Food Analytical Methods (2022) 15:1508–1520

https://doi.org/10.1007/s12161-021-02192-0

Monitoring the Profile of Volatile Compounds During the Storage

of Extra Virgin Olive Oils Produced in Brazil from the Koroneiki Variety
Using the HS‑SPME Technique
Nathália S. Brilhante1 · Adelia F. Faria‑Machado2 · Rosemar Antoniassi2 · Paola E. Gama · Humberto R. Bizzo1,2

Received: 21 July 2021 / Accepted: 7 December 2021 / Published online: 25 January 2022
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022

Abstract
This work aimed to evaluate changes in the profile of volatile compounds of extra virgin olive oils (EVOOs) from the states
of Minas Gerais and Rio Grande do Sul (Brazil) during 8 months of storage. The samples of Koroneiki olive oils were char-
acterized for fatty acid composition, acidity, peroxide value, and ultraviolet absorbance. Volatile compounds were identified
and estimated by headspace solid-phase microextraction (HS-SPME), gas chromatography coupled to mass spectrometry
(GC–MS), and flame ionization detection (GC-FID). The results obtained for fatty acid composition, free fatty acids, peroxide
value, and ultraviolet absorbance are characteristic of extra virgin olive oils. The main volatile compounds identified at zero
storage time were hexanal, (E)-2-hexenal, (Z)-3-hexenol, (Z)-3-hexenyl acetate, and hexyl acetate, all from the lipoxygenase
pathway (LOX) and responsible for the characteristic green odor of EVOOs. The major compound found was (E)-2-hexenal,
which also had the greatest decrease during storage. In the same period, (E)-2-heptenal, octanal, nonanal, and (E)-2-decenal,
compounds associated to undesirable sensory characteristics in olive oils, were identified mainly in the eighth month of stor-
age. Through PCA and fatty acid composition, the samples could be grouped according to their geographic origin.

Keywords EVOOs · Koroneiki · Volatile compounds · HS-SPME · GC–MS · GC-FID

Introduction compounds, the latter being responsible for the aroma of

olive oil (Mariotti and Peri 2014).
Olive oil is mainly composed of triglycerides, about 97 to The Koroneiki cultivar is native from Greece, predomi-
99% by weight. The most abundant fatty acid in the triglyc- nantly from the regions of Kalamata, Peloponnisos, Zak-
erides is oleic (C18:1), associated with the high nutritional inthos, and Crete. These regions are known for producing
value of the oil. In small amounts, linoleic (C18:2), linolenic high-quality olive oils and represent around 60% of olive
(C18:3), and some saturated fatty acids, such as palmitic oil production in Greece, the third largest producer in the
(C16:0) and stearic (C18:0), are also present. The remaining Mediterranean, with quantities approaching 400,000 tons
minor compounds are a complex mixture of polar, nonpo- per year. Its fruit is small, and the oil produced has a cer-
lar, and amphiphilic substances, such as tocopherols, phe- tain astringency and is very fruity, with herbal aromas and
nolic compounds, sterols, chlorophyll, carotenoids, terpenic notes of green apple, almonds, and fig. In addition, it is
acids, mono and diglycerides, free fatty acids, and volatile rich in polyphenols and has good stability (Anastasopoulos
et al. 2011; Dabbou et al. 2011b). The quality of olive oil is
related to several factors, such as agronomic, technological,
* Humberto R. Bizzo and storage conditions.
humberto.bizzo@embrapa.br The main producers of olive oil are countries of the Euro-
1
pean Community, situated in the Mediterranean region. For
Postgraduate Program in Food Science, Universidade
Federal do Rio de Janeiro, Rio de Janeiro, RJ 21941‑909,
the 2019/2020 crop, it was estimated that the world produc-
Brazil tion was 3.0 million tons of olive oil, 1.9 million tons com-
2
Embrapa Agroindústria de Alimentos, Avenida das
ing from Europe (IOC 2020). Spain is the largest producer,
Américas, 29501, Rio de Janeiro, RJ 23020‑470, Brazil with 1.2 million tons for the same period. On the other side

13
Food Analytical Methods (2022) 15:1508–1520 1509

of the market, Brazil is the second world’s largest importer, acid, and hexanoic acid are examples of these compounds
with 104,000 tons in the 2019/2020 crop (IOC 2020). (Morales et al. 1997, 2005).
The growing of olives in Brazil and, therefore, the extrac- Regarding Brazilian olive oils, studies were carried out
tion of oil are quite recent and small in relation to the size involving sensory analysis and identification of volatile
of its domestic market, but the quality is generally high and compounds from different cultivars produced in the states
some of the oils have received awards in international quality of Minas Gerais, Rio Grande do Sul, and Paraná (Zago
competitions (NYIOOC 2020). Production is located mainly et al. 2019; da Costa et al. 2020). However, data on Bra-
in Rio Grande do Sul and Minas Gerais states, in South and zilian olive oils and their volatile profiles are still scarce
Southeast Brazil, respectively. Internal production for the and, to date, no studies have been found with olive oils
2020 crop was only 240 tons (IBRAOLIVA 2020), a very produced in Brazil regarding composition changes in the
small fraction (0.2%) of the total imported. volatiles during storage.
The parameters of identity and quality of olive oil are Thus, the present study had four goals: first, to charac-
defined by instructions from the Ministry of Agriculture, terize the extra virgin olive oils from the cultivar Koro-
Livestock and Supply—MAPA, which in turn follows neiki produced in Rio Grande do Sul and Minas Gerais
European Union regulations (The European Commission regarding the parameters of identity and quality required
2016; Brasil 2018). by the legislation; second, determine the profile of vola-
Extra virgin olive oil (EVOO) has a complex aroma. tile compounds in fresh oils; third, check for changes in
More than one hundred volatile substances have already the profile of the volatile compounds in the samples over
been identified by gas chromatography and mass spectrom- 8 months of storage; Finally, evaluate if there are differ-
etry, including aldehydes, alcohols, esters, hydrocarbons, ences between the volatiles regarding the region the oils
ketones, and furans. The aroma is related to the genetics of are produced in Brazil.
the olive tree (cultivar/variety), the soil, and the climate from
which it is grown and the conditions of the oil extraction
process (Mariotti & Peri 2014).
This unique aroma is attributed to the volatile compounds Materials and Methods
produced from the oxidation of linoleic and linolenic fatty
acids. Oxidation occurs through a series of enzymatic chain Chemicals
reactions, known as the lipoxygenase pathway (LOX), and
occurs throughout the senescence of the fruit, processing, The reagents and solvents used such as methanol, ammo-
and storage of the oil. Alcohols, aldehydes, and esters with nium chloride, sulfuric acid, ethanol, sodium hydroxide,
five and six carbons, especially those of six carbons with potassium iodide, sodium thiosulfate, phenolphthalein,
unsaturated and saturated linear chains, make up the main starch, chloroform, and acetic acid were of analytical
fraction of volatiles in olive oils. These compounds give the grade and purchased from Sigma-Aldrich (St. Louis,
food a green odor (Olías et al. 1993; Kotsiou and Tasioula- Missouri, USA), Merck (Darmstadt, Germany), and
Margari 2015). Tedia (Fairfield, Ohio, USA). The potassium hydro-
The moment the oil is extracted and stored, non-enzy- gen phthalate (≥ 99.95%) and potassium dichromate
matic reactions of auto-oxidation of fatty acids also begin. (≥ 99.5%) were purchased from Sigma-Aldrich. The
These reactions are influenced by storage conditions, such isooctane of spectrophotometric grade and hexane for gas
as exposure to light, availability of oxygen and temperature, chromatography analysis were purchased from Tedia. All
and by the composition of the oil itself. While linoleic and standards used in chromatographic analysis were from
linolenic acids contribute to the oxidation of EVOO, minor Sigma-Aldrich.
substances, such as tocopherols and phenolic compounds,
act as antioxidants (Velasco and Dobarganes 2002; Kotsiou
and Tasioula-Margari 2015). Sampling
Oxidation reactions reduce the nutritional value of
EVOO and lead to the formation of hydroperoxide deg- Nine samples of extra virgin monovarietal olive oils from
radation products, which give off-flavors to the oil. The the cultivar Koroneiki were acquired at the local market.
main defect developed during storage related to oxidation Five of them produced in Minas Gerais and four in Rio
reactions is the rancid aroma. It is felt when saturated and Grande do Sul, from the 2018 harvest. The origin of each
monounsaturated aldehydes with five to ten carbon atoms, sample is shown in Fig. 1. The samples were kept in their
dienes with six to nine carbons, and carboxylic acids are original commercial 500-mL dark green or amber glass
formed. (E)-2-heptenal, (E)-2-octenal, (E)-2-decenal, bottles of the same batch and were stored in a freezer
pentanal, heptanal, octanal, nonanal, acetic acid, butanoic (− 20 °C) until the beginning of the analyses.

