AReviewofThermo-OxidativeDegradationof …the volatile compounds formed in thermo-oxidation...

A Review of Thermo-Oxidative Degradation ofFood Lipids Studied by 1H NMR Spectroscopy:Influence of Degradative Conditions and FoodLipid NatureAndrea Martınez-Yusta, Encarnacion Goicoechea, and Marıa D. Guillen

Abstract: This review summarizes present-day knowledge provided by proton nuclear magnetic resonance (1H NMR)concerning food lipid thermo-oxidative degradation. The food lipids considered include edible oils and fats of animaland vegetable origin. The thermo-oxidation processes of food lipids of very different composition, occurring at low,intermediate, or high temperatures, with different food lipid surfaces exposed to oxygen, are reviewed. Mention ismade of the influence of both food lipid nature and degradative conditions on the thermo-oxidation process. Interest isfocused not only on the evolution of the compounds that degrade, but also on the intermediate or primary oxidationcompounds formed, as well as on the secondary ones, from both qualitative and quantitative points of view. Very valuablequalitative and quantitative information is provided by 1H NMR, which can be useful for metabolomic and lipidomicstudies. The chemical shift assignments of spectral signals of protons of primary (hydroperoxides and hydroxides associatedwith conjugated dienes) and secondary, or further (aldehydes, epoxides, among which 9,10-epoxy-12-octadecenoate[leukotoxin] can be cited, alcohols, ketones) oxidation compounds is summarized. It is worth noting the ability of 1HNMR to detect toxic oxygenated α,β-unsaturated aldehydes, like 4-hydroperoxy-, 4,5-epoxy-, and 4-hydroxy-2-alkenals,which can be generated in the degradation of food lipids having omega-3 and omega-6 polyunsaturated groups in bothbiological systems and foodstuffs. They are considered as genotoxic and cytotoxic, and are potential causative agents ofcancer, atherosclerosis, and Parkinson’s and Alzheimer’s diseases.

Keywords: characterization and quantification, food lipids, oxidation compounds, proton nuclear magnetic resonance(1H NMR), thermo-oxidation

IntroductionLipid oxidation involves a large group of complex reactions in-

duced by oxygen in the presence of initiators, among which areheat, free radicals, light, photosensitizing pigments, and metal ions.It has been described that this process takes place through a se-quential free radical chain reaction mechanism, which implies thewell-known 3 stages of initiation, propagation, and termination(Frankel 2005).

In short, the initiation stage involves the homolytic breakdownof hydrogen in the α-position in relation to the double bond in anunsaturated acyl group, forming unstable free radicals that stabi-lize by abstracting a hydrogen free radical from another chemicalspecies. It has been postulated that hydroperoxides and hydroxidesare formed at the propagation stage, and that they in turn gen-erate free radicals; the formation of hydroperoxides or hydroxides

MS 20140255 Submitted 16/2/2014, Accepted 25/4/2014. Authors Martinez-Yusta, Goicoechea, and Guillen are with Dept. of Food Technology, Lascaray ResearchCenter, Faculty of Pharmacy, Univ. of the Basque Country (UPV/EHU), Vitoria,Spain. Direct inquiries to author Guillen (E-mail: [email protected]).

can occur simultaneously with the formation of conjugated doublebonds. The termination stage takes place when the collision of freeradicals provokes the formation of stable molecules, thus makingthe free radicals disappear. The variety of compounds which mayform in these reactions is great and their formation occurs sequen-tially; for this reason, some of them are called primary oxidationcompounds, and those derived from them are called secondaryoxidation compounds, and there may also be tertiary, or further,oxidation compounds formed.

Pioneer studies carried out with standard compounds made avery important contribution to our knowledge of these processes.In these studies, extraction and separation techniques, and in manycases proton nuclear magnetic resonance (1H NMR) spectroscopy,were used to identify and quantify some of the compounds formed(Perkins and Anfinsen 1971 [using a Varian A-60 spectrome-ter]; Gardner and Weisleder 1972 [using a 100-MHz spectrom-eter]; Wineburg and Swern 1974 [using a Varian XL-100 spec-trometer]; Neff and others 1981, 1982 [using a Bruker WH-90spectrometer]; Neff and others 1983, 1988, 1990; Neff and Frankel1984 [using a Bruker WM-300 spectrometer]; Frankel and others

838 Comprehensive Reviews in Food Science and Food Safety � Vol. 13, 2014C© 2014 Institute of Food Technologists®

doi: 10.1111/1541-4337.12090

Food lipid thermo-oxidation studied by 1H NMR . . .

1982 [using a Bruker WH-90 spectrometer]; Frankel and others1990 [using a Bruker WM-300 spectrometer]; Yamauchi and oth-ers 1983 [using a Hitachi R-22 90 MHz spectrometer]; Coxon andothers 1984 [using a 300-MHz spectrometer]; Porter and others1994 [using a 300-MHz spectrometer]; Butovich 2006; Butovichand others 2006 [using a Varian 400-MHz spectrometer]; Dobsonand others 2013 [using a Bruker Ascend 500 MHz or a JEOLEclipse 400-MHz spectrometer]). In recent years, Hamalainen andKamal-Eldin (2005) have reviewed the 1H and 13C NMR spec-tral data of some of these standards of lipid oxidation compounds.Gas chromatography (GC) has also been essential in the study ofthe volatile compounds formed in thermo-oxidation processes ofstandard lipids (Thompson and others 1978; Frankel and others1981, 1982, 1988), as has high-performance liquid chromatogra-phy (HPLC) in the study of the polymers, oligomers, and dimersformed in thermo-degradation processes (Dobarganes and Perez-Camino 1987; Marquez-Ruiz and others 1995; Marquez-Ruizand Dobarganes 2005; Dobarganes and Marquez-Ruiz 2007).

As it is known that certain compounds or functional groups canbe formed in food lipid thermo-oxidation processes, many studieshave been devoted to the determination of their concentrations.Thus, parameters such as peroxide value and conjugated dieneshave been employed to evaluate the occurrence of primary oxida-tion compounds; however, it has been recently shown that certainsecondary oxidation compounds can also support hydroperoxygroups and conjugated dienic systems (Guillen and Ruiz 2004b,2005a, 2005b, 2005c). Similarly, anisidine value and thiobarbituricacid reactive substances (TBARS) tests have been used to evaluatethe occurrence of secondary oxidation products; however, the use-fulness of both methods has also been subject to criticism (Guillen-Sans and Guzman-Chozas 1998; Laguerre and others 2007). Theseclassical methods give concentration values of certain compoundsor functional groups in a generic way, but do not provide infor-mation about the specific nature of the compounds involved ineach determination, which sometimes could give rise to inaccu-rate conclusions (Frankel 2005). Thus, these classical methods donot shed further light on the thermo-oxidative mechanisms of oilsand caution is required in their interpretation.

Techniques that study food lipids as a whole have hardly everbeen used in the analysis of food lipid thermo-oxidation. Two ofthe most important ones are Fourier transform infrared (FTIR)spectroscopy (Guillen and Cabo 1997, 1999, 2000) and NMRspectroscopy. However, as they are fast and do not require mod-ification or special preparation of the sample, studies aimed atevaluating their usefulness in this field have recently been carriedout.

1H NMR spectroscopy has been gaining increasing interest be-cause it is used in metabolomic and lipidomic approaches. How-ever, in many of these studies the use of 1H NMR spectra islimited to the introduction of spectral profile data in statisticalanalysis in order to extract information on processes or on samplesunder the light shed by the statistics. Nevertheless, to extract therelevant information contained in the 1H NMR spectra a correctassignment of the signals to the corresponding protons is essential.This assignment allows a correct identification of molecules orfunctional groups involved, and also the development of adequatequantitative approaches that permit one to obtain not only qual-itative but also the quantitative information which is essential todraw sound conclusions.

In this context, a review of the use of 1H NMR spectroscopy inthe study of food lipid thermo-oxidation is presented here. Thisreview covers not only the influence of the food lipid nature and

of the degradation conditions on the evolution of lipid thermo-oxidation processes, but also the nature and concentration of thecompounds formed which may be detected by this technique.The aim of this review is to help in providing valuable toolsto improve the conclusions in studies in which lipids and theirthermo-oxidation processes are involved.

1H NMR as a Tool for the Study of Food Lipid Thermo-Oxidation Processes

As mentioned, this spectroscopic technique is able to study thefood lipid sample as a whole, before and after submission to anydegradative process. For this reason, if the information provided issound, it will be a very useful technique.

Characterization of food lipids before submission to oxida-tive conditions

Several studies have shown the usefulness of this technique inthe characterization of nonoxidized food lipids. Since the pioneerstudies of Hopkins and Bernstein (1959) and also of Johnson andShoolery (1962) using equipment of 40 MHz and of 60 MHz,respectively, many advances have been made. In spite of the lim-ited resolution of the spectra obtained with this equipment, theabove-mentioned authors indicated that this technique could beuseful not only for identification, but also for quantification of oilcomponents.

Concerning food lipid main components. Nowadays, due to sev-eral well-received contributions, the assignment of the signals of1H NMR spectra to the protons of the main nonoxidized foodlipid components, namely of triglycerides, and also of some minorfood lipid components, is well established (Hopkins and Bernstein1959; Johnson and Shoolery 1962; Gunstone 1990; Wollenberg1991; Aursand and others 1993; Sacchi and others 1993, 1998;Segre and Mannina 1997; Igarashi and others 2000; Guillen andRuiz 2001, 2003a; Siddiqui and others 2003; Guillen and others2008; Mannina and others 2008; Vidal and others 2012). As an ex-ample, Figure 1 shows a region of the 1H NMR spectra, recordedwith 400 MHz equipment, of lipids of animal origin such as porkadipose tissue (PAT), farmed sea bass (Dicentrarchus labrax; SB), andof 3 vegetable oils, namely virgin linseed (VL), extra-virgin olive(EVO), and sunflower (SF), dissolved in deuterated chloroform.All signals observable in this figure are due to protons of severaltriglycerides of food lipids. Table 1 gives their assignments to thecorresponding protons, letters in Figure 1 and in Table 1 being inagreement. A thorough explanation of each of these signals andits assignment to the different kinds of protons of triglycerideshas been carried out previously (Guillen and Ruiz 2001, 2003a;Guillen and others 2008; Vidal and others 2012).

All spectra given in Figure 1 have been represented on the samescale. The different intensity of the common signals in all spec-tra given, and the presence or absence of others, allows an initialcharacterization of these food lipids by simple and direct obser-vation; that is to say, without further enlargement of the spectra.This task is even easier when some of the signals are enlarged,as shown in Figure 2. Thus, the difference between PAT lipidsand vegetable oils is clear, due to the high content of saturated(S) groups in the former, which is observable by the intensity ofthe signal at 1.259 ppm. It is also evident that SF oil does notcontain linolenic (Ln) acyl groups (absence of signals at 0.972,2.093, 2.111, and 2.801 ppm), but has a very important concen-tration of linoleic (L) acyl groups (intensity of signals at 0.889,2.036, 2.056, and 2.765 ppm). Similarly, it can be seen with-out further enlargement that PAT and EVO oil have very small

C© 2014 Institute of Food Technologists® Vol. 13, 2014 � Comprehensive Reviews in Food Science and Food Safety 839


Figure 1–1H NMR spectra of 2 lipid samples extracted from animal matrixes, pork adipose tissue (PAT) and sea bass (D. labrax; SB), together with 3vegetable oils, which are virgin linseed (VL), extra-virgin olive (EVO), and sunflower (SF). These spectra were acquired in a Bruker Avance 400spectrometer operating at 400 MHz. Signal letters agree with those in Table 1.

concentrations of Ln and L acyl groups, but important ones of oleic(O; intensity of signals at 0.879, 1.270, 2.002, and 2.020 ppm), this latter acyl group being higher in EVO oil than in the lipidsof PAT. Finally, the abundance of ω-3 acyl groups in linseed oiland the lipids of sea bass is shown by the intensity of the sig-nals at 0.972, 2.093, and 2.111 ppm and between 2.773 and2.895 ppm. In addition, the high proportion of Ln acyl groupsin linseed oil is seen by the intensity of the signal at 2.801 ppm,whereas in sea bass lipids that of eicosapentaenoic (EPA) plusarachidonic (ARA) and of docosahexaenoic (DHA) acyl groups isevidenced by the signals at 1.661 to 1.743 and at 2.364 to 2.421ppm, respectively. It must be pointed out that farmed fish usuallycontains lower concentrations of ARA acyl groups than do wildspecimens.

A much more accurate characterization can be carried out bythe determination of the molar percentages of the different kindsof acyl groups. Taking into account that the area of the above-mentioned signals is proportional to the number of protons thatgenerate them, and that the proportionality constant is the same inall the cases, several approaches have been developed to determinethe molar percentage of several kinds of acyl groups in food lipids.These have been thoroughly explained in previous papers (Aur-sand and others 1993; Igarashi and others 2000; Guillen and Ruiz2003b, 2004a, 2004b, 2005a, 2005b, 2005c, 2006, 2008; Guillenand Goicoechea 2007, 2009; Guillen and others 2008; Guillen andUriarte 2009, 2012a, 2012b, 2012c; Goicoechea and Guillen2010; Vidal and others 2012; Sopelana and others 2013; Martınez-Yusta and Guillen 2014a, 2014b). One of them, which is useful forvegetable oils and food lipids of animal origin, except fish lipids,involves the following equations:

Linolenic groups Ln (%) = 100(AJ/3AK) (1)

Linoleic groups L (%) = 100(2AI /3AK) (2)

Oleic (or monounsaturated) groups O (or MU) (%)

= 100[(AF − 2AI − AJ)/3AK] (3)

Saturated groups S (%) = 100[1 − (AF/3AK)] (4)

In these equations AJ, AI, and AF are the area of the signal J(bis-allylic protons of Ln groups), signal I (bis-allylic protons of Lgroups), and signal F (mono-allylic protons of all unsaturated acylgroups), respectively, indicated in Table 1 and Figure 1, and AK

is the area of signal K due to the protons of carbons 1 and 3of the glycerol backbone of triglycerides. Using these equations,the molar percentages of the above-mentioned acyl groups in theseveral lipids were determined and these are given in Table 2;the results are in agreement with those expected for oils and fatsof this nature. The accuracy of these determinations made from1H NMR spectral data has been demonstrated by using standardmixtures of triglycerides, giving satisfactory results (Guillen andRuiz 2003b).

