Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 May 1.
Published in final edited form as: Food Chem. 2014 Nov 8;0:310–318. doi: 10.1016/j.foodchem.2014.11.015

Quantitative analysis of volatiles in edible oils following accelerated oxidation using broad spectrum isotope standards

Pilar Gómez-Cortés a, Gavin L Sacks b,c, J Thomas Brenna a,b,*
PMCID: PMC4274607  NIHMSID: NIHMS641534  PMID: 25529686

Abstract

Analysis of food volatiles generated by processing are widely reported but comparisons across studies is challenging in part because most reports are inherently semi-quantitative for most analytes due to limited availability of chemical standards. We recently introduced a novel strategy for creation of broad spectrum isotopic standards for accurate quantitative food chemical analysis. Here we apply the principle to quantification of 25 volatiles in seven thermally oxidized edible oils. After extended oxidation, total volatiles of high n-3 oils (flax, fish, cod liver) were 120-170 mg/kg while low n-3 vegetable oils were <50 mg/kg. Separate experiments on thermal degradation of d5-ethyl linolenate indicate that off-aroma volatiles originate throughout the n-3 molecule and not solely the n-3 terminal end. These data represent the first report using broad-spectrum isotopically labeled standards for quantitative characterization of processing-induced volatile generation across related foodstuffs, and verify the origin of specific volatiles from parent n-3 fatty acids.

Keywords: n-3 fatty acids, broad spectrum isotopic standards, terminal end, off-flavours, autoxidation, HS-SPME-GC/TOF-MS

1. INTRODUCTION

Oxidative stability is the major challenge in the incorporation of omega-3 fatty acids (n-3 FA) into food products (Decker, Akoh, & Wilkes, 2012). These lipids contain numerous 1,4-cis-pentadiene systems that are prone to hydrogen abstraction and subsequent free-radical oxidation reactions. Thus, their presence significantly shortens the shelf-life of the supplemented food and can even give rise to uncontrolled oxidation problems (Kolanowski, Jaworska, Laufenberg, & Weissbrodt, 2007). Most vegetable oils, except the rarely used flax oil, contain a relatively low proportion of n-3 polyunsaturated fatty acids (n-3 PUFA). Conversely, marine oils contain high amounts of these fatty acids (FA), especially eicosapentaenoic acid (EPA, 20:5 n-3) and docosahexaenoic acid (DHA, 22:6 n-3) which can account for one third or more of total FA.

Lipid autoxidation also causes significant changes to the sensory properties and consumer acceptance of food products including odour, flavour, colour and texture (Jacobsen, 2010). While hydroperoxides, the primary products of lipid autoxidation, are odourless and tasteless, their degradation leads to the formation of complex mixtures of low-molecular-weight compounds with distinctive aromas. (Shahidi & Pegg, 1994). Principally, these include alkanes, alkenes, aldehydes, ketones, alcohols, esters, epoxides, and FA. Those of greatest importance to the aroma of oils rich in n-3 PUFA appear to be medium-chain unsaturated aldehydes and ketones (Ho & Chen, 1994; Genot, Meynier, & Riaublanc, 2003).

While a few comparative studies on volatile profiles of oils exist, these are compromised by the fact that analyses relied on a limited number of internal standards (usually, only one) to quantify a diverse range of compounds (Bendini, Barbieri, Valli, Buchecker, Canavari, & Toschi, 2011; Mildner-Szkudlarz, Jelen, Zawirska-Wojtasiak, & Wasowicz, 2003; Uriarte, Goicoechea, & Guillen, 2011). Quantitative analyses of volatiles by headspace solid phase microextraction (HS-SPME) GC-MS require the use of internal standards to correct matrix effects, losses of analytes during sample preparation and GC-MS response. Ideally, these standards are stable isotopically labeled analogues of the compounds of interest, as these possess near-identical physiochemical properties to the analytes (Grosch, 2001). Assembling an array of standards presents a daunting challenge to studies of complex foodstuffs such as oxidized oils which may possess hundreds of analytes of possible interest. Because of the cost associated with preparing labeled standards, a more typical approach for analysis of edible oils is to use a small number of non-native compounds as internal standards. This approach may result in poor accuracy both because of differences in MS response between the analyte and standard, but also because of compound-dependent variation in extraction efficiency across matrices. For example, we recently reported that matrix effects observed in HS-SPME-GC-MS analyses of soybean oil could result in errors of over an order of magnitude for some oil-derived volatiles when a single internal standard is employed, even if calibration curves were run with external standards (Gómez-Cortés, Brenna, & Sacks, 2012). As a result, volatile concentrations reported in previous profiling studies may be inaccurate.

We recently presented a method to overcome the limitations for quantitative analysis of oil autoxidation by creating a broad-spectrum mixture of isotopically labeled standard prepared from controlled oxidation of [U-13C]-linolenic acid (Gómez-Cortés, Brenna, & Sacks, 2012). Here, we quantify 25 volatiles derived from n-3 FA in oils with various initial FA compositions using headspace solid phase microextraction gas chromatography time-of-flight mass spectrometry (HS-SPME GC/TOF-MS) using our novel uniformly isotopically labeled [U-13C] standards, and apply the method for the first time using a familiar and important food chemical problem. Additionally, we investigated whether these volatiles are derived solely from the n-3 terminal end or from interior portions of the fatty acid.

2. MATERIALS AND METHODS

2.1. Chemicals and Oils

Polyethylene glycol (PEG 400) and non-labeled ethyl linolenate standard (≥ 98%) were purchased from Sigma-Aldrich (St Louis, MO, USA). The isotopically labeled [17,17',18,18',18”-d5]-ethyl linolenate (98%) and [U-13C]-linolenic acid (99%) standards were obtained from Cambridge Isotope Laboratories (Andover, MA, USA). The seven oils used in the study are as follows. (1) Partially hydrogenated vegetable oil (PHVO), specifically soybean oil, and (2) low α-linolenic acid soybean oil (LSO) containing nominally 55% linoleic acid and 3% α-linolenic acid were gifts of Bunge Ltd (White Plains, NY, USA); (3) refined commodity soybean oil (CSO) with nominally 55% linoleic acid and 7% linolenic acid (Wegman's Food Markets, Inc, Rochester, NY, USA); (4) expeller pressed refined walnut oil (WO; Spectrum Naturals, Lake Success, NY, USA); (5) cold pressed unrefined, unfiltered flax oil (FO; Barlean's Ferndale, WA, USA); (6) cod liver oil (CLO; Solgar, Inc, Leonia, NJ, USA) containing 28 mg DHA and 28 mg EPA per 460 mg, and (7) refined fish oil (FO; Wegman's Food Markets, Inc, Rochester, NY, USA) were obtained in the form of capsules which were carefully opened and oil removed. The oil is a blend of anchovy, herring, and sardine oils. All oils were used as set forth below without further processing. Detailed fatty acid profiles of all oils prior to oxidation are provided in Supplementary Table S1, determined as described below.