13
1510 Food Analytical Methods (2022) 15:1508–1520

Fig. 1  Map of Brazil with the

geographical location of the
analyzed olive oil samples. MG,
State of Minas Gerais; RS, State
of Rio Grande do Sul

Analysis of Olive Oils Regarding Quality Parameters s Path length of the quartz cell in centimeter
ΔK = K270 − [1∕2(K274 + K266)]
The analyses of the quality parameters of EVOOs, as
established in the Brazilian legislation (Brasil 2012),
include the absorbance in UV (K270, K232 and ΔK), free
fatty acids (FFA), and peroxide value (PV), and were car- Free Fatty Acids (FFA)
ried out according to the AOCS official methods (2009),
as detailed below. Two grams of oil was weighed into an Erlenmeyer flask and
30 mL of 95% ethanol, previously neutralized with sodium
hydroxide, and 0.5 mL of alcoholic phenolphthalein were
Absorbance in UV added. To solubilize the free fatty acids, this mixture was
placed on a plate for gentle heating, without boiling. Then,
For each sample, 100 mg of the oil was weighed and titration with 0.01 N sodium hydroxide was performed until
diluted with isooctane (spectrophotometric grade), in a the solution turned slightly pink, persistent for 30 s. The
10-mL volumetric flask. sodium hydroxide solution was standardized with potas-
The UV readings were performed in an Agilent 8453 sium hydrogen phthalate, dried in an oven at 120 °C for 2 h.
UV–Visible spectrophotometer (Agilent Technologies, The procedure was performed in triplicate, and the results
Santa Clara, CA) using quartz cuvettes with optical path expressed in percentage of free fatty acids. The formula used
length of 10 mm, first with the pure solvent and then with to calculate the free fatty acid content, with % expressed in
the solution containing the samples. The reading with pure oleic acid, is presented below:
isooctane was made from 200 to 800 nm. The readings with
mL of NaOH × molarity × 28.2
the solution containing the sample were made at the wave- FFA (% of oleic acid) =
sample weight (g)
lengths of 232, 266, 270, and 274 nm. All analyses were per-
formed in triplicate. The formulas used to obtain the specific
extinction value (K) and ΔK are presented below:
Absorbance(nm)
Peroxide Value (PV)
K(nm) =
c×s
Titrations were performed with an automatic titrator Titronic
300 (SI Analytics, Germany) and a stirring plate, while
c Concentration of the solution in g/100 mL always vigorously shaking the samples. A blank test was

13
Food Analytical Methods (2022) 15:1508–1520 1511

performed before each sample. The sodium thiosulfate solu- Analysis of the volatile fraction by headspace
tion was standardized with potassium dichromate, dried in solid‑phase microextraction (HS‑SPME) followed
an oven at 120 °C for 2 h. by GC–MS and GC‑FID
Two grams of sample was weighted directly into an
iodine flask and 30 mL of glacial acetic acid-chloroform The headspace solid-phase microextraction (HS-SPME) of vola-
solution 3:2 (v/v) was added. The flask was capped and tile compounds was carried out with 1.0 g of olive oil. Samples
the mixture homogenized until complete dissolution. were weighted into a 4-mL vial, and n-tetradecane was added
Afterwards, 0.5 mL of potassium iodide solution was as internal standard (ISTD), at a concentration of 0.5 mg/g.
added. After 60 s, with occasional stirring, 30 mL of Vials were closed with a septum cap and conditioned at 40 °C
distilled water was added, and the sample was titrated for 10 min. A SPME holder with a divinylbenzene-carboxen-
with 0.01 N sodium thiosulfate solution until the appear- polydimethylsiloxane (DVB/CAR/PDMS) 50/30 µm StableFlex
ance of a yellow color. Then, 0.5 mL of starch solution fiber from Supelco (Bellefonte, PA) was used. The fiber was
was added as an indicator, turning the solution dark blue. exposed to the headspace of the sample for 40 min at the same
Titration continued until the sample became colorless. temperature. The fiber was then transferred to a gas chroma-
The analysis was performed in triplicate. The peroxide tograph injector, heated at 250 °C, for thermal desorption for
index, in milliequivalents of peroxides per kilogram of 3 min, in splitless mode (Romero et al. 2015; Faria-Machado
olive oil, was calculated using the formula shown below: et al. 2017).
( ) The volatile compounds were then analyzed by GC-FID
mEq (a − b) × sodium thiosulfate molarity × 1000
PV = in an Agilent model 7890A equipment fitted with a DB-5
kg sample weight (g)
column (30 m × 0.25 mm × 0.25 µm) and a pre-column
(1.0 m × 0.25 mm), without stationary phase. To focusing,
a Volume in milliliter of the thiosulfate solution used in the oven was cooled to subambient temperature with liquid
the sample nitrogen during the desorption period (3 min). The tempera-
b Volume in milliliter of the thiosulfate solution used in ture was raised ballistically to 40 °C, and then up to 190 °C,
the blank at 3 °C/min, and kept for 10 min at 190 °C. Hydrogen was
used as carrier gas with a flow rate of 1.5 mL/min. The
Fatty Acid Composition injector and detector temperatures were 250 °C and 280 °C,
respectively. Peak areas were normalized with the use of
Fatty acid methyl esters (FAMEs) were obtained after an internal standard and by correcting them with average
saponification of the oil with methanolic KOH, followed response factors, determined with three standards for each
by methylation using a methanol/sulfuric acid mixture. class of compounds analyzed in the samples (Brevard et al.
The resulting esters were recovered with hexane. The 2011). The chemical classes considered were as follows:
solution was then washed with distilled water and dried alcohols (1-hexanol 99%, (Z)-3-hexenol 98%, and 3-methyl-
with sodium sulfate before analysis (Antoniassi et al. 1-butanol 99%); aldehydes (hexanal 98%, (E)-2-hexenal
2018). 99.5%, and decanal 92%); esters (hexyl acetate 99%, ethyl
Gas chromatography was performed in an Agilent octanoate 98%, and ethyl isovalerate 98%); ketones (3-pen-
model 7890A gas chromatograph (Wilmington, DE) tatone 99%, 3-methyl-2-butanone 99%, and 2-pentanone
equipped with a flame ionization detector (GC-FID, sam- 99.5%); terpenes (limonene 97%, p-cymene 97%, and a
pling rate: 5 Hz) and fitted with a HP FFAP capillary mixture of (Z)-β-ocimene and (E)-β-ocimene 90%); alkanes
column (stationary phase: nitroterephthalic acid modi- (decane 99%, tetradecane 99%, and hexadecane 99%), alk-
fied polyethylene glycol, 25 m × 0.2 mm × 0.30 μm), with enes (sativene 98%, α-pinene 98%, and camphene 95%); and
an initial temperature of 150 °C for 1 min. Temperature alkadienes (terpinolene 90%, α-terpinene 99%, and limonene
was then raised from 150 to 180 °C at 30 °C/min; then, 97%); all standards were from Sigma-Aldrich. Analyses
from 180 to 200 °C at 20 °C/min; finally, from 200 to were performed in triplicate.
230 °C at 3 °C/min, and kept at 230 °C for 10 min. The To determine the linearity and the limit of quantification
split/splitless type injector was kept at 250 °C, and 1.0 (LOQ) of the method, a calibration curve was constructed
µL of a 1% sample solution was injected with a split with n-tetradecane as internal standard, and diluted in
ratio of 50:1. The detector temperature was 280 °C and refined and deodorized sunflower oil, under the same con-
the carrier gas (hydrogen) flow was 2.5 mL/min. The ditions used for the samples. Sunflower oil was chosen to
identification of the fatty acid methyl esters was carried compensate matrix effects. Eight points in the concentration
out by comparing their retention times with those from range from 0.0011 to 4.56 mg of n-tetradecane per gram of
NU-CHEK standards (Elysian, MN). oil were analyzed. All volatile data should be considered