Similarly, the molar percentage of certain acyl groups of fishlipids can also be determined from 1H NMR spectra using dif-ferent approaches, which have been very thoroughly explainedin previous papers (Guillen and Ruiz 2004a; Guillen and others2008; Vidal and others 2012); one of these possible approachesinvolves these equations:

Total ω−3 groups ω−3(%) = 100(4AB)/3(AH+2AG) (5)

Docosahexaenoic groups DHA (%) = 100AH/(AH+2AG) (6)

840 Comprehensive Reviews in Food Science and Food Safety � Vol. 13, 2014 C© 2014 Institute of Food Technologists®


Table 1–Chemical shift assignments and multiplicities of the 1H NMR signals in CDCl3 of the main acyl groups and minor compounds of food lipids. Thesignal letters agree with those given in Figures 1 to 4.

ChemicalSignal shift (ppm) Multiplicity Functional group

Main acyl groupsA 0.879 t −CH3 Saturated, monounsaturated ω-9 and/or ω-7 acyl groupsA 0.889 t −CH3 Unsaturated ω-6 acyl groupsB 0.972 t −CH3 Unsaturated ω-3 acyl groupsC 1.221 to 1.419 ma −(CH2)n− Acyl groupsD 1.522 to 1.700 m −OCO−CH2−CH2− Acyl groups except for DHA, EPA and ARA acyl groupsE 1.661 to 1.743 m −OCO−CH2−CH2− EPA and ARA acyl groupsF 1.941 to 2.139 mb −CH2−CH=CH− Acyl groups except for -CH2- of DHA acyl group in β-position

in relation to carbonyl groupG 2.305 dt −OCO−CH2− Acyl groups except for DHA acyl groupsH 2.364 to 2.421 m −OCO−CH2−CH2− DHA acyl groupsI 2.765 t =HC−CH2−CH= Diunsaturated ω-6 acyl groupsJ 2.801 t =HC−CH2−CH= Triunsaturated ω-3 acyl groupsJ’ 2.773 to 2.895 m =HC−CH2−CH= DHA, EPA, ARA and other polyunsaturated ω-3 acyl groupsK 4.139, 4.303 dd,dd −CH2OCOR Glyceryl groupsL 5.225 to 5.296 m >CHOCOR Glyceryl groupsM 5.296 to 5.470 m −CH=CH− Acyl groups

Minor compoundsN 0.540 s −CH3 (C−18) �7-avenasterolO 0.681 s −CH3 (C−18) CholesterolP 0.684 s −CH3 (C−18) β-sitosterol and �5-campesterolQ 0.704 s −CH3 (C−18) �5-stigmasterol and brassicasterolR 1.670 s −CH3 SqualeneS 3.350 s −N(CH3)3 PhosphatidylcholineT 3.725 tc −CH2OH 1,2-diglyceridesU 4.949 to 5.072 dq,dq −CH=CH2 Unsaturated ω-1 acyl groupsV 5.749 to 5.858 m −CH=CH2 Unsaturated ω-1 acyl groups

aOverlapping of multiplets of methylenic protons in the different acyl groups either in β-position, or further, in relation to double bonds, or in γ -position, or further, in relation to the carbonyl group.bOverlapping of multiplets of the α-methylenic protons in relation to a single double bond of the different unsaturated acyl groups.cStandard compounds (dioleoin and dipalmitin) give a triplet centered at 3.73 ppm, and in complex mixtures the signal generated is a doublet centered at 3.72 ppm instead of a triplet (Sopelana and others2013).t, triplet; m, multiplet; DHA, docosahexaenoic acyl groups; EPA, eicosapentaenoic acyl groups; ARA, arachidonic acyl groups; d, doublet; s, singlet; q, quartet.

Eicosapentaenoic plus arachidonic groups EPA + ARA (%)= 100(2AE)/(AH+2AG)

(7)

Diunsaturated ω−6 (mainly linoleic) groups DUω − 6 (%)= 100(2AI)/(AH+2AG) (8)

Total unsaturated groups U (%) = 100(2AF+AH)/(2AH+4AG)(9)

In these equations AB is the area of signal B due to methylicprotons of ω-3 acyl groups, AH is the area of signal H due tomethylenic protons in α- and β-positions in relation to the car-bonyl group of DHA groups, AG is the area of the signal G relatedto the methylenic protons in the α-position in relation to carbonylgroups, except those of DHA groups, and finally AI and AF wereas defined. All these signals are indicated in Figure 1 and in Table 1.

Concerning food lipid minor components. In addition to maincomponents, nonoxidized food lipids also contain minor compo-nents, which are also important because some of them can showantioxidant ability or interesting nutritional or health-related prop-erties. If these compounds are present in high enough concentra-tions for them to be detected by 1HNMR, and if the signals oftheir protons do not overlap with those of the main lipid compo-nents, then this technique provides information about them. Someof these detected in the above-mentioned food lipids are shown inFigure 3. The assignment of these signals is also given in Table 1. Ithas been found that food lipids of animal origin, like those of PATand sea bass, contain cholesterol (signal O) containing approximateconcentrations near 93 and 734 mg/100 g, respectively (Chiz-zolini and others 1999; Orban and others 2003). Vegetable oilscontain sterols, with β-sitosterol plus �5-campesterol (signal P),�5-stigmasterol plus brassicasterol (signal Q) being present, as well

as �7-avenasterol (signal N) in SF oil (Kurata and others 2005;Zhang and others 2006a; Sopelana and others 2013). The con-tents of sterols in vegetable oils can be very varied, even in thoseof the same vegetable origin; concentrations ranging from 0 to206 mg/100 g of individual sterols have been found in differentvegetable oils (Schwartz and others 2008). However, it must betaken into account that when operating at 400 MHz there is anoverlap between the signal of �5-stigmasterol and brassicasteroland the side band of the methylic protons of linoleic, oleic, andsaturated acyl groups.

Other very interesting minor components are squalene(signal R), present mainly in EVO oil (Mannina and others 2003,2008), and phosphatidylcholine (signal S), present in many lipidsof animal origin like sea bass lipids (Siddiqui and others 2003).Furthermore, some lipids, such as VL, EVO, and SF oils, as wellas PAT, also contain 1,2-diglycerides (signal T; Sacchi and others1996; Segre and Mannina 1997). In addition, it is worth men-tioning the occurrence in sea bass lipids of ω-1 acyl groups whichgive typical signals U and V. Although these had been alreadyobserved in previous studies on salmon lipids (Guillen and Ruiz2004a), they have recently been identified (Shimizu and others1991; Fiori and others 2012; Vidal and others 2012).

The concentration of these compounds in the correspond-ing food lipids can be determined from the area of the above-mentioned signals. It is possible to do this determination eitherin an absolute way, using a standard compound, or in relation toother food lipid components present in the same sample.

Characterization of food lipids after submission to thermo-oxidative conditions

During thermo-oxidative processes changes are produced inthe original food lipid components and, as mentioned before,



Figure 2–Enlargement of 1H NMR spectral signals corresponding to the main acyl groups of the triglycerides of pork adipose tissue lipids (PAT),farmed sea bass lipids (SB), virgin linseed oil (VL), extra-virgin olive oil (EVO), and sunflower oil (SF). These spectra were acquired in a Bruker Avance400 spectrometer operating at 400 MHz. Signal letters agree with those in Table 1.

new compounds are generated. Those formed in the 1st stepare named primary oxidation compounds and the ones derivedfrom them are called secondary or further oxidation compounds.First, the evolution of the original food lipid components will becommented on, followed by the formation and evolution of thedifferent oxidation compounds.

Evolution of main and minor original lipid components. Underthermo-oxidative conditions both main and minor componentscan evolve to give new compounds, and this is reflected in theintensity of the signals of their protons, which undergo changes,which are observable in the region of the spectra in which thesesignals appear. Figure 4(a), as an example, shows the spectral re-gions of the original SF oil and of the same oil after submission,for different heating times, to 2 different temperatures (70 °C and190 °C) under different processing conditions (10 g of oil in Petridishes and 4 L of oil in a deep-fryer, respectively; Goicoechea andGuillen 2010; Guillen and Uriarte 2012c). It can be observed inthe spectrum of SF oil submitted to 70 °C for 264 h (11 d), un-der certain oxidative conditions, that the intensity of the signals of

linoleic (L) protons (signals at 0.889, 2.036, 2.056, and 2.765 ppm)have drastically decreased; however, in the spectrum of SF oil sub-mitted to 190 °C for 40 h under other thermo-oxidative condi-tions, although the intensity of the same signals due to linoleic (L)groups has been reduced in relation to its original intensity, theyhave not disappeared from the spectra. As Figure 4(a) shows, theevolution of the different acyl groups throughout the processingtime, under any processing conditions, can be observed directlyin the 1H NMR spectra. Furthermore, this evolution can alsobe quantified using the appropriate equations, developed for eachspecific case on the basis of the same principles as mentioned fornonoxidized lipids (Guillen and Uriarte 2012a, 2012b, 2012c).Figure 4(b) shows the quantification of the evolution throughoutprocessing time of the molar percentage of the several kinds ofacyl groups in SF oil submitted to the above-mentioned thermo-oxidative conditions. It can be observed that great differences arefound, depending on the thermo-oxidation conditions.

Similarly, if minor lipid components are affected during thethermo-oxidative degradation, the intensity of their proton



Table 2–Composition of food lipid samples shown in Figures 1 to 3, obtained from 1H NMR data and expressed as molar percentage of the differentkinds of acyl groups (data taken from Guillen and Uriarte 2012a, 2012b, 2012c; Vidal and others 2012; Martınez-Yusta and Guillen 2014a, 2014b).

Food lipid Ln or ω-3 (%) L (%) O or MU (%) S (%)

Pork adipose tissue lipids (PAT) 1.67 ± 0.08 10.21 ± 0.02 46.83 ± 0.31 41.29 ± 0.41Sea bass lipids (SB) 22.10 ± 1.60

(DHA 8.09 ± 1.45) 14.28 ± 2.92 36.36 ± 2.97 26.46 ± 2.16(EPA 9.49 ± 1.59)

Virgin linseed oil (VL) 50.37 ± 0.50 17.85 ± 0.03 22.56 ± 0.37 9.20 ± 0.19Extra-virgin olive oil (EVO) 0.49 ± 0.02 4.49 ± 0.28 82.06 ± 0.98 12.93 ± 0.65Sunflower oil (SF) - 55.05 ± 4.31 34.80 ± 2.83 10.03 ± 0.95

Acyl groups: Ln, linolenic; ω-3, omega-3; L, linoleic; O or MU, oleic or monounsaturated; S, saturated; DHA, docosahexaenoic; EPA, eicosapentaenoic.

Figure 3–Enlargement of 1H NMR spectral signals related to minor compounds present in pork adipose tissue lipids (PAT), farmed sea bass lipids (SB),virgin linseed oil (VL), extra-virgin olive oil (EVO), and sunflower oil (SF). These spectra were acquired in a Bruker Avance 400 spectrometer operatingat 400 MHz. Signal letters agree with those in Table 1.

signals in the spectra decreases or increases with the processingtime, and the evolution of their concentration can also be deter-mined, if the signals of their protons do not overlap with othersignals.

Primary oxidation compounds and their evolution throughoutthe thermo-oxidation process. The degradation of the acyl groupsof food lipids provokes the formation of new compounds, the onesformed at the 1st step being named primary oxidation compounds.Their natures can vary, and among them some acyl chains support-ing hydroperoxy or hydroxy groups have been found, as well asconjugated dienic systems, which can also have either Z,E or E,Eisomerism. The possibility that some of the hydroperoxy groupscan be supported on saturated and unsaturated acyl chains with-out conjugated dienic systems exists, as well as the possibility of2 hydroperoxy groups being supported in the same acyl chain,that is to say the occurrence of di-hydroperoxides in the same acylchain (Coxon and others 1984; Neff and others 1988; Schneider

and others 2004, 2005, 2008; Sun and Salomon 2004; Zhang andothers 2006b). The formation of each of these functional groupsdepends on both the food lipid nature and the thermo-oxidationconditions. In fact, under certain degradative conditions theoccurrence of these primary oxidation compounds is not detectedby this technique, which indicates either that they are not formedor that, if they are formed, their degradation rate is too high forthem to be detected by 1H NMR (Guillen and Uriarte 2009,2012a, 2012b, 2012c).

Some of the protons of these compounds give signals in the 1HNMR spectra that do not overlap with those of acyl groups, thusmaking their identification and also their quantification possible.The identification of these signals has been based on previousstudies carried out with standard compounds (Neff and others1990; Frankel and others, 1990; Porter and others 1994; Neffand El-Agaimy 1996; Kuklev and others 1997; Jie and Lam 2004;Lin and others 2007; Pajunen and others 2008). Table 3 gives the



Figure 4–(a) 1H NMR spectra of sunflower oil (SF) submitted to 70 °C (10 g oil, Petri dish) and 190 °C (4 L oil, deep-fryer) for different periods. Thesespectra were acquired in a Bruker Avance 400 spectrometer operating at 400 MHz. (b) Evolution in both processes of the molar percentages oflinoleic (L), monounsaturated (MU), and saturated + modified (S+M) acyl groups along the heating time (Goicoechea and Guillen 2010; Guillen andUriarte 2012c).

chemical shifts of the signals of the protons of the above-mentionedprimary oxidation compounds. In addition, Figure 5(a) gives the1H NMR spectral region of nonoxidized corn oil and of this sameoil after storage at room temperature in a closed receptacle for 121mo (Guillen and Goicoechea 2009), of SF oil submitted to 70 °Cfor 72 h and to 100 °C for 9 h (Goicoechea and Guillen 2010).It can be observed that, depending on the degradative conditions,different kinds of primary oxidation compounds or functionalgroups can be formed or not.