2.2. Oil FA Composition Analysis

FA compositions of the seven oils prior to oxidation were determined by the one step extraction/methylation method described by previously (Zhou, Nijland, Miller, Ford, Nathanielsz, & Brenna, 2008). Analyses were performed in triplicate on a HP 5890 Series II gas chromatograph coupled to a flame ionization detector (GC-FID) (Hewlett Packard, Palo Alto, CA, USA) equipped with a BPX70 fused-silica capillary column (25m × 0.22 mm i.d. × 0.25 μm film thickness; SGE Inc., Austin, TX, USA). The column temperature program was as follows: the initial temperature of 80°C was ramped up to 170°C at 30°C/min, 2 min hold, then increased to 240°C at 10°C/min, 14 min hold. The injector was at 250°C in splitless mode and hydrogen was used as carrier gas at a flow rate of 1mL/min.

2.3. Oil Oxidation

Oils were oxidized following the protocol described in the AOCS (1999)Oven Storage Test Cg 5-97. Ten millilitres of each oil were placed in an amber 20 mL solid-phase microextraction (SPME) vial, sparged with O2 for 5 min, sealed with Teflon and vortexed. Oils were incubated at 60°C in the dark for 0, 3, 6, 9, 12 and 15 days. A separate SPME vial was used for each sampling time. After oxidation, the headspace (HS) of the container was sparged with N2 and stored at −80°C until volatile analysis (AOCS, 1999).

2.4. HS-SPME-GC/TOF-MS Volatile Analyses

Volatiles were extracted by HS-SPME using a LEAP CombiPAL autosampler (Carrboro, NC, USA) as described previously (Gómez-Cortés, Brenna, & Sacks, 2012). Samples were incubated for 10 min at an agitation rate of 300 rpm and with an incubation temperature of 50°C prior to fibre insertion. A 2 cm 50/30 μm divinylbenzene/carboxen/poly(dimethylsiloxane) SPME fibre (Supelco, Bellefonte, PA, USA) was then introduced into the HS and the vial was agitated at 100 rpm for 20 min at 50°C.

Following HS-SPME, volatiles were thermally desorbed from the fiber into the injector of an Agilent 6890 gas chromatograph coupled to a time-of-flight mass spectrometer (GC/TOF-MS, Pegasus 4D, LECO Corp., St. Joseph, MI, USA). SPME injections were splitless with 5 min of desorption at 250°C. The GC column was a DB-FFAP capillary column (30 m × 0.25 mm i.d. × 0.25 μm film thickness; Agilent Technologies Inc., New Castle, DE, USA). The column temperature program was as follows: initial hold for 3 min at 40°C, followed by a 5°C/min ramp to 185°C and then, 8°C/min ramp to 240°C, 5 min hold. Helium was the carrier gas at a flow rate of 1 mL/min, and the detector temperature was 200°C. The TOF-MS was operated in electron impact mode with an ionization energy of 70 eV. The electron multiplier was set to 1500 V. MS data were stored at an effective acquisition rate of 5 spectra/s over a mass range of m/z 35-400, and data processing was carried out by the native LECO ChromaTOF software.

For quantitation purposes, a mixture of [U-13C] volatile standards was generated via [U-13C]-linolenic acid degradation as described in previously (Gómez-Cortés, Brenna, & Sacks, 2012). A 30 μL sub-sample of the [U-13C] mixture was added to each oxidized oil aliquot (220 μL) in a 20 mL amber SPME vial. All samples were analyzed in quadruplicate by HS-SPME-GC/TOF-MS to quantify the volatiles derived from n-3 FA oxidation.

Odour activity values (OAV) were calculated for seven selected volatiles (acetaldehyde, propanal, 2,3-pentanedione, hexanal, (E)-2-pentenal, acetic acid and (E,E)-2,4-heptadienal) based on the availability of their odours thresholds in oil in the literature. OAV was calculated as (odourant concentration) / (odourant threshold in oil).

2.5. Characterizing Provenance of n-3 Derived Volatiles using Deuterated Ethyl Linolenate

To characterize the volatiles arising from the terminal end of n-3 FA, 5 mg of either unlabeled ethyl linolenate or [17,17’,18,18’,18”-d5]-ethyl linolenate standards were added to an amber 20 mL SPME vial and dissolved into 500 mg of PEG 400. Both vials (i.e. unlabeled 18:3 and d5-18:3) were sparged with O2 for 5 min, sealed with Teflon, vortexed and incubated at 60°C in the dark. After 72h of oxidation, vials were sparged with N2 and directly analyzed under the conditions described in section 2.4.

2.6. Statistical Analysis

Statistical analysis was conducted with JMP Version 9 (SAS Institute, Cary, NC). Initial exploration of the data was performed by principal component analysis (PCA) on all the time points (0, 3, 6, 9, 12 and 15 days) of the 7 oils (PHVO, LSO, CSO, WO, Flax oil, FO and CLO). A direct visual scree test performed on the eigenvalue distribution showed two factors were sufficient. Paired comparisons, using Tukey's HSD test, were then used to compare oils volatile composition. P < 0.05 was considered to be statistically significant.

