Sunflower oil and virgin coconut oil were obtained from a local market in Badulla, Sri Lanka. F. inermis (Lovi) fruits were collected from the Western province of Sri Lanka. All of the reagents and solvents used were of analytical reagent grade.

Freeze-dried fruit peel was ground to a fine powder using a laboratory grinder. Dried, powdered peel was sieved (mesh size; 1 mm) and stored at -18 °C temperature until extraction. Freeze dried peel powder (20 g) was dissolved in 200 mL of 70% aqueous ethanol and extracted using an ultrasonic cleaner (model UC -10 A, 40 kHz, Heater 50 w, Tank 2 L, Tank size L 180 * W 165* H 180 cm, China) at 40 ± 5 °C for 30 minutes. The extraction procedure was repeated twice. The combined filtrate was concentrated in vacuo at 40 ± 5 ºC. Concentrated crude extract was stored at -18 °C until analysis.

Total phenol content was determined according to the method of ISO 14502-1 with slight modifications. First, crude extract was dissolved in 70% methanol to obtain a 60 mg·L^-1 solution. Ten-fold diluted Folin-Ciocalteu reagent (1.0 mL) was added to 200 μL plant extract and kept for 8 min. Then 800 μL of 7.5% (w/v) Na₂CO₃ (aq) were added and kept for 1 h. The absorbance was measured at a wavelength of 765 nm (UV /Visible SP-UV 50 VDB spectrophotometer). Gallic acid was used as a standard and a calibration curve was plotted using the absorbance values for gallic acid solution series. The phenolic content of plant extracts was calculated using the calibration curve and expressed as mg of gallic acid equivalents (GAE)/ g extract.

The crude extracts (150-325 ppm) were tested using the 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay (Ali et al., 2018). Each solution (50 µL) was mixed with 50 µL of 4 × 10^-4M methanolic DPPH Solution and kept in the dark at room temperature for 20 min. The absorbance was measured at 517 nm using a microplate spectrophtometer (Multiskan Go, Version100.40, Thermo Fisher Scientific) against a reaction blank (50 µL of Methanol and 50 µL of the crude extract). A mixture of 50 µL of 4 × 10^-4M DPPH solution and 50 µL of methanol were used as a reaction control. Percent inhibition was calculated for each concentration and IC₅₀was calculated for crude extract. The following equation was used to calculate percent inhibition:

Oil samples with crude extract were prepared at three different concentrations viz 500 ppm, 1000 ppm and 2000 ppm along with control (no added peel extract) and α-tocopherol (500 ppm) was used as positive control. First, oil sample was pre heated in a water bath at 50 °C for 3 hours. The crude extract of fruit peel was dissolved in a small amount of absolute ethanol (150 μL) to facilitate a uniform dispersion of extract in the oil. Oil samples with different concentrations of extracts were kept in an ultrasonic water bath at 60 °C for 30 min to obtain a homogenous dispersed sample. A storage study of the oils was conducted as per the method described by Besbes et al., 2004 with slight modifications. Treated oils were completely filled into glass bottles (volume 25 mL) wrapped with aluminum foil, sealed properly and stored in an oven at 65 ± 5 °C for 21 days. Oil samples were drawn for analysis at intervals of three days.

Peroxide values for the oil samples drawn at 3-day intervals were determined according to the standard procedures described as AOAC 965.33 (2016). Five grams of oil sample were dissolved in 30 mL of acetic acid: chloroform solution (3:2). Then, 0.5 mL of saturated KI solution were added and left to stand with occasional shaking for one minute and placed in the dark for 5 min. Then, 30 mL of distilled water were added and the mixture was titrated slowly with 0.01M sodium thiosulfate. A blank was made according to same procedure without the addition of oil. The peroxide value (meq O₂·kg^-1 of oil) of all samples was calculated according to the following equation:

The free fatty acid contents of the oil samples drawn at 3-day intervals were determined according to the procedure described as AOAC 940.28 (2016). An accurately weighed oil sample was dissolved in 20 mL of previously neutralized absolute ethanol. The mixture was titrated with 0.05 M sodium hydroxide with the addition of phenolphthalein as indicator with vigorous shaking until a permanent faint pink color appeared and persisted for more than 1 min. The percentage of free fatty acids of each oil sample was calculated as oleic acid using the following equation:

The absorption spectra of oil samples before and after 21 days of storage were measured on a FTIR spectrometer (Bruker, Alpha-T, German) using Deuterated Triglycine Sulphate (DTGS) as detector. They were interfaced to a computer operating under Windows-based, and connected to OPUS (Version 7.5) software. The samples were placed in contact with attenuated total reflectance (ATR) element (platinum) at controlled ambient temperature. FTIR spectra were collected in the frequency of 4000-500 cm^-1 by co-adding 64 scans and at resolution of 4 cm^-1. All spectra were rationed against a background of air spectrum.

