Monitoring oxidative stability and changes in key volatile compounds in edible oils during ambient storage through HS-SPME/GC–MS

ABSTRACT

Headspace solid-phase microextraction/gas chromatography–mass spectrometry (HS-SPME/GC–MS) analysis combined with ‘relative odour activity value (ROAV)’ was used to monitor changes in key volatile compounds in peanut oil, soybean oil, rapeseed oil, and linseed oil during ambient storage. Volatile composition and oxidation process were compared among edible oil samples. The differences in the volatile contents of edible oils led to their characteristic flavour. Aldehydes featured a relatively high content and low odour threshold and mainly contributed to the flavour of edible oils. The key flavour compounds included pentanal, hexanal, octanal, nonanal, trans-2-heptenal, and benzaldehyde, which are important oxidative degradation products of oleic acid and linoleic acid. The formation of key volatile oxidation compounds was affected by different oxidation processes during ambient storage. Certain aldehydes increased with oxidation level, whereas other aldehydes initially increased then decreased. Correlation analysis showed that the concentrations of several volatile compounds progressively increased during oxidation. The key volatile oxidation compounds formed during oil storage at ambient temperature are partly different from those generated at high temperatures. Volatile oxidation compounds can be a marker for monitoring the oxidation degree of edible oils during ambient storage.

KEYWORDS:

• Edible oils
• Volatile compounds
• Ambient storage
• HS-SPME/GC–MS analysis
• Relative odour activity value

Introduction

Edible oils tend to undergo oxidative rancidity during long-term ambient storage; oxidative rancidity affects the chemical, sensory, and nutritional properties of the oil and determines its shelf life.^[1] Odour is an important index used to evaluate the quality of edible oils and is related to lipid oxidation.^[2] Edible oils mainly contain fatty acids, both saturated and unsaturated, which are bound to glycerol as triacylglycerols. Unsaturated fatty acids contain allyl groups, which are highly susceptible to free-radical reactions, and undergo decomposition in the presence of oxygen even at low temperatures.^[3] Autoxidation of lipids in air is a non-enzymatic autocatalytic process caused by free-radical chain reactions.^[4] In autoxidation, the reaction of unsaturated fatty acids and triglycerides with oxygen forms hydroperoxides as primary products, which are eventually broken down into volatile compounds.^[5,6] These compounds, also called secondary products, include aldehydes, ketones, alcohols, acids, hydrocarbons, furanones, and lactones.^[2,7] The formation of hydroperoxides can be evaluated through classical analytical methods using peroxide value and UV absorbance at 232 nm. Volatile compounds in oils can be decomposed using monohydroperoxides as precursors through different mechanisms; this process generates a complex mixture of volatile and nonvolatile substances (conjugated compounds, high-molecular-weight aldehydes, and others). These substances are traditionally measured through different analytical methods, such as determination of thiobarbituric acid number, p-anisidine value, UV absorbance at 268 nm, conjugated diene value, and conjugated triene value.^[8,9] However, these methods cannot characterise the oxidative state of oils. To obtain comprehensive information regarding the oxidative state of oils, scholars employed solid-phase microextraction (SPME) for determining the composition and the quantity of volatile materials.^[10]

Analytical techniques using headspace are important for determining volatile compounds. Headspace solid-phase microextraction (HS-SPME) is a fast, sensitive, solventless, and economical method used to prepare samples for gas chromatography analysis.^[11] This technique is widely used to determine volatile compounds in fresh and oxidised edible oil samples.^[12–16] Previous studies used HS-SPME under accelerated oxidation instead of storage conditions.^[3,17,18] The nature and proportions of compounds oxidised during oil storage at ambient temperature significantly vary from those generated at higher temperatures.^[19,20] In contrast to that under accelerated oxidative conditions, the shelf life or storage of edible oils at ambient temperature has been less investigated because of the slow process.^[1,21] Goicoechea and Guillén applied SPME followed by gas chromatography–mass spectrometry (GC–MS) to analyse the headspace composition of sunflower and corn oil samples stored at ambient temperature; the analysis was performed in closed receptacles by using limited amount of air for different durations.^[20,21] To the best of our knowledge, limited information is available regarding changes in volatile compounds during ambient storage. This work aims to propose a method based on HS-SPME coupled with GC–MS for rapid analysis and characterisation of volatile compounds produced through lipid oxidation of different edible oils.

