lable at ScienceDirect

Water Research 114 (2017) 1e13

Contents lists avai

Water Research

journal homepage: www.elsevier .com/locate/watres

Treatment efficiency and economic feasibility of biological oxidation,membrane filtration and separation processes, and advancedoxidation for the purification and valorization of olive mill wastewater

L. Ioannou-Ttofa a, I. Michael-Kordatou a, S.C. Fattas a, A. Eusebio b, B. Ribeiro b, M. Rusan c,A.R.B. Amer c, S. Zuraiqi c, M. Waismand d, C. Linder d, Z. Wiesman d, J. Gilron d,D. Fatta-Kassinos a, e, *

a Nireas-International Water Research Center, University of Cyprus, P.O. Box 20537, CY-1678, Nicosia, Cyprusb Bioenergy Unit, National Laboratory of Energy and Geology, Estrada do Paço do Lumiar, 22, 1649-038, Lisboa, Portugalc Department of Natural Resources and Environment, Faculty of Agriculture, Jordan University of Science and Technology, Irbid, Jordand Ben-Gurion University of the Negev, Departments of Biotechnology, Energy and Environmental Engineering, P.O. Box 653, Beer-Sheva, 84105, Israele Department of Civil and Environmental Engineering, University of Cyprus, P.O. Box 20537, CY-1678, Nicosia, Cyprus

a r t i c l e i n f o

Article history:Received 30 August 2016Received in revised form1 February 2017Accepted 10 February 2017Available online 12 February 2017

Keywords:Biological treatmentCost estimationMembrane filtrationOlive mill wastewaterPhenolic compounds recoverySolar photo-Fenton oxidation

* Corresponding author. Nireas-International Wateof Cyprus, P.O. Box 20537, CY-1678, Nicosia, Cyprus.

E-mail address: dfatta@ucy.ac.cy (D. Fatta-Kassino

http://dx.doi.org/10.1016/j.watres.2017.02.0200043-1354/© 2017 Elsevier Ltd. All rights reserved.

a b s t r a c t

Olive mill wastewater (OMW) is a major waste stream resulting from numerous operations that occurduring the production stages of olive oil. The resulting effluent contains various organic and inorganiccontaminants and its environmental impact can be notable. The present work aims at investigating theefficiency of (i) jet-loop reactor with ultrafiltration (UF) membrane system (Jacto.MBR), (ii) solar photo-Fenton oxidation after coagulation/flocculation pre-treatment and (iii) integrated membrane filtrationprocesses (i.e. UF/nanofiltration (NF)) used for the treatment of OMW. According to the results, the ef-ficiency of the biological treatment was high, equal to 90% COD and 80% total phenolic compounds (TPh)removal. A COD removal higher than 94% was achieved by applying the solar photo-Fenton oxidationprocess as post-treatment of coagulation/flocculation of OMW, while the phenolic fraction wascompletely eliminated. The combined UF/NF process resulted in very high conductivity and COD removal,up to 90% and 95%, respectively, while TPh were concentrated in the NF concentrate stream (i.e. 93%concentration). Quite important is the fact that the NF concentrate, a valuable and polyphenol richstream, can be further valorized in various industries (e.g. food, pharmaceutical, etc.). The above treat-ment processes were found also to be able to reduce the initial OMW phytotoxicity at greenhouse ex-periments; with the effluent stream of solar photo-Fenton process to be the least phytotoxic compared tothe other treated effluents. A SWOT (Strength, Weakness, Opportunities, Threats) analysis was per-formed, in order to determine both the strengths of each technology, as well as the possible obstaclesthat need to overcome for achieving the desired levels of treatment. Finally, an economic evaluation ofthe tested technologies was performed in an effort to measure the applicability and viability of thesesystems at real scale; highlighting that the cost cannot be regarded as a 'cut off criterion', since the mostcost-effective option in not always the optimum one.

© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

The olive oil industry is an agro-industrial sector of high rele-vance in the economy of the European Union (Yay et al., 2012);

r Research Center, University

s).

especially in Mediterranean countries, such as Cyprus, Greece, Italy,Israel, Jordan, Portugal, Spain and Tunisia. Although olive oil is aproduct of exceptional nutritional value, its production is associ-ated with several adverse effects to the environment, mainly due tothe formation of a high volume of OMW. Approximately 30 millionm3 of OMW effluents are produced annually in the Mediterraneanarea, characterized by a high pollutant load (Chiavola et al., 2014).The Mediterranean countries, in particular, are facing serious