Mean values of all data were obtained from triplicate determinations. Values were expressed are mean ± SD. Data were subjected to analysis of variance (ANOVA) according to the GLM (general linear model). Data processing was done by the Minitab 16.1 version of the computer software package.

The antioxidant efficacy of plant extracts can be tested using a wide variety of methods. In the present study, the free radical scavenging capacity of FIPE was evaluated by DPPH assay and expressed as IC₅₀. Results of the present study revealed that the antioxidant activity of FIPE (227.14 ± 4.12 µg·mL^-1) was significantly (p < 0.05) lower than that of α-tocopherol (29.80 ± 3.22 µg·mL^-1). Alakolanga et al., (2015) observed comparatively high antioxidant activity (66.2 ppm) in the ethyl acetate extract of F. inermis whole fruit. However, the methanolic extract of the reported study exhibited comparatively low antioxidant activity (212.95 ppm). The disparity of the results may be due to the difference in the extraction technique employed and the fruit material used in the two studies.

In the present study sonication was used as the extraction technique to extract natural antioxidants from the peel of lovi fruit as it is reported to enhance the extraction of bioactive compounds from plants and their products (Annegowda et al., 2010). The present study used 30 min sonication time for the extraction as prolonged sonication led to the decomposition of the bioactive constituents, and most of the active constituents inside the cells are released into the solvent during the first few minutes (Annegowda et al., 2010).

The total phenol content of FIPE was determine using the Folin-Ciocalteau colorimetric method. The total phenolic content of a sample was calculated based on the calibration curve prepared using gallic acid as standard. The total phenol content of FIPE was 4.87 ± 0.01 mg GAE per g extract. However, the study carried out by Alakolanga and others (2015) reported the polyphenol content of lovi fruit as 1.28 g gallic acid equivalents per 100 g of fresh fruit.

The extraction of phenolic compounds from plant materials is influenced by several factors: chemical nature of the plant materials, method of extraction, particle size of the sample (plant material), storage time and conditions, and the presence of interfering substances. The solubility of phenolic compounds is governed by the type of solvent (polarity of the solvent) used, the degree of polymerization of phenolics, the interaction of phenolics with other food constituents and the formation of insoluble complexes (Naczk and Shahidi, 2004). Thus, considering all these factors and the literature (Liew et al., 2018), the present study used 70% ethanol as the solvent for extraction.

Peroxide value measures the formation of peroxide in the early stages of lipid oxidation. The present study measured the variation in peroxide value to determine the effect of FIPE addition on the oxidative stability of VCO and SO during storage at 65 ± 5 °C. The peroxide values for the VCO samples treated with FIPE are shown in Figure 1 along with the control and positive control α-tocopherol. The PV of all the oil samples increased during storage and thereby indicate the progression of lipid oxidation as PV is a good indicator of the extent of primary oxidative products formed in oils. Several studies have proven that the PV of VCO increased during storage at 63 °C (Rohman et al., 2011). However, the initial PV of the reported study was lower (0.42 ± 0.009 meq O₂·kg^-1 oil) than that of the present study (1.20 ± 0.00 meq O₂·kg^-1 oil). The peroxide values of treated samples and control samples showed no significant difference (p >0.05) up to the third day of storage. From day 6 onwards the PV of the control and treated samples showed significant difference (p < 0.05); the PV value of the control sample was significantly (p < 0.05) higher than the 2000 ppm level of the extract-added VCO sample from day 6 onward. However, there was no significant difference between the control and 500 ppm and 1000 ppm levels of FIPE-added VCO samples after day 6 or day 18. The highest peroxide value (1.93 ± 0.12 meq O₂·kg^-1 oil) was observed for the control sample after 21 days storage.

Meanwhile, the lowest peroxide value (1.19 ± 0.01 meq O₂·kg^-1 oil) was reported for the VCO sample with 2000 ppm level of extract after 21 days of storage (Figure 1). There was no significant difference between the oil treated with α- tocopherol (1.39 ± 0.00 meq O₂·kg^-1 oil) and the oil treated with the 2000 ppm level of FIPE after 21 days of storage, which indicated excellent antioxidant capacity at higher concentration levels of FIPE.