13
1512 Food Analytical Methods (2022) 15:1508–1520

semi-quantitative, as fiber saturation can occur for the most Statistical Analysis
concentrated compounds.
Mass spectrometry was performed in an Agilent 5975C XLSTAT 2021 was used to perform one-way analysis of
mass selective detector coupled to an Agilent 7890A GC, in variance (ANOVA) followed by Tukey’s test and principal
electron ionization mode (70 eV), with the same column and component analysis (PCA). PCA was performed to visualize
operated under the same conditions as those for GC-FID, global clustering and separation trends. Pearson correlation
except for the carrier gas, which was helium, with a flow rate coefficients were calculated using the software Statgraph
of 1.0 mL/min. The MS scan range was 40–550 m/z with a 6.0.
scanning rate of 2.86 scan/s. Data handling was performed
using MSD ChemStation E.02.02.1431 software. The iden-
tification was carried out by comparison of the mass spectra Results and Discussion
obtained with those from the Wiley 6th edition Resource
Library and from the literature (Adams 2007). The linear Analyses of Olive Oils Regarding Quality Parameters
retention indices of the volatile compounds were calculated
from their retention times and those from a homologous The values found for the analysis of free fatty acids, perox-
series of n-alkanes ­(C7 –C26), injected under the same con- ide value, and absorbance in UV are in accordance with the
ditions (van Den Dool and Kratz 1963). The values obtained quality parameters required by the legislation (Brasil 2012)
were compared with data from the literature (Adams 2007). and all oils were classified as extra virgin (Table 1). The
free fatty acids content was below 0.80%, indicating that
the fruits used for olive oil production were healthy and that
Storage Test the conditions of harvest, transport, storage, and production
were controlled, so that triacylglycerols hydrolysis was very
For the storage test, each sample was divided into five 10-mL low (Boskou 2006; Brasil 2012, 2018).
amber glass vials, completely filled, and stored for a period All samples analyzed showed a peroxide value below
of 0 to 8 months, protected from light and at a temperature 20 meq/kg, which indicates a low concentration of hydrop-
of 20 ± 2 °C. Every 2 months, one of the stored vials of eroxides and a low degree of oxidation. The UV absorb-
each sample was opened, and the volatile compounds were ance results obtained for all samples were less than 0.22
identified and estimated as described previously. Therefore, for 270 nm, 2.5 for 232 nm, and 0.01 for delta K in agree-
the volatile compounds in each sample were identified and ment with the parameters for extra virgin olive oil cat-
estimated at 0, 2, 4, 6, and 8 months of storage. The tempera- egory. Similar data have already been reported for other
ture of 20 ± 2 °C was chosen to reproduce the most common Brazilian olive oils (Zago et al. 2019; da Costa et al. 2020).
storage conditions in the market. The values obtained in the analysis of absorbance in UV

Table 1  Peroxide value (PV), free fatty acids (FFA), and absorbance in UV of the extra virgin olive oil samples analyzed
Olive oil Origin PV (meq/kg) FFA (%) Ultraviolet absorbance
sample
City State K232 K270 ΔK

1 Pinheiro Machado RS 9.7 ± 0.37cA 0.30 ± 0.004cA 1.57 ± 0.03abA 0.15 ± 0.001bA 0.006
2 Canguçu RS 12.7 ± 0.13aA 0.49 ± 0.005aA 1.46 ± 0.005cdA 0.14 ± 0.004cA 0.003
3 Cachoeira do Sul RS 7.9 ± 0.03dA 0.22 ± 0.004eA 1.44 ± 0.01deA 0.10 ± 0.0004fA 0.003
4 Caçapava do Sul RS 7.3 ± 0.14eA 0.30 ± 0.005cA 1.57 ± 0.02abA 0.14 ± 0.002cA 0.006
5 Alagoa MG 7.3 ± 0.23eA 0.19 ± 0.001fB 1.54 ± 0.028abcA 0.12 ± 0.003dA 0.003
6 Maria da Fé MG 11.5 ± 0.24bA 0.28 ± 0.004 dB 1.36 ± 0.01eA 0.18 ± 0.002aA 0.003
7 Aiuruoca MG 10.9 ± 0.29bA 0.32 ± 0.005bB 1.43 ± 0.005deA 0.14 ± 0.001cA 0.005
8 Consolação MG 6.6 ± 0.07fA 0.18 ± 0.003fgB 1.49 ± 0.01bcdA 0.11 ± 0.0003eA 0.003
9 Baependi MG 4.8 ± 0.09gA 0.17 ± 0.003gB 1.62 ± 0.08aA 0.15 ± 0.01bA 0.004

PV, peroxide value; FFA, free fatty acids; K232 and K270, absorbance in UV at 232 and 270 nm, respectively
PV, FFA, K232, and K270 values are in mean (n = 3) ± SD
Means of samples followed by different lowercase letters in the same column are significantly different (p < 0.05, one-way ANOVA, followed by
Tukey post-test)
Means for regions followed by different uppercase letters in the same column are significantly different (p < 0.05, one-way ANOVA, followed by
Tukey post-test)

13
Food Analytical Methods (2022) 15:1508–1520 1513

are indicators of oxidation of olive oils and wavelengths standards, since the soil and climate characteristics may
of 232 and 270 corresponding to the maximum absorption influence their identity parameters.
of conjugated dienes and trienes formed in the autoxida- As expected, the fatty acid present in the highest con-
tion of hydroperoxides (Boskou 2006). According to Caipo centration in all samples was oleic (C18:1), with a signifi-
et al. (2021), K270 measures secondary oxidation prod- cant difference (p < 0.05) in concentration between the oils
ucts (aldehydes and ketones) produced by the breakdown from Minas Gerais (MG) and Rio Grande do Sul (RS). Olive
of hydroperoxides, reflecting a more advanced oxidation oils from MG showed a higher concentration, ranging from
state than PV or K232. Delta K measures the difference 78.4 to 81.1% oleic acid, while olive oils from RS varied
between absorbance at 270 nm and 266–274 nm and was from 74.1 to 76.3%. Kotsiou and Tasioula-Margari (2015)
calculated to verify if there was adulteration with other reported close values, from 71.32 to 76.29%, for Greek oils
refined or pomace oils (Boskou 2006). from the cultivar Koroneiki, while Dabbou and collaborators
There was significant difference (p < 0.05) among (2009) found lower data, 60.1%, for oils of the same cultivar
samples for PV, FFA, K232, and K270 but the difference from Tunisia. Variations in fatty acid composition of olive
(p < 0.05) between regions was observed only for FFA oils from the same cultivar were attributed to the composi-
(Table 1). tion of the soil, climatic conditions, harvesting techniques,
Although the Brazilian legislation is based on the Euro- and cultivation practices from each region studied (Dabbou
pean one, there is not, to date, an official sensory panel et al. 2011a; Kosma et al. 2017).
established in the country. Therefore, sensory evaluation is Additionally, significant differences (p < 0.05) were
not yet applied for classification and labelling, only chemical observed between the samples for the acids C16:0, C16:1,
analyses. According to the results from free fatty acids, UV C18:0, C18:1, C18:2, C18:3, C20:0, and C20:1. Regarding
absorbance, and fatty acids content, all the samples analyzed the geographical origin, significant differences (p < 0.05) for
are classified as “Virgin Olive Oil,” type “Extra Virgin/Vir- the content of C16:0, C18:0, C18:2, and C20:0 acids. The
gin” (Brasil 2012, 2018; The European Commission 2016). fatty acid compositions of all samples evaluated were con-
The fatty acid composition of the analyzed samples is sistent with the ranges specified in both Brazilian and EU
shown in Table 2. This is an identity parameter normally regulations (The European Commission 2016; Brasil 2018).
used to test if a product is pure olive oil or whether it is It is noteworthy that the concentration of C18:3 was below
mixed with other oils. Herein, the fatty acid composition of 1%. This is a concern in many producing countries, where
olive oils was determined, mainly, to verify if the olive oils this parameter is often outside the limits of the European
produced in Brazil have profiles consistent with international regulation.