Materials and methods

Materials and reagents

Peanut oil, soybean oil, rapeseed oil, and linseed oils without any antioxidants and with different proportions of oleic, linoleic, and linolenic acids were purchased from local markets in Yangling, Shaanxi, China. Isopropyl alcohol, toluene, potassium hydroxide, acetic acid, chloroform, potassium iodide, and sodium thiosulfate were obtained from Tianjin Chemical Company Ltd. All reagents and chemicals used were of analytical grade.

Edible oil sample preparation

Peanut, soybean, rapeseed, and linseed oil samples were pre-treated using 100–200 mesh active silica gel columns to remove hydroperoxides, free fatty acids (FFAs), and other oxygenated compounds. The prepared samples were used for measurement of peroxide value [PV, AOCS Official Method Cd 8b-90],^[22] acid value [AV, AOCS Official Method Cd 3a-63],^[23] iodine value [IV, AOCS Official Method Cd 1d-92],^[24] and saponification value [SV, AOCS Official Method TI 1a-64]^[25] to obtain samples with known PVs, AVs, IVs, and SVs. According to IUPAC standard methods 2.301 and 2.302, the fatty acid composition was analysed by gas chromatography using Agilent 7820, equipped with a FID and a TRACE™ TR-FAME Column (60 m × 0.25 μm, Thermo Fisher). The injection block temperature was set at 250°C. The oven temperature was kept at 60°C for 3 min, then programmed as follows, 60–175°C at 5°C /min, 175°C for 15 min, 175–220°C at 2°C /min, and finally 220°C for 10 min. The carrier gas was nitrogen with a flow rate of 25 mL/min, and the split rate was 1/100.

Identification was performed by GC retention times of standard compounds. Quantification was performed by using internal standard calibration, using tridecanoic acid methyl ester as internal standard (700 mg/Kg⁻¹).

Oxidation treatment

Each of the four oil samples (30 mL) was deposited into a 100 mL beaker to monitor changes in volatile compounds in edible oil. The prepared samples were exposed to visible light for approximately 12 h per day at ambient temperature (23 ± 2°C; 60 ± 3% relative humidity). At fixed intervals (24 h), PV was measured from the prepared samples through the standard method. Four kinds of oil samples with PVs of 3, 20, 40, 60, 80, and 100 meq/kg were obtained. All of the prepared samples were sealed, wrapped with aluminium foil, and kept at −18°C in a refrigerator until further GC–MS analysis.

Determination of volatile compounds

SPME sampling: Samples (5 ± 0.01 g) were placed in a 20 mL vial sealed with an aluminium crimp cap equipped with a needle-pierceable polytetrafluroethylene/ silicone septum. The divinylbenzene/carboxene/polydimethylsiloxane (DVB/CAR/PDMS, 50 μm/30 μm coating, 1 cm length) fibre (Supelco Inc., Bellefonte, PA, USA) was used for headspace sampling. The sample vials were equilibrated in the incubator at 50°C for 30 min under agitation at 500 rpm. The vial was inserted with fibre previously conditioned by heating in a gas chromatograph injection port at 250°C for 60 min. The fibre was kept 1 cm above the samples so that it is exposed to headspace of the vial. After 30 min of extraction at 50°C, the fibre was immediately desorbed into the GC–MS injection port at 250°C for 3 min (splitless mode).

GC–MS determination: Volatile compounds were analysed on GC–MS apparatus (Shimadzu-QP2010, Kyoto, Japan) operating in electron ionization mode (EI, 70 eV). Ion source and GC–MS transfer line temperature were set to 230°C and 250°C, respectively. The DB-17MS column (60 m × 0.25 mm × 0.25μm) was programmed from 40°C (with holding for 3 min) to 120°C at 4°C/min⁻¹ and to 240°C at 6°C/min⁻¹ (with holding for 9 min). Helium was used as carrier gas (1.0 mL/min⁻¹).

Volatile compounds in edible oil samples were identified in full scan mode (m/z 30–550) by NIST mass spectral library (NIST14, version 2.2, National Institute of Standards and Technology, Gaithersburg, MD, USA). Volatile compounds were tentatively identified using the GC–MS spectra. Compounds with ≤80% similarity to the NIST library were not considered. In addition, identification was performed by matching their Kovats indices (KI) determined relative to the retention time of a series of n-alkanes (C8–C20) with linear interpolation, with those of authentic compounds or literature data.^[26] The chromatographic responses of the detected volatile compounds (peak area counts) were monitored and compared among the studied samples. The relative content of the components were determined with peak area normalisation method.

