Green olive fruits (Domat variety), with an average weight of 160-180 fruit/kg were harvested from the Olive Research Institutes orchard in Kemalpaşa, Izmir, Turkey, in October 2013 at 1 maturity index ánd by hand. The olives were then processed according to Spanish style table olive preparation method. The olives remained in the lye solution (1.9% NaOH) until 2/3 of the fruit flesh had been penetrated. Then the olives were washed with tap water for 18 h, to completely remove NaOH. After this pre-treatment, the olives were placed in brine (7 g/100 mL of NaCl) for three months at ambient temperature for fermentation to take place.

At the end of fermentation, the pH value (3.78), free acidity (0.88%) of the brines and L* (56.19), a* (0.59), b* (37.59), and hardness (535.32 g) of the olive fruits was measured.

An ozonation system was designed to apply ozone effectively to the table olives (Figure 1). This system was formed by a perforated nickel chromium basket where olive samples were placed and an application tank (50 L, KenaTek Machine, Izmir, Turkey). Two small diffusers were installed into the application tank to ensure a homogeneous distribution of the ozone gas. The concentration of ozone in the tank was measured by an ozone measurement sensor (Electrode SZ 283, ITALY) in the range between 0-20 ppm. Ozone gas was generated using a corona discharge ozone generator which generates ozone from the air stream (TEKNOZON the model TKZ-25 g, Izmir, Turkey).

To determine the effect of ozone application on the physical, microbiological, and sensory properties of table olives, the experiment was designed to be one-factor (14 levels by adding control and pasteurization groups for concentration * time combination). Table olives were treated with 0.5, 1, 2, and 4 ppm ozone for 5, 10, and 20 minutes before packaging. 5 kg of olives were treated with ozone each time. The water temperature used during the ozonation applications was 20 °C ± 1 °C. At the end of each treatment, the water was refreshed and about 30 liters of water were added to the tank. To disinfect the water, it was treated with 0.5 ppm ozone for 5 minutes. Each application was repeated twice. The ozonated olives were packed in glass jars which had been sterilized at 121 °C for 15 min, in portions of approximately 540 g and covered with 460 mL brine prepared with ozonated water (4% NaCl and 0.6% citric acid) and stored for 6 months at 20 ± 2 °C. Additional groups of not treated samples (without ozone treatment) and pasteurized samples (80 °C, 15 minutes) were added to the experiment following storage under the same conditions as the ozonated samples. During the storage period (180 day) a pH measurement, microbiological analysis, water activity determination, color determination and sensory analysis were carried out at day 0 (right after the ozone treatment) and at monthly intervals.

pH measurement. The pH of the brines was measured by WTW 330 (Germany) (Anon., 2015).

Water activity (a _w ) measurement. The water activity of the table olives was measured by Pre Aqua Lab (2365 NE, USD).

Color. The color of the table olives was measured by colorimeter (Minolta, model CR-400 chromameter, Japan).

For each sample 20 olives were analyzed to evaluate the skin color. The color of the olive surface was evaluated in terms of lightness (L*, where L=0 is black and L=100 is white), greenness to redness (a*, a+ is red and a- is green) and blueness to yellowness (b*, b+ is yellow and b- is blue) (Panagou, 2004).

Hardness. Hardness and its evolution in olives were determined through a texture analyzer (TA.XT plus, Stable Micro Systems, Godalming, UK) equipped with a 2 mm diameter cylinder probe. By puncturing the olives, the target value of 4 mmthe maximum force (N, g) applied against the time (sec) was measured. The settled parameters of the test were: pre-test speed 2 mm·sec^-1, test speed 0.5 mm·sec^-1, post-test speed 4 mm·sec^-1, and load cell 30 kg (Romeo et al., 2012).

The automated system TEMPO (Biomerieux) was used for the enumeration of TAMB (total aerobic mesophilic bacteria), lactic acid bacteria (LAB), yeasts and moulds (YM), Escherichia coli (EC) and Enterobacteriaceae count (EB) in the olive samples. TEMPO uses a dehydrated culture medium and an enumeration card containing 48 wells across 3 different dilutions for the automatic determination of the most probable number (MPN) (User’s Manual, Application Guide of TEMPO^®, Biomerieux, Available from:

https://techlib.biomerieux.com/wcm/techlib/techlib/applications/guidedSearch/cleverLink/CleverLink.jsp?productnumber=tempo./Accessed 06.08.12).

Samples of table olive pulp (10 g) and sterile peptone saline diluent (90 mL) were mixed in plastic bags with a filtered compartment. Each of samples was homogenized for 2 min. in a stomacher (easy MIX AES Chemunex 1068, France). After homogenization, for LAB and TAMB, 0.1 mL and for YM, EB and EC, 1 mL of the homogenized sample suspension was taken using a sterile pipette from the filtered bag and transferred into the vial containing culture media previously mixed with 3 mL (for YM, EB, EC) and 3.9 mL (for TAMB, LAB) of sterile distilled water. The vials were mixed with a vortex for 4 seconds. The inoculated media were moved into the TEMPO cards. Cards and reagents were then loaded into the TEMPO Preparation Station according to the manufacturer’s instructions, to automatically fill the cards with the appropriate sample dilution. The cards were incubated for 48 h at 30 °C (TAMB and LAB tests), 72 h at 25 °C (YM test) 24 h at 30 °C (EC test) and 24 h at 35 °C (EB test). After incubation, the cards were placed in an automated reader which detects fluorescence. The results were expressed in colony-forming units per 1 g of sample (CFU·g^-1).