The evolution of these primary oxidation compounds or func-tional groups throughout a thermo-oxidative process can be ob-served directly in the spectra and quantified, as mentioned before.This quantification can be carried out either in an absolute wayby using a standard compound or in relation to main functionalgroups present in the food lipids, such as triglycerides. As will beseen in subsequent sections, great differences have been found inrelation to the formation, or not, of primary oxidation compoundsin high enough concentrations for detection by this technique, inrelation to their nature and also to their formation and degradationrates. These differences between primary oxidation compoundsare functions of both the food lipid nature and of the conditionsunder which the thermo-oxidation occurs.

Secondary or further oxidation compounds and evolutionthroughout the thermo-oxidation process. Primary oxidationcompounds are unstable and degrade to generate secondary oxi-dation compounds, which can have very varied structures. Someof them have protons whose signals in the 1H NMR spectrado not overlap with any other, making it possible to identifyand quantify them. Table 3 gives the chemical shifts of someof the protons of different secondary oxidation compounds orfunctional groups which can be present in oxidized food lipids.Among them, there are different kinds of aldehydes, alcohols,epoxides, and ketones. The formation of these compounds alsodepends on both the original food lipid composition and theconditions of the thermo-oxidation process. Thus, aldehydes in-cluding n-alkanals, (E)-2-alkenals, (E,E)-2,4-alkadienals, (Z,E)-2,4-alkadienals, 4,5-epoxy-2-alkenals, 4-hydroxy-(E)-alkenals, 4-hydroperoxy-(E)-alkenals, as well as 4-oxo-alkanals can be found(Guillen and Ruiz 2004a, 2004b, 2005a, 2005b, 2005c, 2006,2008; Guillen and Goicoechea 2007, 2009; Guillen and Uriarte2009, 2012a, 2012b, 2012c; Goicoechea and Guillen 2010;Martınez-Yusta and Guillen 2014a,2014b). It should be noticedthat although aldehydes are generally considered as secondary oxi-dation compounds, all α,β-unsaturated aldehydes have conjugated



Table 3–Chemical shifts and multiplicities of the 1H NMR signals in CDCl3 of some protons of primary and secondary (or further) oxidation compoundsgenerated in edible oils submitted to different degradative conditions (data taken from Porter and others 1994; Kuklev and others 1997; Jie and Lam2004; Guillen and Ruiz 2005a, 2005b, 2005c; Guillen and Goicoechea 2007; Lin and others 2007; Guillen and Uriarte 2009, 2012a, 2012b, 2012c;Goicoechea and Guillen 2010; Martınez-Yusta and Guillen 2014a, 2014b). The signal letters agree with those given in Figures 5 to 9 and 11.

ChemicalSignal shift

(ppm)Multiplicity Functional group

Primary oxidation compoundsa 6.58 dddd −CH=CH−CH=CH− (Z,E)-conjugated doublea 6.00 ddtd bonds associated with

5.56 ddm hydroperoxides (OOH)5.51 dtm

b 6.45 ddd −CH=CH−CH=CH− (Z,E)-conjugated double5.94 dd bonds associated with5.64 dd hydroxides (OH)5.40 ddt

c 6.27 ddm −CH=CH−CH=CH− (E,E)-conjugated doublec 6.06 ddtd bonds associated withc 5.76 dtm hydroperoxides (OOH)

5.47 ddmd 5.72 m −CHOOH−CH = CH− Double bond associated with

hydroperoxides (OOH)e 8.3 to 8.9 — −OOH Hydroperoxide group

Secondary or further oxidation compoundsAldehydes

f 9.49 d −CHO (E)-2-alkenalsg 9.52 d −CHO (E,E)-2,4-alkadienalsh 9.55 d −CHO 4,5-epoxy-2-alkenalsi 9.57 d −CHO 4-hydroxy-(E)-2-alkenalsj 9.58 d −CHO 4-hydroperoxy-(E)-2-alkenalsk 9.60 d −CHO (Z,E)-2,4-alkadienalsl 9.75 t −CHO n-alkanalsm 9.78 t −CHO 4-oxo-alkanalsn 9.79 t −CHO n-alkanals of low molecular

weight (ethanal andpropanal)

Alcoholso 3.43 m −CHOH−CHOH− 9,10-dihydroxy-12-

octadecenoate(leukotoxindiol)

p 3.54–3.59 m −CHOH− secondary alcoholsq 3.62 t −CH2OH− primary alcohols

Epoxidesr 2.63 m −CHOHC− (E)-9,10-epoxystearates 2.88 m −CHOHC− (Z)-9,10-epoxystearatet 2.90 m −CHOHC− 9,10-epoxy-octadecanoate;

9,10-epoxy-12-octadecenoate(leukotoxin); 12,13-epoxy-9-octadecenoate(isoleukotoxin)

u 2.90 m −CHOHC−CHOHC− 9,10–12,13-diepoxy-octadecanoate

v 3.10 m −CHOHC−CH2−CHOHC− 9,10–12,13-diepoxy-octadecanoate

Ketones and unidentifiedw 6.08 dt O=C<CH=CH− Double bond conjugated with

a ketow 6.82 m groupx 7.50 ?? ?? Unidentifiedy 8.10 ?? ?? Unidentified

d, doublet; t, triplet; m, multiplet.

dienic systems and some of them also contain hydroperoxy groups.This fact indicates that neither the hydroperoxy group nor conju-gated dienic systems are exclusively present in primary oxidationcompounds, as mentioned before.

Concerning alcohols, primary and secondary ones, togetherwith dialcohols, have been found (Guillen and Uriarte 2009,2012a, 2012b, 2012c; Goicoechea and Guillen 2010; Martınez-Yusta and Guillen 2014a, 2014b). With regard to epoxides, thosederived from oleic and from linoleic acyl groups have been de-tected. The latter may be in the form of not only mono- but alsoof diepoxides (Goicoechea and Guillen 2010; Guillen and Uriarte

2012a; Martınez-Yusta and Guillen 2014a). And finally, ketoneshaving a conjugated dienic system (keto group conjugated with adouble bond) derived from oleic acyl groups have also been found.Although signals of these latter ketones were observed previously(Guillen and Ruiz 2005a, 2005c), this review is where they havebeen assigned for the first time on the basis of data for pure com-pounds provided by other authors (Kuklev and others 1997; Jieand Lam 2004; Lin and others 2007). However, it should be re-membered that none of these compounds can be formed in theoxidation of every food lipid, or under every thermo-oxidativecondition. Figure 5(b) shows the spectral regions in which some



Figure 5–(a) 1H NMR spectra of primary oxidation compounds formed in corn oil (CO), nonoxidized (0 h), and stored at room temperature in closedreceptacles for 121 mo, sunflower oil (SF) submitted (10 g oil, Petri dish) to 70 °C for 72 h and to 100 °C for 9 h. (b) 1H NMR spectra of secondary orfurther oxidation compounds formed in fish lipids extracted from salmon (SA) submitted to 50 °C for 768 h, sunflower oil (SF), and extra-virgin olive oil(EVO) submitted (10 g oil, Petri dish) to 70 °C for 216 h and 1560 h, respectively, and to 190 °C (4 L oil, deep-fryer) for 32.5 h (Guillen and Ruiz2004a, 2005a; Guillen and Goicoechea 2009; Goicoechea and Guillen 2010; Guillen and Uriarte 2012a, 2012c). These spectra were acquired in aBruker Avance 400 spectrometer operating at 400 MHz. Signal letters agree with those in Table 3.

of these compounds appear in salmon lipids from fillets submittedto 50 °C for 768 h (extraction using CS2 as solvent and assisted byan ultrasound bath), SF oil and EVO oil submitted to 70 °C, andto deep-frying conditions at 190 °C for different periods of time(Guillen and Ruiz 2004a, 2005a; Goicoechea and Guillen 2010;Guillen and Uriarte 2012a, 2012c).

The formation of all these compounds and the evolution oftheir concentration throughout the thermo-oxidative degradationprocess can be observed directly in the 1H NMR spectra andcan be quantified either in an absolute way, in relation to a stan-dard compound or in relation to a main functional group suchas triglycerides, this latter method being much easier and faster(Guillen and Goicoechea 2007, 2009; Guillen and Uriarte 2009,2012a, 2012b, 2012c; Goicoechea and Guillen 2010; Martınez-Yusta and Guillen 2014a, 2014b).

It is worth noting the capacity for 1H NMR spectroscopy todetect and quantify oxygenated α,β-unsaturated aldehydes, suchas 4,5-epoxy-, 4-hydroperoxy-, and 4-hydroxy-(E)-2-alkenals,which are considered as genotoxic and cytotoxic, and poten-tial causal agents of cancer, atherosclerosis, Parkinson’s disease,

or Alzheimer’s disease, among others (Esterbauer and others 1991;Guillen and Goicoechea 2008). They are generated in the degra-dation of ω-3 and ω-6 polyunsaturated acyl groups, fatty acids,and esters, both in biological systems and foodstuffs, and can beabsorbed through the diet. Similarly, it should be noticed that thistechnique permits detection and quantification of 9,10-epoxy-12-octadecenoate (leukotoxin) and 12,13-epoxy-9-octadecenoate(isoleukotoxin) groups, derived from linoleic acyl groups andbonded to triglyceride backbones (Goicoechea and Guillen 2010).The former is thus named as it can be formed endogenously, causesdegeneration and necrosis of leukocytes, and has been associatedwith multiple organ failure, breast cancer, cell proliferation in vitro,and disruption of reproductive functions in rats (Markaverich andothers 2005; Thompson and Hammock 2007).

In the thermo-oxidation of food lipids of a specific composi-tion, the types and concentrations of primary and also of secondaryoxidation compounds formed, and also the timing of their forma-tion, depend on the conditions under which the process takesplace. It is also true that under the same thermo-oxidative condi-tions the types and concentrations of primary and also of secondary



oxidation compounds formed and also the timing of their forma-tion depend on the food lipid nature. Both food lipid composi-tion and conditions of the thermo-oxidation process govern thephysico-chemical rules under which the process evolves. It is ob-vious that not all intermediate or all final oxidation compoundsmentioned are formed in any particular thermo-oxidation processor from any food lipid.

Evolution of food lipid oxidation under various degradativeconditions as seen in terms of 1H NMR

The first studies by 1H NMR (400 and 600 MHz) on the oxida-tion process of food lipids, and not on pure standard compounds,were carried out by Grootveld and coworkers, who submitted10 to 25 g of oil samples to 180 °C for 90 min, 20 to 60 mLof oil samples to pan-frying conditions with food (180 °C, 30min), and also analyzed several repeatedly used culinary frying oils,among them lard, beef fat, ghee, corn, SF, soybean, grape seed, co-conut, rapeseed, peanut, and EVO oils (Claxson and others 1994;Haywood and others 1995; Sheerin and others 1997; Silwood andGrootveld 1999). In short, they concluded that thermal stressingof culinary oils rich in polyunsaturated acyl groups generated highlevels of n-alkanals, (E)-2-alkenals, and (E,E)-2,4-alkadienals, andthe tentatively identified 4-hydroxy-(E)-2-alkenals, via decompo-sition of their precursor hydroperoxides associated with conjugateddienes. When oils with a low unsaturation degree and lard weresubjected to the above-mentioned heat treatments, only low con-centrations of these aldehydes were produced. The authors alsodiscussed the dietary, physiological, and toxicological significanceof these results regarding frying practices. Furthermore, using1,3,5-trichlorobenzene as an external standard, the concentration(mol/kg oil) of total unsaturated and saturated aldehydes generatedin corn oil, submitted to 180 °C for 90 min in vessels of differ-ent size, was determined. It was concluded that, as expected, thelonger the heating time and the bigger the vessel size (and there-fore the exposure to oxygen), the higher the concentration of bothsaturated and unsaturated aldehydes (Haywood and others 1995).

In this section the information obtained by 1H NMR on themechanisms of lipid oxidation that takes place under differentconditions is reviewed, with regard to degradation of main com-ponents, namely the acyl groups of triglycerides, and also to theformation of oxidation compounds. Special attention is paid to theinfluence of both degradative conditions and food lipid nature.Table 4 summarizes the different food lipid oxidation processesstudied by 1H NMR and reviewed in the following sections.

Evolution of the oxidation of refined SF and corn oils at roomtemperature in closed receptacles. As for the experimental con-ditions of these studies, different amounts of refined SF and cornoils, both rich in linoleic acyl groups, were stored separately inclosed receptacles, in the presence of different air volumes andhaving different air/oil contact surfaces. These oil samples weremaintained at room temperature for different periods of time upto 10 y (Guillen and Goicoechea 2007, 2009). Under these con-ditions the oils underwent oxidation very slowly.

A study by 1H NMR of the different samples evidenced thatsome of them were nonoxidized, others were at incipient or in-termediate oxidation stages, and some others at very advancedoxidation stages. The 1H NMR spectra showed that under theseconditions the oxidation process of the above-mentioned oils hasa prolonged period of initiation (induction period), after whichthe formation of primary oxidation compounds begins as a con-sequence of the degradation of the acyl groups, in this casemainly linoleic. Figure 6(a) shows the enlarged regions of the

1H NMR spectra of several SF oil samples in different oxidationstages (samples S6, S7, S17, S18, and S23), in which the evo-lution of the primary oxidation compounds is shown. It can beobserved that this oxidation process evolves, forming primary ox-idation compounds having initially hydroperoxy groups as wellas (Z,E)-conjugated dienic systems. At more advanced oxidationstages, hydroperoxy groups associated with (E,E)-conjugated di-enic systems and hydroxy groups supported on chains also having(Z,E)-conjugated dienic systems appear, both in smaller concen-trations than hydroperoxy-(Z,E)-conjugated dienic systems. Ac-cording to other authors, hydroperoxy-(Z,E)-conjugated dienicsystems are considered to be formed under kinetic control at lowtemperatures, because they are formed earlier (Schneider 2009).However, with prolonged autoxidation or at higher temperatures,the double-bond configuration of the conjugated hydroperoxidesrearranges from (Z,E) to (E,E). The apparent driving force forthis transformation is the greater thermodynamic stability of the(E,E)-conjugated double bond in comparison with the (Z,E).Hydroperoxy-(E,E)-conjugated dienic systems could be formedat the expense of the (Z,E), and therefore be considered to formunder thermodynamic control, because they are formed later andare more stable.