Relative odour activity value determination: Relative odour activity value (ROAV) was determined to measure the contribution of each volatile compound towards the entire aroma profile. ROAV was calculated using the following equation^[27]:

OAV_i was determined according to OAV_i = C_i /OT_i, where C_i is the concentration of compound i in the sample, and OT_i is its odour detection threshold concentration measured in air found in literature. OAV_max was determined as the maximum OAV_i of all compounds in the sample. Table 4 shows the OT values selected for ROAV determination. Compounds with ROAV equal to or higher than 1 significantly contribute to aroma and are considered key volatile components. Compounds with 0.1 ≤ ROAV <1 also contribute to the whole flavour.

Table 1. Quality indices and fatty acid compositions of edible oils (mean ±SD, n = 3).

Oil sample	PV(meq/kg)	AV(mg/g)	SV (mg/g)	IV (g/100g)	Fatty acid composition (relative %)
					Palmitic	Stearic	Oleic	Linoleic	Linolenic
Rapeseed oil	-	-	175.12 ± 0.1	113.71 ± 0.01	4.34 ± 0.00	2.03 ± 0.00	67.18 ± 0.01	20.56 ± 0.01	5.90 ± 0.00
Peanut oil	-	-	187.03 ± 0.1	105.51 ± 0.05	11.08 ± 0.01	3.95 ± 0.01	42.57 ± 0.01	37.16 ± 0.01	0.00 ± 0.00
Soybean oil	-	-	191.22 ± 0.3	137.53 ± 0.01	10.63 ± 0.01	4.08 ± 0.00	24.47 ± 0.01	54.92 ± 0.01	5.90 ± 0.01
Linseed oil	-	-	195.02 ± 0.2	178.38 ± 0.05	5.08 ± 0.01	3.20 ± 0.00	30.98 ± 0.01	16.87 ± 0.01	43.88 ± 0.01

“ –” means less than detection limit

Table 2. Comparison of the time when edible oils reached different PVs (mean ± SD, n = 3).

Oil sample	Oxidation time (days)
	3 meq/kg	20 meq/kg	40 meq/kg	60(meq/kg	80 meq/kg	100 meq/kg
Rapeseed oil	2 ± 0.20	45.75 ± 0.20	90.5 ± 0.20	125 ± 0.20	150.5 ± 0.10	170.5 ± 0.30
Peanut oil	1.5 ± 0.10	25.5 ± 0.10	58 ± 0.10	85 ± 0.10	105.5 ± 0.20	119.25 ± 0.20
Soybean oil	1 ± 0.10	17.5 ± 0.10	36 ± 0.10	53 ± 0.10	67 ± 0.10	78.5 ± 0.10
Linseed oil	0.5 ± 0. 05	7.50 ± 0.10	15 ± 0.10	21 ± 0.10	26 ± 0.10	30.25 ± 0.10

Table 3. Changes in key volatile compounds in rapeseed oil during ambient storage (mean ± SD, n = 3).