3. RESULTS AND DISCUSSION

3.1. FA Composition and Total Volatile Production from Commercial Oils

Table S1 (see the Supplementary Data) shows the FA composition of the 7 oils selected for the autoxidation study. The oils were chosen based on differences in their n-3 content. The FA profiles coincide with the common composition of these oils and are in agreement with those reported previously (Moffat & McGill, 1993; Tompkins & Perkins, 2000; Uriarte, Goicoechea, & Guillen, 2011). As expected, the only n-3 FA detected in vegetable oils corresponds to α-linolenic acid (ALA, cis-9 cis-12 cis-15 18:3) which accounted for 0.5, 3, 7, 12 and 56% of the total fat in PHVO, LSO, CSO, WO and flax oil, respectively (Table S1, Supplementary Data). The low oxidative stability associated with ALA (i.e. in comparison to linoleic or oleic acids) is commonly related to its higher number of double bonds, as an increase in the number of double bonds increases the rate of autoxidation of the oils, a chemical reaction with very low activation energy (Zhu & Sevilla, 1990). For example, due to the sensitivity to oxidation of soybean oil, in order to increase the storage stability and improve its performance in frying applications, the FA profile of soybean oil is normally altered. By lowering the ALA levels to approximately 3%, low-ALA soybeans create a more stable oil and eliminate the need for partial hydrogenation, a process that creates trans fats, which have been linked to heart disease and other health concerns (Dhaka, Gulia, Ahlawat, & Khatkar, 2011). On the other hand, flax oil and in a lesser extent WO are expected to be particularly susceptible to oxidation as their concentrations of ALA were roughly eight-fold and two-fold greater than in CSO. The present study also included two marine oils (FO and CLO) that are well known for their high levels of n-3 PUFA and their susceptibility to autoxidation (De Leonardis & Macciola, 2006). Although total n-3 FA contents were lower than in flax oil, these marine oils were characterized by major levels of highly unsaturated EPA (20 and 10% of total FA methyl esters in FO and CLO, respectively) and DHA (11 and 8% of total FA methyl esters in FO and CLO, respectively).

Figure S1 (see Supplementary Data) presents the total concentration of volatile compounds derived from n-3 FA when these 7 oils were subjected to thermal oxidation. Increasing oxidation time was correlated with greater volatile production, and the highest volatile production was observed at 15 d for all oils. As expected, the total volatile concentration was much greater in oils with high levels of n-3 FA such as flax oil, FO or CLO than in vegetable oils with lower levels of n-3 FA. PHVO was the least susceptible to oxidation and generated the lowest levels of volatiles (Figure S1, Supplementary Data), followed by LSO. Although flax oil had the greatest levels of total n-3 FA, marine lipids yielded more volatiles and thus were more prone to degradation. Furthermore, a sharp increase between day 0 and day 3 was observed in total volatile concentration of FO and CLO (Figure S1, Supplementary Data).

The greater oxidizability of EPA- and DHA-containing marine oils as compared to vegetable oils can be rationalized by differences in the dissociation energy of the carbon-hydrogen (C-H) bonds and their susceptibility to proton abstraction. In general, the dissociation energy of the C-H bond at bis-allylic methylene groups is lower than at allylic methylene groups (50 kcal/mole vs 75 kcal/mole), whereas the dissociation energy of the C-H bond at methylene groups without adjacent double bonds is markedly higher (approximately 100 kcal/mol) (Min & Boff, 2002). Hence, PUFA are more easily oxidized than monounsaturated and saturated FA (Eder & Ringseis, 2010). Among PUFA, increasing the number of bis-allylic methylene groups increases susceptibility to oxidation; for the FA series linoleic acid (18:2), ALA, arachidonic acid (20:4), EPA, DHA, each bis-allylic methylene group increases the relative rate of oxidation two-fold (Cosgrove, Church, & Pryor, 1987). Our work shows that the degree of volatile production is related to the presence of FA with greater number of bis-allylic positions, which is in turn critically related to the risk of autoxidative rancidity and spoilage.

3.2. Quantitative Distribution of n-3 Volatiles

Table 1 presents the quantitative distribution of volatile compounds derived from n-3 FA oxidation in the 7 oils after 15 days of autoxidation. These data is very valuable because most HS-SPME-GC/MS oil oxidation studies only provide semi-quantitations of the volatile components (Jelen, Mildner-Szkudlarz, Jasinska, & Wasowicz, 2007; Petersen, Kleeberg, Jahreis, & Fritsche, 2012; Uriarte, Goicoechea, & Guillen, 2011). Quantitative analyses in our current work were performed using a previously developed uniformly labeled [U-13C] volatile mixture, derived from thermal degradation of a [U-13C]–ALA standard (Gómez-Cortés, Brenna, & Sacks, 2012). In total, 25 compounds were identified and quantified (Table 1). The majority of the volatiles were aldehydes (four alkanals, three alkenals and one dienal) and ketones (three saturated, three unsaturated, three hydroxy ketones and one α-diketone). Two unsaturated alcohols, two carboxylic acids, two alkyl furanones and one alkyl furan were also found. These compounds represent characteristic groups of FA secondary oxidation products, resulting mainly from autoxidation of ALA (Gómez-Cortés, Brenna, & Sacks, 2012).

Table 1.

Quantification of volatile compounds in oils with different n-3 fatty acid content after 15 days of oxidation. Data presented as mean ± SD; units are ng compound per gram oil.