Table 2  Fatty acid composition (% of total fatty acids) of the samples of extra virgin olive oil analyzed
Olive 1 2 3 4 5 6 7 8 9
oil
sample

State RS RS RS RS MG MG MG MG MG
C16:0 12.16 ± 0.25bA 12.34 ± 0.05bA 13.07 ± 0.03aA 13.31 ± 0.05aA 11.01 ± 0.38 dB 9.85 ± 0.01eB 9.92 ± 0.11eB 10.99 ± 0.06cdB 11.14 ± 0.11cB
C16:1 0.77 ± 0.02dA 0.87 ± 0.00bA 0.82 ± 0.00cA 0.92 ± 0.01aA 0.75 ± 0.07fA 0.74 ± 0.00eA 0.74 ± 0.01eA 0.77 ± 0.00dA 0.87 ± 0.01bA
C17:0 < 0.1 nd < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 nd nd
C17:1 < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 < 0.1
C18:0 2.7 ± 0.07aA 2.4 ± 0.01cA 2.3 ± 0.04cdA 2.5 ± 0.03bA 2.1 ± 0.07deB 2.1 ± 0.03efB 2.1 ± 0.02fB 2.0 ± 0.01fB 2.3 ± 0.02 dB
C18:1 76.3 ± 0.26eA 75.0 ± 0.06fA 76.2 ± 0.24eA 74.1 ± 0.18gA 78.4 ± 0.30 dB 80.4 ± 0.06bB 81.1 ± 0.19aB 79.0 ± 0.05 dB 79.6 ± 0.10cB
cA aA dA bA fB gB hB eB
C18:2 5.1 ± 0.01 6.7 ± 0.00 4.7 ± 0.01 6.3 ± 0.01 4.4 ± 0.11 3.9 ± 0.00 3.6 ± 0.00 4.6 ± 0.00 3.5 ± 0.00iB
C18:3 0.74 ± 0.00fA 0.78 ± 0.00eA 0.81 ± 0.00dA 0.83 ± 0.00cA 0.84 ± 0.02bA 0.89 ± 0.00aA 0.72 ± 0.00gA 0.67 ± 0.00iA 0.69 ± 0.00hA
C20:0 0.45 ± 0.01aA 0.43 ± 0.00bA 0.44 ± 0.00abA 0.44 ± 0.00abA 0.43 ± 0.02abB 0.41 ± 0.00cB 0.38 ± 0.00 dB 0.38 ± 0.00 dB 0.41 ± 0.01cB
C20:1 0.32 ± 0.00dA 0.33 ± 0.00cdA 0.36 ± 0.01bA 0.32 ± 0.00dA 0.39 ± 0.01aA 0.40 ± 0.00aA 0.34 ± 0.00cA 0.36 ± 0.00bA 0.32 ± 0.00dA
A A
C22:0 < 0.1 < 0.1 < 0.1 < 0.1 0.16 ± 0.01 0.15 ± 0.00 < 0.1 < 0.1 < 0.1
C24:0 < 0.1 nd < 0.1 < 0.1 < 0.1 < 0.1 < 0.1 nd nd

Means of samples followed by different lowercase letters in the same row are significantly different (p < 0.05, one-way ANOVA, followed by
Tukey post-test)
Means for regions followed by different uppercase letters in the same row are significantly different (p < 0.05, one-way ANOVA, followed by
Tukey post-test)
Fatty acid composition values are in mean ± SD (n = 3)
nd, not detected

13
1514 Food Analytical Methods (2022) 15:1508–1520

Profile of Volatile Compounds from Olive Oil other factors, on the vapor pressure of the volatiles present
Samples in the sample matrix, and these relations are expressed by
an equilibrium constant (KD), making quantitation a com-
The LOQ of the method was established as the first level of plex, challenging, and, sometimes, tricky task (Sgorbini
concentration of the analytical curve; therefore, the LOQ et al. 2019). For a HS-SPME, the partition between the
was 0.0011 mg/g or 1.1 mg/kg. As for the linearity of the headspace and the fiber also plays an important role (Zhang
method, the r was 0.9950, which indicates a strong corre- and Pawliszyn 1993). Fiber saturation can lead to large
lation between the concentration of n-tetradecane and the errors in quantitation, as has been recently demonstrated to
response signal (area) provided by GC-FID in the studied olive oil analyses (Stilo et al. 2021). Although the volatiles
concentration range (available in supplementary material in the samples were found in quantities of 0.1 to 0.01% of
Fig. SI1). the upper side of the calibration curve, the linearity was
The identifications of volatile compounds in extra virgin checked using only the internal standard. Therefore, all
olive oil samples are shown in Table 3. In total, 19 different figures in the tables shall be regarded as rough estimates.
compounds were identified at zero storage time. A typical According to literature data (Kotsiou and Tasioula-Margari
chromatogram of the volatiles from the samples is shown in 2015), the volatile compounds identified in the samples con-
Fig. 2, the peak numbers corresponding to the compounds tribute positively to the aroma of olive oils, corroborating
in Table 3. the results obtained in the identity and quality analyses,
Whenever a headspace sampling method is used, informa- except for nonanal which appears in a low concentration
tion is gathered from the volatile fraction, not from the sample (less than 1.0 mg/kg) in samples 1, 2, 3, and 9. The odor
itself. The composition of this fraction is dependent, amidst descriptors for nonanal are rancid, fat, citrus, and green,

Table 3  Volatile composition (estimate, in mg/kg) in extra virgin olive oil samples from RS and MG at zero storage time
Peak Compound RI lit RI exp Rio Grande do Sul (RS) Minas Gerais (MG)
1 2 3 4 5 6 7 8 9

1 Hexanal* 801 800 nq nq nq nq nq nq nq nq nq

2 (E)-2-Hexenal 846 853 2.45e 15.56c 39.58a 8.20d 17.98bc 39.25a 14.00c 21.56b 13.33c
3 (Z)-3-Hexenol* 850 871 nq nq nq nq nq nq nq nq nq
4 (E,E)-2,4-Hexadienal 907 912 3.71b < 1.0 < 1.0 < 1.0 < 1.0 1.76d 4.72a 3.09c 1.20e
5 (Z)-3-Ethyl-1,5-octadieneǂ # 934 < 1.0 2.22e 1.15f 1.10f 2.42d 3.13c < 1.0 5.30a 4.04b
6 (E)-3-Ethyl-1,5-octadiene 947 942 < 1.0 1.22e < 1.0 < 1.0 1.37d 1.69c < 1.0 3.70a 1.87b
7 5-Ethyl-2(5H)-furanone 968 969 nd 2.57c nd 1.55d nd nd < 1.0 3.02b 5.38a
8 6-Methyl-5-hepten-2-one 986 987 < 1.0 1.32b nd nd 1.44b 2.07a 1.19c < 1.0 2.18a
9 4,8-Dimethyl-1,7-nonadieneǂ # 1000 1.20 g 3.51d 1.97f nd 3.71c 3.31e < 1.0 5.18b 7.31a
10 (Z)-3-Hexenyl acetate 1004 1008 < 1.0 1.65d < 1.0 < 1.0 5.86a 5.40a 2.73c 1.48d 3.44b
11 Hexyl acetate 1014 1013 < 1.0 3.88a 2.47b < 1.0 < 1.0 < 1.0 < 1.0 1.86c 2.48b
12 p-Cymene 1020 1022 < 1.0 < 1.0 < 1.0 < 1.0 nd < 1.0 < 1.0 < 1.0 < 1.0
13 Limonene 1024 1026 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 1.16 < 1.0
14 (E)-β-Ocimene 1044 1047 nd < 1.0 nd < 1.0 < 1.0 < 1.0 nd < 1.0 < 1.0
15 Nonanal 1100 1100 < 1.0 < 1.0 < 1.0 nd nd nd nd nd < 1.0
16 (E)-4,8-Dimethyl-1,3,7-nonatriene 1116 1116 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 nd nd nd < 1.0
17 n-Dodecane 1200 1200 nd nd nd < 1.0 nd nd nd nd < 1.0
18 1-Methylcyclodecane 1202 1206 nd 2.70a 2.16b < 1.0 nd < 1.0 nd nd 1.46c
19 n-Tridecane 1300 1300 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 < 1.0 nd < 1.0

