Ultrasound-Assisted Production of Virgin Olive Oil: Effects on Bioactive Compounds, Oxidative Stability, and Antioxidant Capacity

1. Introduction

Virgin olive oil (VOO) enjoys great popularity as one of the main representatives of the Mediterranean diet due to its specific sensory properties and proven positive effects on human health [1]. In the olive fruit, the oil is located in the cytoplasm inside of the specialized leucoplasts called elaioplasts. During the crushing phase of VOO production, mechanical force is applied to olives, breaking down their structure and releasing oil from the disrupted cells. However, some of the cells remain intact and the oil which is released from disrupted cells can form an emulsion with the aqueous phase further complicating separation [2]. These factors contribute to relatively low oil yield (Y) in conventional VOO production which are addressed during the malaxation phase. Malaxation, a controlled slow mixing of the crushed olive paste at regulated temperatures (typically up to 28 °C) and controlled duration (up to 60 minutes), increases Y and influences the overall quality of VOO [3]. During malaxation, the oil droplets bind to each other through hydrophobic interactions (coalescence phenomenon), while chemical substances (triacylglycerols, fatty acids, pigments and phenols) released from the cells during the crushing phase are distributed between the oil and water phases and transformed by the activity of the endogenous enzymes of the olive fruit (EE) [2,4]. The enzymes of the lipoxygenase pathway (acyl hydralase, lipoxygenase (LOX), hydroperoxide lyase, alcohol dehydrogenase and alcohol acyl transferase) are involved in the formation of the desirable C5 and C6 volatile compounds responsible for the positive sensory properties of VOO. Qualitatively and quantitatively, the main volatile compounds contributing to the VOO aroma are hexanal, (E)-2-hexenal, (E)-2-hexen-1-ol, hexan-1-ol and (Z)-3-hexenyl acetate [5]. The composition and quantity of phenols are directly influenced by β-glucosidase (β-GLU), which hydrolyses phenolic glycosides, producing simple sugars and phenolic aglycones that are more soluble in oil and contribute to the sensory and nutritional value of VOO. On the other hand, the oxidoreductases polyphenol oxidase (PPO) and peroxidase (POX) are responsible for the degradation of phenols. POX oxidizes phenols at the expense of hydroperoxides, while PPO oxidizes phenols converting phenols into benzene-1,2-diols (o-diphenols or catechins) through hydroxylation and into benzo-1,2-quinones (o-quinones) through their oxidation, producing melanins with amines or phenols through a series of enzymatic browning reactions [6]. To improve the nutritional and sensory profile of VOO, the ultimate goal would be to influence the activity of EE by increasing the activity of LOX and β-GLU and decreasing the activity of POX and PPO.

Although essential for enhancing Y and overall quality of produced VOO, malaxation is also the limiting factor in the VOO production process. It disrupts process continuity, and significantly contributes to production costs due to the substantial energy required to maintain the olive paste temperature [7]. For these reasons, the shortening of the malaxation phase by optimizing the process parameters of conventional VOO production (temperature and time) and, more recently, the application of innovative technologies such as flash thermal treatment, ultrasound (US) and pulsed electric field, which could potentially replace malaxation, are some of the possibilities investigated [8]. Application of high-power US in the production of VOO induces both thermal and mechanical effects [9]. The thermal effect results from the conversion of US wave energy into heat, which reduces the viscosity of the olive paste and increases coalescence of oil droplets. The mechanical effect is primarily driven by acoustic cavitation, characterized by the formation, growth to a critical size and implosion of gas bubbles, causing localized increases in temperature and pressure [2]. Cavitation induces destabilization of the plant cell structure and disruption of the cell membrane and wall, thereby enhancing mass transfer and subsequently increasing Y and bioactive compounds [3,9,10]. The incorporation of US in the VOO production process offers significant advantages in terms of environmental sustainability. It contributes to pollution reduction by decreasing the volume of by-products and facilitating their disposal, due to the lower concentration of phenolic compounds, which can be ecotoxic at high levels. Furthermore, US enhances energy efficiency and reduces production costs by shortening or potentially eliminating the malaxation phase [7].

To date, the application of ultrasound has been studied on Spanish, Italian, Turkish and Tunisian olive varieties [9,10,11,12,13,14,15,16,17,18]. However, autochthonous Croatian varieties originate from different geographical and climatic Mediterranean regions - Istria (45°04’ N; 13°79’ E) and Dalmatia (43°55’ N; 16°35’ E) [19] and have not yet been investigated in this context. Varieties Istarska Bjelica, Rosulja, Oblica and Levantinka are the leading Croatian varieties and represent the core of national olive oil production. These varieties are genetically and chemically distinct types of olives that have not yet been sufficiently researched in the context of new processing technologies. Examining the response of different varieties to ultrasound contributes to a broader understanding of cultivar-specific reactions to innovative processing methods and offers insights applicable to similar Mediterranean growing regions, supporting the development of optimized, variety-specific approaches for VOO production. Therefore, the aim of this study was to evaluate the effects of ultrasound application prior to malaxation on Y, oxidative stability (OSI), antioxidant capacity (AC) and the composition of phenols, tocopherols, volatile compounds and fatty acids in VOOs obtained from Istarska Bjelica, Rosulja, Oblica and Levantinka.

2. Materials and Methods

2.1. Plant Material

Olive fruits (Olea europaea L.) of four different autochthonous Croatian varieties were used for this experiment. Oblica and Levantinka came from the Dalmatia region from the orchard of the Institute for Adriatic Crops and Karst Reclamation in Split, while Rosulja and Istarska Bjelica came from the Istria region from the agricultural holding in Bale. The olive fruits were harvested in mid-October and early November 2022, with a ripening index of 1.75 for Oblica, 2.01 for Levantinka, 2.19 for Rosulja and 1.32 for Istarska Bjelica. The ripening index of the harvested olives was determined based on the degree of pigmentation of the skin and pulp according to the method of the International Olive Council [20].

2.2. Virgin Olive Oil Production

The olives were processed within 24 hours of harvest, and VOO was produced on laboratory scale using an Abencor MC2 system (Ingenierías y Sistemas, Seville, Spain) which simulates the industrial process of VOO production. After cleaning and washing, approximately 800 g of olives were crushed into olive paste using a metal hammer mill (MM-100) equipped with a 4.5 mm sieve and operating at 3000 rpm. The paste was then subjected to US treatment, following the parameters defined by the central composite plan varying the duration and US power. The Sonorex Digiplus US bath (BANDELIN electronic, Berlin, Germany) with a maximum power of 640 W, a voltage of 120 - 240 V and a frequency of 20 kHz was used for the treatment. During the US treatment, the olive paste was constantly stirred to ensure uniform exposure. After the US pretreatment, malaxation was carried out at 27 °C for 40 minutes in a thermos beater (TM-100). The oil was separated on a CF-100 vertical centrifuge (90 s, 3500 rpm) and additionally clarified by a second centrifugation (4 min, 5000 rpm, 18 °C) on a Rotina 320R centrifuge (Hettich, Tuttlingen, Germany). The oils obtained were stored in dark glass bottles under a nitrogen atmosphere at a temperature not exceeding 20 °C until further analysis.

