Olive mill wastewater (OMW) generated in two-phase olive oil production consists of wastewater derived from the washing of the olives (OWW) and wastewater that leaves the vertical centrifuge machines (OOWW). Figure 1 shows the flow diagram of the continuous production process of olive oil by the two-phase extraction system, resulting in the generation of these effluents. The OOWW generated in the two-phase olive oil production process is the major waste product of this industry and one of the most serious environmental problems in Mediterranean countries. Spain, Italy, Portugal, Greece and the North African countries Syria, Algeria, Turkey, Morocco, Tunisia, Libya, and Lebanon are the largest producers in the world. Other countries such as France, Serbia and Montenegro, Macedonia, Cyprus, Egypt, Israel, and Jordan also produce a considerable annual amount of olive oil according to the data registered by International Olive Oil Council (IOOC, 2016). What is more, the production of olive oil is growing in countries such as China, USA, Australia and the Middle East. The current growth of the olive oil production sector is parallel to the development and modernization of olive oil factories, due to the greater global demand of olive oil.

OOWW is a heavily contaminated liquid stream that has an acidic pH, black color, very high chemical oxygen demand (COD) and a high concentration of microbial growth-inhibiting components, such as tannins and phenols (Elkacmi et al., 2016). Because of its concentration of phenolic compounds, OOWW is highly polluting and phytotoxic, which impedes its biological degradation and produces harmful effects on the flora and fauna of disposed areas (Schieber et al., 2011).

On the other hand, the phenolic compounds from OOWW have a high antioxidant activity (Obied et al., 2005). The extraction of these biologically active compounds could convert this polluting and toxic waste into a source of antioxidants of great value for the pharmaceutical and food industries, in the prevention of different human diseases as well as for the inhibition of food oxidation (De Marco et al., 2007).

Phenolic compounds are composed of aromatic benzene ring/s, hydroxyl group/s and functional side chain/s. The most abundant and interesting phenolic compounds present in OOWW from the two-phase olive oil system are hydroxytyrosol and tyrosol (Takaç and Karakaya, 2009). One of the explanations for the formation of hydroxytyrosol is the hydrolysis of oleuropein (the main biophenol in many olive varieties) due to the action of esterases and acid hydrolysis of secoiridoid derivatives caused by the addition of HCl to the OOWW during oil extraction.

Hydroxytyrosol is one of the strongest antioxidants and it is, in part, responsible for the bitter taste of extra virgin olive oil. Research shows its antioxidant and beneficial properties for health as well as its good bioavailability (Torrecilla, 2010). Hydroxytyrosol has amphiphilic character which enables its uptake in the intestines and protects both lipid and aqueous cellular compartments. The strong antioxidant efficacy of hidroxytyrosol, due to the presence of the o-dihydroxyphenyl moiety, is a result of its high capacity for free radical scavenging during oxidation processes and to its reducing power on Fe³⁺ (Torres de Pinedo, 2005).

Hydroxytyrosol has the capacity to scavenge peroxyl radicals, which inhibits the oxidation of low-density lipoproteins (LDL), and carries cholesterol (LDL-C), a fundamental stage in the emergence of atherosclerosis and other cardiovascular diseases (Vázquez-Velasco et al. 2011). Moreover, this compound may produce a pro-apoptotic effect intervening in the expression of genes related to the tumor cell propagation of promyelocytes (HL60 cells) (Fabiani et al. 2011), as well as the inhibition of the development of some carcinogenic human cells (Bouallagui et al., 2011; Bulotta et al., 2011).

The antimicrobial properties of oleuropein, tyrosol and hydroxytyrosol have been tested in vitro against viruses, bacteria and protozoa (Bisignano et al., 1999).

The relation of hydroxytyrosol to platelet function has been investigated and it was observed that it impedes the chemically induced aggregation, the accumulation of the thromboxane pro-aggregant agent in human serum, the production of the pro-inflammatory leukotriene molecules and the activity of arachidonate lipoxygenase (Visioli et al., 2002).

Hydroxytyrosol and its derivatives also have applications as microbicides to avoid HIV-infection, and other sexually transmitted diseases generated by fungi, bacteria or viruses (Gómez-Acebo et al., 2011).