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e132

problems in managing OMW.The chemical composition of OMW is highly variable depending

on various factors, such as the olive type, the geographical andclimatic conditions, the tree age, the cultivation system, the degreeof maturity of the fruit, the type of oil extraction process applied,the use of pesticides and fertilizers, etc. (Kallel et al., 2009; Yayet al., 2012). More specifically, OMW is characterized by a highsuspended solids content (e.g. TSS between 6 and 70 g/L), darkbrown color, characteristic unpleasant odor, low pH, high turbidityand high organic load (e.g. COD between 30 and 318 g/L) (Yalılı Kılıçet al., 2013; Amor et al., 2015). In addition, OMW contains manycomplex organic substances (i.e. phenolic compounds (TPh be-tween 0.5 and 24 g/L), polysaccharides, sugars, proteins, lipids,tannins, pectin, organic acids, etc.), which are generally resistant tobiodegradation. Furthermore, these persistent organic substancescan lead to various negative adverse effects on the environment,such as foul odors, threat to the aquatic life, saturation of soil,changes in soil quality, discoloring of natural waters, eutrophicationof surface waters, pollution of superficial groundwater and toxicity(i.e. eco- and phytotoxicity) (Ca~nizares et al., 2007; Yay et al., 2012;Amor et al., 2015). Although the phenolic fraction accounts for arelatively minor contribution to COD (approx. 14%) (Michailideset al., 2011), it is generally considered that phenolic compoundsare responsible for the dark brown color, phytotoxic effects andantibacterial activity of OMW (Khoufi et al., 2009; Yalılı Kılıç et al.,2013).

The most common practice for the management of OMW in-cludes the use of evaporation ponds and the subsequent dischargeof solids in landfills and/or on soil. While evaporation ponds offer agood way of reducing the liquid portion of this effluent, they do notcontribute to the reduction of its toxicity, while they simulta-neously impart an odor problem to the areas where such waste isstored (Michael et al., 2014). Moreover, the direct spreading of rawOMW for agricultural purposes could be considered as: (i) aneconomic mean to solve the OMW pollution-related problems,especially in replacing chemical fertilization, and (ii) a low-costsource for water, especially in the countries of the Mediterraneanregion that are facing serious and prolonged water scarcity events(El Hajjouji et al., 2014). The impact of OMW on soil propertiesdepends on the relative amounts of beneficial and toxic organic andinorganic compounds existing in it. Consequently, several chemicaland biochemical soil properties have shown variable susceptibilityto applied OMWeffluents (Piotrowska et al., 2006). In addition, theimpact of OMW on soil microflora can include the temporary soilenrichment of easily degradable carbon, which stimulates micro-flora development, and the enrichment of the soil with OMWcompounds that are toxic to certain microorganisms. A number ofstudies reveal that the effects of OMW spreading on soil charac-teristics depend on many factors, including pedologic and climaticcharacteristics, the amount of OMW and the method of spreading(Mekki et al., 2006; Mechri et al., 2007). Moreover, the main socialimpacts associated with OMW production and managementinclude among others, odor complaints from nearby residents(Azbar et al., 2004) and negative impact on property values andannual property yield (both residential and agricultural properties).From all the above, it can be concluded that uncontrolled dischargeof OMW can pose a serious environmental risk, and as a conse-quence appropriate treatment is required.

A number of OMW treatment methods have been employed sofar and these can be divided into four general categories: (i)physicochemical methods (e.g. sedimentation, coagulation/floccu-lation, etc.), (ii) biological processes (e.g. aerobic activated sludge,anaerobic digestion, etc.), (iii) membrane filtration and separationprocesses (e.g. microfiltration (MF), UF, NF, reverse osmosis (RO)),and (iv) advanced chemical oxidation processes (e.g.

heterogeneous photocatalysis, ozonation, photo-Fenton oxidation,etc.) (Paraskeva and Diamadopoulos, 2006; Eus�ebio et al., 2007;Cassano et al., 2011; Michael et al., 2014).

It is well known that the simple physical processes (i.e. dilution,evaporation, sedimentation, coagulation/flocculation, filtration andcentrifugation) are not able, if applied alone, to reduce the organicload and the toxicity of OMW to acceptable limits (Paraskeva andDiamadopoulos, 2006).

On the other hand, conventional biological processes haveshown satisfactory efficiencies in terms of OMW purification,especially regarding the biodegradable organic content (Chiavolaet al., 2014). More specifically, the aerobic biological treatmentsystems could become an interesting alternative due to their fastprocess kinetic and high removal rates. On the other hand, anaer-obic bioreactors have been used to convert OMW organic contentinto biogas that can be used for energy production; however, cur-rent methods for the anaerobic treatment are only applicable tohighly diluted OMW (Marques, 2001), and methanogenic bacteriaare often inhibited by the high phenolic content of this type ofeffluent (Eus�ebio et al., 2007).