2.3. Oil Yield

The oil yield (Y, %) was calculated using the following equation (1) according to Peres et al. [21]:

$$Y={\displaystyle \frac{Vvoo\times \rho }{m}}\times 100,$$

(1)

where Vvoo is the volume of VOO produced (mL), ρ is the density of VOO (0.915 g/mL) and m is the mass of olive paste used for oil extraction (g).

2.4. Basic Quality Parameters

The basic quality parameters (peroxide value, free acidity and UV absorption characteristics - K₂₃₂, K₂₆₈ and ΔK) of the oils produced were measured according to the ISO standard methods [22] and the method of the International Olive Council [23,24].

2.5. Determination of Fatty Acid Composition

The fatty acid composition was determined by gas chromatography (GC) according to ISO standard method [25]. A solution of methyl pentadecanoate in octane (C15 methyl ester, γ = 1.25 μg/mL) was used as an internal standard, replacing isooctane in the sample preparation protocol. Prepared fatty acid methyl esters (FAMEs) were analyzed using a gas chromatograph equipped with a flame ionization detector (GC-FID; Agilent 6890N Network GC System, Agilent Technologies, Santa Clara, CA, USA), following the procedure described by Kraljić et al. [26]. Fatty acids were identified by comparing their retention times with those of a commercial standard mixture (FAME 37, Supelco, Bellefonte, PA, USA). Quantification of individual fatty acids was performed using internal standard method.

2.6. Enzyme Activity Assay

2.6.1. Olive Paste and Acetone Powders

Olive paste was used to isolate and determine LOX activity, while acetone powder was prepared and used for extraction of β-GLU, PPO, and POX. After malaxation, 30 g of olive paste was sealed in a plastic bag, frozen in liquid nitrogen and stored in a freezer until further analysis. An additional 90 g of the paste was used for the preparation of acetone powder following the modified method described by Romero-Segura et al. [27]. To prepare the powder, 1200 mL of cold acetone (- 20 °C) was added to the paste and the mixture was homogenized for 2 minutes using the GLH 850 homogenizer (Omni International, Kennesaw, GA, USA) at 15,000 rpm. The homogenate was vacuum filtered and the resulting powder was subjected to two additional homogenizations, each with 250 mL of cold acetone. The final residue was rinsed with 50 mL of diethyl ether, sealed in a plastic bag, frozen in liquid nitrogen and stored in a freezer until further analysis.

2.6.2. Lipoxygenase Extraction and Activity Assay

LOX was isolated according method by Luaces et al. [28] with minor modifications. Briefly, 5 g of olive paste was homogenized in 20 mL of extraction buffer (100 mM phosphate buffer, pH 6.7) using GLH 850 homogenizer at 11,000 rpm during two 1-minute intervals. The homogenate was vacuum filtered through a double layer of Miracloth and subsequently centrifuged at 27,000 g for 30 minutes at 4 °C. The resulting supernatant was used as crude enzyme extract.

LOX activity was determined according to the protocol described by Soldo et al. [29]. Linoleic acid (25 mM), prepared as described in our previous work [30], was used as a substrate. The reaction mixture consisted of 4.0 mL of 0.1 M sodium phosphate buffer (pH 6.0), 500 μL of linoleic acid, and 500 μL of enzyme extract. The reaction was magnetically stirred at 25 °C for 30 minutes and terminated by adjusting the pH to 2.0 with 3 M HCl. Butylated hydroxytoluene (BHT; 1 mL of 0.25 mM) was added as an internal standard.

The hydroperoxides were extracted in three successive extractions with 10 mL hexane:isopropanol solution (95:5 v/v). Following the final extraction, the mixture was centrifuged at 5000 rpm for 10 minutes. The hexane phases were combined and evaporated to dryness at 40 °C under reduced pressure (<200 mbar) using a rotary evaporator (Rotavapor, Heidolph Instruments GmbH & Co., Schwabach, Germany). The dried residue was reconstituted in 1.5 mL of acetonitrile:water (67:33, v/v) and sonicated for 10 seconds in an ultrasonic bath.

For analysis, 10 μL of the prepared extract was injected into a HPLC system (Agilent Technologies 1200 Series, Santa Clara, CA, USA) equipped with a diode array detector (DAD). Separation was achieved using a reverse-phase C18 column (Luna, 250 mm × 4.6 mm, 5 μm particle size, 100 Å pore size; Phenomenex, Torrance, CA, USA) maintained at 35 °C. The mobile phase consisted of solvent A (0.25% acetic acid in water) and solvent B (acetonitrile), with a constant flow rate of 1.0 mL/min. The following gradient program was applied: 63% B at 0 min, held until 17 min; increased linearly to 80% B by 20 min; maintained at 80% B until 32 min; and returned to initial conditions by 35 min, followed by a 10-minute equilibration period. Detection was performed at 234 nm. Hydroperoxy-octadecatrienoic acids (HPOTs) were identified by comparing retention times and UV spectra with those of a commercial 13(S)-hydroperoxy-9(Z),11(E),15(Z)- octadecatrienoic acid standard. Quantification was carried out using an internal standard method based on the response factor ratio between HPOTs and BHT. LOX activity was defined as the amount of HPOT in μmol synthesized by 1 mg of enzyme per minute at 25 °C.

2.6.3. β-Glucosidase Extraction and Activity Assay

β-GLU was isolated, and its activity was determined according to the modified method of Romero-Segura et al. [27]. Enzyme extracts were prepared by mixing 3 g of acetone powder and 12 mL of extraction buffer (5 mM ethylenediaminetetraacetic acid (EDTA), 1 mM phenylmethanesulfonylfluoride (PMSF) and 0.25 % (w/v) dithiothreitol in 100 mM borate buffer, pH 9) on a magnetic stirrer in an ice bath for 1 h. The resulting homogenate was then centrifuged at 27,000 g for 30 min at 4 °C and the clear supernatant was used as crude extract for determination of enzyme activity. The reaction medium for β-GLU activity consisted of 1.5 mL oleuropein (5 mM in 50 mM sodium acetate buffer pH 5.5) and 500 μl enzyme extract. The reaction lasted 30 minutes at 45 °C, after which it was stopped by adding 1.5 mL methanol. The prepared samples were injected (20 μl) into an HPLC system (Agilent Technologies LC 1200, Santa Clara, CA, USA) and analyzed isostatically on a non-polar C18 column (Luna 250 mm × 4.6 mm, 5 μm, 100 Å, Phenomenex, Torrance, CA, USA). A 0.1 % solution of formic acid in water/methanol (V/V, 1:1) at a constant flow rate of 1 mL/min was used as the mobile phase. The chromatograms were recorded using a DAD detector at 280 nm. The oleuropein concentration before and after the reaction was calculated using the corresponding calibration curve. The β-GLU activity was defined as the amount of residual oleuropein in μmol after reduction by 1 mg of enzyme per minute.