Volatile compounds	Relative peak area (%)
	3 meq/kg	20 meq/kg	40 meq/kg	60 meq/kg	80 meq/kg	100 meq/kg
Pentanal	0.15 ± 0.008	0.28 ± 0.01	0.63 ± 0.03	0.93 ± 0.04	1.74 ± 0.07	1.07 ± 0.04
Hexanal	1.34 ± 0.06	4.4 ± 0.2	7.07 ± 0.3	7.33 ± 0.3	7.63 ± 0.3	9.74 ± 0.4
Heptanal	0.1 ± 0.006	0.24 ± 0.012	0.45 ± 0.02	0.69 ± 0.03	0.91 ± 0.04	1.11 ± 0.05
Octanal	0.97 ± 0.04	1.26 ± 0.05	2.44 ± 0.1	1.99 ± 0.05	1.65 ± 0.07	1.41 ± 0.06
Nonanal	3.1 ± 0.1	6.55 ± 0.3	11.94 ± 0.5	6.5 ± 0.3	2.32 ± 0.1	3.02 ± 0.2
Decanal	0.19 ± 0.006	0.56 ± 0.02	1.31 ± 0.05	1.06 ± 0.05	0.65 ± 0.03	0.3 ± 0.01
2-Butenal	0.1 ± 0.004	0.32 ± 0.015	0.56 ± 0.02	0.78 ± 0.03	1.02 ± 0.04	2.48 ± 0.1
(E)-2-pentenal	0.5 ± 0.002	0.86 ± 0.04	1.35 ± 0.07	2.43 ± 0.11	3.43 ± 0.13	1.21 ± 0.05
(E)-2-heptenal	0.56 ± 0.03	0.87 ± 0.04	1.45 ± 0.07	2.06 ± 0.1	2.34 ± 0.11	2.41 ± 0.11
E-2-decenal	0.02 ± 0.001	0.04 ± 0.002	0.08 ± 0.004	0.11 ± 0.005	0.13 ± 0.005	0.15 ± 0.006
(E,E)-2,4 -Heptadienal	0.1 ± 0.005	0.15 ± 0.006	0.42 ± 0.02	0.51 ± 0.02	0.71 ± 0.034	0.84 ± 0.04
Benzaldehyde	0.92 ± 0.045	1.12 ± 0.06	1.12 ± 0.06	0.71 ± 0.034	nd	nd
1-Pentanol	nd	nd	0.25 ± 0.01	0.33 ± 0.012	0.38 ± 0.015	0.41 ± 0.02
1-Octanol	0.1 ± 0.005	0.13 ± 0.006	0.15 ± 0.006	0.17 ± 0.007	0.19 ± 0.008	0.21 ± 0.01
Acetic acid	nd	2.3 ± 0.1	3.26 ± 0.14	3.01 ± 0.14	2.73 ± 0.12	2.62 ± 0.13
Hexanoic acid	0.16 ± 0.008	2.62 ± 0.12	1.67 ± 0.08	0.78 ± 0.035	0.32 ± 0.015	0.28 ± 0.014
Nonanoic acid	0.29 ± 0.014	0.48 ± 0.02	1.1 ± 0.05	0.95 ± 0.04	0.87 ± 0.04	0.86 ± 0.04
Dodecane	0.88 ± 0.04	2.13 ± 0.1	3.13 ± 0.12	1.12 ± 0.06	1.09 ± 0.05	1.41 ± 0.07
D-Limonene	0.72 ± 0.035	1.68 ± 0.08	1.17 ± 0.06	0.64 ± 0.03	1.48 ± 0.07	1.6 ± 0.08

nd, non-detectable

Table 4. Odour contribution of volatile components of four edible oil samples.

Key compounds	Odour threshold (ng/L)	ROAV				Sensory description
		Rapeseed oil	Peanut oil	Soybean oil	Linseed oil
Pentanal	38	1.53	1.13	11.54	1.96	Pungent, almond, malt
Hexanal	51	10.17	100	11.39	23.73	Grass, green, rancid, tallowy, fat
Heptanal	46	0.84	0.75	3.34	nd	Rancid, fatty, citrus
Octanal	9.3	40.37	26.73	41.24	42.82	Fatty, soapy, green, oily, fresh
Nonanal	12	100	85.48	100	100	Citrus, green, fatty, oily, soapy, tallowy, fruity
Decanal	4.3	17.1	nd	19.11	40.75	Floral, orange peel, soapy, tallowy, green, fresh
2-Butenal	350	0.11	nd	nd	0.61	Pungent
(E)-2-pentenal	1500	0.13	nd	nd	0.14	Strawberry, fruity, tomato, pungent, apple
(E)-2-heptenal	88	2.46	1.07	3.61	1.21	Fatty, soapy, almond, oily
(E)-2-octenal	20	nd	2.14	4.66	nd	Green, nutty, fatty, oily
E-2-decenal	3.2	2.42	2.68	10.27	nd	Tallowy, orange-like
(E,E)-2,4-heptadienal	38	1.02	nd	nd	2.80	Fatty
Benzaldehyde	186	1.91	3.82	1.80	2.65	Almond, burnt sugar
1-Pentanol	3590	nd	0.016	0.11	0.05	Balsamic
1-Hexanol	360	nd	nd	nd	0.5	Green, floral
1-Octanol	73	0.53	0.43	0.52	nd	Chemical, metallic, burnt
Acetic acid	1384	nd	0.12	nd	0.26	Sour
Hexanoic acid	3000	0.02	0.02	0.06	0.11	Sour, fatty, sweaty, cheese
Nonanoic acid	12	9.35	nd	10.96	nd	Green, fatty
Dodecane	5300	0.064	0.14	0.026	0.15	Gasoline-like
D-Limonene	718	0.38	0.57	0.2	1.06	Lemon, orange
2-Pentylfuran	181	nd	1.07	2.75	1.47	Butter, green beans

Mean values of odour thresholds in air^[36,39] sensory description referred to literature^[40,41]; nd, non-detectable. Only flavour components with ROAV ≥ 0.1 in one of the four edible oil samples are presented.