Compound (ng/g oil) PHVO1 LSO2 CSO3 WO4 Flax Oil FO5 CLO6
Acetaldehyde - 3571 ± 131c 5826 ± 702bc 9063 ± 1362b 13972 ± 4084a 15354 ± 1688a 18696 ± 2352a
Propanal 289 ± 56d 2109 ± 204d 3008 ± 182d 6444 ± 582c 12729 ± 2973b 14269 ± 1853ab 16587 ± 1475a
2-Propanone 273 ± 26d 48 ± 4d 121 ± 4c 76 ± 4d 438 ± 29a 314 ± 27b 293 ± 4b
2-Propenal - 416 ± 15d 694 ± 40d 954 ± 102d 1773 ± 224c 4071 ± 75b 5847 ± 696a
Butanal 70 ± 36c 40 ± 4c 81 ± 17c 185 ± 24c 584 ± 99b 1870 ± 248a 1674 ± 126a
2-Butanone 6.0 ± 0.3d 13.2 ± 0.4d 24 ± 1d 65 ± 4c 227 ± 18a 227 ± 16a 194 ± 4b
2-Ethylfuran 2.1 ± 0.3e 5.1 ± 0.5e 26 ± 2de 99 ± 4d 700 ± 17c 2336 ± 48b 4364 ± 76a
2-Pentanone 19 ± 5d 10.34 ± 0.04d 21 ± 7d 22 ± 14d 174 ± 12a 92 ± 13b 60 ± 9c
1-Penten-3-one 3.5 ± 0.3e 25.5 ± 0.5e 59 ± 2de 124 ± 4d 496 ± 30c 933 ± 77a 816 ± 61b
(E)-2-Butenal - 117.8 ± 0.7f 323 ± 3e 750 ± 3d 2707 ± 29a 2357 ± 71b 2004 ± 73c
2,3-Pentanedione - 41 ± 6d 39 ± 6d 83 ± 4c 78 ± 8c 107 ± 6b 192 ± 9a
Hexanal 485 ± 83d 1375 ± 95b 734 ± 63c 668 ± 60c 1318 ± 62b 1353 ± 79b 1753 ± 57a
(E)-3-Penten-2-one - 3.7 ± 0.2e 7.1 ± 0.7e 37 ± 1d 440 ± 13a 322 ± 2b 269 ± 2c
(E)-2-Pentenal 38 ± 6g 241 ± 4f 520 ± 21e 1062 ± 39d 3649 ± 38a 2782 ± 17c 2841 ± 23b
1-Penten-3-ol - - - - - 55 ± 5b 63 ± 3a
3-Hexen-2-one - 1.6 ± 0.2e 7.2 ± 0.8e 27 ± 1d 219 ± 8a 133 ± 2b 101 ± 1c
1 -Hydroxy-2-propanone - 41 ± 2d 26.7 ± 0.9e 48 ± 3d 132 ± 8c 197 ± 3b 231 ± 3a
(E)-2-Penten-1-ol - - - - 328 ± 43b 466 ± 80a 506 ± 45a
2-Hydroxy-3-pentanone - 28 ± 1e 30 ± 1e 95 ± 6d 623 ± 32a 296 ± 18c 470 ± 7b
1-Hydroxy-2-butanone - 58 ± 4e 160 ± 18d 463 ± 24c 1473 ± 58a 726 ± 16b 773 ± 40b
Acetic acid 2626 ± 811e 2299 ± 243e 2718 ± 19e 4725 ± 153d 8617 ± 508c 16599 ± 578b 20026 ± 596a
(E,E)-2,4-Heptadienal 725 ± 194f 10437 ± 225e 18129 ± 2283d 24108 ± 2175c 61902 ± 1474b 94749 ± 5127a 56485 ± 2647b
Propanoic acid 442 ± 89e 540 ± 22e 673 ± 12e 1895 ± 35d 5993 ± 401b 5462 ± 197c 7431 ± 242a
5-Methyl-2(5H)-Furanone - 129 ± 15d 225 ± 46cd 424 ± 136c 1029 ± 80b 1384 ± 135a 1469 ± 106a
5-Ethyl-2(5H)-Furanone - 43 ± 3d 80 ± 8d 233 ± 21c 822 ± 84b 839 ± 92b 1100 ± 64a
Total Volatiles 4978 ± 974f 21592 ± 519e 33533 ± 2931e 51652 ± 3565d 120430 ± 8035c 167294 ± 9616a 144244 ± 6810b
Open in a new tab
1

PHVO, partially hydrogenated vegetable oil.

2

LSO, low α-linolenic acid soybean oil.

3

CSO, commodity soybean oil.

4

WO, walnut oil.

5

FO, fish oil.

6

CLO, cod liver oil.

a

Means within a row with different superscripts differ significantly.

b

Means within a row with different superscripts differ significantly.

c

Means within a row with different superscripts differ significantly.

d

Means within a row with different superscripts differ significantly.

e

Means within a row with different superscripts differ significantly.

f

Means within a row with different superscripts differ significantly.

g

Means within a row with different superscripts differ significantly.

To obtain an overview of the similarities and differences among the autoxidized oils, a PCA was performed on the volatile profiles for each oil at every time point (0, 3, 6, 9, 12 and 15 days). The PCA scores plot is shown in Figure 1 (top). The first two principal components accounted for 82.3 and 8.28%. Oils separated based on a combination of storage time and oil type in the first dimension. Low n-3 oils segregated to the left, and higher n-3 oils (flax oil, FO or CLO) to the right along PC1, with further discrimination of the high n-3 oils by storage time. In the loadings plot (Figure 1, bottom), all volatiles were located on the right side of the plot and thus PC1 is separating based on total volatile production, similar to data in Figure S1 (Supplementary Data). PC2 differentiated the high ALA flax oil from the high DHA and EPA marine oils (FO and CLO). Based on the loadings plot, flax oil was characterized by higher concentrations of saturated, unsaturated and hydroxymethyl ketones (2-propanone, 2-pentanone, 2-hydroxy-3-pentanone, 1-hydroxy-2-butanone, (E)-3-penten-2-one and 3-hexen-2-one) while the marine oils were characterized by higher concentrations of a diverse range of compound, including aldehydes (1-penten-3-ol, 2-ethylfuran, butanal, 2-propenal and 2,3-pentanedione; Figure 1).

Figure 1.

Figure 1

Open in a new tab

Scores plot (top) and loadings plot (bottom) for principal component analysis carried out on the volatile data profile of all oils at all time points.

The differences between the volatile profiles of oils were further investigated by ANOVA of each volatile for the seven oils at the final time point, day 15 (Table 1). Compounds that discriminated marine and flax oils in the loading plot of PC2 in Figure 1 generally differed significantly. For example, butanal was 3-fold higher in the marine oils, while 2-pentanone was 50% higher in flax oil. To our knowledge, these differences in volatile profile have not been previously reported, although the reasons for their appearance are unclear.

From a quantitative point of view, the most prominent volatile degradation products from n-3 FA were (E,E)-2,4-heptadienal, acetaldehyde, acetic acid, propanal, propanoic acid and 2-propenal (Table 1). The high concentration of aldehydes could explain the off-flavour and limited frying live usually associated with n-3 oils. Aldehydes tend to be more potent, more stinky and can react with amino groups to produce imino Schiff bases, which themselves polymerize by aldol condensation to dimers and complex high-molecular-weight brown macromolecules known as melanoidins (Eder & Ringseis, 2010). This reaction significantly contributes to the development of the typical flavour of fried foods and causes significant loses of labile amino acids reducing the nutritive value of food proteins (Eder & Ringseis, 2010). Furthermore, aldehydes are longer lived than free radicals and its increased dietary intake may potentially adversely affect health (Dyall, 2011). On the other hand, 2-propenal is known for reacting with phenolic hydroxyl groups from other food products to yield bitter tasting compounds (Novo, Quirós, Morales, & González, 2012).