2.6.4. Polyphenol Oxidase Extraction and Activity Assay

The enzyme extracts were prepared according to the modified method of Peres et al. [31]. Enzyme extracts were prepared by mixing 3 g of acetone powder, 2.5 % polyvinylpyrrolidone and 12 mL of extraction buffer (0.05 M potassium phosphate and 1 M KCl, pH 6.2) on a magnetic stirrer in an ice bath for 1 h. The resulting homogenate was then centrifuged at 27,000 g for 30 min at 4 °C and the clear supernatant was used as crude extract for determination of enzyme activity. The reaction medium consisted of 2,5 mL of catechol (c = 30 mM) and 500 μL of enzyme extract. The activity of PPO was determined spectrophotometrically by measuring the increase in absorbance at 420 nm associated with the oxidation of catechol after one minute of reaction at room temperature according to the method described by Caponio et al. [32]. One unit of PPO activity was defined as the amount of o-quinone in μmol (ε = 1623 M/cm) produced by 1 mg of enzyme per minute.

2.6.5. Peroxidase Extraction and Activity Assay

The enzyme extracts were prepared in the same way as described in section 2.6.4. POX activity in the extracts was determined spectrophotometrically based on the increase in absorbance at 470 nm, which is due to the formation of tetraguaiacol (ε = 26,600 M/cm) by peroxidation of guaiacol according to the modified method of Caponio et al. [32]. 2 mL of guaiacol (c = 30 mM) and 1 mL of H202 (30%, c = 4 mM) were added simultaneously and the absorbance was measured after 10 minutes of reaction at room temperature. One unit of POX activity was defined as the amount of tetraguaiacol in μmol produced by 1 mg of enzyme per minute.

2.6.6. Total Protein Content

The amount of total proteins in enzyme extracts was determined using the Bradford reagent with crystalline bovine serum albumin (BSA) as the standard protein [33]. In brief, 300 μL of the enzyme extract was added to 1.2 mL of Bradford reagent, and the absorbance was measured against a blank after a 5-minute reaction at room temperature at 595 nm.

2.7. Determination of Volatile Compounds

The composition and content of volatile compounds were determined by gas chromatography coupled with mass spectrometry (GC/MS) using the 6890N Network GC System and the 5973 inert Mass Selective Detector (Agilent Technologies, Santa Clara, CA, SAD) according to the method of Kraljić et al. [26] with some minor modifications. Samples were prepared in a 20-mL glass vial by weighting 0.015 g of internal standard (0.15 % solution of 4-methyl-2-pentanol in refined sunflower oil) and 10 g of VOO. The vial was sealed with a silicone septum and placed in a heating block at 40 °C with a magnetic stirrer (Pierce Reacti-Therm Heating/Stirring Module, Artisan, Champaign, IL, USA). After 10 minutes of sample conditioning, the SPME fiber (a 2 cm long Divinylbenzene/Carboxen/Polydimethylsiloxane fiber; Supelco, Bellefonte, PA, USA) was exposed to the headspace for 30 minutes and immediately desorbed in the GC injector in splitless mode at 260 °C for 1 minute. The volatile compounds were separated using a HP-5 capillary column (30 m×0.25 mm×0.25 μm; Agilent Technologies, Santa Clara, USA) and helium as the carrier gas at a flow rate of 1.5 mL/min. The initial oven temperature of 30 °C was maintained for 3 minutes. Then the temperature increased to 150 °C at a rate of 5 °C/min and further to 250 °C at a rate of 20 °C/min. The final temperature was maintained for 5 minutes. The mass spectrometer was operated in electron ionization (EI) mode scanning the mass range m/z 50-550. The ion source and transfer line temperatures were set at 250 °C and 260 °C, respectively. The mixture of standard n-alkanes (C8–C20) in pure dichloromethane was run under the same conditions as the samples to calculate the Kovats retention index values. The volatile compounds were identified by comparing their retention times and mass spectra with those of the standards and with the mass spectra in the National Institute of Standards and Technology (NIST) mass spectral library databases. They were also identified by comparing Kovats indices from the same library or from the literature. The quantification of the compounds was performed considering the response and the concentration of the internal standard.

2.8. Determination of Phenolic Compounds

The composition of the phenolic compounds was determined using the RP-HPLC method. The extraction of phenolic compounds was carried out according to the standard method of the International Olive Council [34] with syringic acid as internal standard. The prepared methanolic extracts (20 μl) were analyzed on HPLC according to the method described by Škevin et al. [35]. The concentration of total and individual phenols was expressed as mg tyrosol/kg oil.

2.9. Determination of Tocopherols

The tocopherol content was analyzed according to ISO standard method [36]. Separation and detection of tocopherols was performed according to Kraljić et al. [26]. The retention time of the commercial standards was used to identify the tocopherols in the sample. The concentration of each tocopherol detected in the sample was expressed by the concentration of α-tocopherol (mg/kg).

2.10. Determination of Oxidative Stability

The oxidative stability was determined by measuring the induction time on a differential scanning calorimeter (DSC) 214 Polyma (NETZSCH-Gerätebau GmbH, Selb, Germany) according to the modified method by Tan et al. [37]. 4.0 ± 0.3 mg of VOO was weighed into an aluminum pan and hermetically sealed with a lid with a pinhole in the middle. It was then placed in the sample chamber of the instrument. The device was calibrated with high-purity indium and the baseline was determined with an empty aluminum pan. The isothermal method was used. The sample was heated to 140 °C at a rate of 20 °C/min under constant nitrogen flow (40 mL/min). After 5 minutes of equilibrium at 140 °C, nitrogen was replaced by purified oxygen (99.95 %, 100 mL/min). Nitrogen was used as a protective gas throughout the analysis at a flow rate of 60 mL/min. The DSC curves were evaluated using the NETZSCH Proteus Thermal Analysis 8.0.1 program. The induction time was determined at the intersection of the extrapolated baseline and the tangent of the exothermic part of the recorded exotherm.