Results with different lowercase letters for the same compound are significantly different (p < 0.05, one-way ANOVA, followed by Tukey’s post-
test)
Mean values of triplicate data
Standard deviations were < 0.00 for all estimated compounds
*
Identification confirmed by co-injection of authentic standards
ǂ
Tentative identification
#
Retention index not found in the literature
RI exp, experimental retention index; RI lit, retention index available in the literature (Adams 2007); nq, not quantified; nd, not detected

13
Food Analytical Methods (2022) 15:1508–1520 1515

Fig. 2  Typical chromatogram of volatile compounds of extra virgin olive oil obtained by HS-SPME followed by GC–MS analysis. The numbers
from 1 to 19 correspond to the compounds identified in Table 3

and its perception limit is 0.15 mg/kg (Romero et al. 2015). fourteen samples of Greek extra virgin olive oils from
Therefore, the concentrations found in the samples are very the cultivar Koroneiki and, regarding the composition of
close to the limit of perception, which may cause them not (E)-2-hexenal, found a variation of 13.59 to 35.28 mg/kg.
to be noticed in the oils. Thus, if the samples were submitted In a study carried out with seven monovarietal olive oils
to sensory analysis, it is probable that, in relation to odor, (Arbequina, Barnea, Frantoio, Koroneiki, Leccino, Manza-
the median of the defect would be equal to 0 and the median nilla, and Picual) from Chile, the average of (E)-2-hexenal
of the fruitiness would be greater than 0 for all of them. found was 7.3 mg/kg in the samples (García-González et al.
These parameters are essential for olive oils to be classified 2010), while analyses carried out on extra virgin olive oils
as extra virgin in countries where sensory tests are carried produced in Tunisia showed lower values of (E)-2-hexenal
out (Brasil 2012; Romero et al. 2015). Costa et al. (2020) for three samples of the Koroneiki cultivar, around 4.00 mg/
also identified nonanal in their samples of Brazilian oils. kg (Dabbou et al. 2009). The samples analyzed in this work
However, during the sensory analysis, they did not conclude showed great variability, as well as the results found in the
that this compound had interfered in the aroma of the oils. literature. Although the oils are all extra virgin from the
Zago et al. (2019) detected nonanal in only two of the twelve Koroneiki cultivar, it is assumed that this large difference is
samples of Brazilian oils analyzed, and also indicated that due to the place of cultivation and climatic, processing, and
this low incidence in the samples is related to the improve- storage conditions.
ment in agricultural and processing conditions employed in Despite the wide variation in the amount of (E)-2-hexe-
the country, which resulted in the production of high-quality nal, in all samples this compound was well above its percep-
oils, both in chemical and sensory aspects. tion limit, which is 0.42 mg/kg. This aldehyde contributes
Through the analysis of variance, it was possible to positively to the odor of olive oils and has been associated
observe that all samples are different from each other in with notes of almond, green, bitter, and fruity (Kotsiou and
terms of the concentration of volatile compounds (p < 0.05). Tasioula-Margari 2015; Zago et al. 2019; da Costa et al.
As expected, compounds from the lipoxygenase pathway 2020). Another compound that contributes positively is
were identified: hexanal, (E)-2-hexenal, (Z)-3-hexenol, (Z)- 6-methyl-5-hepten-2-one, with a perception limit of 1.0 mg/
3-hexenyl acetate, and hexyl acetate. All of them contribute kg and sensory descriptors of green, fruity, grass, and pun-
to the characteristic green aroma of extra virgin olive oils gent (Romero et al. 2015). In samples 5, 6, 7, and 9, the
(Olías et al. 1993; Kotsiou and Tasioula-Margari 2015). compound was detected in a concentration above this limit.
Except for sample 1, the compound present in the high- Other compounds found in the samples that possibly
est concentration was (E)-2-hexenal, in both RS and MG contribute positively to the odor are as follows: (E,E)-
samples. Kotsiou and Tasioula-Margari (2015) analyzed 2,4-hexadienal, with sensory descriptors of green, citrus,

13
1516 Food Analytical Methods (2022) 15:1508–1520

and pungent, and terpenes, such as limonene, which appear In this work, it was observed a great decrease of the com-
in all samples and have a citric character (Kotsiou & pounds that contribute positively to the odor of extra virgin
Tasioula-Margari, 2015). The other terpenes that appear in olive oils, mainly in the concentration of (E)-2-hexenal.
the samples are p-cymene, (E)-β-ocimene, and (E)-4,8-di- Sample 6 showed the greatest variation of this substance
methyl-1,3,7-nonatriene. The latter is a homoterpene com- during storage, which was reduced to circa 20% of the
monly produced by plants after an attack by herbivores, original after 8 months. It is assumed that this difference,
which was found in six of the nine samples (Pinto-Zevallos as described by Kotsiou and Tasioula-Margari (2015), may
et al. 2013). Some hydrocarbons were also found in the have occurred due to the conditions of processing and stor-
samples. Although other authors have also identified these age of olive oils.
compounds in olive oils, the role of these substances in the In contrast, Fregapane et al. (2006) evaluated the influ-
odor of the food has not been established yet (Kotsiou and ence of the filtration process on the stability of Spanish and
Tasioula-Margari 2015; Romero et al. 2015). Italian virgin olive oils from the cultivars Picual, Cornicabra,
Another substance found in samples 2, 4, 7, 8, and 9 and Arbequina over 12 months of storage. Sensorial analysis
that showed great variation was 5-ethyl-2(5H)-furanone. indicated a great decrease in the fruity attribute during stor-
Kandylis et al. (2011) also found this compound in Tuni- age for all samples. Even though the identification and quan-
sian olive oils at a concentration of 14.85 µg/kg, but did tification of volatile compounds were not carried out in the
not describe its influence on the product. Buttery and study described by those authors, the decrease in the fruity
Takeoka (2004) identified 5-ethyl-2(5H)-furanone in toma- attribute may indicate a reduction in volatile compounds that
toes and their leaves as an unusual volatile component, positively influence the aroma of olive oils throughout the
having also indicated that this is a substance already iden- storage period, similarly to what was found in our study
tified in raspberries, asparagus, and bread, and that it is (Fregapane et al. 2006).
probably formed from the auto-oxidation of (Z)-3-hexenal It is important to note that, in both studies mentioned
(Buttery and Takeoka 2004; Kandylis et al. 2011). above, the storage temperatures were different, which may
It was not possible to estimate hexanal, as this sub- have influenced the changes observed here. While Kotsiou
stance coeluted with others not identified. Its presence and Tasioula-Margari (2015) maintained a temperature of up
in the sample was confirmed through its retention index to 25 °C during the summer months and 12 ± 6 °C in other
and the co-injection of an authentic standard. The same periods of the year, Fregapane et al. (2006) kept the tempera-
occurred with (Z)-3-hexenol. ture at 25 °C during the 12 months of the study, which was
Over the 8 months of storage, 17 to 21 volatile com- closer to the conditions of the present study.
pounds were identified in the samples analyzed (available
in supplementary material Tables SI2 – SI10). The main Volatile Compounds Responsible for Sensory
observations made were regarding the compounds from Defects
LOX (Table 4) and the compounds responsible for sen-
sory defects in EVOOS. Although the total amount of (E)- During the storage period, it was possible to identify (E)-
2-hexenal varied greatly between samples, in all of them, 2-heptenal, octanal, nonanal, and (E)-2-decenal, all asso-
a decrease during the storage period was observed. The ciated to negative sensory characteristics in olive oils. In
behaviors in the variations of the concentration of (Z)- supplementary material (Table SI11), there is a description
3-henexyl acetate and hexyl acetate were different between of the sensory properties for these compounds and their per-
the samples (Table 4). All of these compounds are pro- ception limits in olive oil.
duced through LOX, as well as hexanal and (Z)-3-hexenol, In all the samples analyzed in this work, at least one of
and have been identified throughout all months of storage the substances associated with sensory defects was iden-
(Olías et al. 1993). tified during the storage period. Despite appearing in low
Kotsiou and Tasioula-Margari (2015) evaluated the concentrations, they are likely to have a negative influence
changes in the profile of volatile compounds in Greek extra on the sensory characteristics of the oils analyzed because
virgin olive oils during an 18-month period. Contrary to their limits of perception are low (Romero et al. 2015). Kot-
what was observed in this work, the researchers observed siou and Tasioula-Margari (2015) identified and quantified
an increase in LOX compounds in the first 6 months, and (E)-2-heptenal in stored olive oil samples in concentrations
a decrease of 10%, considered non-significant, in the fol- ranging from 0.05 mg/kg in time 0 to 0.20 mg/kg after
lowing months. In addition, when LOX compounds were 18 months. Although it was present in low concentration, the
observed separately, only (Z)-3-hexenal and hexanal showed researchers stated that this substance had a negative effect
significant changes. While the first one was reduced by up to on the odor of olive oils.
100%, the second increased by up to 80% among the samples It must be highlighted that, except for samples 2 and 4,
analyzed. in which (E)-2-heptenal was identified from the second and