During the storage period, the olives were submitted to a sensory analysis by a trained taste panel from the Olive Research Institute (Izmir, Turkey) according to the methodology described in the IOC method (IOC, 2011). The panel consisted of 8 different judges, 5 women and 3 men, between 31-50 years old. The evaluated sensory characteristics were: a) negative sensations or defects (abnormal fermentations and other defects such as butyric, putrid and zapatería, winey-vinegary, soapy, metallic, cooking effects, rancid, musty and earthy defects); b) gustatory sensations (salty, bitter, acidic) and c) kinesthetic sensations (hardness, fibrousness, crunchiness). The samples were coded with random three-digit numbers. Three to four olives were given to the panelists in individual booths in a sensory laboratory. Between samples, water was used to clean out the pallet.

All of these analyses were carried out in duplicate for each sample. To elaborate the sensory data, the method for calculating the means and the confidence intervals was used, as detailed in Annex 1 (COI/OT/MO/No 1/Rev.2 Annex 1 Method for calculating the means and the confidence intervals) (IOC, 2011), taking into account the attributes with a robust coefficient of variation of 20% or less. The computer program for carrying out the calculations was as presented in Annex 3 (COI/OT/MO/No 1/Rev.2 Annex 3 Sensory analysis of table olives computer program) (IOC, 2011). For classification purposes, only the mean of the defect predominantly perceived (DPP) was considered, in other words, perceived with the greatest intensity.

According to the DPP intensity, the samples were classified into four categories:

A statistical analysis of the obtained data was carried out using the SPSS version 20.0 (SPSS, Inc., Chicago, IL, USA). Significant differences among treatments were determined by variance analysis and means were separated by the Duncan test with a 95% significance level.

At the end of fermentation, the pH decreased to 3.78, and thus the acidity 0.88% increased, which ensures the preservation of the product. The initial pH values of ozonated olives were slightly lower than the values of pasteurized and control samples (Figure 2 (a)). The concentration-time combination and storage time effect on the pH value were statistically significant (p < 0.05). The values obtained for the pH values of the cover brines obtained at the end of the storage period were in agreement with the IOC limits (IOC, 2004). However, the final pH value for the control group was close to the limit of 4.3. At the end of the storage period, the pH values for 1-10 and 1-20 applications were smaller than the other applications. The pH values of the brines were very similar and did not show significant differences among treatments. The brine pH of the control group was the highest at the end of storage and was in a different group from the other applications. In line with the findings of this study, Alexandre et al., (2005) reported that the pH of green basil and bell pepper was significantly affected by ozone application.

Ozone application did not affect the a_w value of the table olives. There was no significant difference between the control group, pasteurized olives or ozonated olives. A slight increase was determined until the 90^th day for aw values and after the 90^th day, there was a slight decrease in aw values. At the end of storage, the a_w values for the table olives increased compared to the beginning of storage.

The hardness of olives is one of the key quality parameters for commercial value. As there is no standard for hardness, it is important for the olives to preserve their original firmness as much as possible during storage (Romeo et al., 2009). In fact, a non-appropriate fruit texture can be one of the main reasons for rejection by the consumer. In addition, a lack of fruit hardness may cause high economic losses to the processing industry because of difficulties in fruit pitting and stuffing after lactic fermentation (Garrido Fernandez et al., 1997). In our work ozone application significantly affected the hardness of the olives (Fig. 3). The hardest olives were detected in pasteurized olives (310.9) and the softest olives in the control group (219.8 g). The hardness of ozonated olives was found to be softer than pasteurized olives and harder than the control group. Compared with the control group, ozone treatment increased the hardness of the olives.

Koyuncu et al., (2008) reported that the hardness of cherries treated with ozone increased slightly during storage. The hardness did not reveal significant differences among ozonated groups, but showed a decrease in the values during the storage period. The greatest loss of hardness in the olives occurred on the 60^th day. The decrease in hardness during the storage is in accordance with data reported by Sánchez-Gómez et al., (2013).

The role of color is a highly significant attribute in the quality evaluation and choice made by consumers (Romeo et al., 2012).

While the effect of ozone application was not significant for the a* values of the olives (Table 1 (b)), it was found to be significant for L* (Table 1 (a)) and b* (Table 1 (c)) values. In our study, the L* values of the ozonized olives were found to be lower than the control group and higher than the pasteurized olives (except for 2-20 applications). As the lowest L* value was measured in pasteurized olives with 54.72, it can be said that the pasteurization process caused some darkening of the olives. The L* values for the ozonated olives were lower than the control group. This can be attributed to the fact that ozone is relatively blackish due to its oxidizing properties. At the end of storage, the L* value for the olives had increased compared to the beginning of storage. In other words, the olives had been tinted.