2.11. Determination of Antioxidative Capacity

The antioxidant capacity of VOO was determined by measuring the percentage of reduction of 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical using electron paramagnetic resonance (EPR) according to the modified method of Markić et al [39]. The measurements were carried out at room temperature using the Magnettech MS – 5000 (Freiberg instruments GmbH, Freiberg, Germany) with the following parameters: frequency of magnetic field modulation - 100 kHz, magnetic field strength - 331–343 mT, field sweep range - 12 mT, sweep time - 30 s, microwave power - 10 mW and modulation amplitude - 0.2 mT. 20 μl of VOO was added to 980 μl of DPPH˙ solution (0.15 mM in ethyl acetate), mixed vigorously and EPR spectra were recorded after exactly 30 minutes of the reaction in the dark. The blank sample was prepared by adding 20 μl of ethyl acetate instead of the VOO sample. The amplitudes of the EPR signals of the DPPH˙ were calculated by integration of the EPR spectra using EW (EPRWare) Scientific Software Service. The reduction of amplitude of the DPPH˙ (%) was calculated according to the following formula (2):

$$DPPH\dot{}reduction\left(\%\right)={\displaystyle \frac{A0-A30}{A0}}\times 100,$$

(2)

where A0 is amplitude of the blank sample and A30 is the amplitude of the EPR signal of DPPH˙ in the VOO solution measured after 30 minutes of reaction time.

2.12. Statistical Analysis

The VOO samples were produced with the application of US pretreatment according to the central composite experiment design with two independent variables: treatment duration (min) and ultrasonic bath power (W). The experimental matrix, shown in Table 1, consisted of 10 experimental conditions, including a control with no US application (0 min, 0 W) and a central point which was replicated 5 times to enable the estimation of experimental error and assess the reproducibility of the system. The treatment duration ranged from 3 to 17 minutes, while US power varied between 256 and 640 W (axial points). A one-way analysis of variance (ANOVA) was performed to test the influence of variety. Significant differences between the control samples of different olive varieties were determined using the (two-tailed) Tukey‘s post hoc test at p ≤ 0.05.

Response surface methodology (RSM) was employed to understand the effects of US pretreatment on key quality and compositional parameters of VOO. A two-factor interaction (2FI) models were developed to describe the influence of US power and treatment time across different olive varieties on Y, OSI, AC, phenolic and tocopherol content, and linoleic acid concentration. The empirical equation (first order multiple regression equation) was developed (3):

$$\begin{gathered} \mathrm{R}\mathrm{e}\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{e}=\mathsf{\beta }0+\mathsf{\beta }1\times \mathrm{A}+\mathsf{\beta }2\times \mathrm{B}+\mathsf{\beta }3\times \mathrm{C}1+\mathsf{\beta }4\times \mathrm{C}2+\mathsf{\beta }5\times \mathrm{C}3+\mathsf{\beta }6\times \mathrm{C}4+\mathsf{\beta }7\times \mathrm{A}\times \mathrm{B}+ \\ \mathsf{\beta }8\times \mathrm{A}\times \mathrm{C}1+\mathsf{\beta }9\times \mathrm{A}\times \mathrm{C}2+\mathsf{\beta }10\times \mathrm{A}\times \mathrm{C}3+\mathsf{\beta }11\times \mathrm{A}\times \mathrm{C}4+\mathsf{\beta }12\times \mathrm{B}\times \mathrm{C}1+\mathsf{\beta }13\times \mathrm{B}\times \mathrm{C}2+ \\ \mathsf{\beta }14\times \mathrm{B}\times \mathrm{C}3+\mathsf{\beta }15\times \mathrm{B}\times \mathrm{C}4, \end{gathered}$$

(3)

where A=time, B=power, C₁=Istarska Bjelica variety, C₂=Rosulja, C₃=Oblica and C₄=Levantinka. The predicted response correlated to the set of regression coefficients (β): the intercept (β0), linear (β1-β6), and interaction coefficients (β7-β15). Analyses were performed using XLSTAT statistical software (Lumivero, Denver, CO, USA). The results are shown as mean ± standard deviation. All analyses were performed in at least two replicates.

3. Results and Discussion

The application of US in VOO production has received growing attention as a technology potentially capable of enhancing the release of intracellular components and improving oil functionality [16]. In this study, the effects of US pretreatment in the production of VOO in laboratory conditions were evaluated across four Croatian olive varieties, focusing on both compositional changes and biochemical responses. All samples, regardless of treatment, met the official criteria for extra virgin olive oil under EU regulations [38]. Peroxide values ranged from 1.0 to 3.0 meq O₂/kg, free fatty acids from 0.19 to 0.42% oleic acid, and UV absorption indices (K₂₃₂, K₂₆₈, ∆K) remained within acceptable limits. As in previous studies on both industrial and laboratory scale [9,10,15,16,17], no significant differences were observed between control and US-treated samples for these basic quality parameters. Therefore, they are not further discussed in this paper, allowing greater emphasis on the more sensitive indicators of oil quality and technological response. Furthermore, given that olive variety strongly determines the chemical composition, phenolic content, and antioxidant capacity of VOO, and thus its specific response to US-assisted extraction [39], it is essential to first discuss the intrinsic differences among the studied varieties.

3.1. Influence of Olive Variety

Differences among Istarska Bjelica, Rosulja, Oblica, and Levantinka were significant for nearly every chemical and biochemical parameter measured, confirming previous studies highlighting the importance of genotype in oil quality outcomes [19,40].

3.1.1. Oil Yield

From the results shown in Figure 1 and models detailed in Table 6, it is evident that the olive variety was the main factor that significantly influenced Y (p ≤ 0.001). Specifically, Levantinka exhibited the highest Y, while Rosulja showed the lowest. Similar results were reported in the literature for Levantinka (20 %) and Rosulja (5 and 8 %) [41,42,43]. In contrast, Y of 18–21 % and about 21 % for laboratory production were usually reported for Oblica and Istarska Bjelica, respectively [19,41,44], which corresponds to values 1.5 to 2 times higher than those observed in this study. However, lower Y comparable with our results were reported in some of the previous studies [42,45,46]. The lower Y can be most probable attributed to the pronounced drought in 2022 during the pit hardening and intensive fruit growth compared to previous years, and also to differences in ripening indexes, with Istarska Bjelica and Oblica having somewhat lower ripening indexes than the other two varieties [41,46,47,48]. Moreover, excess rain was observed near harvest in Istria that could have increased fruit moisture and diluted Y during extraction [49]. Nevertheless, it is well established that Y is affected by factors such as fruit ripeness, climatic conditions, agronomic practices, and technological parameters applied during processing [50,51].

3.1.2. Fatty Acid Composition

The fatty acid composition of VOO is strongly influenced by the variety, the degree of ripeness and the geographical origin, especially the latitude and climatic conditions [52]. Table 2 shows the main fatty acids in the VOO samples and includes fatty acid groups (saturated fatty acids – SFA, monounsaturated fatty acids – MUFA and polyunsaturated fatty acids - PUFA).