13
Food Analytical Methods (2022) 15:1508–1520 1517

Table 4  Changes in the Olive oil Compound Storage (months)

concentrations of volatile sample
compounds (estimate, mg/kg of 0 2 4 6 8
sample) from the lipoxygenase
c a b d
pathway (LOX) in olive 1 (E)-2-Hexenal 2.45 6.3 4.28 1.39 < 1.0
oil samples during storage (Z)-3-Hexenyl acetate < 1.0 < 1.0 1.69b 2.19a 2.13a
(months) Hexyl acetate < 1.0 < 1.0 < 1.0 < 1.0 4.63a
2 (E)-2-Hexenal 15.56a 8.55b 2.79d 3.42c 3.46c
(Z)-3-Hexenyl acetate 1.65d nq 2.67c 4.36a 3.64b
Hexyl acetate 3.88a 3.55b nq 2.13c 1.44d
3 (E)-2-Hexenal 39.58a 36.87a 18.31c 22.00b 13.88d
(Z)-3-Hexenyl acetate < 1.0 < 1.0 1.70a 1.89b 2.22c
Hexyl acetate 2.47b 2.9a < 1.0 < 1.0 < 1.0
4 (E)-2-Hexenal 8.2a 6.31b 2.65d nq 3.12c
(Z)-3-Hexenyl acetate < 1.0 1.17b 1.17b nq 2.37a
Hexyl acetate < 1.0 < 1.0 < 1.0 nq 2.91a
5 (E)-2-Hexenal 17.98a 11.58c 10.73d 14.70b 6.18e
(Z)-3-Hexenyl acetate 5.86e 6.63c 9.01a 6.13d 7.19b
Hexyl acetate < 1.0 2.33a 1.50b < 1.0 < 1.0
6 (E)-2-Hexenal 39.25a 33.18b 28.47c 23.29d 7.66e
(Z)-3-Hexenyl acetate 5.40a 2.86b nq 2.07c 1.56d
Hexyl acetate < 1.0 < 1.0 < 1.0 < 1.0 < 1.0
7 (E)-2-Hexenal 14.00a 3.66c 4.85b 1.97e 2.59d
(Z)-3-Hexenyl acetate 2.73d 7.78a nq 7.47b 5.96c
Hexyl acetate < 1.0 < 1.0 < 1.0 1.17b 1.35a
8 (E)-2-Hexenal 21.56a 4.98b 2.91c 2.36d 1.89e
(Z)-3-Hexenyl acetate 1.48a < 1.0 < 1.0 < 1.0 < 1.0
Hexyl acetate 1.86a < 1.0 < 1.0 < 1.0 1.53b
9 (E)-2-Hexenal 13.33a 3.71b 1.19e 1.98c 1.72d
(Z)-3-Hexenyl acetate 3.44a 1.08b 1.19b < 1.0 < 1.0
Hexyl acetate 2.48a 1.19c 1.35b 1.32b 1.25c

Results with different lowercase letters on the same line are significantly different (p < 0.05, one-way
ANOVA, followed by Tukey’s post-test)
Mean values of triplicate data
Standard deviations were < 0.00 for all estimated compounds
nq, not quantified

Fig. 3  Differentiation of the RS

and MG regions using multi-
variate analysis. A PCA score
plot for samples; B PCA score
plot of the chemical parameters
evaluated for olive oil (C16:0,
C16:1, C18:0, C18:1, C18:2,
C18:3, (Z)-3-hexenyl acetate,
(E)-2-hexenal, hexyl acetate,
free fatty acids, specific extinc-
tion at 270 nm and peroxide
value). All values used were the
mean of triplicate data

13
1518 Food Analytical Methods (2022) 15:1508–1520

fourth months of storage, respectively, in all other samples

Hexyl acetate
this compound was only detected at the eighth month. As

0.5138**

0.3846*
− 0.3343

− 0.3036
0.3553

0.0482

0.3704

0.1202

0.1331
for octanal, nonanal, and (E)-2-decenal, it is not possible to

− 0.262

− 0.222
state whether they influenced the odor of the studied samples
because, just like their perception limits, their concentra-
tion in the samples is below the quantification limit of the

(Z)-3-Hexenyl acetate
method used.
The development of compounds from the oxidation of

− 0.7236**
− 0.5848**
− 0.4855**

− 0.5575**
olive oils is directly linked to the processing and storage

− 0.3343

− 0.2704
0.4431*
conditions of this food. Other factors that influence the for-

0.0162

0.3737
0.3138
mation of undesirable compounds are the cultivar and the
maturation stage of the olives. While the high concentration

(E)-2-Hexenal
of linoleic acid present in some cultivars increases the rate

− 0.521**
of oxidation, the composition and the number of phenolic

0.6233**
− 0.3085

− 0.2063
0.4267*
− 0.188
− 0.244

− 0.123
0.1228
compounds can decrease this process. The cultivar Koro-
neiki is known to have a low concentration of linoleic acid
and a high concentration of phenolic compounds. Thus, it

− 0.3827*

− 0.1456
is expected that oils produced from this cultivar, under con-

0.3814*
0.1080
0.2607

0.1865

0.2995
0.037
K270
trolled processing conditions, have lower concentrations of
compounds that negatively influence the odor of the food

0.6516**
during storage (Boskou 2006; Vieira Neto et al. 2008; Kot-

− 0.4217*

0.841**
0.2199
0.2881
0.3245

0.1443
siou and Tasioula-Margari 2015).

Differences Between Olive Oils Produced in Distinct FFA

− 0.1602
− 0.1837

− 0.0347
Regions

0.0538

0.2842
0.2940
Table 5  Pearson’s correlation coefficients data for chemical parameters evaluated for olive oil samples
PV

The main fatty acids, quality parameters, and main volatile

compounds were evaluated through a principal component
0.5712** − 0.1685

− 0.91** − 0.2125
0.7501** 0.0819

0.1202

analysis (PCA) to provide information about the pattern of 0.1853

C18:3

similarity of the observations (samples) and the regions.