The variation in the PV of sunflower oil treated with FIPE is illustrated in Figure 2. The PVs of all the tested oil samples were increased with time, which indicated the generation of hydroperoxides in sunflower oil (Szydłowska-Czerniak and Rabiej, 2018). The PVs of sunflower oil samples were higher than that of VCO owing to the high amount of unsaturated fatty acids present in SO which are more prone to oxidation (Wójcickiwt et al., 2015). The PV of oil samples treated with the 2000 ppm level of FIPE after 21 days of storage was 12.38 ± 0.46 meq/kg; while the control and α-tocopherol showed 28.30 ± 0.58 meq O₂·kg^-1 and 17.94 ± 0.35 meq O₂·kg^-1, respectively. However, there was no significant difference between the control and treatments until day 6. From day 9 onwards the PV of SO treated with the 2000 ppm level of FIPE was significantly (p<0.05) lower than that of the positive control, α-tocopherol. Moreover, there was no significant difference between SO with 500 ppm or 1000 ppm level of FIPE addition. However, the PV of the positive control was significantly lower than those of 500 ppm and 1000 ppm levels of FIPE-added SO samples, thus proving the efficacy of 2000 ppm level of FIPE addition in controlling the oxidation of SO.

Free fatty acid content is one of the important quality parameters in oils, as FFA is responsible for the development of undesirable flavor and aroma in oils. These free fatty acids formed in oils from hydrolysis, oxidation due to free radical formation and cleavage of double bonds during frying (Ma et al., 2015). The free fatty acid content of VCO used for the study was relatively low (0.26 ± 0.04%) indicating that the initial quality of VCO used for the study was good. However, the FFA content in the present study was higher than that of the reported study (0.16 ± 0.002%) (Rohman et al., 2011). The variation in FFA contents in the VCO samples treated with FIPE are shown in Figure 1 along with the control and positive control, α-tocopherol. The FFA content in VCO treated with FIPE at the 2000 ppm level was significantly (p < 0.05) lower than that of the control after day 3. However, there was no significant difference between the FFA content of the 2000 ppm level of extract-added oil and the positive control after 3 days of storage. Initially, FFA content was not significantly different from the control and treated oil samples. However, with time the FFA contents in the oil samples were increased as shown in Figure 1. After 21 days of storage the VCO (control) sample exhibited the highest FFA content (0.36 ± 0.00 %); while the 2000 ppm level of extract-added oil sample contained the lowest FFA content (0.28 ± 0.00%). In addition, after 21 days of storage the FFA content of 2000 ppm level of extract-added sample was significantly lower than that of the positive control (0.29 ± 0.00%).

The initial FFA content of SO (0.15 ± 0.00%) used for the study was lower than that of VCO (0.26 ± 0.04%). A comparatively high initial FFA content of VCO could be attributed to the presence of free fatty acids, chlorophyll or products of oxidative degradation which impair the quality of the oil (Ma et al., 2015). Many studies were conducted to assess the effect of the addition of natural antioxidants on the oxidative stability of SO (Carelli et al., 2005). The initial FFA content in the reported study (0.16%) was comparable to the present study. The FFA content of SO treated with the 2000 ppm level of extract was significantly (p < 0.05) lower than that of the control from day 3 onwards. However, there was no significant difference between oil treated with 2000 ppm level of extract and α-tocopherol until 21 days of storage. After 21 days of storage SO treated with the 2000 ppm level of FIPE exhibited the lowest FFA content (0.10 ± 0.04%); while oil treated with 500 ppm and 1000 ppm levels of extracts showed comparatively high FFA content (Figure 2). Moreover, there was no significant difference between the control and SO treated with 500 ppm and 1000 ppm extract levels after 21 days of storage.

The influx of research studies was carried out on the potential applications of FTIR in monitoring the oxidation of edible oils (Van de et al., 1994). Thus, the present study used FTIR spectroscopy to monitor the oxidative changes occurring in treated and non-treated sunflower oil and virgin coconut oil after 21 days of storage. Figure 3 illustrates the FTIR spectra of non-treated VCO, VCO treated with the 2000 ppm level of FIPE and VCO treated with α-tocopherol. Among other edible oils VCO has a unique spectrum, as shown in Figure 3. In the spectra of VCO no peaks are identified at the region of 3008 cm^-1 and 1655 cm^-1 which is in agreement with previous studies (Rohman and CheMan, 2011). Peaks in the above-mentioned region represent the unsaturated double bonds and are used to express the degree of unsaturation of triglycerides. Since VCO contains very low levels of unsaturated fatty acids, the peak is hardly noticeable in this region (Rohman and CheMan, 2011).