The results for fatty acid composition of the VOOs produced in this study were consistent with previous studies [19,45,53]. MUFAs dominated the fatty acid profile, accounting for 64–75% of total fatty acids, of which oleic acid accounted for approximately 98%. Due to the high content of oleic acid, VOO carries the health claim “high unsaturated fat”, which indicates that this food contributes to the maintenance of normal blood cholesterol levels [54]. The detailed fatty acid profiles of different olive varieties further elucidate the variability in oil composition and its implications for nutritional quality and oxidative stability. The MUFA/PUFA ratio serves as a critical indicator of the oxidative stability of VOOs, reflecting their resistance to oxidative degradation and, consequently, shelf life under appropriate storage conditions [19]. In the present work, Levantinka and Rosulja exhibited MUFA/PUFA ratio of around 10, Istarska Bjelica 8, and Oblica had the lowest ratio of around 4, indicating greater susceptibility to oxidation. This is due to the fact that Oblica exhibited the lowest level of oleic acid, while displaying the highest concentrations of linoleic fatty acid. Notably, the study of Žanetić et al. [19] reported a MUFA/PUFA ratio for Oblica approximately twice as high, while the ratio for Levantinka was consistent with these findings. On the other hand, the research of Jukić Špika et al. [53] reported the same ratio for Oblica as in this research.

3.1.3. Endogenous Enzymes

Olive paste was used to isolate and determine LOX activity, while for β-GLU, POX and PPO proteins were first concentrated in an acetonic powder which was then used for their isolation (Table 3). EE play a very important role in the production of VOO, as they are responsible for its quality and nutritional value. As mentioned, during the crushing phase of VOO production, the structure of the plant cells breaks down and the chemical substances such as triacylglycerols, fatty acids, pigments and phenols (substrates for EE) are released. In the next phase of production (the malaxation phase), the enzymes are activated and convert these compounds into volatile and phenolic compounds, which are responsible for the nutritional, sensory and antioxidant properties and oxidative stability of the VOOs [2].

Olive variety had a significant influence on the overall enzyme activity. In the control samples, the highest LOX activity was found in Rosulja, which had the highest maturity index (2.19) compared to the other three olive varieties. This is consistent with the literature [55,56], which shows a significant increase in LOX activity in olive fruits as ripening progresses. Compared to the research of Soldo et al. [55], this study found a 2 to 3 times higher LOX activity in Oblica and Levantinka, respectively, which is related to their higher ripening indexes. The activity of β-GLU was detected exclusively in Rosulja, which exhibited the highest maturity index among the studied varieties. This is in agreement with the study of Susamci et al. [57] who observed that β-GLU activity increases with fruit ripening; however, this pattern can vary considerably depending on the olive variety. In this study, β-GLU activity was measured in acetone powders, following the procedure described by Romero-Segura et al. [27]. While acetone precipitation effectively concentrates proteins, it can also lead to partial inactivation due to the high sensitivity of β-GLU to organic solvents [58,59]. This likely explains the absence of detectable activity in the other varieties. For more accurate assessment in future studies, it would be advisable to determine β-GLU activity directly in olive paste, as is commonly done for LOX analysis. No PPO activity was detected in the control sample in Rosulja, while the highest activity was recorded in Istarska Bjelica, the opposite of what was the case with LOX. This is also confirmed by the literature, which states that the activity of PPO and POX decreases as the olive fruit ripens [60]. POX activity, on the other hand, was only detected in Levantinka, but it was very low, similar to previous reports [31,32].

3.1.4. Volatile Compounds

According to the literature [59], volatile compounds can be divided into products of the lipoxygenase pathway (ΣLOX), oxidation products (ΣOX) and products of microbiological activity (ΣMBA). The composition of volatile components of produced VOO is presented in Table 4. Of the detected compounds, 2-methylbutanal is identified as sole product of microbiological activity responsible for the fusty defect in VOO. 2,4-Hexadienal, 4-oxohex-2-enal, pentanal and nonanal are products of oxidation (ΣOX), while all other detected compounds originate from the lipoxygenase pathway (ΣLOX). The products of the lipoxygenase pathway contribute to the specific aroma and flavor of the VOO being responsible for the green and fruity aroma, while the oxidation products are responsible for the development of the undesirable off-flavor of rancidity. VOO is very resistant to non-enzymatic oxidation as it contains oleic acid and various antioxidants. The oxidation of VOO mostly occurs through autoxidation, a chain reaction triggered by radicals that leads to the loss of antioxidants and the formation of toxic free radicals and hydroperoxides [61]. The content of volatile compounds strongly depends on the variety, the degree of ripeness of the fruit, the geographical area of cultivation, the agronomic factors and the processing method [55]. In the control samples, aldehydes dominated among the volatile compounds in all olive varieties, with the exception of the Levantinka variety, in which the mass fraction of aldehydes and alcohols was approximately equal. Aldehydes were more abundant in the varieties from the Istria region, while alcohols were more abundant in the varieties from the Dalmatia region. The highest content of ketones was found in Istarska Bjelica and the highest content of esters in Levantinka. In addition, the varieties from the Istria region had the highest content of compounds from the LOX pathway and also the highest content of OX products, while the highest content of MBA products was found in Levantinka and Istarska Bjelica. Compared to previous reports [62], Rosulja showed a 5-fold lower total amount of aldehydes than in our study, while other volatile groups were present in the same amount. In contrast, in the study by Žanetić et al. [19], Oblica had a threefold higher content of aldehydes, a twofold higher content of alcohols and a slightly higher content of esters, while Levantinka had a fivefold higher content of aldehydes, an approximately equal content of alcohols and a fourfold lower amount of esters.

3.1.5. Phenolic Compounds and Tocopherols

Several phenolic compounds have been detected in the VOOs. These included phenolic alcohols (hydroxytyrosol and tyrosol), p-coumaric acid, hydroxytyrosol acetate, oleacein, methyl hemiacetal of oleocanthal, oleocanthal, oleuropein and ligstroside aglycones, as shown in Table 5. The phenolic composition is a characteristic of the olive variety but depends on many different factors such as the maturity index, the geographical area of cultivation, the climate and the agricultural and technological conditions [63]. Among the control samples of the four varieties analyzed, Istarska Bjelica proved to be the variety with the highest content of phenolic compounds. This variety had the highest concentration of total phenolic compounds (TPC), oleocanthal and ligstroside aglycones, while Levantinka had the highest concentration of phenolic alcohols. In addition, Rosulja had the highest concentration of methyl hemiacetal of oleocanthal, oleuropein aglycones and oleacein. Oblica, on the other hand, had the lowest TPC, oleuropein aglycones, methyl hemiacetal of oleocanthal, phenolic alcohols and oleacein, while having the highest concentration of p-coumaric acid. According to the research of Šarolić et al. [64] Levantinka exhibited half the TPC content as in this study. Oblica, on the other hand, showed a content twice as high. In the research of Franić et al. [45], Levantinka had a lower TPC content (279 mg/kg), similar to Oblica and Rosulja in this study, while Istarska Bjelica had a content twice as high as in this study (582 mg/kg).