The two principal components, PC1 and PC2, explained
0.589**

PV, peroxide value; FFA, free fatty acids; K270, absorbance in UV at 270 nm (nm)
60.7% of the total data variation (Fig. 3). Through the PCA,
C18:2

it was possible to observe that the samples could be grouped

by geographic regions, RS and MG, since they are in the
− 0.9417**
− 0.6487**
− 0.7561**

opposite sides of the plot, as is shown in Fig. 3A. The vari-

ables that most contributed to this distinction were fatty
Asterisks indicate significant values at * p ≤ 0.05 and ** p ≤ 0.01
C18:1

acids C16:0, C16:1, C18:0, C18:1, and C18:2 and volatile

compounds (Z)-3-hexenyl acetate, (E)-2-hexenal, and hexyl
0.7175**

acetate, the latter two with a smaller contribution. Free fatty

0.4365*
C18:0

acids, UV absorbance at 270 nm, peroxide value, and C18:3

did not contribute to the clustering (Fig. 3B). It was possi-
ble to observe the strong negative correlation among C18:1
0.7187**

and C18:2 and C18:0, as well as between (Z)-3-hexenyl

C16:1

acetate, (E)-2-hexenal, and hexyl acetate. The samples from

MG state were most correlated with C18:1, (Z)-3-hexenyl
(Z)-3-Hexenyl acetate (mg/kg)

acetate, and (E)-2-hexenal while those from RS state were

most correlated with C18:2, C18:0, C16:0, C16:1, and hexyl
(E)-2-Hexenal (mg/kg)

acetate. These results corroborated the significant difference

(p < 0.05) observed between regions by analysis of variance
for C16:0, C16:1, C18:0, C18:1, C18:2, and (Z)-3-hexenyl
acetate.
In addition, the correlation analyses (Table 5) revealed
C16:0
C16:1
C18:0
C18:1
C18:2
C18:3

K270
Trait

FFA

that, in relation to fatty acids, there was a negative correlation

PV

13
Food Analytical Methods (2022) 15:1508–1520 1519

among C18:1 and C16:0, C16:1, C18:0, and C18:2 (Pearson, Supplementary Information The online version contains supplemen-
p < 0.01), then the increase of C18:1 observed for samples tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 12161-0​ 21-0​ 2192-0.
of MG state resulted in the reduction of those fatty acids.
Acknowledgements N. S. B. thanks CNPq and CAPES for MSc.
C18:3 did not correlate with any of the other fatty acids, (133417/2018-2) and DSc. (88887.474267/2020-00) scholarships,
but it did have a positive correlation with (E)-2-hexenal and respectively. H. R. B. thanks CNPq for productivity scholarship
(Z)-3-hexenyl acetate (Pearson, p < 0.05). This means that (310248/2018-3). The authors thank Mrs. Marcelly C. S. Santos for
samples with a higher concentration of C18:3 also presented technical assistance and Mr. Andre L. N. Gomes for drawing the map
of Brazil (Fig. 1).
a higher amount of these volatiles, which was expected,
since these compounds are produced from C18:3 through Author Contribution Nathalia S. Brilhante: formal analyses, investiga-
the LOX pathway. C18:1 also showed a positive correlation tion, data curation, writing (original draft preparation); Adelia F. Faria-
with (E)-2-hexenal, while C18:2 correlated negatively with Machado: conceptualization, funding acquisition, data curation, writing
(Z)-3-hexenyl acetate (p < 0.01). Another expected positive (review); Rosemar Antoniassi: data curation, software, visualization,
writing (review); Paola E. Gama: formal analyses, data curation; Hum-
correlation was that of PV with K270 and FFA, which are berto R. Bizzo: conceptualization, funding acquisition, data curation,
parameters of primary and secondary oxidation products and writing (review and editing), supervision.
triacylglycerol hydrolysis, respectively, and indicate how the
raw material was treated during harvesting and processing. Funding This work was supported by the Conselho Nacional de
The variation observed between regions suggests that Desenvolvimento Científico e Tecnológico (CNPq), Coordenação de
Aperfeiçoamento de Pessoal de Nível Superior (CAPES), and Empresa
the fatty acids profile could be used as a marker for geo- Brasileira de Pesquisa Agropecuária (Embrapa).
graphical indication and PCA was a useful tool for clus-
tering. However, it is necessary that a larger number of Data Availability All data generated or analyzed during this study are
samples be analyzed, including those from different cul- included in this published article and in its supplementary informa-
tivars and crops, to verify whether the variation in the tion files.
composition of fatty acids remains.
Declarations

Ethics Approval This article does not contain any studies with human
participants or animals.
Conclusions
Consent to Participate Not applicable.
The results obtained for fatty acid composition, free fatty
acids, peroxide value, and ultraviolet absorbance are char- Conflict of Interest Nathalia S. Brilhante declares that she has no con-
acteristic for extra virgin olive oils. The absence of volatile flict of interest. Adelia F. Faria-Machado declares that she has no con-
flict of interest. Rosemar Antoniassi declares that she has no conflict
compounds associated to sensorial defects at the beginning of interest. Paola E. Gama declares that she has no conflict of interest.
of the storage period is in agreement with the previous Humberto R. Bizzo declares that he has no conflict of interest.
parameters analyzed. The main volatile compounds iden-
tified in the samples at zero storage time were hexanal,
(E)-2-hexenal, (Z)-3-hexenol, (Z)-3-hexenyl acetate, and
hexyl acetate. Through analysis of variance, it was possible References
to observe that all samples were different from each other
in terms of the composition of volatiles. During the storage Adams RP (2007) Identification of essential oil components by gas
period, there was a decrease in compounds that contrib- chromatography/mass spectrometry. Allured Publishing Co.,
Carol Stream
ute positively to the odor of olive oils, the concentration American Oil Chemists’ Society (2009) Official methods and recom-
of (E)-2-hexenal being the one that decreased the most. mended practices of the American Oil Chemists’ Society. AOCS,
In the same period, compounds that negatively influence Urbana
the sensory characteristics of the food were still identi- Anastasopoulos E, Kalogeropoulos N, Kaliora AC et al (2011)
Quality indices, polyphenols, terpenic acids, squalene, fatty
fied: (E)-2-heptenal, octanal, nonanal, and (E)-2-decenal, acid profile, and sterols in virgin olive oil produced by
which were mainly produced in the eighth month of stor- organic versus non-organic cultivation method. WIT Trans
age. Through the evaluation of the chemical parameters of Ecol Environ 152:135–143. https://​d oi.​o rg/​1 0.​2 495/​F ENV1​
olive oils by PCA, the samples of Brazilian Koroneiki extra 10141
Antoniassi R, Wilhelm AE, Faria-Machado AF, Guedes AMM,
virgin olive oils could be grouped by geographic regions. Bizzo HR (2018) Otimização do Método Hartman e Lago de
The variables that most contributed to this distinction were Preparação de Ésteres Metílicos de Ácidos Graxos. Boletim
fatty acids C16:0, C16:1, C18:0, C18:1, and C18:2 and vol- de Pesquisa e Desenvolvimento 26:1–20. https://​ainfo.​cnptia.​
atile compounds (Z)-3-hexenyl acetate, (E)-2-hexenal, and embra​p a.​b r/​d igit​a l/​b itst​ream/​i tem/​1 98136/​1 /​B olet​i m-​P D-​2 6-​
metodo-​Hartm​an-e-​L ago.​pdf. Accessed 17 July 2021
hexyl acetate, the latter two with a smaller contribution.