Although the antioxidant capacity of tyrosol is much lower than that of hydroxytyrosol, tyrosol is an important antioxidant, and it is extracted rapidly from the intestinal tract. In addition, it can pass the blood brain barrier in large quantities. Researchers found tyrosol in the cerebrospinal fluid of rats that had taken tyrosol orally. Because it increases the total antioxidant levels of the brain, it can be used to treat disorders like Alzheimer’s and Parkinson’s diseases and can also be used to treat stroke and loss of memory (Geelings et al., 2003). Catechol and its derivatives are also phenolic antioxidants found in olive oil that have a great interest in the pharmaceutical and food industries.

In this framework, there are some studies which were aimed at the removal of the phenolic concentration present in two-phase OMW, e.g. by the Fenton process (Cañizares et al. 2009; Hodaifa et al., 2013), or found in other types of wastewater (Jiménez et al., 2017), with the aim to reuse the regenerated water.

On the other hand, for the recovery of these valuable phenolic compounds on an industrial scale, La Scalia et al. (2017) tested membrane filtration and reverse osmosis processes, whereas Zagklis et al. (2015) used filtration though membranes followed by resin adsorption/desorption, and vacuum distillation for the final concentration of phenols. Zafra et al. (2006) only used a liquid–liquid micro-extraction with ethyl acetate, and Elkacmi et al. (2017), after OMW delipidation with hexane, also extracted the phenolic compounds with ethyl acetate.

For the quantification of phenolic compounds, there are spectrophotometric methods such as the 4-aminoantipyrine method or the Folin-Ciocalteu method, which are generally used to quantify total phenol concentrations. These methods are easy but they are not precise, and it is not possible to quantify them individually because they simply evaluate the total phenolic concentration (Brune et al., 1991).

The aim of this study was the identification of the phenolic compounds present in the OMW from the two-phase olive oil extraction system. Specifically, OOWW and OWW generated in this system have been studied for the possible use of these liquid residues before conventional treatments as a source of natural antioxidants of interest in the pharmaceutical and food industries. For this goal, the optimization of different extraction methods was carried out, since in complex matrices, such as these OMW, the key to the determination is the preliminary sample processing, which requires a previous separation of the phenolic fraction from the other constituents present in the samples.

In this work, a comparison between liquid-liquid extraction (LLE) and solid phase extraction (SPE) methods was made. Furthermore, the effect of a previous centrifugation step or hexane extraction in the efficiency was studied, and other components present in these wastewaters were identified. After the application of these extraction methods for phenol identification in OOWW, the best method obtained was also used to identify the main interesting compounds in OWW.

A particularity in this study has been the utilization of OMW generated by the two-phase olive oil system, which is a more recent process which involves a reduction in wastewater volume but an increased concentration in organic matter in comparison with the OMW obtained in the three-phase process (Borja et al. 2006) whose samples have been used in other works (Allouche et al., 2004; De Marco et al., 2007; Zagklis et al., 2015). Moreover, OOWW from vertical centrifuge machines generated by the two-phase olive oil extraction system and wastewater derived from the washing of the olives were used in this work, in comparison with other studies where phenolic compounds have been identified in OMW in general, consisting of a mixture of washing oil wastewater, olive washing wastewater and other residues generated in the oil industry. In contrast with other works, DeMarco et al. (2007) use OMW generated from the three-phase olive oil production process; other differences are that they do not specify the type of OMW; before the liquid–liquid extraction with ethyl acetate, the pre-treatment of the sample is different from this work, where whether a previous centrifugation is more appropriate, an extraction with hexane or both were studied. The compounds extracted in hexane were also identified in this work. In addition, for the identification of the compounds, these authors use HPLC. As in the case of DeMarco and colleagues, Allouche et al. (2004) use wastewater generated from the three-phase olive oil process, they do not specify the origin and only liquid-liquid extraction is performed. Another example is Deeb et al. (2012) who do not make a comparison between different extraction methods.