According to EU Commission Regulation [54], olive oil may bear the health claim “Olive oil polyphenols contribute to the protection of blood lipids from oxidative stress” if the oil contains at least 5 mg of hydroxytyrosol and its derivatives per 20 g of olive oil. Of the varieties in our experiment, only Rosulja VOO can carry this health claim.

The concentration of α-tocopherol (the main tocopherol in VOOs) is also shown in Table 5. The variety had the greatest influence on its concentration. In the control samples, the highest content was found in Levantinka (367 mg/kg) and the lowest in Oblica and Rosulja (273 and 272 mg/kg, respectively). In the research of Jukić Špika et al. [65], Oblica had similar contents (284 and 263 mg/kg). For other varieties, lower levels of α-tocopherol were found in the literature than in this study. The study of Šarolić et al. [64] found a lower α-tocopherol content in Oblica (213 mg/kg) and Levantinka (222 mg/kg), which can be attributed to the higher ripening indexes as the study of Morello et al. [66] demonstrated that the α-tocopherol concentration decreases with increasing ripeness of the olive fruit. In the research of Koprivnjak et al. [67] both varieties from Istria also had a lower content than in our experiment, with Istarska Bjelica having a three times lower content. According to the EU Regulation on nutrition and health claims made on foods [68,69] food may bear the claim “Vitamin E contributes to the protection of cells from oxidative stress” if they contain at least 1.8 mg vitamin E per 100 g, which is the case for all VOOs in this study.

3.1.6. Oxidative Stability and Antioxidant Capacity

The oxidative stability of VOOs was described by the induction time, which is defined as the time in minutes during which the oil is resistant to oxidation at high temperatures. The higher the induction time, the higher the concentration of compounds with antioxidant properties in the oil, which is why the oil deteriorates more slowly and is more stable during storage. Components of the oil that contribute to oxidative stability are phenols, tocopherols and fatty acid composition [70]. Among the control samples (Figure 2), Rosulja had the highest oxidative stability, while Oblica had the lowest, corresponding to the lowest TPC content and the highest content of PUFAs, more specifically linoleic acid. In the research of Bilušić et al. [71] Oblica also had the shortest induction time among other Croatian varieties in that study.

The antioxidant capacity of the oil is described as the potential of the oil to protect the consumer from a range of oxidative reactions that lead to oxidative stress and the development of various diseases. Compounds with antioxidant properties such as phenols and tocopherols are responsible for the antioxidant capacity of VOO [11]. The antioxidant capacity in this study was assessed in vitro and is expressed as a percentage of the reduction of stable DPPH radical. The higher the percentage of reduction, the higher the concentration of antioxidants in the oil, so the oil is more stable over time, degrades more slowly and is healthier for consumers given the numerous proven health benefits of the above compounds [72]. As can be seen in Figure 3, the highest antioxidant capacity among the control samples was found in Rosulja and the lowest in Oblica, following the pattern of oxidative stability. Rosulja showed similar AC (75 %) in the research of Koprivnjak et al. [43], while Istarska Bjelica had higher AC (78 %) than in our study.

3.2. Influence of Ultrasound

It was previously reported in several studies that US affected Y, OSI, AC, phenolic and tocopherol content [9,10,12,15,16,17,73,74,75]. However, negligible effect of US on volatile compounds has been reported [14,15,17,76] or that it led to a slight decrease [9,16,17], which was explained by the increase in temperature and the inactivation of enzymes of the LOX pathway by acoustic cavitation. In addition, Yahyaoui et al. [18] reported no significant impact of US on LOX activity, while β-GLU showed a slight decrease in its activity. On the other hand, US significantly affected PPO and POX activities [2,3]. Mild or short US stimulated activity due to heat and micro-mixing effects, while higher power decreased enzyme activity or inhibited the enzyme. These results highlight the complexity of the effects of US technology on both the production yield and the quality of virgin olive oil.

To better understand the effects of US treatment on key quality and compositional parameters of VOO produced from Croatian olive varieties, RSM was employed. This statistical tool enables the evaluation of multiple variables and their interactions simultaneously, offering an efficient approach to model complex systems such as olive paste processing. In this study, a two-factor interaction (2FI) models were developed to describe the influence of US power (256 – 640 W) and treatment time (3 – 17 min). Results in Table 6 present model parameters for Y, OSI, AC, phenolic and tocopherol content, and linoleic acid concentration. The models for other parameters investigated were also developed, but were not significant, which is why they are not presented and discussed further here.

Table 6. Two-factor interaction model parameters (regression coefficients, p-value, coefficient of determination (R2) and lack of fit) for oil yield (Y), oxidative stability index (OSI), antioxidant capacity (AC), content of linoleic acid, total phenolic compounds and α-tocopherol content affected by US treatment duration (min), power (W) and olive variety.

Model parameter	Y (%)	OSI (min)	AC (% of DPPH˙ reduction)	Linoleic acid (mg/g)	TPC (mg/kg)	α-tocopherol (mg/kg)
Intercept	12.923	130.94	53.037	93.6	296.586	287.106
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001
Time	0.159	-3.908	-2.012	-0.166	-12.653	-2.557
p-value	0.093	0.004	<0,0001	0.603	0.000	0.375
Power	0.105	-0.879	-0.453	-0.649	-3.305	-0.228
p-value	0.247	0.477	0.272	0.037	0.275	0.935
Variety-Istarska Bjelica	-2.167	24.252	8.156	-9.462	147.029	13.317
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	0.001
Variety-Levantinka	7.05	7.46	-3.413	-27.823	-41.035	35.01
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001
Variety-Oblica	0.9	-92.602	-22.275	68.004	-179.161	-33.683
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001
Variety-Rosulja	-5.783	60.89	17.533	-30.719	73.168	-14.644
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	0
Time *Power	-0.184	-1.325	0.4	1.066	-3.992	16.688
p-value	0.164	0.462	0.505	0.019	0.365	<0,0001
Time *Variety-Istarska Bjelica	0.217	-1.066	-0.403	0.053	12.657	-6.509
p-value	0.182	0.63	0.585	0.923	0.021	0.193
Time *Variety-Levantinka	0.064	-1.879	-1.322	0.191	-8.421	-5.966
p-value	0.691	0.398	0.076	0.729	0.122	0.233
Time *Variety-Oblica	-0.131	1.542	-0.204	-1.087	-0.853	1.509
p-value	0.419	0.487	0.783	0.051	0.875	0.762
Time *Variety-Rosulja	-0.151	1.403	1.929	0.842	-3.382	10.966
p-value	0.352	0.527	0.01	0.129	0.532	0.03
Power *Variety-Istarska Bjelica	-0.276	-7.309	-2.812	0.493	-15.387	0.169
p-value	0.082	0.001	0	0.356	0.004	0.972
Power *Variety-Levantinka	0.115	5.803	1.782	1.131	8.789	-11.743
p-value	0.461	0.009	0.014	0.036	0.095	0.016
Power *Variety-Oblica	0.363	3.574	2.376	-3.343	9.826	-1.184
p-value	0.024	0.1	0.001	<0,0001	0.062	0.806
Power *Variety-Rosulja	-0.202	-2.068	-1.347	1.719	-3.229	12.757
p-value	0.199	0.336	0.061	0.002	0.537	0.009
R2	0.991	0.988	0.958	0.997	0.966	0.653
R2 adjusted	0.988	0.985	0.952	0.996	0.962	0.607

Table 6. Two-factor interaction model parameters (regression coefficients, p-value, coefficient of determination (R2) and lack of fit) for oil yield (Y), oxidative stability index (OSI), antioxidant capacity (AC), content of linoleic acid, total phenolic compounds and α-tocopherol content affected by US treatment duration (min), power (W) and olive variety.