The volatile compounds quantified in the present study have a large range of sensory thresholds. To better evaluate the potential contribution of these compounds to aroma in oil samples, OAV were calculated (Table 2). In flax oil, FO and CLO, the seven odourants were present in concentrations above their odour thresholds. By far the highest OAV was calculated for the sweet burning-smelling acetaldehyde. Although (E,E)-2,4-heptadienal was the major volatile compound from n-3 autoxidized PUFA oils, it also has a high threshold (10000 ng/g of oil, Table 2) and is thus less likely to contribute to the aroma. (E)-2-pentenal and “sharp-irritating” propanal were below thresholds in PHVO, LSO and CSO (OAV < 1) with supra- or peri-threshold concentrations only observed in the high n-3 oils (Table 2). Surprisingly, the butter-like 2,3-pentanedione had a high OAV, due to its low odour threshold (0.3 ng/g of oil, Table 2). The OAV data should be interpreted with care as they ignore additive and masking effects that often occur. Moreover, these OAV were obtained using odour thresholds in oil and cannot be extrapolated to food systems with more complex matrices. Oxidized oils contain a complex mixture of different volatiles and the relationship between their concentration and the sensory impact is still poorly understood (Jacobsen, 2010). Additionally, the SPME-GC/TOF-MS method employed is less sensitive to volatiles with more than eight or more carbons due to their limited volatility, and several longer chain volatiles (e.g. 1-octen-3-one, 2,4,7-decatrienal) have been reported to be important to the aroma of oxidized n-3 oils (Genot, Meynier, & Riaublanc, 2003; Venkateshwarlu, Let, Meyer, & Jacobsen, 2004). Finally, the presence of other ingredients during cooking could significantly alter the odour impression by reason of dilution, masking, or the formation of new volatiles (Urbach & Gordon, 1995).

Table 2.

Odour thresholds and odour activity values (OAV) of selected odour-active compounds in oil after 15 days of autoxidation; units are ng compound per gram oil.

Odourant Odour threshold (ng/g oil) OAV1
PHVO2 LSO3 CSO4 WO5 Flax Oil FO6 CLO7
Acetaldehyde 0.22a - 16231 26482 41197 63535 69791 84980
Propanal 3600b 0.1 0.6 0.8 1.8 3.5 4.0 4.6
2,3-Pentanedione 0.3c - 137 129 277 261 358 640
Hexanal 276c 1.8 5.0 2.7 2.4 4.8 4.9 6.4
(E)-2-Pentenal 2300d 0.0 0.1 0.2 0.5 1.6 1.2 1.2
Acetic acid 124c 21 19 22 38 69 134 161
(E,E)-2,4-Heptadienal 10000d 0.1 1.0 1.8 2.4 6.2 9.5 5.6
Open in a new tab
1

OAV = (odourant concentration) / (odourant threshold in oil).

2

PHVO, partially hydrogenated vegetable oil.

3

LSO, low α-linolenic acid soybean oil.

4

CSO, commodity soybean oil.

5

WO, walnut oil.

6

FO, fish oil.

7

CLO, cod liver oil.

3.3. Origin of Volatiles from n-3 FA Oxidation

It is well known that ALA autoxidation occurs at a higher rate than linoleic (18:2 n-6) and oleic (18:1 n-9) acids degradation (Hsieh & Kinsella, 1989; Lea, 1952). Although the degree of unsaturation is a major determining factor in the rate of lipid oxidation, it is unclear how the n-3 terminal end contributes to volatile formation. To address the molecular origin of the volatile compounds quantified in the present study, unlabeled ethyl linolenate and [17, 17', 18, 18', 18”-d5]-ethyl linolenate standards were purposively degraded by heat treatment and the mass spectra (MS) of corresponding volatiles compared. As an example, Figure 2 shows the MS of four ALA degradation metabolites that derive from the n-3 terminal end. The d5-propanal MS has prominent ions at m/z 62 and 63 (M+), 5 amu higher than the respective ions m/z 57 and 58 (M+) from unlabeled propanal. A similar shift was also detected for M+ in unlabeled (E)-2-pentenal and d5-(E)-2-pentenal (m/z 84 and 89, respectively), for unlabeled 2,3-pentanedione and d5-2,3-pentanedione (m/z 100 and 105, respectively), and for unlabeled 5-ethyl-2(5H)-furanone and d5-5-ethyl-2(5H)-furanone (m/z 112 and 117, respectively). After degradation of [17, 17', 18, 18', 18”-d5]-ethyl linolenate standard, these unlabeled volatiles (propanal, (E)-2-pentenal, 2,3-pentanedione, and 5-ethyl-2(5H)-furanone) were not detected in the chromatogram. Thus, we can assert that they derive almost exclusively from the n-3 terminal end of the molecule.

Figure 2.

Figure 2

Open in a new tab

Comparison of mass spectra of four volatile compounds obtained by thermal degradation of unlabeled ethyl linolenate and [17, 17', 18, 18', 18”-d5]-ethyl linolenate standards: (A) Propanal, (A’) d5-Propanal, (B) (E)-2-Pentenal, (B’) d5-(E)-2-Pentenal, (C) 2,3-Pentanedione, (C’) d5-2,3-Pentanedione, (D) 5-Ethyl-2(5H)-Furanone and (D’) d5-5-Ethyl-2(5H)-Furanone.

In contrast, Figure 3 presents the partial chromatogram and the MS of two volatiles whose origin is the terminal end and other parts of the n-3 molecule at different ratios (butanal and 2-butanone). After [17, 17', 18, 18', 18”-d5]-ethyl linolenate autoxidation, two peaks were detected for each compound (d5-labeled and unlabeled). Both d5-butanal and d5-2-butanone MS have the molecular ion at m/z 77, 5 amu higher than their respective unlabeled compounds (M+ = 72). Unlabeled butanal is characterized by the McLafferty rearrangement (ML, m/z 44) and loss of the methyl group (m/z 57) while these ions are detected at m/z 45 and 59 in the deuterated form. On the other hand, unlabeled 2-butanone presents ions at m/z 43 (base peak, loss of ethyl group) and m/z 57 (loss of methyl) which were also observed in d5-2-butanone (m/z 43 and 62, respectively). The equivalents of terminal end contributing to each volatile were calculated by comparing the relative peak heights (unlabeled analyte / d5-analyte × 100%) measured by HS-SPME-GC/TOF-MS (Table 3). Accordingly, the amount of butanal and 2-butanone that derives from the ALA n-3 terminal end would be 65 and 45%, respectively (Figure 3 and Table 3).