Peaks appearing at 721, 962, 1744, 2852, 2927 and 3300 cm^-1 were selected to compare the differences between the treated and non-treated VCO samples used for the study. All the VCO samples failed to show a peak at around 3444cm^-1 (Figure 03), hence no evidence was detected for the presence of hydroperoxides or free fatty acids which are formed in the thermal oxidation of oils. The carbonyl absorption of the triacylglyceol linkage is noticeable at 1744 cm^-1. However, a noticeable difference was not observed in VCO samples at this frequency (Figure 4 (a)). High intensity at 1744 cm^-1 represents the high amounts of carbonyl compounds such as aldehydes, esters, ketones, and lactones present in the oils which formed during oxidation (Smith et al., 2005). Moreover, added FIPE and α-tocopherol were not effective in lowering the intensity of carbonyl compounds (Figure 4 (a)). The peak at 962 cm^-1 represents the bending vibration of CH functional groups of trans olefins. This peak serves as indicator of oxidative stability as the peak provides the information relevant to gradual cis to trans isomerization in the oxidized oil (Marina et al., 2013). There was a minor difference in the peak intensities of VCO at 962 cm^-1; the highest peak intensity represented the VCO treated with α- tocopherol followed by the VCO treated with the 2000 ppm level of FIPE and VCO without any treatment (Figure 4 (b)). The peak revealed in the 721 cm^-1 region also displayed similar variation as the intensity variation in peak appeared in the 962 cm^-1 region; the VCO sample stored for 21 days without any treatment exhibited the lowest intensity. The peaks appeared in the 2927 cm^-1 and 2852 cm^-1 regions also exhibited minor differences (Figure 4 (c)). One of the important regions of the FTIR spectra for VCO is the region of 1500 to 1000 cm^-1, which is also known as the finger print region (Marina et al., 2013). There is no noticeable difference revealed in the spectra of treated and non-treated VCO in this region (Figure 4 (d)).

Figure 5 illustrates the FTIR spectra of non-treated SO, SO treated with the 2000 ppm level of lovi peel extract and SO treated with α-tocopherol. In the FTIR spectra of SO no peak is revealed at 3444 cm^-1, which represent the hydroperoxide functional group. With the progression of oxidation hydroperoxide formation increases and in relation to that, the intensity of the band revealed at 3444 cm^-1 also increases gradually (Hamed and Allam, 2006). However, the spectra of SO revealed a peak at 3008 cm^-1, which represents the unsaturated double bond (Rohman and Man, 2011). Since SO contains comparatively high levels of unsaturated fatty acids the peak appeared in this region, but that peak is not revealed in the spectra of VCO.

Figure 6 illustrates the magnification of spectral ranges 2000-1000 cm^-1, 1200-500 cm^-1, 3500-2500 cm^-1 and 1500-900 cm^-1 in the FTIR spectra of treated and non-treated SO samples. Similar to the VCO a significant difference was not observed between the treated and non-treated SO samples at 1744 cm^-1 (Figure 6 (a)). In non-oxidized oil this peak solely represents the ester carbonyl functional group of triglycerides and when the oil is oxidized, this peak overlaps with aldehydes, thereby producing a decrease in intensity (Guillén and Goicoechea, 2007). Figure 6 (b) illustrates the differences in peak intensities between treated and non-treated SO samples at 962 cm^-1. The positive control, α-tocopherol, exhibits the lowest peak intensity. While the FIPE-added sample shows the highest peak intensity. This peak represents the gradual cis to trans isomerization in the oxidized oil, which increases with the progression of oxidation (Marina et al., 2013). Thus, α-tocopherol is effective in controlling the oxidation of SO since it exhibits the lowest peak intensity. The peak at 722 cm⁻¹ was decreased with the progression of oxidation due to the loss of cis double bonds (Rohman and Che Man, 2013). The Figure 6 (b) illustrates the differences in peak intensities at 722 cm⁻¹. As illustrated in the figure, the SO sample treated with FIPE extract exhibits the highest peak intensity and thereby represents the lowest rate of oxidation. Two peaks at 2853 and 2923 cm⁻¹ remained unchanged. However, minor differences were observed in the peak intensities of treated and non-treated SO samples in the finger print region of the FTIR spectra (Figure 6 (d)).