Statistical analysis

All analyses were performed in triplicate, and the mean values and standard deviations (SDs) were obtained to express the results. Pearson’s correlation analysis was performed using SPSS version 20.0 (SPSS/IBM, Armonk, NY, USA).

Results

Quality indices and fatty acid compositions of different edible oils

Table 1 shows the quality indices and fatty acid compositions of the prepared peanut, soybean, rapeseed, and linseed oil samples. The PVs and AVs of all the processed edible oils using the silica gel column cannot be detected, indicating that hydroperoxides and FFAs were nearly completely removed. Linseed oil contains large amounts of linolenic acid (43.88%), SV (195.02 mg/g), and IV (178.38 g/100 g). Soybean oil possesses SV of 191.22 mg/g, IV of 137.53 g/100 g, high unsaturation degree, and large amounts of linoleic acid (54.92%). Peanut oil contains oleic acid (42.57%) and linoleic acid (37.16%). Peanut oil is one of the most stable edible oils to oxidation partly due to its fatty acid composition, which is low in 18:3ω3.^[28] Rapeseed oil contains a large amount of oleic acid (67.18%). The unsaturation degree of fatty acids affects the stability of oils; that is, fatty acids with many double bonds can be easily attacked by free radicals.^[29] For example, the autoxidation rate of oleic, linoleic, and linolenic methyl esters is 1:40:100.^[30,31] During oxidation, different types of unsaturated fatty acids produced different volatile components, which contribute to the unique odour of different edible oil samples.

Linoleic acid is an essential precursor of volatile compounds in edible oils and can be easily oxidised to produce hexanal, pentanal, heptanal, and trans-2-heptenal.^[17,32,33] Oleic acid is also an important oxidation precursor of volatile compounds, particularly nonanal and octanal, which are derived from the oxidative degradation of oleic acid.^[17,33] The oxidative degradation rate of linolenic acid is faster than that of linoleic acid and oleic acid, and linolenic acid is an important source of trans, trans-2,4-heptadienal, and trans-2-hexenal in the oxidation products of edible oils.^[17,33] The four selected edible oil samples show strong representation of oil types.

Comparison of oxidation processes among edible oil samples

Table 2 shows the comparison of the time when PV reached 3, 20, 40, 60, 80, and 100 meq/kg in rapeseed, soybean, peanut, and linseed oil samples oxidised during ambient storage, respectively. As shown in Table 2, the time when edible oil reached different PVs and the time when they reached the same PV significantly varied during ambient storage. The different durations when PV reached 3, 20, 40, 60, 80, and 100 meq/kg primarily indicate the degree of oxidative stability (rapeseed oil > peanut oil > soybean oil > linseed oil). Linseed oil can be easily oxidised because it is rich in unsaturated fatty acids (linolenic acid). Rapeseed oil exhibits improved oxidative stability than soybean and peanut oils because of its higher oleic acid and lower linolenic acid contents.

Changes in volatile compounds in edible oils during ambient storage

Table 3 and Tables S1−S3 show the relative peak area of volatile compounds in the prepared rapeseed, peanut, soybean, and linseed oil samples at PVs = 3, 20, 40, 60, 80, and 100 meq/kg. Table 3 and Tables S1–S3 show the changes in volatile compounds in the four edible oil samples stored at ambient temperature. Aldehydes are the most abundant volatile compounds associated with stored/rancid oils and are considered the main secondary oxidation compounds formed during lipid oxidation. The samples predominantly contain pentanal, hexanal, heptanal, octanal, nonanal, and decanal.

The autoxidation of linoleic acid involves hydrogen abstraction on the doubly reactive allylic C-11 and formation of a pentadienyl radical. The intermediate radical reacts with oxygen to produce a mixture of conjugated 9- and 13-diene hydroperoxides.^[33] The cleavage of 13-hydroperoxide produces hexanal and pentanal, and the breakdown of linoleic acid 11-hydroperoxide generates heptanal. The linoleic acid content of these edible oils follows the trend of soybean oil > peanut oil > rapeseed oil > linseed oil. Soybean, peanut, rapeseed, and linseed oil samples contain high hexanal amounts of 2.05%, 2.00%, 1.74%, and 0.98%, respectively, consistent with the order of linoleic acid content. Furthermore, pentanal and heptanal exhibit the same trend. Hence, oils with high linoleic acid content also possess high amounts of the three saturated linear aldehydes.