Figure 3.

Figure 3

Open in a new tab

Mass spectra and partial GC chromatogram displaying m/z 72 and m/z 77 illustrating separation of d5-Butanal and d5-2-Butanone from their parallel unlabeled compounds.

* ML, McLafferty rearrangement.

Table 3.

Retention index (RI), aroma properties and molecular origin of volatile compounds arising from oxidation of n-3 fatty acids (FA). Units of Equivalents (rightmost column) are percent of compound originating from the terminal end.

Compound RI1 Odour descriptor Equivalents of terminal end in fragment (%)
Acetaldehyde 812 Apple skin, fruity, green leaves, sweet, sweet burninga,b 20
Propanal 833 Sharp-irritatingc 97
2-Propanone 842 Pungent, irritatingb 0
2-Propenal 857 Irritatingb 0
Butanal 873 Burnt, irritating, green, penetrating pungent, varnishb 65
2-Butanone 890 - 45
2-Ethylfuran 940 Flowerc 100
2-Pentanone 968 Acetone, sweet fruity ketoneb 96
1-Penten-3-one 1018 Pungent, rancid green, glue, sharp fishyc 100
(E)-2-Butenal 1039 Old cheesec 100
2,3-Pentanedione 1059 Butter-likea 100
Hexanal 1080 Pungent, green, fresh grass, fattya,c 48
(E)-3-Penten-2-one 1125 - ?
(E)-2-Pentenal 1128 Pungent, glue, green, grassy, applec 100
1-Penten-3-ol 1156 Sweetc 100
3-Hexen-2-one 1214 - 91
1 -Hydroxy-2-propanone 1300 - 0
(E)-2-Penten-1-ol 1308 Greenc 100
2-Hydroxy-3-pentanone 1355 - 100
1 -Hydroxy-2-butanone 1373 - 100
Acetic acid 1450 Vinegaryd 2
(E,E)-2,4-Heptadienal 1492 Nasty, green, fatty, fatty-oily, rancid hazelnutsa,c ?
Propanoic acid 1538 Goaty, vinegar, cooked, potato-liked,e 100
5-Methyl-2(5H)-Furanone 1684 Caramel, fruity, phenolicse ?
5-Ethyl-2(5H)-Furanone 1765 - 100
Open in a new tab

- Data not found. ? No conclusive data were obtained from the present study.

1

Retention index on DB-FFAP.

b

The LRI and Odour Database.

The n-3 terminal end was the precursor of most of the volatiles quantified in the present study (at least 14 from the 25). Although volatiles with 5 carbon atoms derived almost exclusively from the ALA n-3 terminal end, most 3- and 4-carbon compounds contained <100%, indicating that these smaller molecules can be derived from interior regions of ALA (Table 3).

The formation of short chain aldehydes from interior sites of ALA can be rationalized based on the mechanism of unsaturated FA autoxidation (Ho & Chen, 1994). ALA autoxidation generates 9-, 12-, 13- and 16- conjugated diene-triene hydroperoxide isomers in various proportions (30, 12, 12 and 46%, respectively) (Frankel, 1984). Thermal decomposition of the hydroperoxide involves the homolytic cleavage of the O-O bond to produce alkoxyl and hydroxyl radicals. The alkoxyl radical undergo C-C bond scissions on either sides to produce and alkyl radical on one side and a vinyl radical on the other which lead to a large number of volatile products including alkanes, alkenes, aldehydes, ketones, alcohols, esters, and acids. Many aldehydes could be produced by scission of the lipid molecules on either side of the radical, which would explain why they were not derived entirely from the n-3 terminal end (e.g. acetaldehyde, 2-propenal, butanal and hexanal, Table 3). Regarding other aldehydes, propanal is well known to be specifically formed from the oxidation of n-3 PUFA (Liang, Wang, Simon, Shahidi, & Ho, 2007). (E)-2-butenal and (E)-2-pentenal were fully derived from the n-3 terminal end and their levels in all oils were closely related (R2 = 0.987). These three compounds all share structural similarity to the n-3 terminal end.

Peroxyl radical cyclization is believed to be the mechanism for formation of cyclic volatiles in the autoxidation of FA having 3 or more double bonds (Porter, Caldwell, & Mills, 1995). Adams et al. proposed a mechanism for the formation of 2-alkylfurans from the corresponding (E)-2-alkenals (Adams, Bouckaert, Van Lancker, De Meulenaer, & De Kimpe, 2011), e.g 2-pentenal would be the precursor of 2-methylfuran and 2-hexenal would be the precursor of 2-ethylfuran. We observed that 2-pentenal and 2-ethylfuran originated exclusively from the ALA n-3 terminal end, most likely following the pathway hypothesized by Adams et al (Adams, Bouckaert, Van Lancker, De Meulenaer, & De Kimpe, 2011). 5-ethyl-2(5H)-furanone was also derived from the terminal end (Table 3). In this case, it would be generated via oxidation of (Z)-3-hexenal, although this compound was not quantified in our study. This mechanism involves the conversion of (Z)-3-hexenal to the peracid followed by the oxidation of the (Z)-3-double to form an epoxide. Then, in an intramolecular process, the peracid would open the epoxide ring to first form the hydroxy lactone and then dehydrate to yield 5-ethyl-2(5H)-furanone (Buttery & Takeoka, 2004). Although we could not determine the source of 5-methyl-2(5H)-furanone due to its low signal, it would also derive from the ALA n-3 terminal end due to its close relation detected with 5-ethyl-2(5H)-furanone (R2 = 0.921).

Our results indicate that cyclic volatiles and volatiles with 5 carbon atoms derive almost exclusively from the terminal end of n-3 FA. However, the origin of lower molecular weight compounds (with 4 or 3 carbon atoms) is more variable, with compounds sharing structural similarity to the n-3 terminal end being more likely to derive from the terminal end. However, more FA autoxidation studies quantifying volatiles and high molecular weight compounds are needed to determine whether the susceptibility to oxidation of oils rich in n-3 PUFA is solely due to the terminal end or to the total number of bis-allylic groups in the molecules.

