR integrals of the fatty acid profile of
edible oils with similar profile.
Tab. 2 Classification of oils with similar fatty acid profiles.
Adulteration control –
is it real oil?
A series of 46 blends with a total weight of 300 mg was prepared from selected olive (n=5) and sunflower (n=5) oil
samples. The amount of sunflower oil in these mixtures was 0%, 20%, 40%, 60%, 80% and 100% (w/w). In each mixture the
samples from both origins were selected randomly in order to assure representativeness of multivariate regression models.
Recorded NMR spectra preprocessed by bucketing were correlated with “ground truth” data of the blends composition using partial
least squares regression (PLS).
Scatter plot of PLS scores for the first and the second PLS dimensions clearly demonstrated the discrimination of all spiking groups (Fig. 8).
Fig. 8 PLS scores on the first and the second PLS components for cross validation for olive-sunflower blends. Ellipsoids show 95%
probability for each “spiking group” (defined as an average sunflower oil content in each group).
Fig. 9 Predicted-reference plot and 95% prediction bands (solid lines) obtained from spiking experiments.
Fig. 9 showed the predicted-reference plot along with 95% prediction bands for sunflower oil content in blends obtained by PLS regression. Root mean square error of regression was equal to 1.1% w/w. Limit of detection (LOD) estimated using the slope and standard deviation of response of this curve was found to be 1.1%. Quantification of olive oil adulteration is possible starting from 3.5% w/w.
 Diehl, B., Skiera, C.; Nuclear Magnetic Resonance (NMR)-Spektrokopie. [book auth.] B. Matthäus and H.-J. Fiebig. Speiseöle und -fette; Recht, Sensorik, Analytik. EU : Erling, 2013, pp. 169-180.
 Diehl, B.; Multinuclear high-resolution nuclear magnetic resonance spectroscopy. [book auth.] R. J. Hamilton. Lipid Analysis in Oils and Fats. : Blackie Academic & Professional, 1998, pp. 87-135.
 Skiera, C., et al.; 1H NMR approach as an alternative to the classical p-anisidine value method. Eur Food Res Technol. 2012, 235, pp. 1101–1105.
 Skiera, C., et al.; 1H NMR spectroscopy as a new tool in the assessment of the oxidative state in edible oils. J Am Oil Chem Soc. 2012, 89, pp. 1383-1391.
 Skiera, C., et al.; Determination of free fatty acids in edible oils by 1H NMR spectroscopy. Lipid Technol. 2012, 24, pp. 279-281.
 Guillén, M., Ruiz, A.; Edible oils: dicrimination by 1H nuclear magnetic resonance. J. Sci. Food Agric. 2003, 83, pp. 338-346.