For oleic acid, the hydrogen abstraction on C-8 and C-11 produces two allylic radicals, which react with oxygen to produce a mixture of 8-, 9-, 10-, and 11-allylic hydroperoxides.^[33] Cleavage of 8-hydroperoxide, 10-hydroperoxide, and 11-hydroperoxide produces decanal, nonanal, and octanal, respectively. The oleic acid content of the samples follows the order of rapeseed oil > peanut oil > linseed oil > soybean oil; these oil samples also contain high octanal amounts of 2.44%, 2.43%, 1.39%, and 1.38%, respectively. Linseed and soybean oils contain similar levels of oleic acid; hence, their octanal content is also the same.

The relative contents of hexanal and heptanal in rapeseed oil increased with oxidation level; meanwhile, the relative contents of pentanal, octanal, nonanal, and decanal initially increased then decreased. In the later stage of oxidation, the contents of these aldehydes in rapeseed oil decreased probably because they were converted into other oxidation compounds. For peanut oil, the amounts of pentanal, hexanal, heptanal, and nonanal were low when PV reached 3 meq/kg but increased when PV reached about 100 meq/kg. The highest amount of octanal was detected when PV reached 40 meq/kg. For soybean oil, the relative contents of hexanal, heptanal, and nonanal increased during oxidation. The highest amounts of pentanal and octanal were detected when PV reached 60 meq/kg. Decanal reached the relatively highest at 20 meq/kg but was not detected with further increase in PV. For linseed oil, the contents of heptanal and nonanal contents increased, and those of the four other saturated linear aldehydes increased to the highest value then decreased.

Unsaturated monoaldehydes such as trans-2-pentenal, trans-2-heptenal, trans-2-octenal, and trans-2-decenal were also detected in all of the four edible oil samples. 2-Butenal was only detected in rapeseed and linseed oil samples. trans-2-Heptenal, which was formed by decomposition of linoleic acid 12-hydroperoxide, was found at high levels in soybean oil (2.91%) and at low levels in linseed oil (0.97%).^[33] The relative content of trans-2-heptenal increased in rapeseed oil, soybean oil, and rapeseed oil. trans-2-Octenal was formed from decomposition of linoleic acid 11-hydroperoxide. The content of octenal increased with oxidation and reached 2.49% and 2.48% in soybean and peanut oils, respectively. However, octenal was not found in the two other oil samples.

Only (E,E)-2,4-heptadienal, a dienal alkenal, was detected in the tested edible oil samples. (E,E)-2,4-Heptadienal was formed by decomposition of linolenic acid 12-hydroperoxid. The content of (E,E)-2,4-heptadienal increased with oxidation and reached 2.79% and 0.84% in linseed and rapeseed oils, respectively. However, this was not found in peanut oil and only detected in low amount in soybean oil when PV reached 100 meq/kg.

Edible oil samples stored at ambient temperature contain not only saturated and unsaturated aliphatic compounds but also aromatic aldehydes, such as benzaldehyde. Benzaldehyde, which is derived from Strecker degradation of aromatic amino acids, provides almond-like aroma. The highest content of benzaldehyde reached 7.67% in linseed oil at PV = 40 meq/kg and 1.12% and 1.89% in rapeseed and soybean oils, respectively, at PV = 20 meq/kg. The benzaldehyde content in peanut oil showed a decreasing trend with oxidation from the highest value of 2.49%.

Alcohols are generally derived from degradation of the secondary hydroperoxide of fatty acids or reduction of carbonyl compounds.^[33] 1-Pentanol is the most abundant alcohol and is derived from linoleic groups. 1-Pentanol was detected at high levels in soybean and peanut oils. Meanwhile, the amounts of 1-pentanol and 1-octanol increased with oxidation in rapeseed, peanut, and soybean oil samples.

Saturated fatty acids such as acetic acid, hexanoic acid, and nonanoic acid were detected in the edible oil samples. These acids were produced through oxidation of their corresponding aldehydes.^[20] Nonanoic acid presents a low odour threshold and contributes to the whole flavour profile of peanut and linseed oils. The proportions of nonanoic acid increased with increasing oxidation degree of these oil samples.

Alkylfuran derivatives were also detected in the headspace of the edible oil samples. 2-Pentylfuran was the main alkylfuran detected in the samples. The conjugated diene radical, which was generated from cleavage of the 9-hydroxy radical of linoleate, may react with oxygen to produce vinyl hydroperoxide, which will then undergo cyclisation via the alkoxy radical to form 2-pentylfuran.^[34] High amount of 2-pentylfuran was detected in edible oil samples with relatively high linoleic acid content and contributed to butter and green bean aroma. As shown in Table 3 and Tables S1–S3, 2-pentylfuran was present at relatively high amounts of 1.15% and 1.12% in soybean and peanut oil samples, respectively, at about 60 meq/kg. Dodecane contributed to the whole flavour in peanut and linseed oils and showed an increasing trend in these samples. D-Limonene was the key volatile compound detected in linseed oil, with a relative content of 1.99% at PV = 80 meq/kg; this compound contributed to the whole flavour in linseed oil. The relative content of D-limonene initially increased and then decreased in all oil samples.