Many different molecular structures can be associated with theabove-mentioned functional groups. Figure 6(b) shows some ofthe structures, which are present in oxidation compounds, for ex-ample, 13-hydroperoxyoctadeca-9,11-dienoate, 9-hydroperoxy-,and 9-hydroxyoctadeca-10,12-dienoates, all derived from linoleicacyl groups and able to be formed in this process.

It is noteworthy that, of all the oxidation processes analyzed,only in those that occur under these conditions (room tempera-ture and closed receptacles) has the formation of hydroxy groupssupported on long acyl chains having also a conjugated dienicsystem been observed; this could be related to the limited con-centration of oxygen in the closed receptacles. As far as we know,only 1 previous paper has reported the formation, in SF oil afterprolonged storage, of (Z,E)-conjugated diene primary oxidationproducts having hydroxy groups (Mikolajczak and others 1968).The formation of this kind of compounds is known to occur inin vivo oxidation processes, and, in fact, they are considered asclinical markers of lipid peroxidation and oxidative stress, usuallydetected in urine and serum (Spindler and others 1997; Browneand Armstrong 2000). Similarly, the formation of either (Z,E)- or(E,E)-conjugated dienic systems related to hydroperoxy groups ishighly dependent on the conditions to which the oil is submitted.Very different situations are possible, as the following sections willshow.

All the above-mentioned compounds are usually called primaryoxidation products; however, as Figure 6(a) shows (see S17 and S18samples), the formation of the hydroxy-(Z,E)-conjugated dienicsystems, and of hydroperoxy-(E,E)-conjugated dienic systems, oc-curs simultaneously with that of aldehydes (signals between 9.4 and9.8 ppm). It can be observed in Figure 6(a) that the appearance ofthese 3 kinds of groups is incipient in S17 sample and that theirsignals are clearly observable in the S18 sample and onwards.

The conditions under which these oils are stored affect not onlythe nature and concentration of the hydroxy and hydroperoxycompounds mentioned above, but also those derived from them,like aldehydes and epoxides. As Figure 7(a) shows, the main alde-hydes formed were n-alkanals, (E)-2-alkenals, and 4-hydroxy-(E)-2-alkenals. In smaller concentrations (E,E)-2,4-alkadienals and4-hydroperoxy-(E)-2-alkenals and, in very small proportions, 4,5-epoxy-2-alkenals were generated. Molecular structures of some of



Tabl

e4–

Sum

mar

yof

stud

ies

onfo

odlip

idox

idat

ion

proc

esse

sca

rrie

dou

tby

1H

NM

R.

Tim

eun

der

Air

/oi

lox

idat

ive

cont

act

Food

lipid

sTe

mpe

ratu

re(°

C)co

ndit

ions

Rece

ptac

leO

ilvo

lum

esu

rfac

eH

eati

ngso

urce

Aer

atio

nRe

fere

nce

Sunfl

ower

Room

tem

pera

ture

Up

to10

yCl

osed

rece

ptac

les

Var

ied

Und

eter

min

ed—

Not

Gui

llen

and

Goi

coec

hea

(200

7)Co

rnRo

omte

mpe

ratu

reU

pto

10y

Clos

edre

cept

acle

sV

arie

dU

ndet

erm

ined

—N

otG

uille

nan

dG

oico

eche

a(2

009)

Fish

lipid

s(Sa

lmo

sala

r)50

Up

to32

dSa

lmon

tissu

e—

—H

eate

dai

rFo

rced

Gui

llen

and

Ruiz

(200

4a)

Sesa

me

70U

ntil

poly

mer

izat

ion

(20

d)O

pen

Petr

idis

hes

10g

50.3

cm2

Hea

ted

air

Forc

edG

uille

nan

dRu

iz(2

004b

)

Virg

inol

ive,

haze

lnut

,an

dpe

anut

70U

ntil

poly

mer

izat

ion

(72,

20,a

nd51

d)O

pen

Petr

idis

hes

10g

50.3

cm2

Hea

ted

air

Forc

edG

uille

nan

dRu

iz(2

005a

)

Corn

,and

sunfl

ower

70U

ntil

poly

mer

izat

ion

(11

and

12d)

Ope

nPe

trid

ishe

s10

g50

.3cm

2H

eate

dai

rFo

rced

Gui

llen

and

Ruiz

(200

5b)

Rape

seed

,wal

nut,

and

linse

ed70

Unt

ilpo

lym

eriz

atio

n(1

3,5,

and

4d)

Ope

nPe

trid

ishe

s10

g50

.3cm

2H

eate

dai

rFo

rced

Gui

llen

and

Ruiz

(200

5c)

Sunfl

ower

100

Unt

ilpo

lym

eriz

atio

n(5

5h)

Ope

nPe

trid

ishe

s10

g50

.3cm

2H

eate

dai

rFo

rced

Goi

coec

hea

and

Gui

llen

(201

0)

Virg

inol

ive,

oliv

e,ha

zeln

ut,c

orn,

sunfl

ower

,soy

bean

,an

dlin

seed

190

Unt

ilpo

lym

eriz

atio

n(li

nsee

d7

han

dth

eot

hers

12h)

Ope

nPe

trid

ishe

s10

g50

.3cm

2H

eate

dai

rFo

rced

Gui

llen

and

Ruiz

(200

8)

Virg

inol

ive,

corn

,and

linse

ed19

0U

pto

4h

Ope

nPe

trid

ishe

s10

g50

.3cm

2M

icro

wav

eN

atur

alG

uille

nan

dRu

iz(2

006)

Sunfl

ower

190

Up

to32

hD

omes

ticfr

yer(

20.5

×23

.5×

13.5

cm)

2L

481.

7cm

2El

ectr

icre

sist

ance

Nat

ural

Gui

llen

and

Uria

rte

(200

9)Ex

tra-

virg

inol

ive

190

Up

to40

hIn

dust

rialf

ryer

(12

×30

×17

cm)

4L

360

cm2

Elec

tric

resi

stan

ceN

atur

alG

uille

nan

dU

riart

e(2

012a

)V

irgin

linse

ed19

0U

pto

20h

Indu

stria

lfry

er(1

2×

30×

17cm

)4

L36

0cm

2El

ectr

icre

sist

ance

Nat

ural

Gui

llen

and

Uria

rte

(201

2b)

Sunfl

ower

190

Up

to40

hIn

dust

rialf

ryer

(12

×30

×17

cm)

4L

360

cm2

Elec

tric

resi

stan

ceN

atur

alG

uille

nan

dU

riart

e(2

012c

)



Figure 6–(a) Some expanded regions of the 1H NMR spectra of sunflower oil (SF) stored at room temperature in closed receptacles for differentperiods (Guillen and Goicoechea 2007). These spectra were acquired in a Bruker Avance 400 spectrometer operating at 400 MHz. Signal letters agreewith those in Table 3. (b) Some of the primary oxidation compounds derived from linoleic acyl groups that can be detected by 1H NMR.

these compounds are given in Figure 7(b). It should be commentedon that both the nature and concentration of aldehydes formedfrom the same oil are very different depending on the conditionsunder which the oxidation process takes place. It is noticeable thatunder the conditions of this study (room temperature and lim-ited amount of air in closed receptacles) 4-hydroxy-(E)-2-alkenalsare formed in significant concentration and that 4-hydroperoxy-(E)-2-alkenals are also formed, but in much smaller concentra-tions than the former, both known to be genotoxic and cytotoxiccompounds. The fact that the concentration of 4-hydroxy-(E)-2-alkenals is higher than that of 4-hydroperoxy-(E)-2-alkenals couldbe due to either the limited concentration of oxygen in this oxida-tion process, which could favor the formation of hydroxy over thatof hydroperoxy aldehydes, or to the presence of hydroxy-(Z,E)-conjugated diene systems among the intermediate oxidation prod-ucts, which can evolve directly to give 4-hydroxy-(E)-2-alkenals,as has been described by Schneider and others (2004).

It is also worth mentioning that together with the formation ofaldehydes and of the other compounds mentioned, the generationof epoxy groups supported on acyl chains also occurs. Some theo-retical mechanisms proposed for the formation of epoxides involvethe participation of hydroperoxy groups and of double bonds andentail the disappearance of both groups (Lercker and others 2003).The protons of the carbon atoms bonded to the epoxy groupgive signals in the spectral region between 2.8 and 3.2 ppm, beingclearly observable in increasing concentration in S17, S18, and S23samples (see Figure 7a). They are derived from linoleic groups andare composed not only of monoepoxides (signals at 2.9 ppm) butalso of diepoxides (signals at 2.9 and 3.1 ppm), these latter being inthe samples at the most advanced oxidation stages. A wide varietyof structures can be found among epoxides derived from linoleic

acyl groups; it is worth noting the monoepoxide 9,10-epoxy-12-octadecenoate, whose hydrolysis gives leukotoxin, and its isomer,the monoepoxide 12,13-epoxy-9-octadecenoate whose hydrolysisgives isoleukotoxin. As mentioned before, in recent years a greatdeal of attention has been paid to these 2 compounds as a result oftheir biological activity (Markaverich and others 2005; Thompsonand Hammock 2007). According to 1H NMR data provided byother authors, the 2 protons corresponding to the carbon atomsin positions 9 and 10 of leukotoxin diol give a multiplet signal at3.43 ppm (Aerts and Jacobs 2004); in S23 spectra a broad signalcan be found there (see Figure 7a) that appears simultaneouslywith that of epoxy groups, and that could be assigned to the dihy-droxyleukotoxin derivative, as Table 3 shows. Like aldehydes, theformation of epoxides, their nature, and concentration are highlydependent on the oil nature and degradation conditions, as willbe seen later.

It is also remarkable that in these polyunsaturated oils (cornand SF) at room temperature all kinds of oxidation-derived com-pounds are present simultaneously for a very prolonged period,except at the early stages, when only hydroperoxy groups and(Z,E)-conjugated dienic systems are present (see Figure 6a and7a). This may be because at low temperatures the hydroperoxideshave longer average lives than at higher temperatures.

Evolution of the thermo-oxidation of fish lipids contained in dry-salted and unsalted salmon (Salmo salar) fillets submitted to 50 °Cwith aeration in a convection oven. This study was carried outon lipids contained in fillets of dry-salted and unsalted salmon(Salmon salar) submitted to thermo-oxidative conditions at 50 °Cwith aeration in a convection oven (Guillen and Ruiz 2004a).The lipids were extracted periodically after submitting fillets tothe above-mentioned degradative conditions for different periods



Figure 7–(a) Some expanded regions of the 1H NMR spectra of sunflower oil (SF) stored at room temperature in closed receptacles for differentperiods (Guillen and Goicoechea 2007). These spectra were acquired in a Bruker Avance 400 spectrometer operating at 400 MHz. Signal letters agreewith those in Table 3. (b) Some of the secondary or further oxidation compounds derived from linoleic acyl groups that can be detected by 1H NMR.

and were studied by 1H NMR; the extraction was carried out atlow temperature in order to avoid additional oxidation. Regardingsalmon lipid nature, their important content in ω-3 acyl groups(around 25% molar percentage), mainly DHA and EPA groups,and their low oxidative stability are well known. Nevertheless,their high carotenoid content, which is able to exhibit antioxidantability, is also well known.

In this case the oxidation process was fairly different to thatoccurring in the previous room temperature studies, not only re-garding the conditions to which the lipids were submitted, butalso their special composition and the environment in which theselipids were surrounded by water, proteins, and so on. Figure 8shows the spectral regions of the unsalted samples in which the1H NMR signals of the protons of primary and secondary ox-idation compounds appear. It is noteworthy that in spite of thelow temperature, neither the signals of hydroperoxides nor thoseof the conjugated double bonds are clearly observed. This couldbe due to the high degree of unsaturation of salmon lipids, whichgenerate very reactive and unstable primary oxidation compounds,which disappear very quickly. After 17 d under degradative condi-tions very weak signals of saturated aldehydic protons around 9.75ppm appear. These have greater intensity in the spectra of moreoxidized samples, signals of saturated aldehydes of small size likeethanal and propanal at 9.79 ppm also being observable, as wellas other less well-resolved signals between 9.48 and 9.60 ppm.As mentioned before, the signals of protons of oxygenated and

nonoxygenated α,β-unsaturated aldehydes appear in this latter re-gion. A poor resolution of signals in this region has also beenobserved in other lipids rich in ω-3 acyl groups oxidized at inter-mediate temperatures, as will be commented on later.

In summary, the oxidation process in this case displayed a veryprolonged period of initiation, probably due to the high contentof carotenoids and other compounds with antioxidant ability inthe salmon. In spite of these, it was demonstrated that althoughthe dry-salting process does not provoke any immediate oxidationof fish lipids, it reduces fish oxidative stability in an importantway, with the lipids of salted fillets being oxidized more quicklythan those of unsalted fillets. Signals of protons of hydroperox-ides or of conjugated dienic system were not observed beforethe appearance of signals of aldehydic protons in any of them;this could be due to the fact that they are very unstable and dis-appear very quickly. The aldehydes formed were n-alkanals ofvarious sizes and nonoxygenated and oxygenated α,β-unsaturatedaldehydes; the signals of these latter appeared overlapped. More-over, the presence of amino groups of proteins or of other com-pounds of similar nature to the compounds formed in fish lipidoxidation, either primary or secondary, can provoke the estab-lishment of reactions that reduce their presence in the system,such as the Maillard reaction. However, it should be mentionedthat both facts, the nondetection of hydroperoxides or conju-gated dienic systems and the overlapping of oxygenated α,β-unsaturated aldehydes, are also observed in studies of the oxidation



Figure 8–Expanded regions of the 1H NMR spectra of the lipids extracted from unsalted salmon fillets that were submitted to 50 °C for days (Guillenand Ruiz 2004a). These spectra were acquired in a Varian Plus spectrometer operating at 300 MHz. Signal letters agree with those in Table 3.

process at low temperature of oils rich in ω-3 groups (Guillen andRuiz 2005c).