Model parameter	Y (%)	OSI (min)	AC (% of DPPH˙ reduction)	Linoleic acid (mg/g)	TPC (mg/kg)	α-tocopherol (mg/kg)
Intercept	12.923	130.94	53.037	93.6	296.586	287.106
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001
Time	0.159	-3.908	-2.012	-0.166	-12.653	-2.557
p-value	0.093	0.004	<0,0001	0.603	0.000	0.375
Power	0.105	-0.879	-0.453	-0.649	-3.305	-0.228
p-value	0.247	0.477	0.272	0.037	0.275	0.935
Variety-Istarska Bjelica	-2.167	24.252	8.156	-9.462	147.029	13.317
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	0.001
Variety-Levantinka	7.05	7.46	-3.413	-27.823	-41.035	35.01
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001
Variety-Oblica	0.9	-92.602	-22.275	68.004	-179.161	-33.683
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001
Variety-Rosulja	-5.783	60.89	17.533	-30.719	73.168	-14.644
p-value	<0,0001	<0,0001	<0,0001	<0,0001	<0,0001	0
Time *Power	-0.184	-1.325	0.4	1.066	-3.992	16.688
p-value	0.164	0.462	0.505	0.019	0.365	<0,0001
Time *Variety-Istarska Bjelica	0.217	-1.066	-0.403	0.053	12.657	-6.509
p-value	0.182	0.63	0.585	0.923	0.021	0.193
Time *Variety-Levantinka	0.064	-1.879	-1.322	0.191	-8.421	-5.966
p-value	0.691	0.398	0.076	0.729	0.122	0.233
Time *Variety-Oblica	-0.131	1.542	-0.204	-1.087	-0.853	1.509
p-value	0.419	0.487	0.783	0.051	0.875	0.762
Time *Variety-Rosulja	-0.151	1.403	1.929	0.842	-3.382	10.966
p-value	0.352	0.527	0.01	0.129	0.532	0.03
Power *Variety-Istarska Bjelica	-0.276	-7.309	-2.812	0.493	-15.387	0.169
p-value	0.082	0.001	0	0.356	0.004	0.972
Power *Variety-Levantinka	0.115	5.803	1.782	1.131	8.789	-11.743
p-value	0.461	0.009	0.014	0.036	0.095	0.016
Power *Variety-Oblica	0.363	3.574	2.376	-3.343	9.826	-1.184
p-value	0.024	0.1	0.001	<0,0001	0.062	0.806
Power *Variety-Rosulja	-0.202	-2.068	-1.347	1.719	-3.229	12.757
p-value	0.199	0.336	0.061	0.002	0.537	0.009
R2	0.991	0.988	0.958	0.997	0.966	0.653
R2 adjusted	0.988	0.985	0.952	0.996	0.962	0.607

As mentioned earlier, Y was primarily influenced by olive variety. Although a slight increase in Y was observed with longer pretreatment times, neither treatment time (p = 0.093) nor power (p = 0.247) had a statistically significant main effect on Y (Table 6, Figure 4), suggesting that US treatment alone did not consistently enhance oil recovery across all conditions. However, a significant interaction between US power and the Oblica variety suggests that applying US can moderately enhance oil recovery in specific varieties, most likely by promoting cell wall disruption and facilitating oil release. Compared to the control, the application of 640 W US power for 10 min increased Y from Oblica by about 4%. These findings are consistent with previous reports emphasizing that varietal differences in cell structure and moisture or oil content play a dominant role in extractability [77]. Similar improvements in Y have been reported in other varieties subjected to US pretreatment before malaxation under industrial conditions, such as a 3.4% increase in Ogliarola Garganica with a maturity index of 2.82 [16], and increases of 4.4% and 4.6% in Nocellara and Coratina, respectively [15]. Additionally, a 4.3% increase was observed in Chemlali after a 10-minute US treatment [73].

The oxidative stability index (OSI) is a critical parameter reflecting the resistance of VOO to lipid oxidation and serves as a strong predictor of shelf life. The model (Table 6 and Figure 5) developed for oxidative stability index (OSI) exhibited a high level of predictive accuracy (R² = 0.988; adjusted R² = 0.985), indicating that the combined effects of US parameters and olive variety explain nearly all the observed variability in OSI. Among the US factors, treatment time had a statistically significant negative effect on OSI (p = 0.004), suggesting that prolonged sonication promotes oxidative degradation. This aligns with previous reports demonstrating that acoustic cavitation during US processing can induce localized heating and the formation of reactive oxygen species, accelerating oxidative processes [78]. In contrast, US power alone did not have a significant effect (p = 0.477), indicating that the duration of exposure is a more critical factor for oxidative damage than power intensity. These findings are supported by Iqdiam et al. [12], who observed a significant decrease in OSI after 8 and 10 minutes of both direct and indirect high-power US treatment compared to shorter exposure times (0–6 min), for two different olive varieties. They reported that 10 minutes of direct high-power US caused an approximate 8% reduction in induction time, with no significant changes observed for shorter durations (p > 0.05). This indicates that US begins to compromise oxidative stability only beyond a critical exposure threshold.

However, the negative interaction of power in Istarska Bjelica observed (Figure 5a) here suggests that even phenolic-rich cultivars [79] may be susceptible to oxidation when exposed to excessive US energy, confirming the narrow margin between beneficial extraction and detrimental degradation. Interestingly, the interaction of US power in Levantinka (Figure 5d) was significant and positive, suggesting that in this variety, moderate US energy may help release more antioxidants without causing substantial oxidative stress. This agrees with results from Servili et al. [15], who reported that under optimized conditions, in their case increased pressure, US can selectively enhance antioxidant extraction and improve oil quality. Together, these findings underscore the importance of variety-specific optimization of US parameters. Short to moderate treatment durations may offer benefits, while prolonged sonication appears to pose a risk to oxidative stability, particularly in more sensitive varieties.