Volatile composition was compared in one kind of edible oil at various PVs. The headspace volatile composition of the oil samples varied, ranging from the characteristics of non-oxidised oils to those of oils with high oxidation level. The composition and content of volatile compounds also varied in different oil samples.

Key odour compounds of edible oils

Volatile compounds exhibit different odour thresholds, leading to different levels of sensitivity in humans; as such, the relative content of these compounds cannot reflect their true contribution to the whole aroma profile. Therefore, ROAV was used to detect the contribution of volatile compounds to the entire aroma profile. Table 4 shows the key odour compounds selected from different edible oil samples at PV = 3 meq/kg.

As shown in Table 4, nonanal exhibits low odour threshold and high relative content in rapeseed, soybean, and linseed oils and thus greatly contributes to the whole flavour. The ROAV of nonanal was defined as 100, and those of the other volatile flavour components were calculated in comparison with nonanal. In peanut oil, hexanal presents low odour threshold and high relative content; therefore, the ROAV of hexanal was defined as 100. Table 4 shows the ROAVs of volatile flavour components in the four edible oil samples and their sensory descriptions.

The different compositions of volatile compounds in the four edible oil samples could be due to their different fatty acid compositions, which lead to their characteristic flavour. Rapeseed oil contains the following 10 key volatile components (ROAV ≥ 1): pentanal, hexanal, octanal, nonanal, decanal, (E)-2-heptenal, E-2-decenal, (E,E)-2,4-heptadienal, benzaldehyde, and nonanoic acid. Moreover, heptanal, 2-butenal, (E)-2-pentenal, 1-octanol, and D-limonene contribute to the whole flavour (0.1 ≤ ROAV < 1). Peanut oil was found to contain nine key flavour components (ROAV≥1): pentanal, hexanal, octanal, nonanal, (E)-2-heptenal, (E)-2-octenal, E-2-decenal, benzaldehyde, 2-pentylfuran and five main flavour components (0.1≤ ROAV<1): heptanal, 1-octanol, acetic acid, dodecane and D-limonene. Soybean oil was found to contain 12 key flavour components (ROAV≥1): pentanal, hexanal, heptanal, octanal, nonanal, decanal, (E)-2-heptenal, (E)-2-octenal, E-2-decenal, benzaldehyde, nonanoic acid, 2-pentylfuran and three main flavour components (0.1≤ ROAV<1): 1-pentanol, 1-octanol, and D-limonene. Linseed oil was found to contain 10 key flavour components (ROAV≥1): pentanal, hexanal, octanal, nonanal, decanal, (E)-2-heptenal, (E,E)-2,4-heptadienal, benzaldehyde, D-limonene, 2-pentylfuran and six main flavour components (0.1≤ ROAV<1): 2-butenal, (E)-2-pentenal, 1-hexanol, acetic acid, hexanoic acid, and dodecane. As shown in Table 4, aldehydes present high relative content and low odour threshold in all of the four edible oil samples, thereby significantly contributing to their whole flavour profile. Twenty-two typical key volatile compounds were identified by HS-SPME/GC–MS analysis in the four oil samples. The similar key flavour compounds among the samples include pentanal, hexanal, octanal, nonanal, trans-2-heptenal, and benzaldehyde. These volatile components are important oxidative degradation products of oleic acid and linoleic acid.^[33]

Key volatile compounds associated with PV

As shown in Table 5, the amounts of volatile oxidation compounds including hexanal, heptanal, (E)-2-heptenal, E-2-decenal, E,E-2,4-heptadienal, 1-pentanol, and 1-octanol increased linearly with PV for rapeseed oil (r > 0.917, P < 0.01; Table 5). This relation can be used to identify the different oxidation degrees of rapeseed oil samples. For rapeseed, peanut, and soybean oil samples, the amounts of volatile oxidation compounds including hexanal, heptanal, E-2-decenal, 1-pentanol, and 1-octanol increased linearly with PV. Similarly, (E)-2-pentenal, (E)-2-heptenal, (E,E)-2,4-heptadienal,1-pentanol, and nonanoic acid exhibited close relationship to PV in linseed oil. A combined assessment of these five substances can provide a guide for monitoring linseed oil. The correlation analysis indicated that several volatile oxidation compounds are significantly related to the oxidation degree of edible oil. The concentrations of these volatile compounds were high and progressively increased in the headspace of the samples during oxidation.