Evolution of the thermo-oxidation of edible oils of very differ-ent composition submitted to 70 °C with aeration in a convectionoven. In these studies vegetable oils rich in oleic (olive, hazel-nut, peanut), linoleic (sesame, SF, corn), and linolenic (rapeseed,walnut, linseed) acyl groups were submitted to accelerated storageconditions at 70 °C with aeration (10 g of the oil in Petri dishes;Guillen and Ruiz 2004b, 2005a, 2005b, 2005c; Goicoechea andGuillen 2010). In these studies the thermo-oxidative conditionswere the same for all the oils; the only difference was the oil com-position. The processes were studied until the samples were totallypolymerized and it was not possible to take samples to acquire the1H NMR spectra. In all the oils studied the signals of the protonsof polyunsaturated (linoleic and linolenic) groups disappeared atthe end of the process.

Figure 9 shows some enlarged regions of the 1H NMR spectra of3 oils (SF, EVO, and VL), selected as representatives of different acylgroup compositions, in which the signals of protons of oxidationcompounds appear.

Oils rich in linoleic acyl groups: Under these degradativeconditions the evolution of oils rich in linoleic groups, such asrefined SF, is fairly different from that occurring at room temper-ature in closed receptacles, although the nature of the compoundsformed is relatively similar, with certain exceptions. After 11 dunder these degradative conditions this oil is totally polymerized.On day 1 signals of hydroperoxy groups associated with (Z,E)-conjugated dienic systems are observable in the 1H NMR spectra,and their formation is favored in contrast to those associated with(E,E)-conjugated dienic systems until day 5; afterwards, the latterare in much higher concentrations than the former (Goicoecheaand Guillen 2010). In agreement with observations made at roomtemperature, the hydroperoxide signal near 8.46 ppm is associatedwith (E,E)-dienes and those at 8.50 ppm and at 8.53 are asso-ciated with (Z,E)-dienes. Furthermore, the formation, at roomtemperature in closed receptacles, of hydroxy groups associatedwith conjugated dienic systems, is not observed here.

As Figure 9(a) shows, the concentration of these primary oxi-dation compounds, especially the hydroperoxy-(E,E)-conjugateddienic systems, increases as the process advances, reaching a max-imum value after 7 d under degradative conditions; afterwardstheir sharp degradation occurs. Like at room temperature, thedegradation of primary oxidation compounds provokes the for-

mation of aldehydes, epoxides, and alcohols. Nevertheless, somedifferences may be observed. Thus, in relation to aldehydes itis particularly worth mentioning that the main aldehydes ini-tially formed are 4-hydroperoxy-(E)-2-alkenals, with n-alkanalsand (E)-2-alkenals in smaller concentrations. As the process ad-vances (E,E)-2,4-alkadienals and 4-hydroxy-(E)-alkenals appear:the concentration of (E,E)-2,4-alkadienals remains small through-out the entire degradation process, whereas that of 4-hydroxy-(E)-alkenals increases significantly and becomes higher than thatof 4-hydroperoxy-(E)-alkenals at the end of the process; it seemsthat the formation of the former involves the disappearance of thelatter. The concentration of 4,5-epoxy-2-alkenals is even smallerthan that of (E,E)-2,4-alkadienals and they are only observable atadvanced oxidation stages.

As mentioned, epoxides are also formed in the oxidation of SFoil at 70 °C. A comparison of Figure 9(a) and 7(a) indicates thatat 70 °C the same epoxides and dialcohols, as at room tempera-ture in closed receptacles, are formed but their concentrations aredifferent. It is clear that the formation of diepoxides and also ofdialcohols is more favored at 70 °C, probably due to the greaterpresence of oxygen in the system. The same can be said of alcohols,which give proton signals at 3.59 and 3.62 ppm.

In summary, in comparison with room temperature the ef-fect of the temperature (70 °C) in this process provokes the for-mation of greater concentrations of hydroperoxides in chains of(E,E)-conjugated dienic systems and the higher availability of oxy-gen gives rise to the formation of greater concentrations of themost oxygenated final compounds in the groups of aldehydes (thetoxic and reactive 4-hydroperoxy-(E)-2-alkenals, 4-hydroxy-(E)-2-alkenals), epoxides (diepoxides), and alcohols. This higher tem-perature also significantly affects the rate of the formation of all ofthese compounds.

Oils rich in oleic acyl groups: The evolution of oils richin oleic acyl groups, such as EVO oil, under these same degrada-tive conditions, shares some features with the above-mentionedlinoleic-rich oils, but great differences also exist. This is summa-rized in Figure 9(b). First, the induction period is much moreprolonged (day 26) than in SF oil, and after 72 d under thesedegradative conditions, EVO oil polymerizes totally, and the maincompounds formed, even though some of them are the same,are somewhat different. As can be observed in Figure 9(b), theformation of hydroperoxy groups begins after a prolonged periodunder degradative conditions. The antioxidant compounds present



Figure 9–Expanded regions of the 1H NMR spectra of oils submitted to accelerated storage conditions in Petri dishes (10 g) in a convection oven at70 °C with aeration until polymerization: (a) refined sunflower (SF), (b) extra-virgin olive (EVO), and (c) virgin linseed (VL; Guillen and Ruiz 2005a,2005b; Goicoechea and Guillen 2010). These spectra were acquired in a Bruker Avance 400 spectrometer operating at 400 MHz. Signal letters agreewith those in Table 3.

in the oil and the lower unsaturation degree of the acyl groupsare the reason for this. Among the hydroperoxy groups formed,some are derived from linoleic groups, which are the same as thosegenerated in SF oil; however, in this case their concentration isvery small. In addition, there are other ones in higher concentra-tions derived from oleic, some of whose structures are given inFigure 10; they can be distinguished from the ones derived fromlinoleic by the signal of their olefinic protons at 5.72 ppm (Porterand others 1994). In any case, under the same degradative condi-tions the formation of hydroperoxides in EVO oil is much smallerthan in SF oil.

The difference between the compounds formed from olive oiland SF oil is even more noticeable in the groups of aldehydes,epoxides, and ketones. Thus, regarding aldehydes, at 70 °C withaeration in EVO oil practically only n-alkanals and (E)-2-alkenalsare formed, with 4-hydroperoxy- and 4-hydroxy-(E)-2-alkenalsin very small concentrations, although these were the main onesformed in SF oil. As can be observed in Figure 9(b), at advancedoxidation stages these oxygenated aldehydes are almost unappre-ciable in the spectra.

From these results it is evident that the determination of theconcentration of the aldehydic functional group, in a global way,in different oils submitted to degradation is a very crude way todefine or to establish oil safety.

As in SF oil, epoxy groups are also formed in EVO oil; nev-ertheless, in this latter case the main ones are those derivedfrom oleic groups, namely (E)-9,10-epoxystearate groups (sig-nal at 2.63 ppm) and (Z)-9,10-epoxystearate groups (signal at2.88 ppm; see Figure 9b and 10). By contrast, in SF oil leukotoxin(9,10-epoxy-12-octadecenoate) and/or isoleukotoxin (12,13-epoxy-9-octadecenoate) were formed, in addition to diepoxides,all of them derived from linoleic acyl groups.

Another distinctive characteristic of oils rich in oleic groupsoxidized at intermediate temperatures with aeration is the forma-tion, in important concentrations, of keto groups conjugated witha double bond (Gibson 1948), this latter giving the characteristicsignals at 6.08 and 6.82 ppm in the spectral region of protonssupported on conjugated systems (see Figure 9b and 10). Thesesignals, which had been observed in previous studies (Guillen andRuiz 2005a, 2005c) and were unidentified at the time, can now



Figure 10–Some of the primary oxidation compounds that can be detected by 1H NMR derived from: (a) oleic and (b) linolenic acyl groups.

be identified. This identification has been carried out by compar-ison with data on similar oxidation compounds provided by otherauthors (Kuklev and others 1997; Jie and Lam 2004; Lin and oth-ers 2007); these keto groups, like epoxides, are derived from thehydroperoxides of oleic groups.

Again, this shows that the determination, in a global way, ofconjugated dienic systems is a very crude way of establishing theoxidation degree, because not only primary but also secondarycompounds, these of very different natures, have conjugated dienicsystems.

In summary, under the same conditions of accelerated stor-age, the different composition of EVO oil and SF oil determinethat: the rate of oxidation in the former is much smaller thanin the latter; the main hydroperoxides and epoxides formed inthe former are those derived from oleic acyl groups, whereas inthe latter they are those derived from linoleic, with leukotoxin(9,10-epoxy-12-octadecenoate), isoleukotoxin (12,13-epoxy-9-octadecenoate), and diepoxides in important concentrations; themain aldehydes formed in EVO oil are n-alkanals and (E)-2-alkenals, with very small amounts of the genotoxic and cytotoxic 4-hydroperoxy-(E)-2-alkenals and 4-hydroxy-(E)-2-alkenals, whichare the main ones in oxidized SF oil; and finally that in the oxi-dized EVO oil, by contrast with oxidized SF oil, chains having ketogroups conjugated with a double bond are present in significantconcentrations.

Oils rich in linolenic acyl groups: The evolution of thethermo-oxidative process of oils rich in linolenic groups, such asVL oil submitted to 70 °C with aeration, is also very different fromthat of the previously mentioned oils, as may be expected. Therate of degradation is very high and in 4 d under these degradativeconditions linseed oil is totally polymerized.

As Figure 9(c) shows, the spectra of linseed oil do not show well-resolved signals either of protons of hydroperoxides or of those ofconjugated dienic systems related to them, these signals beingof low intensity in comparison with those observed in SF oil (seeFigure 9a). Their low concentration could be due to the instabilityof these hydroperoxides, which do not accumulate and evolve veryquickly to give other compounds. It must be remembered that inthe thermo-oxidation at 50 °C of fish lipids, which are also richin ω-3 groups, hydroperoxides, and the dienic conjugated systemsassociated with them, were not clearly observed (see Figure 8;Guillen and Ruiz 2004a).

The low resolution of these signals could be due to the overlap-ping of those of several different structures of hydroperoxides andconjugated dienic systems, supported by octadecadienoates andoctadecatrienoates, derived from linoleic and linolenic acyl groups,respectively. Frankel and others (1990) submitted pure trilinoleninto 40 °C and studied the primary oxidation products generated bydifferent techniques, among them 1H NMR. The main com-pounds identified were several hydroperoxy-octadecatrienoatesshowing 2 of the double bonds conjugated (see Figure 10b). Thecharacteristic proton signals of these (Z,E)-conjugated dienes ofoctadecatrienoates appeared at 6.55, 6.00, 5.56, and 5.42 ppm,with chemical shifts that were very similar to those of the (Z,E)-conjugated dienes present in hydroperoxides derived from linoleic(see Table 3).

It is noteworthy that many other different structures of oxidationproducts derived from pure linolenic acid or ester were identifiedby 1H NMR, such as dihydroperoxides or dihydroxides associatedwith conjugated diene–triene systems, hydroxy or hydroperoxy-epidioxides (also called hydroxy or hydroperoxy-cyclic peroxides),and so on (Gardner and Weisleder 1972; Neff and others 1981;



Coxon and others 1984; Frankel and others 1990). This greatcomplexity of structures could explain the above-mentioned low-signal resolution in certain regions of the spectra.

Regarding secondary oxidation compounds, as in SF oil, 4-hydroperoxy-, and 4-hydroxy-(E)-2-alkenals, are formed in im-portant concentrations, in addition to n-alkanals, (E)-2-alkenals,(Z,E)-2,4-alkadienals, 4,5-epoxy-2-alkenals, and their precursors(E,E)-2,4-alkadienals. The signals of these aldehydes greatly over-lap, as was observed in the oxidation of the fish lipids mentionedbefore. It is remarkable that from 2.75 d onwards, when VL oilstarted to polymerize, some signals appeared at 9.38 to 9.42 ppm.Although these signals were previously considered as unknown(Guillen and Ruiz 2005c), now they can be tentatively at-tributed to the hydroperoxy proton (OOH) present in the above-mentioned hydroperoxy-cyclic peroxides, which are secondary orfurther oxidation compounds derived from linolenate (Neff andothers 1981).

As far as epoxides are concerned, mono- and diepoxides canbe tentatively identified, because proton signals near 3 ppm arehighly overlapped, probably due to the great variety of structuresmentioned before. However, they are in lower concentrations thanin SF oil submitted to similar conditions (see Figure 9a and c).

In summary, under the same thermo-oxidative conditions (10 gof oil at 70 °C with aeration) oil composition decisively influencesnot only the rate of the oxidation process, but also the compoundsformed, which are very different from 1 oil to another. The oilstaken as examples in the previous paragraphs show significantabundances of the 3 kinds of acyl groups: linoleic, oleic, andlinolenic. The performance in the thermo-oxidation processesof other oils having intermediate acyl group composition is alsointermediate, in comparison with that exhibited by the above-mentioned oils.

Evolution of the thermo-oxidation process of refined SF oil sub-mitted to 100 °C with aeration in a convection oven. In this study,as above, samples of 10 g of SF oil placed in open Petri disheswere submitted to 100 °C with aeration in a convection ovenand the evolution of their composition was studied by 1H NMR(Goicoechea and Guillen 2010). As expected, under these condi-tions the induction period was shorter than in SF oil submittedto 70 °C, and the formation of hydroperoxides associated with(E,E)-conjugated dienic systems predominated from the begin-ning of the process, attributable to the increase of the temperatureby 30 °C (Schneider 2009).