The regression model for the content of linoleic acid (Table 6, Figure 6), displayed excellent predictive power (R² = 0.997; adjusted R² = 0.996). The models were created with mass fractions in mg/g of oil (not in percentages) so that small changes in concentration can be recognized and a higher precision and resolution of the model is achieved. US power emerged as a significant negative factor (p = 0.037), indicating that higher sonication intensity reduces linoleic acid levels, likely through oxidative degradation triggered by cavitation-generated radicals. In contrast, treatment time did not have a statistically significant main effect (p = 0.603), suggesting that long exposures alone may not significantly alter linoleic acid content unless combined with high power. This aligns with the observed decline in OSI, suggesting that US-induced oxidation may target unsaturated lipids like linoleic acid.

Interestingly, in the Oblica variety (Figure 6c), treatment with high-power ultrasound (p < 0.001) resulted in a significant decrease in linoleic acid content. This finding is not unexpected, as Oblica typically contains at least twice the amount of linoleic acid compared to the other varieties studied (Table 2). Linoleic acid is considerably more susceptible to oxidation than oleic acid, and high-power ultrasound (20 kHz, 150 W) has been shown to accelerate oxidative processes in oils. As reported by Chemat et al. [80], sonication of refined sunflower oil for just 2 minutes increased peroxide values from 5.38 to 6.33 meq O₂/kg, indicating the onset of oxidation. Although there were no significant immediate changes in fatty acid composition, the volatile profile in their study showed the formation of hexanal and limonene resulting from sonication.

The regression model for total phenolic compounds (Table 6, Figure 7) showed a strong fit (R² = 0.966; adjusted R² = 0.962), confirming that besides variety, US processing influences the phenolic profile of VOO. Among US parameters, treatment duration had a highly significant negative effect (–12.653 mg/kg per minute, p < 0.001), while US power was not a significant predictor. However, the interaction between sonication time and the variety Istarska Bjelica was significant and positive (p = 0.021), suggesting that this variety is less sensitive to phenolic losses at longer treatment times. Reports on US implementation in VOO production show that its effect can be highly dependent on the variety and treatment conditions. Several authors have also reported a significant increase in phenolic compounds following US treatment which was specific to chemical structure. Tamborrino et al. [9] observed a 41 % increase in TPC and 68 % increase in oleacein in the Peranzana variety, while oleocanthal increased only slightly (~10 %), and phenolic alcohols decreased. Oils extracted with the application of US from Ogliarola garganica variety analyzed by Taticchi et al. [16] showed TPC increase of 10 % which was related to oleuropein aglycon species. Servilli et al. [15] reported a 28 % increase in TPC in the Nocellara variety and notable increases in oleocanthal in both Nocellara and Coratina varieties (39 % and 19 %, respectively), along with increases in ligstroside and tyrosol. Conversely, Gila et al. [17] found that phenolic content was generally not affected by US treatment and even observed a slight decrease in Arbequina oils, which was attributed to moderate temperature rises during treatment.

On the other hand, Iqdiam et al. [12] reported that after 10 minutes of direct high-power US treatment, total polyphenol levels were significantly reduced, by approximately 19% in Arbequina and 22% in Frantoio, indicating a substantial loss of antioxidants under extended sonication. This is consistent with the findings of Jiménez et al. [74], who reported that US exposure can induce phenolic degradation. These differing findings can be explained by the dual role of US. On one hand, US enhances the extraction of phenolic compounds by disrupting cellular structures and facilitating the release of bound phenolics [3]. On the other hand, it may simultaneously promote oxidative processes, leading to the consumption of these compounds as they act as primary antioxidants to neutralize US-induced free radicals [81]. This mechanism is supported by the observed decreases in OSI and linoleic acid content, suggesting a coupled degradation pathway whereby phenolics are depleted to counteract oxidation, subsequently facilitating the degradation of unsaturated fatty acids and reducing overall oil stability. The results in Istarska Bjelica imply that brief US treatments can still effectively enhance phenolic content without triggering significant oxidative losses, underscoring the importance of carefully optimizing sonication duration to balance these competing effects.

The regression model for α-tocopherol content (Table 6, Figure 8) (R² = 0.653; adjusted R² = 0.607) explained a moderate proportion of the variability, reflecting the complex behavior of tocopherols under US processing. Treatment time and US power alone were not significant, but their interaction was highly significant (p < 0.0001), indicating that the combination of higher US intensity and longer exposure can strongly affect tocopherol levels probably due to enhanced release from cellular matrices, as observed in Rosulja (24% increase). Taticchi et al. [16] and Pagano et al. [10] also observed an increase in α-tocopherol concentration (11%) in Ogliarola garganica, while Gila et al. [17] reported no effect of sonication process on tocopherol content. In present study, US showed declined α-tocopherol content in all investigated varieties except aforementioned Rosulja (Figure 8).

The regression model for antioxidant capacity (AC), expressed as % DPPH radical reduction (Table 6 and Figure 9), demonstrated strong explanatory power (R² = 0.958; adjusted R² = 0.952). Time had a highly significant negative effect (p < 0.0001), indicating that longer ultrasound (US) exposure reduced AC. This effect is likely due to temperature increases during sonication, which can lead to the degradation of antioxidant compounds [17]. In contrast, ultrasound power alone did not have a statistically significant effect (p = 0.272). However, the interaction between power and the Istarska Bjelica variety showed a significant negative influence (p < 0.001), suggesting that higher US power reduces AC in this specific cultivar. These findings align well with the observed effects of US parameters on total phenolic content (TPC), the primary antioxidants in virgin olive oil (VOO). Similarly, Aydar et al. [75] reported a decrease in AC with prolonged US treatment in the Edremit cultivar, attributing this to a reduction in TPC. Conversely, in the Rosulja variety, the interaction with treatment time had a significant positive effect on AC, potentially linked to an increase in α-tocopherol content observed with extended US exposure.

4. Conclusions

This study investigated the effects of ultrasound (US) pretreatment on oil yield, endogenous enzyme activity, and virgin olive oil (VOO) quality in four Croatian olive varieties: Istarska Bjelica, Rosulja, Oblica, and Levantinka. US was applied to olive paste prior to malaxation, with treatment durations ranging from 3 to 17 minutes and power levels between 256 and 640 W.

The effects of US were variety-dependent, with treatment time having a stronger influence than power. A moderate increase in oil yield (~4%) was observed only in Oblica. In general, prolonged US exposure reduced TPC and, consequently OSI and AC, except in Rosulja, where AC increased. Higher US power levels were associated with a decrease in linoleic acid content, indicating enhanced oxidative degradation, which also contributed to the reduction in AC observed in Istarska Bjelica. α-Tocopherol content was generally negatively affected by US, except in Rosulja, where extreme US conditions led to a significant increase, likely contributing to the observed rise in its AC.

Other quality parameters remained largely unaffected by US pretreatment. Given the risk of oxidative degradation and loss of bioactive compounds with extended sonication, short-to-medium US durations at low-to-moderate power are recommended. Optimization of treatment parameters should be tailored to each olive variety to ensure the preservation of VOO quality and functional properties.