Regarding the b* parameter, for the index of blue/yellow coloring, the highest value was measured in the control group (38.46). As the b* values of ozonated olives were lower than the control group, it was concluded that the application of ozone led to a decrease in the yellowness of the olives.

As in the case of texture, it could be indicated that the values for the parameters L*, a*, b * were of the same order as those found by Sánchez-Gómez et al., (2013) for the Spanish style green olives of the cultivars used in Spain.

The results from the microbiological analyses of the olive samples are shown in Table 2. The EB and EC loads in the table olives after fermentation were below the detection limits (< 10). Undesired microorganisms, such as Enterobacteriaceae and Clostridium spp. are present at the beginning of olive fermentation but at the end of the process, when the pH decreases, they are not generally detected (Alves et al., 2012; Randazzo et al., 2012). According to the results of the statistical analysis, the concentration-time combination, storage time and interaction effect on total aerobic bacteria, lactic acid bacteria and moulds and yeasts of the olives were found statistically significant (p < 0.05).

The initial counts of total aerobic mesophilic bacteria, lactic acid bacteria and yeasts and molds in the untreated table olives were 4.96, 2.49 and 3.46 log CFU g^-1, respectively. After ozonation, total aerobic mesophilic bacteria ranged from < 100 to 3.10 log CFU g^-1. At the end of 180 days storage, the total aerobic mesophilic bacteria of the pasteurized olives was below the detection limits (< 100), and the highest total aerobic mesophilic bacteria number was found in 2-20 applications with 6.14 log CFU g^-1. Unexpectedly, at 120 and 150 days of storage, the microbial load was observed in the pasteurized jars. It is thought that it can be caused by contamination from the other jars during analysis.

The differences among the responses observed for ozone concentration due to ozone treatments can be explained by the diversity of native flora enumerated in each case, and consequent different sensitivity to the ozone oxidizing effect. The nature and composition of food surface, the degree of attachment to or association of microorganisms with food and biofilm formation are other possible justifications (Alexandre et al., 2011.The lower microbial inactivation achieved in our study can be explained by the fact that bacterial spores are more resistant to environmental abuses and toxic chemicals than vegetative cells. According to Alexopoulos et al., (2017) no statistical differences in the viable counts of the lactic acid bacteria were observed between the control and the samples treated with ozone. This indicates that ozone did not alter the population of beneficial microflora (Alexopoulos et al., 2017).

At the end of storage, the lowest and the highest lactic acid bacteria were determined in pasteurized olives and in the control group, respectively. Ozonation produced nearly 1.5 log and 2 log decreases in the number of lactic acid bacteria and total bacteria, respectively, due to its strong oxidative power against a wide spectrum of microorganisms (Guzel-Seydim et al., 2004). However, no apparent effect of ozone on the lactic acid bacteria population was observed by Arroyo-Lopez et al., (2006). In line with our research findings, Venturini et al., (2002) reported that ozone treatment prevented microbiological developments which occurred on the fruit during storage and was effective in reducing the yeast-mold load.

It is observed that the application of ozone is particularly effective in reducing the total bacterial load in olives at the beginning of storage. However, apart from pasteurized olives, total aerobic mesophilic bacteria development was observed during storage in both control group and olives subjected to ozone application. The total number of aerobic mesophilic bacteria of the ozonated olives was about 2 log units lower than the control group, and about 4 log units higher than the pasteurized olives. In this case, it can be said that the ozone application significantly reduced the total aerobic mesophilic bacteria growth that occurred during the storage period. Although the O₃ treatments reduced the microbial population in green table olives, the population gradually increased with storage time.

The presence and intensity of organoleptic negative attributes affect the quality of table olives (Lanza., 2018). Although not legally required, table olives may be classified according to the evaluation of negative sensory attributes, performed by a trained sensory panel (IOC, 2011).

As the pH value exceeded 4.0 and the yeast-mould layer developed on the surface after the 30th day in the control group, the sensory analysis was not performed after the 30th day for this group.

The results of the organoleptic assessment of the green Domat table olives after ozonation and during storage period are shown in Table 3.

The saltiness of the ozonated olives was determined to be lower and statistically significant than the control group (Table 3 (a)). As reported in Table 3 (c), the bitterness of ozonated olives showed similarity with the control group, but lower than pasteurized olives and statistically found in different groups (except 1-20 application).

Fruit hardness is an important organoleptic characteristic for table olives (IOC, 2014) and also one of the sensory attributes of greatest importance to consumer acceptance (Lanza, 2018). The application of ozone results in a lower sensation of salinity and hardness and a higher perception of fibrousness and crispness than the control group (Table 3 (a) and (d)). The acidity and bitterness of the olives were similar to the control group (Table 3 (b) an (c)).

All the samples show a mean value of DPP ≤ 3.0 and therefore they were classified as “Extra or Fancy”. It can be said that the application of ozone can improve green table olives with minimal changes in organoleptic properties.