In relation to the generation of aldehydes, it can be seen that thesame kinds as those previously detected at 70 °C are also formedat 100 °C; however, the formation of oxygenated α,β-unsaturatedaldehydes is less favored than at 70 °C. It is worth mentioning thatin the process at 100 °C the detection of aldehydes is producedat lower concentrations of hydroperoxides (77.2 mmol/L oil at9 h) than in the process at 70 °C (346.8 mmol/L oil at 144 h); this proves that the degradation of hydroperoxides at higher tem-peratures is faster than at lower temperatures. This may be oneof the reasons for discrepancies when classical methods are usedto evaluate the oxidation level of edible oils. Furthermore, thesequential formation of 4,5-epoxy-2-alkenals and of 4-hydroxy-(E)-2-alkenals after (E,E)-2,4-alkadienals and 4-hydroperoxy-(E)-2-alkenals, respectively, also observed at 70 °C is maintained, to-gether with the diminution in the concentration of the 2 lattertypes of aldehydes as the concentration of the 2 former increases.

Temperature also affects epoxide formation, favoring thatof monoepoxides (leukotoxin [9,10-epoxy-12-octadecenoate]and isoleukotoxin [12,13-epoxy-9-octadecenoate]) over that of

diepoxides and their diol derivatives, unlike what was seen at70 °C.

In short, the increase of the temperature from 70 to 100 °Cin the SF oil thermo-oxidative process favors, on the one hand,the formation of hydroperoxy groups supported by chains having(E,E)-conjugated dienic systems versus those supported by (Z,E)-dienic systems, and, on the other hand, reduces the formation ofoxygenated α,β-unsaturated aldehydes, as well as of diepoxidesversus monoepoxides.

Evolution of the thermo-oxidation of edible oils of very differentcompositions submitted at 190 °C with aeration in a convectionoven. In this study, samples of 10 g of oil put in open Petri disheswere also submitted to 190 °C with aeration in a convection ovenuntil total polymerization; the oils in this study were rich in oleic(virgin olive, olive, hazelnut), linoleic (corn, SF), and linolenic acylgroups (soybean, linseed; Guillen and Ruiz 2008). The decreas-ing order of resistance to degradation was as follows: virgin-oliveoil > olive oil > hazelnut oil > corn oil > SF oil > soybeanoil > linseed oil, in agreement with the studies carried out at70 °C with aeration (Guillen and Ruiz 2005a, 2005b, 2005c). Ingeneral, the lower the content in polyunsaturated groups in theoil, the higher its resistance to degradation; however, other factors,such as the presence of minor antioxidant components, play a veryimportant role in the degradation rate of the oils, including whenthe oil is submitted to high temperatures. So, the virgin olive oilstudied contains similar proportions of oleic and linoleic groups toolive oil, but its degradation rate is lower, making it evident thatthe differences found in the resistance against degradation amongthe different oils studied are not only due to the structure andcomposition of the main components.

Regarding the formation of oxidation compounds, neither pro-tons of hydroperoxy groups nor protons of conjugated dienic sys-tems bonded to hydroperoxy or hydroxy derivatives were observedin the 1H NMR spectra of the oils kept under these degradativeconditions, while aldehydic signals were the first to be observed.The signals observed between 6.08 and 7.09 ppm are related tothe conjugated double bonds present in aldehydes (Goicoecheaand Guillen 2010). This fact indicates that, if primary oxidationcompounds are formed, they degrade at very high rates and donot accumulate in sufficient amounts to be detected by 1H NMR.

It is noteworthy that oils of different compositions generatedaldehydes of different natures and also in different proportions.Thus, in virgin olive, olive, and hazelnut oil, only significantconcentrations of (E)-2-alkenals and of n-alkanals are generated.However, in oils rich in polyunsaturated groups, in addition to n-alkanals and (E)-2-alkenals, significant proportions of (E,E)-2,4-alkadienals are also formed, as well as 4-hydroxy-(E)-2-alkenals,those tentatively identified as (Z,E)-2,4-alkadienals, and also smallproportions of 4,5-epoxy-2-alkenals. In all the oils, 4-oxo-alkanals,a kind of aldehyde typical of frying temperatures, were tentativelyidentified (Takeoka and others 1995). In none of the oils were4-hydroperoxy-(E)-2-alkenals detected. In comparison with thethermo-oxidation process of these oils at 70 °C with aeration(Guillen and Ruiz 2005a, 2005b, 2005c), the increase of the tem-perature to 190 °C provoked a lesser generation of oxygenatedα,β-unsaturated aldehydes. It is also remarkable that oils rich inoleic acyl groups were not only more resistant to degradation thanoils rich in linoleic and linolenic acyl groups, but also producedlower proportions of toxic aldehydes.

Regarding other secondary or further oxidation compounds,small proton signals at 6.08 ppm and 6.82 ppm related to ketogroups conjugated with a double bond were observed in oils rich in



oleic acyl groups; as mentioned before, these signals also appearedwhen these oils were submitted to 70 °C with aeration, but inmuch greater intensities.

In short, if primary oxidation compounds were formed at190 °C with aeration, they degraded so fast that they cannot beobserved in the 1H NMR spectra. As for aldehydes, in comparisonwith the thermo-oxidation with aeration at lower temperatures,smaller concentrations of oxygenated α,β-unsaturated aldehydeswere generated; 4-hydroperoxy-(E)-2-alkenals were not detected,and 4-oxo-alkanals were.

Evolution of the thermo-oxidation process of virgin olive, refinedcorn, and VL oils submitted to microwave action at 190 °C. Inthis study, virgin olive, refined corn, and VL oils (10 g of oil inPetri dishes) were repeatedly submitted to the effect of microwaveheating at 190 °C without exceeding this temperature (Guillenand Ruiz 2006) for 240 min. In virgin olive oil the highest levelof degradation was undergone by the oleic acyl groups, in cornoil by the linoleic, and in linseed oil by the linolenic acyl groups.In contrast with observations made at lower temperatures, noneof these acyl groups was totally degraded in any of the 3 oils, witha very small remaining molar percentage of linoleic in virgin oliveand in corn oils after 240 min.

The times at which the initiation of the degradation was de-tected in these 3 oils were very similar in this process and whensubmitted to 190 °C with aeration in a convection oven (Guillenand Ruiz 2008), suggesting that the heat supplied to the oil affectsthe rate of its degradation to a higher degree than the means bywhich the heat is produced.

Regarding the formation and occurrence of primary oxidationcompounds, only very slight and occasional hydroperoxide protonsignals appear between 8.4 and 8.6 ppm in the 1H NMR spectra(in virgin olive oil at minute 70 and in corn oil at minute 40),which were not accompanied by signals of protons of associated di-enic conjugated systems. Signals of aldehydes were detected at thesame time as the appearance of these rare incipient signals of hy-droperoxides. In VL oil, hydroperoxides were not clearly detected.The absence of significant concentrations of hydroperoxides in theoils submitted to microwave-heating is remarkable, along with theabsence of related dienic systems, and the rapid generation of alde-hydes. These results are in agreement with those obtained in oilssubmitted to 190 °C with aeration in a convection oven (Guillenand Ruiz 2008).

The different composition of the 3 oils studied (virgin olive,corn, and VL), has, as expected, differences in the proportionsof the aldehydes generated. Thus, of the 3 oils, virgin olive oilproduced the smallest amount of (E,E)-2,4-alkadienals, with (E)-2-alkenals and n-alkanals the main aldehydes generated from thisoil. As the proportion of polyunsaturated acyl groups in the oilrises, so does the proportion of (E,E)-2,4-alkadienals, 4-hydroxy-(E)-2-alkenals, and 4,5-epoxy-2-alkenals, and that of tentativelyidentified (Z,E)-2,4-alkadienals and 4-oxo-alkanals, in such a waythat (E,E)-2,4-alkadienals are produced in higher concentrationsthan (E)-2-alkenals in the degradation of corn and linseed oils.In summary, n-alkanals and (E)-2-alkenals are the main aldehydesgenerated from virgin olive oil, whereas those generated fromcorn and VL oils are (E,E)-2,4-alkadienals. Furthermore, virginolive oil is the only one that does not produce 4-hydroxy-(E)-2-alkenals. These facts are very important because of the differentreactivity and toxicity of the several kinds of aldehydes.

In short, the results obtained after submitting the oils to mi-crowave action at 190 °C are in agreement with those obtainedat 190 °C with aeration in a convection oven. The absence of

significant concentrations of primary oxidation compounds andthe rapid generation of aldehydes, especially nonoxygenated ones,is noteworthy.

Evolution of EVO, refined SF, and VL oils under deep-frying con-ditions at 190 °C in the absence of food. EVO, refined SF, andVL oils were submitted to deep-frying conditions (4 L oil) with-out food at 190 °C for a prolonged period of time (Guillen andUriarte 2012a, 2012b, 2012c). Regarding the degradation of theacyl groups, it can be observed that the highest rate occurred inthe main unsaturated acyl group in the oil, independently of itsunsaturation degree. Furthermore, it can also be seen that the mo-lar percentages of the acyl groups, which are less saturated than themain one, decreased with heating time, whereas those more satu-rated either increased or remained unchanged. Thus, the highestdecrease per hour in VL oil was produced in the molar percent-age of linolenic acyl groups, in SF oil in the molar percentage oflinoleic acyl groups, and in EVO oil in the molar percentage ofoleic acyl groups; these results indicate that the rate of degradationof the various acyl groups, at frying temperature, depends more ontheir concentrations in the oil than on their unsaturation degree,this latter associated with their facility or tendency to oxidize.

Under these conditions formation of primary oxidation com-pounds was not observed. As can be observed in Figure 11, protonsignals of hydroperoxy groups, as well as of conjugated dienes,were not detected, in agreement with what was observed dur-ing thermo-oxidation at 190 °C in a convection oven (Guillenand Ruiz 2008) and at 190 °C under microwave action (Guillenand Ruiz 2006). This fact indicates that in oils submitted to hightemperatures if hydroperoxides are formed, they decompose veryquickly, independently of the heating method, in such a way thatthey are not present in the sample in significant concentrationsat any time. This is in agreement with the absence of significantchanges in the classical peroxide value assay of oils submitted tofrying temperature (Dobarganes and Perez-Camino 1988), andwith the observations by other authors concerning the forma-tion of aldehydes in oils heated at high temperature without pre-vious induction periods (Silvagni and others 2010). Moreover,the classical parameter, conjugated dienes, associated with theconcentration of acyl group chains having conjugated dienic sys-tems in the oils, in other words, primary oxidation compounds,which is determined by absorption at 232 nm in ultraviolet (UV)spectroscopy, is not useful for this purpose in oils submitted to theseconditions (190 °C). This is so because, as Figure 11 shows, theonly signals present at this temperature, in the spectra of protonsbelonging to conjugated dienic systems, are those of unsaturatedaldehydes and of other secondary oxidation compounds.

With regard to the formation of aldehydes under deep-fryingconditions, there are very important differences between oils.Thus, in EVO oil, as Figure 11(b) shows, the most abundant alde-hydes are the less reactive, that is (E)-2-alkenals and n-alkanals, andthe least abundant are the more reactive aldehydes, that is (E,E)-2,4-alkadienals; neither 4-hydroxy-(E)-2-alkenals nor (Z,E)-2,4-alkadienals were detected in this oil (see Figure 11b). However,as Figure 11(a) shows, the heated SF oil was, of the 3 oils, theone which had higher concentrations of the most reactive alde-hydes such as (E,E)-2,4-alkadienals, (Z,E)-2,4-alkadienals, 4,5-epoxy-2-alkenals, and 4-hydroxy-(E)-2-alkenals. And, finally, VLoil contained higher concentrations of (E,E)-2,4-alkadienals thanof (E)-2-alkenals and n-alkanals; however, the difference in con-centration of these kinds of aldehydes was much smaller than inSF oil (see Figure 11c). This may be because the main (E,E)-2,4-alkadienal formed in linseed oil is (E,E)-2,4-heptadienal (Guillen



Figure 11–Expanded regions of the 1H NMR spectra of oils heated in a professional deep-fryer (4 L oil) without food for several hours: (a) refinedsunflower (SF), (b) extra-virgin olive (EVO), and (c) virgin linseed (VL; Guillen and Uriarte 2012a, 2012b, 2012c). These spectra were acquired in aBruker Avance 400 spectrometer operating at 400 MHz. Signal letters agree with those in Table 3.

and Uriarte 2012d), which has a lower boiling point (177 °C)than the frying temperature (190 °C), and for that reason a largequantity of this aldehyde escapes into the atmosphere. However, inSF oil the (E,E)-2,4-alkadienals formed were mainly (E,E)-2,4-decadienal and (Z,E)-2,4-decadienal, both with higher boilingpoints (244 °C) than the frying temperature (190 °C), and forthat reason their ability to escape from the oil matrix into theatmosphere is greatly diminished, so they remain in the oil liquidmatrix to a large extent. Nevertheless, it has to be noted that theformation of aldehydes in thermo-oxidation processes at 190 °Cis affected by the air–oil contact surface in relation to the oil vol-ume (Guillen and Ruiz 2006, 2008; Guillen and Uriarte 2009,2012c).

In short, under frying conditions, once again, EVO oil generateslower concentrations of the most reactive aldehydes than oils richin polyunsaturated groups, such as SF and VL oils.

In addition to aldehydes, other secondary compounds are alsoformed under these conditions and they too are different depend-ing on the oil nature. So, the 1H NMR spectra of SF oil submittedto deep-frying conditions do not have signals of epoxides or of thediol derivatives; the only signals that are visible here are the sidebands of the bis-allylic protons (see Figure 11a). However, whenit comes to EVO oil, the formation of 2 epoxides derived fromoleic acyl groups, such as (E)-9,10-epoxystearate (signal at 2.63ppm) and (Z)-9,10-epoxystearate (signal at 2.88 ppm), is observed

(Du and others 2004; Marmesat and others 2008; Guillen andUriarte 2012a; Martınez-Yusta and Guillen 2014a). And, finally,no signals of epoxidic protons were observed in the spectra of VLoil submitted to the same conditions as above.