Figure 2a shows the peaks corresponding to the compounds found in the ethyl acetate extract after the centrifugation of OOWW using the method comprising centrifugation + LLE with hexane and ethyl acetate, and the retention times of these peaks. The identified compounds were 2-hydroxypropanoic acid at 6.13 min, succinic acid at 9.90 min, tyrosol at 13.22 min, hydroxytyrosol at 15.40 min, alcoholic compounds at 15.55 min, 3, 4-dihydroxybenzyl alcohol polyphenols at 16.56 min and a small amount of oleic acid at 19.89 min. The results obtained after the hexane extract was analysed, are reported in Figure 2b. In this case, the main compounds found were palmitic acid and oleic acid, found at 18.30 min and 19.89 min, respectively.

In this study the liquid-liquid extraction was preferred for its simplicity and convenience (Obied et al., 2005). The extraction with hexane in order to remove the lipidic fraction has been carried out by some authors (Ceccon et al., 2001; De Marco et al., 2007; Lesage-Meessen et al., 2001) whereas others only extracted with ethyl acetate after acidification (Lesage-Meessen et al., 2001). Ethyl acetate was chosen for this study because it is frequently used to extract bio-phenols from aqueous matrices such as OOWW (Lesage-Meessen et al., 2001). There are some studies that indicate the loss of some bio-phenols (Bullock et al. 1996; Lafkaa et al., 2011; Zafra et al., 2006), but in contrast, ethyl acetate was chosen as the most suitable solvent for hydroxytyrosol (Allouche et al., 2004) because it is selective towards low and medium molecular weight phenols. This solvent is also employed in the food industry, in confectionery, beverages, candy and for artificial fruit essences.

Otherwise, the main compounds in OOWW from the two-phase centrifuge process, identified in the ethyl acetate extract after the method of centrifugation + LLE with ethyl acetate, are shown in Figure 3. Comparing both ethyl acetate extracts shown in Figures 2a and 3, it can be noted that these samples were centrifuged but the extraction with hexane did not have any influence. As can be seen by comparing both figures, there was no presence of palmitic acid or any significant differences in the amount of oleic acid, tyrosol and hydroxytyrosol in both chromatograms. This fact is quite relevant, since it pin-points that the extraction with hexane is useless and it could be eliminated to avoid spending this solvent uselessly.

Figure 4a shows the main compounds in the two-phase OOWW centrifuge effluent identified in the ethyl acetate extract resulting from the method consisting of LLE with hexane and ethyl acetate. The comparison of Figure 2a (Ethyl acetate extract with centrifugation) and Figure 4a (Ethyl acetate extract without centrifugation) reveals that the only difference is a major quantity of oleic acid in the sample without centrifuging, but there are no differences in the quantity of phenolic compounds.

Figure 4b shows the results obtained after the analysis of the hexane extract.

Figures 2b and 3b confirm that hexane only extracts oleic acid and palmitic acid. Comparing Figure 2b (Hexane extract with centrifugation) and Figure 4b (Hexane extract without centrifugation), it is noted that there was a major quantity of oleic acid and palmitic acid in the hexane extract of the sample that was not centrifuged.

On another hand, the main compounds found in the OOWW from two-phase centrifuge identified in the ethyl acetate extract after the method comprising just LLE with ethyl acetate are reported in Figure 5. This method required more time to separate the aqueous phase from the ethyl acetate fraction in the first extraction.

Both samples, extracted only with ethyl acetate and without hexane, shown in Figures 3 and 5, have in common that the phenols were the same but the centrifuged sample contained less oleic acid.

Additionally, Figure 6a shows the main compounds in OOWW identified in the ethyl acetate extract after the centrifugation + SPE method, whereas Figure 6b shows the main compounds in OOWW from the centrifuge identified in the ethyl acetate extract after the method of SPE alone.

Comparing the extractions in the solid phase (Figures 6a and 6b), analogous results to those previously observed with the LLE were attained, revealing that in the previously centrifuged sample (Figure 6a) there were no lipids whereas in the not-centrifuged one (Figure 6b) there was oleic acid, oleic acid with a methyl group and palmitic acid, which were retained by the resin.