Table 5. Correlation analysis between PV and key volatile compounds.

Key volatile compounds	Correlation coefficient
	Rapeseed oil	Peanut oil	Soybean oil	Linseed oil
Pentanal	0.852*	0.970**	0.806	0.705
Hexanal	0.935**	0.969**	0.995**	0.688
Heptanal	0.935**	0.999**	0.993**	0.662
Octanal	0.281	0.362	0.472	0.567
Nonanal	−0.291	0.969**	0.981**	0.681
Decanal	0.053	0.000	−0.738	0.321
2-Butenal	0.898*	0.000	0.000	0.906*
(E)-2-Pentenal	0.598	0.000	0.000	0.972**
(E)-2-Heptenal	0.939**	0.557	0.985**	0.964**
(E)-2-Octenal	0.000	0.991**	0.997**	0.000
E-2-Decenal	0.992**	0.990**	0.990**	0.000
(E,E)-2,4–Heptadienal	0.991**	0.000	0.662	0.990**
Benzaldehyde	−0.867*	−0.901*	−0.496	−0.228
1-Pentanol	0.945**	0.914*	0.882*	0.997**
1-Hexanol	0.000	−0.007	0.000	0.403
1-Octanol	0.995**	0.982**	0.996**	0.000
Acetic acid	0.620	0.749	0.623	0.767
Hexanoic acid	−0.414	0.965**	0.962**	0.810
Nonanoic acid	0.662	0.984**	0.752	0.967**
Dodecane	−0.173	−0.817*	−0.524	0.029
D-Limonene	0.379	−0.797	−0.176	−0.396
2-Pentylfuran	0.000	−0.153	−0.493	−0.730

** Correlation is significant at 0.01 level; * Correlation is significant at 0.05 level.

Discussion

A PV of 20 meq/kg is of particular interest because oils with PV higher than this are characterised as oxidised oil and are thus inedible. A PV of 100 meq/kg is important to evaluate the oxidative stability of edible oil by active oxygen method. Thus, these PVs were selected to represent different oxidation degrees in edible oils. In this study, HS-SPME/GC–MS and correlation analyses were used to determine whether volatile oxidation compounds could be a new marker for monitoring the oxidation degree of edible oils.

Aldehydes were found at high relative content and low odour threshold in edible oils and contributed to the entire oil flavour. Pentanal, hexanal, octanal, nonanal, trans-2-heptenal, and benzaldehyde were identified as key flavour compounds in the edible oils tested. The contents of several aldehydes increased with oxidation level. By contrast, the contents of other aldehydes initially increased then decreased possibly because they were converted into other oxidation compounds such as low molecule weight compounds in the later stage of oxidation. Correlation analysis indicated that several volatile oxidation compounds are related to conventional lipid oxidation indices, such as PV; thus, these compounds can be used to identify the oxidation degree of oils. Mildner–Szkudlarz monitored volatile compounds of different oils stored at 60°C for 5 days; results indicated that hexanal, E-2-heptenal, E-2-pentenal, E,E-2,4-heptadienal, 3-ethyl-1,5-octadiene, 3-ethyl-1,5-octadiene isomer, 1-octen-3-ol, and n-decane are the representative volatile compounds in oxidised soybean oils.^[35] However, the later four compounds were not found in the present study. Jeleń reported that the most abundant volatile compounds in oxidised rapeseed oils after 12 days of storage at 60°C included hexanal, 2,4-heptadienal, E-2-heptenal, E-2-pentenal, 1-pentene-3-ol, and unidentified isomers of a compound eluting at RI 947 and 949.^[18] However, the latter three compounds were not detected in the present study. These findings indicate that the key volatile oxidation compounds formed during oil storage at ambient temperature are partly different from those generated at high temperatures.

Conclusion

HS-SPME/GC–MS was successfully applied to monitor changes in the composition of key volatile compounds in edible oils in various oxidation processes during ambient storage. Volatile oxidation compounds can be a marker for monitoring the oxidation degree of edible oils during ambient storage.

Supplemental data

Supplemental data for this article can be accessed on the publisher’s website.

Additional information

Funding

The authors would like to thank the National Natural Science Foundation of China (NO.:31671819) for the financial support.