Concerning alcohols, only signals attributable to primary al-cohols were detected in SF oil spectra, whereas in that of EVOoil signals attributable to primary and to secondary ones wereobserved, the formation of this functional group in the degrada-tion of VL oil under deep-frying conditions being unclear. Theformation of hydroxyl groups in edible oils submitted to fryingtemperature has been reported previously (Schwartz and others1994).

In short, when oils are submitted to deep-frying conditions at190 °C, in the absence of food, their thermo-oxidation evolveswithout the occurrence, in measurable concentrations by this tech-nique, of primary oxidation compounds (hydroperoxy groups andconjugated dienic systems). In comparison with thermo-oxidationprocesses at lower temperatures with aeration, lower concentra-tions of oxygenated secondary oxidation compounds are gener-ated. Differences between the performance of oils as a function oftheir compositions can be observed.

Conclusions1H NMR gives very valuable information about the different

types of protons present in food lipids, which can be supported by



main and minor compounds. This spectroscopic technique is fastand simple, requires no chemical modification of the sample, andprovides a great deal of information on the lipid sample as a whole,this latter fact being very important. It is very useful not only forthe study of the nondegraded food lipids, but also for the study ofthe degraded ones. For this reason it is very useful to follow, in aqualitative and quantitative way, the evolution of food lipids whenthey are submitted to degradative conditions. It allows not onlythe identification of many of the main functional groups formed,but also to quantify them providing, in this way, information ofthe thermodegradative processes.

From the results described here, it can be concluded that the useof classical methods to determine the concentration of functionalgroups in a global way (hydroperoxides, conjugated dienes, alde-hydes, and so on) in food lipids of different compositions whensubmitted to different degradation conditions, is a very crude wayof analyzing their thermo-oxidation processes. It should be takeninto account that each functional group includes compounds ofvery different reactivities and toxicities.

The studies mentioned here have demonstrated that both foodlipid composition and the conditions under which the thermo-oxidation process takes place govern the mechanisms by which theprocess evolves, as well as affecting the nature and concentrationsof intermediate and finally formed compounds.

From the above-mentioned studies it can also be concluded thatpolyunsaturated food lipids submitted to oxidative conditions atlow temperatures are able to generate a great number of differ-ent kinds of very reactive oxygenated compounds, some of whichare known to be toxic, such as 4-hydroperoxy-, and 4-hydroxy-(E)-2-alkenals, or 9,10-epoxy-12-octadecenoate (leukotoxin) and12,13-epoxy-9-octadecenoate (isoleukotoxin) derivatives. As thetemperature increases, the number of different kinds of compoundswhich may be formed from these lipids decreases. At low temper-atures the process is slow; however, at high temperatures it takesplace very quickly. Although the number and concentration of themost reactive compounds which may form from polyunsaturatedlipids at high temperatures is smaller than at low temperatures,they can accumulate with heating time.

Food lipids with small concentrations of polyunsaturated groupsand significant concentrations of compounds that exhibit antiox-idant ability, when submitted to thermo-oxidative conditions atlow temperature, not only generate much lower concentrationsof the most reactive and toxic lipid-derived compounds, but thisformation occurs much more slowly than in the case of the foodlipids, which are rich in polyunsaturated groups. When these lipidsare submitted to high temperatures the rate of their degradationincreases, as does that of those rich in polyunsaturated groups, butthe range of compounds formed is different from those formed inthe latter.

In addition to the influence of both food lipid nature and thetemperature during which the thermo-oxidation evolves, anotherimportant factor is the food lipid surface exposed to the oxygen.When this surface is high in relation to the oil volume involved,then higher concentrations of oxygenated derivatives are formedthan when the exposed surface is small. This fact is very importantbecause some of the oxygenated compounds are very reactive andtoxic.

The results summarized here are worthy of consideration be-cause some of the above-mentioned toxic compounds detected bythis technique during food lipid thermo-oxidation processes havealso been found in human cells and tissues. Finally, it only remainsto be added that the results commented on and discussed here may

be helpful in the interpretation of results obtained in metabolomicand lipidomic studies based on 1H NMR techniques.

AcknowledgmentsThis work was supported by the Spanish Ministry of Economy

and Competitiveness (MINECO, AGL2012-36466), the BasqueGovernment (EJ-GV, GIC10/IT-463–10), and the Univ. of theBasque Country (UPV/EHU, UFI-11/21). A. Martınez-Yustathanks the EJ-GV for a predoctoral contract.

References

Aerts HAJ, Jacobs PA. 2004. Epoxide yield determination of oils and fattyacid methyl esters using 1H NMR. J Am Oil Chem Soc 81:841–6.

Aursand M, Rainuzzo JR, Grasladen H. 1993. Quantitative high-resolution13C and 1H nuclear magnetic resonance of ω-3 fatty acids from whitemuscle of Atlantic salmon (Salmo salar). J Am Oil Chem Soc 70(10):971–81.

Browne RW, Armstrong D. 2000. HPLC analysis of lipid-derivedpolyunsaturated fatty acid peroxidation products in oxidatively modifiedhuman plasma. Clin Chem 46(6, part 1):829–36.

Butovich IA. 2006. A one-step method of 10,17-dihydro(pero)xydocosahexa-4Z,7Z,11E,13Z,15E,19Z-enoic acid synthesisby soybean lipoxygenase. J Lipid Res 47(4):854–63.

Butovich IA, Lukyanova SM, Bachmann C. 2006.Dihydroxydocosahexaenoic acids of the neuroprotectin D family: synthesis,structure, and inhibition of human 5-lipoxygenase. J Lipid Res 47(11):2462–74.

Chizzolini R, Zanardi E, Dorigoni V, Ghidini S. 1999. Calorific value andcholesterol content of normal and low-fat meat and meat products. TrendsFood Sci Tech 10:119–28.

Claxson AWD, Hawkes GE, Richardson DP, Naughton DP, Haywood RM,Chander CL, Atherton M, Lynch EJ, Grootveld MC. 1994. Generation oflipid peroxidation products in culinary oils and fats during episodes ofthermal stressing: a high field 1H NMR study. FEBS Lett 335(1):81–90.

Coxon DT, Peers KE, Rigby NM. 1984. Selective formation ofdihydroperoxides in the α-tocopherol inhibited autoxidation of methyllinolenate. J Chem Soc Chem Comm 1:67–8.

Dobarganes MC, Marquez-Ruiz G. 2007. Formation and analysis of oxidizedmonomeric, dimeric and higher oligomeric triglycerides. In: Erickson MD,editor. Deep frying: chemistry, nutrition and practical applications. 2nd ed.Champaign, Ill.: AOCS Press. p 87–110.

Dobarganes MC, Perez-Camino MC. 1987. Non-polar dimer formationduring thermoxidation of edible fats. Fett Wiss Technol 89(6):216–20.

Dobarganes MC, Perez-Camino MC. 1988. Fatty acid composition: a usefultool for the determination of alteration level in heated fats. Rev Fr CorpsGras 35(2):67–70.

Dobson EP, Barrow CJ, Kralovec JA, Adcock JL. 2013. Controlledformation of mono- and dihydroxy-resolvins from EPA and DHA usingsoybean 15-lipoxygenase. J Lipid Res 54(5):1439–47.

Du G, Tekin A, Hammond EG, Woo LK. 2004. Catalytic epoxidation ofmethyl linoleate. J Am Oil Chem Soc 81(5):477–80.

Esterbauer H, Schaur RJ, Zollner H. 1991. Chemistry and biochemistry of4-hydroxy-2-nonenal, malonaldehyde and related aldehydes. Free RadicalBio Med 11(1):81-128.

Fiori L, Solana M, Tosi P, Manfrini M, Strim C, Guella G. 2012. Lipidprofile of oil from trout (Oncorhynchus mykiss) heads, spines and viscera: troutby-products as a possible source of omega-3 lipids? Food Chem134:1088-95.

Frankel EN. 2005. Lipid oxidation, 2nd ed. Bridgwater, U.K.:Oily Press.Frankel EN, Neff WE, Selke E. 1981. Analysis of autoxidized fats by gaschromatography-mass spectrometry: VII. Volatile thermal decompositionproducts of pure hydroperoxides from autoxidized and photosensitizedoxidized methyl oleate, linoleate and linolenate. Lipids 16(5):279–85.

Frankel EN, Neff WE, Selke E, Weisleder D. 1982. Photosensitizedoxidation of methyl linoleate: secondary and volatile thermal decompositionproducts. Lipids 17(1):11–8.

Frankel EN, Neff WE, Selke E, Brooks DD. 1988. Analysis of autoxidizedfats by gas chromatography-mass spectrometry: X. Volatile thermaldecomposition products of methyl linolenate dimers. Lipids 23(4):295–8.



Frankel EN, Neff WE, Miyashita K. 1990. Autoxidation of polyunsaturatedtriacylglycerols. II. Trilinolenoylglycerol. Lipids 25(1):40–7.

Gardner HW, Weisleder D. 1972. Hydroperoxides from oxidation of linoleicand linolenic acids by soybean lipoxygenase: proof of the trans-11 doublebond. Lipids 7(3):191–3.

Gibson GP. 1948. Autoxidation of methyl oleate and linoleate and thedecomposition of the oxidation products: a theoretical discussion. J ChemSoc 2275–90.

Goicoechea E, Guillen MD. 2010. Analysis of hydroperoxides, aldehydes andepoxides by 1H Nuclear Magnetic Resonance in sunflower oil oxidized at70 and 100 °C. J Agr Food Chem 58(10):6234–45.

Guillen MD, Cabo N. 1997. Infrared spectroscopy in the study of edible oilsand fats. J Sci Food Agr 75(1):1–11.

Guillen MD, Cabo N. 1999. Usefulness of the frequency data of the Fouriertransform infrared spectra to evaluate the degree of oxidation of edible oils. JAgr Food Chem 47(2):709–19.

Guillen MD, Cabo N. 2000. Some of the most significant changes in theFourier transform infrared spectra of edible oils under oxidative conditions. JSci Food Agr 80(14):2028–36.

Guillen MD, Goicoechea E. 2007. Detection of primary and secondaryoxidation products by Fourier transform infrared spectroscopy (FTIR) and1H nuclear magnetic resonance (NMR) in sunflower oil during storage. JAgr Food Chem 55(26):10729–36.

Guillen MD, Goicoechea E. 2008. Toxic oxygenated α,β-unsaturatedaldehydes and their study in foods: a review. Crit Rev Food Sci48(2):119–36.

Guillen MD, Goicoechea E. 2009. Oxidation of corn oil at roomtemperature: primary and secondary oxidation products and determinationof their concentration in the oil liquid matrix from 1H nuclear magneticresonance data. Food Chem 116(1):183–92.

Guillen MD, Ruiz A. 2001. High resolution 1H nuclear magnetic resonancein the study of edible oils and fats. Trends Food Sci Tech 12(9):328–38.

Guillen MD, Ruiz A. 2003a. Edible oils: Discrimination by 1H nuclearmagnetic resonance. J Sci Food Agr 83(4):338–46.

Guillen MD, Ruiz A. 2003b. Rapid simultaneous determination by protonNMR of unsaturation and composition of acyl groups in vegetable oils. EurJ Lipid Sci Tech 105(11):688–96.

Guillen MD, Ruiz A. 2004a. Study of the oxidative stability of salted andunsalted salmon fillets by 1H nuclear magnetic resonance. Food Chem86(2):297–304.

Guillen MD, Ruiz A. 2004b. Formation of hydroperoxy- andhydroxyalkenals during thermal oxidative degradation of sesame oilmonitored by proton NMR. Eur J Lipid Sci Tech 106(10):680–7.

Guillen MD, Ruiz A. 2005a. Study by proton nuclear magnetic resonance ofthe thermal oxidation of oils rich in oleic acyl groups. J Am Oil Chem Soc82(5):349–55.

Guillen MD, Ruiz A. 2005b. Oxidation process of oils with high content oflinoleic acyl groups and formation of toxic hydroperoxy- andhydroxyalkenals. A study by 1H nuclear magnetic resonance. J Sci Food Agr85(14):2413–20.

Guillen MD, Ruiz A. 2005c. Monitoring the oxidation of unsaturated oilsand formation of oxygenated aldehydes by proton NMR. Eur J Lipid SciTech 107(1):36–47.

Guillen MD, Ruiz A. 2006. Study by means of 1H nuclear magneticresonance of the oxidation process undergone by edible oils of differentnatures submitted to microwave action. Food Chem 96(4):665–74.

Guillen MD, Ruiz A. 2008. Monitoring of heat-induced degradation ofedible oils by proton NMR. Eur J Lipid Sci Tech 110(1):52–60.

Guillen MD, Uriarte PS. 2009. Contribution to further understanding of theevolution of sunflower oil submitted to frying temperature in a domesticfryer: study by 1H nuclear magnetic resonance. J Agr Food Chem57(17):7790–9.

Guillen MD, Uriarte PS. 2012a. Study by 1H NMR spectroscopy of theevolution of extra virgin olive oil composition submitted to fryingtemperature in an industrial fryer for a prolonged period of time. FoodChem 134(1):162–72.

Guillen MD, Uriarte PS. 2012b. Monitoring by 1H nuclear magneticresonance of the changes in the composition of virgin linseed oil heated atfrying temperature. Comparison with the evolution of other edible oils.Food Control 28(1):59–68.

Guillen MD, Uriarte PS. 2012c. Simultaneous control of the evolution of thepercentage in weight of polar compounds, iodine value, acyl groups

proportions and aldehydes concentrations in sunflower oil submitted tofrying temperature in an industrial fryer. Food Control 24(2):50–6.

Guillen MD, Uriarte PS. 2012d. Aldehydes contained in edible oils of a verydifferent nature after prolonged heating at frying temperature: presence oftoxic oxygenated α,β-unsaturated aldehydes. Food Chem 131(3):915–26.

Guillen MD, Carton I, Goicoechea E, Uriarte PS. 2008. Characterization ofcod liver oil by spectroscopic techniques. New approaches for thedetermination of compositional parameters, acyl groups, and cholesterolfrom 1H nuclear magnetic resonance and Fourier transform infrared spectraldata. J Agr Food Chem 56(19):9072–9.