4. CONCLUSIONS

In this study, we provide quantitative data of 25 volatile compounds derived from the thermal degradation of seven oils with different n-3 FA content using a new methodology that relies on a range of [U-13C]-labeled compounds generated from the thermal oxidation of [U-13C] linolenic acid. In addition, autoxidation of [17,17',18,18',18”-d5]-ethyl linolenate revealed that the n-3 terminal end of PUFA is not the only responsible of volatiles commonly related with off-aromas.

Supplementary Material

NIHMS641534-supplement.docx (171.6KB, docx)

Highlights.

  • A broad spectrum isotopic standard facilitates quantitative analysis

  • Twenty-five volatiles derived from n-3 polyunsaturated oil degradation were determined.

  • High n-3 oils flax and marine oil generate volatiles at greater rates than low n-3 oils

  • Volatiles originated throughout the n-3 molecule, not just the terminal end.

ACKNOWLEDGEMENTS

This research project was supported by NIH grant RR031264. P. Gómez-Cortés is the recipient of a postdoctoral grant from the Alfonso Martín Escudero Foundation (Madrid, Spain). The authors are indebted to P. Lawrence for providing the oils FA composition and Dr I. Ryona for her valuable assistance with the instrument and helpful discussions.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

APPENDIX

Supplementary data (Figure S1 and Table S1) associated with this article can be found, in the online version, at ScienceDirect: http://www.sciencedirect.com.