13
1520 Food Analytical Methods (2022) 15:1508–1520

Boskou D (2006) Olive oil: chemistry and technology. Academic Press, fruits. JAOCS, J Am Oil Chem Soc 94:1373–1384. https://​doi.​org/​
Cambridge 10.​1007/​s11746-​017-​3036-5
Brasil (2012) Ministério da Agricultura, Pecuária e Abastecimento. Kotsiou K, Tasioula-Margari M (2015) Changes occurring in the volatile
Instrução Normativa n­ o 1, de 30 de janeiro de 2012. Diário Oficial composition of Greek virgin olive oils during storage: oil variety
da União, 149(23), section 1:5–8. Brasilia, DF. https://​pesqu​isa.​in.​ influences stability. Eur J Lipid Sci Technol 117:514–522. https://​
gov.​br/​impre​nsa/​jsp/​visua​liza/​index.​jsp?​data=​01/​02/​2012&​jornal=​ doi.​org/​10.​1002/​ejlt.​20140​0231
1&​pagina=​5&​total​Arqui​vos=​112. Accessed 17 July 2021 Mariotti M, Peri C (2014) The composition and nutritional properties
Brasil (2018) Ministério da Agricultura, Pecuária e Abastecimento. of extra-virgin olive oil. In: Peri C (ed) The extra-virgin olive oil
Instrução Normativa ­no 24, de 18 de junho de 2018. Diário Oficial handbook. Wiley-Blackwell, Hoboken, pp 21–34 https://​doi.​org/​10.​
da União, 155(117), section 1:5–6. https://​pesqu​isa.​in.​gov.​br/​impre​ 1002/​97811​18460​412.
nsa/​jsp/​visua​liza/​index.​jsp?​data=​20/​06/​2018&​jornal=​515&​pagina=​ Morales MT, Rios JJ, Aparicio R (1997) Changes in the volatile com-
5&​total​Arqui​vos=​64 Accessed 17 July 2021 position of virgin olive oil during oxidation: flavors and off-flavors.
Brevard H, Cantergiani E, Cachet T et al (2011) Guidelines for the quan- J Agric Food Chem 45:2666–2673. https://​doi.​org/​10.​1021/​jf960​
titative gas chromatography of volatile flavouring substances, from 585+
the Working Group on Methods of Analysis of the International Morales MT, Luna G, Aparicio R (2005) Comparative study of virgin
Organization of the Flavor Industry (IOFI). Flavour Fragr J 26:297– olive oil sensory defects. Food Chem 91:293–301. https://​doi.​org/​
299. https://​doi.​org/​10.​1002/​ffj.​2061 10.​1016/j.​foodc​hem.​2004.​06.​011
Buttery RG, Takeoka GR (2004) Some unusual minor volatile compo- NYIOOC - New York World Olive Oil Competition (2020) Official index
nents of tomato. J Agric Food Chem 52:6264–6266. https://​doi.​org/​ of the world’s best olive oils for 2019. https://​besto​liveo​ils.​com/​
10.​1021/​jf040​176a search. Accessed 22 April 2021
Caipo L, Sandoval A, Sepúlveda B et al (2021) Effect of storage con- Olías JM, Pérez AG, Ríos JJ, Sanz LC (1993) Aroma of virgin olive oil:
ditions on the quality of Arbequina extra virgin olive oil and the biogenesis of the “green” odor notes. J Agric Food Chem 41:2368–
impact on the composition of flavor-related compounds (phenols and 2373. https://​doi.​org/​10.​1021/​jf000​36a029
volatiles). Foods 10:2161. https://​doi.​org/​10.​3390/​foods​10092​161 Pinto-Zevallos DM, Martins CBC, Pellegrino AC, Zarbin PHG (2013)
da Costa JRO, Dal Bosco SM, de Ramos RC, S, et al (2020) Determina- Compostos orgânicos voláteis na defesa induzida das plantas contra
tion of volatile compounds responsible for sensory characteristics insetos herbívoros. Quim Nova 36:1395–1405. https://​doi.​org/​10.​
from Brazilian extra virgin olive oil using HS-SPME/GC-MS direct 1590/​S0100-​40422​01300​09000​21
method. J Food Sci 85:3764–3775. https://​doi.​org/​10.​1111/​1750-​ Romero I, García-González DL, Aparicio-Ruiz R, Morales MT (2015)
3841.​15467 Validation of SPME-GCMS method for the analysis of virgin olive
Dabbou S, Issaoui M, Servili M et al (2009) Characterisation of virgin oil volatiles responsible for sensory defects. Talanta 134:394–401.
olive oils from European olive cultivars introduced in Tunisia. Eur J https://​doi.​org/​10.​1016/j.​talan​ta.​2014.​11.​032
Lipid Sci Technol 111:392–401. https://​doi.​org/​10.​1002/​ejlt.​20080​ Sgorbini B, Cagliero C, Liberto E, Rubiolo P, Bicchi C, Cordero CEI
0032 (2019) Strategies for accurate quantitation of volatiles from foods
Dabbou S, Brahmi F, Selvaggini R et al (2011a) Contribution of irrigation and plant-origin materials: a challenging task. J Agric Food Chem
and cultivars to volatile profile and sensory attributes of selected vir- 67:1619–1630. https://​doi.​org/​10.​1021/​acs.​jafc.​8b066​01
gin olive oils produced in Tunisia. Int J Food Sci Technol 46:1964– Stilo F, Segura-Borrego MP, Bicchi C, Battaglino S, Callejón-Fernandez
1976. https://​doi.​org/​10.​1111/j.​1365-​2621.​2011.​02708.x RM, Morales ML, Reichenbach SE, McCurry J, Perroni D, Cordero
Dabbou S, Dabbou S, Chehab H et al (2011b) Chemical composition C (2021) Delineating the extra-virgin olive oil aroma blueprint by
of virgin olive oils from Koroneiki cultivar grown in Tunisia with multiple headspace solid phase microextraction and differential-flow
regard to fruit ripening and irrigation regimes. Int J Food Sci Tech- modulated comprehensive two-dimensional gas chromatography. J
nol 46:577–585. https://​doi.​org/​10.​1111/j.​1365-​2621.​2010.​02520.x Chromatogr A 1650:462232. https://​doi.​org/​10.​1016/j.​chroma.​2021.​
Faria-Machado AF, Santos MCS, Antoniassi R, Jorge RO, Bizzo HR 462232
(2017) Analysis of volatile compounds from Brazilian virgin olive The European Commission (2016) Commission Delegated Regulation
oils. Poster presentation at the Simpósio Científico-Técnico Expo- (EU) 2016/2095 amending Regulation (EEC) ­No 2568/91 on the
liva, Jaén, Spain characteristics of olive oil and olive-residue oil and on the relevant
Fregapane G, Lavelli V, León S et al (2006) Effect of filtration on virgin methods of analysis. Off J Eur Union, 59:1–6. http://​data.​europa.​eu/​
olive oil stability during storage. Eur J Lipid Sci Technol 108:134– eli/​reg_​del/​2016/​2095/​oj. Accessed 17 july 2021
142. https://​doi.​org/​10.​1002/​ejlt.​20050​1175 van Den Dool H, Kratz PD (1963) A generalization of the retention index
García-González DL, Romero N, Aparicio R (2010) Comparative study system including linear temperature programmed gas—liquid parti-
of virgin olive oil quality from single varieties cultivated in Chile tion chromatography. J Chromatogr A 11:463–471. https://​doi.​org/​
and Spain. J Agric Food Chem 58:12899–12905. https://​doi.​org/​ 10.​1016/​S0021-​9673(01)​80947-X
10.​1021/​jf103​1313 Velasco J, Dobarganes C (2002) Oxidative stability of virgin olive oil.
Instituto Brasileiro de Olivicultura - IBRAOLIVA (2020) Safra 2020 dos Eur J Lipid Sci Technol 104:661–676. https://​doi.​org/​10.​1002/​1438-​
azeites nacionais chega ao mercado com novidades. https://​www.​ 9312(200210)​104:9/​10%​3c661::​AID-​EJLT6​61%​3e3.0.​CO;2-D
ibrao​liva.​com.​br/​notic​ias/​detal​he/​84/s. Accessed 22 April 2021 Vieira Neto J, Oliveira AF, Oliveira NC, Duarte HSS, Gonçalves ED
International Olive Oil Council - IOC (2020) IOC Olive Oil Dashboard. (2008) Aspectos técnicos da cultura da oliveira. EPAMIG. http://​
https://​www.​inter​natio​nalol​iveoil.​org/​wp-​conte​nt/​uploa​ds/​2020/​11/​ www.​epamig.​br/​downl​oad/​bt88-​aspec​tos-​tecni​cos-​da-​cultu​ra-​da-​
IOC-​Olive-​Oil-​Dashb​oard-​Novem​ber-​2020.​html#​consu​mption-1. olive​ira/. Accessed 17 July 2021
Accessed 22 April 2021 Zago L, Squeo G, Bertoncini EI, et al (2019) Chemical and sensory char-
Kandylis P, Vekiari AS, Kanellaki M et al (2011) Comparative study of acterization of Brazilian virgin olive oils. Food Res Int 126https://​
extra virgin olive oil flavor profile of Koroneiki variety (Olea euro- doi.​org/​10.​1016/j.​foodr​es.​2019.​108588
paea var. Microcarpa alba) cultivated in Greece and Tunisia during Zhang Z, Pawliszyn J (1993) Headspace solid-phase microextraction.
one period of harvesting. LWT - Food Sci Technol 44:1333–1341. Anal Chem 65:1843–1852. https://​doi.​org/​10.​1021/​ac000​62a008
https://​doi.​org/​10.​1016/j.​lwt.​2010.​12.​021
Kosma I, Vatavali K, Kontakos S et al (2017) Geographical differentiation Publisher's Note Springer Nature remains neutral with regard to
of Greek extra virgin olive oil from late-harvested Koroneiki cultivar jurisdictional claims in published maps and institutional affiliations.

13