Guillen-Sans R, Guzman-Chozas M. 1998. The thiobarbituric acid (TBA)reaction in foods: a review. Crit Rev Food Sci 38(4):315–30.

Gunstone FD. 1990. 1H- and 13C-NMR spectra of six n-3 polyene esters.Chem Phys Lipids 56:227–9.

Hamalainen TL, Kamal-Eldin A. 2005. Analysis of lipid oxidation productsby NMR spectroscopy. In: Kamal-Eldin A, Pokorny J, editors. Analysis oflipid oxidation. 1st ed. Champaign, Ill.: AOCS Press. p 70-126.

Haywood RM, Claxson WD, Hawkes GE, Richardson DP, Naughton DP,Coumbarides, G, Hawkes J, Lynch EJ, Grootveld MC. 1995. Detection ofaldehydes and their conjugated hydroperoxydiene precursors inthermally-stressed culinary oils and fats: investigations using high resolutionproton NMR spectroscopy. Free Radical Res 22(5):441-82.

Hopkins CY, Bernstein HJ. 1959. Applications of proton magnetic resonancespectra in fatty acid chemistry. Can J Phys 37:775–82.

Igarashi T, Aursand M, Hirata Y, Gribbestad IS, Wada S, Nonaka M. 2000.Nondestructive quantitative acid and n-3 fatty acids in fish oils byhigh-resolution 1H nuclear magnetic resonance spectroscopy. J Am OilChem Soc 77(7):737–48.

Jie MSFLK, Lam CNW. 2004. Reaction of mono-epoxidized conjugatedlinoleic acid ester with boron trifluoride etherate complex. Lipids39(6):583–7.

Johnson LF, Shoolery JN. 1962. Determination of unsaturation and averagemolecular weight of natural fats by nuclear magnetic resonance (N.M.R.).Anal Chem 34(9):1136–9.

Kuklev DV, Christie WW, Durand T, Rossi JC, Vidal JP, Kasyanov SP,Akulin VN, Bezuglov VV. 1997. Synthesis of keto- and hydroxydienoiccompounds from linoleic acid. Chem Phys Lipids 85(2):125–34.

Kurata S, Yamaguchi K, Ohtsuka S, Nagai M. 2005. Rapid and convenientdiscrimination of various types of fats and oils by proton nuclear magneticresonance spectroscopy. Bunseki Kagaku 54(11):1083–90.

Laguerre M, Lecomte J, Villeneuve P. 2007. Evaluation of the ability ofantioxidants to counteract lipid oxidation: existing methods, new trends andchallenges. Prog Lipid Res 46(5):244–82.

Lercker G, Rodriguez-Estrada MT, Bonoli M. 2003. Analysis of theoxidation products of cis- and trans-octadecenoate methyl esters by capillarygas chromatography-ion-trap mass spectrometry I. Epoxide and dimericcompounds. J Chromatogr A 985:333–42.

Lin D, Zhang J, Sayre LM. 2007. Synthesis of six epoxyketooctadecenoicacid (EKODE) isomers, their generation from nonenzymatic oxidation oflinoleic acid, and their reactivity with imidazole nucleophiles. J Org Chem72(25):9471–80.

Mannina L, Sobolev AP, Segre A. 2003. Olive oil as seen by NMR andchemometrics. Spectros Eur 15(3):6–14.

Mannina L, Sobolev AP, Capitani D, Iaffaldano N, Rosato MP, Ragni P,Reale A, Sorrentino E, D’Amico I, Coppola R. 2008. NMR metabolicprofiling of organic and aqueous sea bass extract: implications in thediscrimination of wild and cultured sea bass. Talanta 77(1):433–44.

Markaverich, BM, Crowley JR, Alejandro MA, Shoulars K, Casajuna N,Mani S, Reyna A, Sharp J. 2005. Leukotoxins diols from ground corncobbedding disrupt estrus cyclicity in rats and stimulate MCF-7 breast cancercell proliferation. Environ Health Persp 113(12):1698–704.

Marmesat S, Velasco J, Dobarganes MC. 2008. Quantitative determination ofepoxy acids, keto acids and hydroxy acids formed in fats and oils at fryingtemperatures. J Chromatogr A 1211(1–2):129-34.

Marquez-Ruiz G, Dobarganes MC. 2005. Analysis of non-volatile lipidoxidation compounds by high-performance size-exclusion chromatography.In: Kamal-Eldin A, Pokorny J, editors. Analysis of lipid oxidation. 1st ed.Champaign, Ill.: AOCS Press. p 40–69.

Marquez-Ruiz G, Tasioula-Margari M, Dobarganes MC. 1995. Quantitationand distribution of altered fatty acids in frying fats. J Am Oil Chem Soc72(10):1171–6.

Martınez-Yusta A, Guillen MD. 2014a. Deep-frying food in extra virginolive oil. A study by 1H nuclear magnetic resonance of the influence of food



nature on the evolving composition of the frying medium. Food Chem150:429–37.

Martınez-Yusta A, Guillen MD. 2014b. A study by 1H nuclear magneticresonance of the influence on the frying medium composition of somesoybean oil-food combinations in deep-frying. Food Res Int 55:347–55.

Mikolajczak KL, Freidinger RM, Smith CR Jr, Wolff IA. 1968. Oxygenatedfatty acids of oil from sunflower seeds after prolonged storage. Lipids3(6):489–94.

Neff WE, El-Agaimy M. 1996. Autoxidation of trioleoylglycerol. OCL-OlCorps Gras Li 3:71–4.

Neff WE, Frankel EN. 1984. Photosensitized oxidation of methyl linolenatemonohydroperoxides: hydroperoxy cyclic peroxides, dihydroperoxides andhydroperoxy bis-cyclic peroxides. Lipids 19(12):952–7.

Neff WE, Frankel EN, Weisleder D. 1981. High-pressure liquidchromatography of autoxidized lipids. II. Hydroperoxy-cyclic peroxides andother secondary products from methyl linolenate. Lipids 16(6):439–48.

Neff WE, Frankel EN, Weisleder D. 1982. Photosensitized oxidation ofmethyl linolenate. Secondary products. Lipids 17(11):780–90.

Neff WE, Frankel EN, Selke E, Weisleder D. 1983. Photosensitizedoxidation of methyl linoleate monohydroperoxides: hydroperoxy cyclicperoxides, dihydroperoxides, keto esters and volatile thermal decompositionproducts. Lipids 18(12):868–76.

Neff WE, Frankel EN, Fujimoto K. 1988. Autoxidative dimerization ofmethyl linolenate and its monohydroperoxides, hydroperoxy epidioxides anddihydroperoxides. J Am Oil Chem Soc 65(4):616–23.

Neff WE, Frankel EN, Miyashita K. 1990. Autoxidation of polyunsaturatedtriacylglycerols: I. Trilinoleoylglycerol. Lipids 25(1):33–9.

Orban E, Nevigato T, Di Lena G, Casini I, Marzetti A. 2003. Differentiationin the lipid quality of wild and farmed sea bass (Dicentrarchus labrax) andgilthead sea bream (Sparus aurata). J Sci Food Agr 68:128–13.

Pajunen TI, Koskela H, Hase T, Hopia A. 2008. NMR properties ofconjugated linoleic acid (CLA) methyl ester hydroperoxides. Chem PhysLipids 154:105–14.

Perkins EG, Anfinsen JR. 1971. Characterization of the nonvolatilecompounds formed during the thermal oxidation of triolein. J Am OilChem Soc 48(10):556–62.

Porter NA, Mills, KA, Carter, RL. 1994. A mechanistic study of oleateautoxidation: competing peroxyl H-atom abstraction and rearrangement. JAm Oil Chem Soc 116(15):6690–6.

Sacchi R, Medina I, Aubourg SP, Addeo F, Paolillo L. 1993. Proton nuclearmagnetic resonance rapid and structure specific determination of ω-3polyunsaturated fatty acids in fish lipids. J Am Oil Chem Soc 70(3):225–8.

Sacchi R, Patumi M, Fontanazza G, Barone P, Fiodiponti P, Mannina L,Rossi E, Segre, AL. 1996. A high-field 1H nuclear magnetic resonancestudy of the minor components in virgin olive oils. J Am Oil Chem Soc73(6):747–58.

Sacchi R, Mannina L, Fiordiponti P, Barone P, Paolillo L, Patumi M, SegreA. 1998. Characterization of Italian extra virgin olive oils using 1H-NMRspectroscopy. J Agr Food Chem 46(10):3947–51.

Schneider C. 2009. An update on products and mechanisms of lipidperoxidation. Mol Nutr Food Res 53(3):315–21.

Schneider C, Porter NA, Brash AR. 2004. Autoxidative transformation ofchiral omega-6 hydroxy linoleic and arachidonic acids to chiral4-hydroxy-2Enonenal. Chem Res Toxicol 17(7):937–41.

Schneider C, Boeglin WE, Yin H, Ste DF, Hachey DL, Porter NA, BrashAR 2005. Synthesis of dihydroperoxides of linoleic and linolenic acids andstudies on their transformation to 4-hydroperoxynonenal. Lipids40(11):1155-62.

Schneider C, Porter NA, Brash AR. 2008. Routes to 4-hydroxynonenal:fundamental issues in the mechanisms of lipid peroxidation. J Biol Chem283(23):15539-43.

Schwartz DP, Rady AH, Castaneda S. 1994. The formation of oxo- andhydroxy-fatty acids in heated fats and oils. J Am Oil Chem Soc 71(4):441–4.

Schwartz H, Ollilainen V, Piironen V, Lampi A-M. 2008. Tocopherol,tocotrienol and plant sterol contents of vegetable oils and industrial fats. JFood Compos Anal 21(2):152–61.

Segre AL, Mannina L. 1997. H-NMR study of edible oils. Rec Res Dev OilChem 1:297–308.

Sheerin AN, Silwood C, Lynch E, Grootveld M. 1997. Production of lipidperoxidation products in culinary oils and fats during episodes of thermalstressing: a high field 1H NMR investigation. Biochem Soc T 25(3):495S.

Shimizu S, Jareonkitmongkol S, Kawashima H, Akimoto K, Yamada H.1991. Production of a novel ω1-eicosapentaenoic acid by Mortierella alpina1S-4 grown on 1- hexadecene. Arch Microbiol 156(3):163–6.

Siddiqui N, Sim J, Silwood CJL, Toms H, Iles RA, Grootveld M. 2003.Multicomponent analysis of encapsulated marine oil supplements usinghigh-resolution 1H and 13C NMR techniques. J Lipid Res 44(12):2406–27.

Silvagni A, Franco L, Bagno A, Rastrelli F. 2010. Thermoinduced lipidoxidation of a culinary oil: a kinetic study of the oxidation products bymagnetic resonance spectroscopies. J Phys Chem A 114(37):10059–65.

Silwood CJL, Grootveld M. 1999. Application of high-resolution,two-dimensional 1H and 13C nuclear magnetic resonance techniques to thecharacterization of lipid oxidation products in autoxidized linoleoyl/linolenoylglycerols. Lipids 34(7):741–56.

Sopelana P, Arizabaleta I, Ibargoitia ML, Guillen MD. 2013.Characterization of the lipidic components of margarines by 1H nuclearmagnetic resonance. Food Chem 141(4):3357–64.

Spindler SA, Clark KS, Callewaert DM, Reddy RG. 1997. Role anddetection of 9- and 13-hydroxyoctadecadienoic acids. Adv Exp Med Biol407:477–8.

Sun M, Salomon RG. 2004. Oxidative fragmentation of hydroxyoctadecadienoates generates biologically active γ -hydroxyalkenals. J Am OilChem Soc 126(18):5699–708.

Takeoka GR, Buttery RG, Perrino CT Jr. 1995. Synthesis and occurrenceof oxoaldehydes in used frying oils. J Agr Food Chem 43(1):22–6.

Thompson DA, Hammock BD. 2007. Dihydroxyoctadecamonoenoate estersinhibit the neutrophil respiratory burst. J Bioscience 32(2):279–91.

Thompson JA, May WA, Paulose MM, Peterson RJ, Chang, SS. 1978.Chemical reactions involved in the deep-fat frying of foods. VII.Identification of volatile decomposition products of trilinolein. J Am OilChem Soc 55(12):897–901.

Vidal NP, Manzanos MJ, Goicoechea E, Guillen MD. 2012. Quality offarmed and wild sea bass lipids studied by 1H NMR: usefulness of thistechnique for differentiation on a qualitative and a quantitative basis. FoodChem 135(3):1583–91.

Wineburg JP, Swern D. 1974. NMR chemical shift reagents in structuraldetermination of lipid derivatives. IV. Methyl cis- and trans-9,10-epoxystearate and methyl erythro- and threo-9,10-dihydroxystearate. J AmOil Chem Soc 51(12):528–33.

Wollenberg K. 1991. Quantitative triacylglycerol analysis of whole vegetableseeds by proton and carbon-13 magic angle sample spinning NMRspectroscopy. J Am Oil Chem Soc 68(6):391–400.

Yamauchi, R, Yamada T, Kato K, Ueno Y. 1983. Monohydroperoxidesformed by autoxidation and photosensitized oxidation of methyleicosapentaenoate. Agr Biol Chem 47(12):2897–902.

Zhang X, Cambrai A, Miesch M, Roussi S, Raul F, Aoude-Werner D,Marchioni E. 2006a. Separation of �5- and �7-phytosterols by adsorptionchromatography and semipreparative reversed-phase high-performanceliquid chromatography for quantitative analysis of phytosterols in foods. JAgr Food Chem 54(4):1196–202.

Zhang W, Sun M, Salomon RG. 2006b. Preparative singlet oxygenation oflinoleate provides doubly allylic dihydroperoxides: putative intermediates inthe generation of biologically active aldehydes in vivo. J Org Chem71(15):5607–15.


AReviewofThermo-OxidativeDegradationof …the volatile compounds formed in thermo-oxidation...

Documents

Transcript of AReviewofThermo-OxidativeDegradationof …the volatile compounds formed in thermo-oxidation...