REFERENCES

  1. Adams A, Bouckaert C, Van Lancker F, De Meulenaer B, De Kimpe N. Amino Acid Catalysis of 2-Alkylfuran Formation from Lipid Oxidation-Derived alpha,beta-Unsaturated Aldehydes. Journal of Agricultural and Food Chemistry. 2011;59(20):11058–11062. doi: 10.1021/jf202448v. [DOI] [PubMed] [Google Scholar]
  2. AOCS . Official Methods and Recommended Practices of the American Oil Chemists' Society. 5th ed. American Oil Chemists' Society; Champaign, IL: 1999. [Google Scholar]
  3. Bendini A, Barbieri S, Valli E, Buchecker K, Canavari M, Toschi TG. Quality evaluation of cold pressed sunflower oils by sensory and chemical analysis. European Journal of Lipid Science and Technology. 2011;113(11):1375–1384. [Google Scholar]
  4. Buttery RG, Takeoka GR. Some unusual minor volatile components of tomato. Journal of Agricultural and Food Chemistry. 2004;52(20):6264–6266. doi: 10.1021/jf040176a. [DOI] [PubMed] [Google Scholar]
  5. Chetschik I, Granvogl M, Schieberle P. Quantitation of Key Peanut Aroma Compounds in Raw Peanuts and Pan-Roasted Peanut Meal. Aroma Reconstitution and Comparison with Commercial Peanut Products. Journal of Agricultural and Food Chemistry. 2010;58(20):11018–11026. doi: 10.1021/jf1026636. [DOI] [PubMed] [Google Scholar]
  6. Cosgrove JP, Church DF, Pryor WA. The kinetics of the autoxidation of polyunsaturated fatty-acids. Lipids. 1987;22(5):299–304. doi: 10.1007/BF02533996. [DOI] [PubMed] [Google Scholar]
  7. Curioni PMG, Bosset JO. Key odorants in various cheese types as determined by gas chromatography-olfactometry. International Dairy Journal. 2002;12(12):959–984. [Google Scholar]
  8. De Leonardis A, Macciola V. Behaviour of cod liver oil during the autoxidation process. European Journal of Lipid Science and Technology. 2006;108(10):871–876. [Google Scholar]
  9. Decker EA, Akoh CC, Wilkes RS. Incorporation of (n-3) Fatty Acids in Foods: Challenges and Opportunities. Journal of Nutrition. 2012;142(3):610S–613S. doi: 10.3945/jn.111.149328. [DOI] [PubMed] [Google Scholar]
  10. Dhaka V, Gulia N, Ahlawat KS, Khatkar BS. Trans fats-sources, health risks and alternative approach - A review. Journal of Food Science and Technology-Mysore. 2011;48(5):534–541. doi: 10.1007/s13197-010-0225-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dyall SC. Methodological issues and inconsistencies in the field of omega-3 fatty acids research. Prostaglandins Leukotrienes and Essential Fatty Acids. 2011;85(5):281–285. doi: 10.1016/j.plefa.2011.04.009. [DOI] [PubMed] [Google Scholar]
  12. Eder K, Ringseis R. Health aspects of oxidized dietary fats. In: Decker EA, Elias RJ, McClements DJ, editors. Oxidation in foods and beverages and antioxidant applications: Understanding mechanisms of oxidation and antioxidant activity. Vol. 1. Woodhead Pub.; 2010. pp. 143–180. [Google Scholar]
  13. Frankel EN. Lipid oxidation - Mechanisms, products and biological significance. Journal of the American Oil Chemists Society. 1984;61(12):1908–1917. [Google Scholar]
  14. Genot C, Meynier A, Riaublanc A. Lipid Oxidation in Emulsions. In: Kamal-Eldin A, editor. Lipid Oxidation Pathways. American Oil Chemists' Society; 2003. pp. 190–244. [Google Scholar]
  15. Gómez-Cortés P, Brenna JT, Sacks GL. Production of Isotopically Labeled Standards from a Uniformly Labeled Precursor for Quantitative Volatile Metabolomic Studies. Analytical Chemistry. 2012;84(12):5400–5406. doi: 10.1021/ac300933d. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Grosch W. Evaluation of the key odorants of foods by dilution experiments, aroma models and omission. Chemical Senses. 2001;26(5):533–545. doi: 10.1093/chemse/26.5.533. [DOI] [PubMed] [Google Scholar]
  17. Ho CT, Chen QY. Ho CT, Hartman TG, editors. Lipids in food flavors - An overview. Lipids in Food Flavors. 1994. pp. 2–14.
  18. Hsieh RJ, Kinsella JE. Oxidation of polyunsaturated fatty acids: mechanisms, products, and inhibition with emphasis on fish. Advances in food and nutrition research. 1989;33:233–341. doi: 10.1016/s1043-4526(08)60129-1. [DOI] [PubMed] [Google Scholar]
  19. Jacobsen C. Understanding and reducing oxidative flavour deterioration in foods. In: Decker EA, Elias RJ, McClements DJ, editors. Oxidation in foods and beverages and antioxidant applications: Understanding mechanisms of oxidation and antioxidant activity. Vol. 1. Woodhead Pub.; 2010. pp. 122–142. [Google Scholar]
  20. Jelen HH, Mildner-Szkudlarz S, Jasinska I, Wasowicz E. A headspace-SPME-MS method for monitoring rapeseed oil autoxidation. Journal of the American Oil Chemists Society. 2007;84(6):509–517. [Google Scholar]
  21. Kolanowski W, Jaworska D, Laufenberg G, Weissbrodt J. Evaluation of sensory quality of instant foods fortified with omega-3 PUFA by addition of fish oil powder. European Food Research and Technology. 2007;225(5-6):715–721. [Google Scholar]
  22. Lea CH. Methods for determining peroxide in lipids. Journal of the Science of Food and Agriculture. 1952;3(12):586–594. [Google Scholar]
  23. Liang CP, Wang MF, Simon JE, Shahidi F, Ho CT. Method Development for Monitoring Seal Blubber Oil Oxidation Based on Propanal and Malondialdehyde Formation. In: Shahidi F, Ho CT, editors. Antioxidant Measurement and Applications. Vol. 956. Amer Chemical Soc.; Washington: 2007. pp. 125–139. [Google Scholar]
  24. Mildner-Szkudlarz S, Jelen HH, Zawirska-Wojtasiak R, Wasowicz E. Application of headspace-solid phase microextraction and multivariate analysis for plant oils differentiation. Food Chemistry. 2003;83(4):515–522. [Google Scholar]
  25. Min DB, Boff JM. Chemistry and Reaction of Singlet Oxygen in Foods. Comprehensive reviews in food science and food safety. 2002;1(2):58–72. doi: 10.1111/j.1541-4337.2002.tb00007.x. [DOI] [PubMed] [Google Scholar]
  26. Moffat CF, McGill AS. Variability of the composition of fish oils: significance for the diet. Proceedings of the Nutrition Society. 1993;52(03):441–456. doi: 10.1079/pns19930085. [DOI] [PubMed] [Google Scholar]
  27. Novo M, Quirós M, Morales P, González R. Wine Technology. In: Sinha NK, Sidhu JS, Barta J, Wu JSB, Cano MP, editors. Handbook of Fruits and Fruit Processing. Wiley-Blackwell; 2012. pp. 461–488. [Google Scholar]
  28. Petersen KD, Kleeberg KK, Jahreis G, Fritsche J. Assessment of the oxidative stability of conventional and high-oleic sunflower oil by means of solid-phase microextraction-gas chromatography. International Journal of Food Sciences and Nutrition. 2012;63(2):160–169. doi: 10.3109/09637486.2011.609158. [DOI] [PubMed] [Google Scholar]
  29. Porter NA, Caldwell SE, Mills KA. Mechanisms of free-radical oxidation of unsaturated lipids. Lipids. 1995;30(4):277–290. doi: 10.1007/BF02536034. [DOI] [PubMed] [Google Scholar]
  30. Rapp A, Mandery H. Wine Aroma. Experientia. 1986;42(8):873–884. [Google Scholar]
  31. Shahidi F, Pegg RB. Ho CT, Hartman TG, editors. Hexanal as an indicator of the flavor deterioration of meat and meat-products. Lipids in Food Flavors. 1994;558:256–279. [Google Scholar]
  32. Tompkins C, Perkins E. Frying performance of low-linolenic acid soybean oil. Journal of the American Oil Chemists' Society. 2000;77(3):223–229. [Google Scholar]
  33. Triqui R, Guth H. Determination of Potent Odorants in Ripened Anchovy (Engraulis encrasicholus L.) by Aroma Extract Dilution Analysis and by Gas Chromatography-Olfactometry of Headspace Samples. In: Shahidi F, Cadwallader K, editors. Flavor and Lipid Chemistry of Seafoods. Vol. 674. American Chemical Society; 1997. pp. 31–38. [Google Scholar]
  34. Urbach G, Gordon MH. Flavours Derived from Fats. In: Moran DPJ, Rajah KK, editors. Fats in Food Products. Springer; US: 1995. pp. 347–405. [Google Scholar]
  35. Uriarte PS, Goicoechea E, Guillen MD. Volatile components of several virgin and refined oils differing in their botanical origin. Journal of the Science of Food and Agriculture. 2011;91(10):1871–1884. doi: 10.1002/jsfa.4400. [DOI] [PubMed] [Google Scholar]
  36. Vazquez-Cruz MA, Jimenez-Garcia SN, Torres-Pacheco I, Guzman-Maldonado SH, Guevara-Gonzalez RG, Miranda-Lopez R. Effect of maturity stage and storage on flavor compounds and sensory description of berrycactus ( Myrtillocactus geometrizans). Journal of Food Science. 2012;77(4):C366–C373. doi: 10.1111/j.1750-3841.2012.02621.x. [DOI] [PubMed] [Google Scholar]
  37. Velasco J, Dobarganes C, Márquez-Ruiz G. Oxidative rancidity in foods and food quality. In: Skibsted LH, Risbo J, Andersen ML, editors. Chemical deterioration and physical instability of food and beverages. Woodhead Pub.; 2010. pp. 3–32. [Google Scholar]
  38. Venkateshwarlu G, Let MB, Meyer AS, Jacobsen C. Chemical and Olfactometric Characterization of Volatile Flavor Compounds in a Fish Oil Enriched Milk Emulsion. Journal of Agricultural and Food Chemistry. 2004;52(2):311–317. doi: 10.1021/jf034833v. [DOI] [PubMed] [Google Scholar]
  39. Zhou YH, Nijland M, Miller M, Ford S, Nathanielsz PW, Brenna JT. The influence of maternal early to mid-gestation nutrient restriction on long chain polyunsaturated fatty acids in fetal sheep. Lipids. 2008;43(6):525–531. doi: 10.1007/s11745-008-3186-1. [DOI] [PubMed] [Google Scholar]
  40. Zhu J, Sevilla MD. Kinetic analysis of free-radical reactions in the low-temperature autoxidation of triglycerides. The Journal of Physical Chemistry. 1990;94(4):1447–1452. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS641534-supplement.docx (171.6KB, docx)

RESOURCES