Finally, the best phenolic extraction method was used for the further extraction and subsequent analysis using olive washing water (OWW). Figure 6c shows the main compounds in OWW identified in the ethyl acetate extract after the method of centrifugation + SPE. In this case, other different phenolic compounds were identified since it is another type of olive mil wastewater.

After the analysis of these OOWW, the compounds extracted by ethyl acetate were 2-hydroxypropanoic acid, butanedioic acid (succinic acid), tyrosol, hydroxytyrosol, alcoholic compounds, 3,4-dihydroxybenzyl alcohol, 3-hydroxyphenil acetic acid and oleic acid. These compounds were the same in all the methods used, with the exception that in methods 5 and 6, 2-hydroxypropanoic acid did not appear since it was not retained by the SPE columns. In method 4 and 6 there was also palmitic acid because in these methods no previous centrifugation or hexane extraction was performed. The compounds found in the hexane extract were palmitic and oleic acids.

The best methods to recover the phenolic compounds were methods 2 and 5 in which a centrifugation was carried out before extraction with ethyl acetate, since the target phenols were barely lost and a cleaner fraction was obtained. Using method 5, another OWW sample was analyzed and the compounds found were cresol, catechol, 4-methylcatechol, hydrocinnamic acid, p-hydroxy-hydrocinnamic acid but only a small quantity of hydroxytyrosol and no tyrosol was found.

Although hydroxytyrosol and tyrosol were described since they are the most commonly referenced components, it is worth highlighting that there are also interesting uses for the other compounds identified, such as hydrocinnamic acid, which is a color stabilizer (Hernández et al. 2006), an antioxidant and radical scavenger (Kikuzaki et al., 2002), an antioxidant together with glutathione (Bouzanquet et al., 2012) and it is a flavor precursor. Succinic acid, which was also identified, is not a phenolic compound but is useful for energy metabolism in plants, and human and animal tissue. A major application is its use in biodegradable plastics and in automotive engineering, nonwovens and fibers, as well as sportswear, furniture and even the construction industry (Succinity, 2016).

In other studies, OMW consisting of a mixture of oil washing waste water, olive washing wastewater and other residues generated in the oil industry were analyzed. This mixture, along with the fact that the chemical composition of OMW depends on the variety of the olives, the stage of maturity, storage time and climate, among other factors (Obied et al., 2005), could explain the identification of different phenolic compounds in other studies.

To sum up, the most important compounds found in OOWW were tyrosol, hydroxytyrosol and succinic acid, whereas the most interesting compounds found in OWW were cresol, catechol, 4-methylcatechol, hydrocinnamic acid, p-hydroxy-hydrocinnamic acid. The recovery of these bio-phenolic fractions in olive mill wastewater will help valorize these effluents, and at the same time minimize a deleterious environmental problem in producing countries.

The goal of this work was to identify the polyphenols present in the wastewater generated from two-phase olive oil production, specifically from the washing of the olives, OOWW, and from the vertical centrifugation, OWW, since it possesses antioxidant properties which are beneficial for health and industrial applications, whereas these effluents represent a serious environmental problem because these compounds are phytotoxic and difficult to degrade.

Before reaching this goal, it is crucial to establish a good method of extraction. After the identification of bio-phenols present in OOWW through simple methods such as liquid-liquid extraction using ethyl acetate, it was concluded that for recovering the phenolic fraction, a previous centrifugation of the sample is more effective than extraction with hexane to eliminate the remaining lipid fraction (palmitic and oleic acids) in the OOWW. With this centrifugation, only lipids are removed but not phenolic compounds. Method 2 is the most appropriate for this: centrifugation and extraction with ethyl acetate, or if a solid phase extraction is preferred, good results are also obtained with a previous centrifugation as described in method 5.

The most important compounds found in OOWW were tyrosol, hydroxytyrosol and succinic acid.

On the other hand, the most interesting compounds found in OWW were determined to be cresol, catechol, 4-methylcatechol, hydrocinnamic acid, p-hydroxy-hydrocinnamic acid.

Once these compounds have been identified on a laboratory scale using easy and cheap methods, other processes such as membrane filtration or resins must be applied for the recovery of polyphenols on an industrial scale.