Furthermore, the use of advanced membrane filtration andseparation processes, such as MF, UF, NF and RO, have been pro-posed to obtain effluent streams from OMW of acceptable qualityfor safe discharge in the environment, tree or land irrigation, oreven for recycling and reuse in the olive mill facilities (Cassanoet al., 2011; Ochando-Pulido et al., 2014; Zagklis et al., 2015;Sanches et al., 2016; Kontos et al., 2016). However, these membraneprocesses suffer from various drawbacks, such as the concentrategenerated that contains high levels of refractory organic pollutantsand inorganic salts which needs further treatment before itsdisposal in the environment, the membrane fouling, the energycost, and the membrane replacement (Ioannou et al., 2013). In thelast years, OMW has received great attention in relation to thepresence of high added-value compounds, such as antioxidantsubstances and phenolic compounds, which can be used in phar-maceutical, nutriceutical and cosmetic applications (Obied et al.,2005). Since the phenolic compounds are highly valuable forvarious applications, but at high concentration are inhibitorymainly to the biological treatment due to their toxicity, their re-covery from various agro-industrial wastewater streams hasgenerated significant interest (Han et al., 2001). With a number ofadvantages, such as high efficiency, simple equipment, convenientoperation, etc., membrane technology has become one of the mostimportant industrial separation techniques and has been appliedextensively to various fields including the recovery of valuableproducts from agro-food and other by-products (El-Abbassi et al.,2011; Zagklis et al., 2015; Hamza and Sayadi, 2015; Kontos et al.,2016).

However, in the last decades, the application of advancedremediation strategies was found to be required mainly to fulfilllegislative requirements for direct disposal into landfills and toreduce/eliminate their toxicity (Hodaifa et al., 2013). Advancedoxidation processes (AOPs) are based on the creation of veryreactive species, such as hydroxyl radicals (HO�), and are well-known for their capability to break down a wide range oforganic/inorganic compounds quickly and non-selectively andeliminate effluents' toxicity, as well (Yalılı Kılıç et al., 2013). AOPshave been extensively studied for the OMW treatment throughozonation, photo-Fenton oxidation, TiO2 photocatalysis, electro-chemical oxidation and wet air oxidation, and their efficiency wasfound to be high (Gernjak et al., 2004; Ca~nizares et al., 2007, 2009;Chatzisymeon et al., 2009; Belaid et al., 2013; Michael et al., 2014).It should be noted that if OMW is properly treated and managed, itcan be beneficial and a source of nutrients (e.g. N, P, K, etc.) essentialto the plants and to the fertility of the soil (Mekki et al., 2013).

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e13 3

The aim of this study was to investigate different treatmentmethods, in order to obtain a suitable alternative option for thetreatment of OMW in an economical and feasible way. More spe-cifically, the technologies that were investigated are: (i) a jet-loopreactor with UF membrane system (Jacto.MBR), (ii) a coagulation/flocculation system followed by solar photo-Fenton oxidation, and(iii) an integrated membrane filtration and separation process (UF/NF), which was applied both for the treatment of OMW and therecovery of phenolic compounds through a novel process, namelymembrane aromatic recovery system (MARS) (Han et al., 2001;Ferreira et al., 2003). Moreover, a SWOT analysis was performed,in order to identify the strengths, weaknesses, opportunities andthreats of the proposed technologies, which is an analysis that stillmissing from the scientific literature for these types of treatmenttechnologies. The importance of a SWOT analysis is highlighting tothe fact that it focuses on identifying both the internal and externalfactors affecting the successful implementation of a treatmenttechnology, helping the olive mill managers to forecast or predictopportunities and threats that could influence the decision-makingprocess regarding the management of their effluents. Finally, theinvestment and the operational costs of all the proposed treatmentalternatives needed for the treatment of 1 m3 of OMW were eval-uated, in order to estimate the applicability and viability of thesesystems at real scale. According to the authors' opinion, the out-comes of this study are highly innovative and important, since theyprovide clear and valuable information for olive mill operators/managers, that can facilitate decisionsmaking with regard towhichtreatment technologies can be applied, according to local condi-tions. It should be also highlighted that this is an integrated studyinvestigating not only the efficiency of the three different treatmenttechnologies at pilot scale, but also their environmental, social andeconomic impacts, which are key elements, still missing from theexisting scientific literature.

2. Materials and methods

2.1. Chemicals

The anionic polyelectrolyte FLOCAN 23 (manufactured by SNFFloerger and purchased from ChemFlo-Hellas) was used as a floc-culation agent for the physicochemical pre-treatment of OMWprior to the advanced chemical oxidation process. Solar photo-Fenton experiments were performed using FeSO4$7H2O (SigmaAldrich), reagent-grade H2O2 (30% w/w, Merck) and H2SO4(95e97%, Merck). In order to prevent interference with analyticalmeasurements, before the COD and DOC analyses, residual H2O2was removed by adding MnO2 (powder, 300e400 mesh, �90%,Sigma-Aldrich). For the toxicity analysis the treated solutions wereneutralized by 2 N NaOH (Merck), and the residual hydrogenperoxide was removed from the treated samples with commer-cially available catalase solution (Micrococcus lysodeikticus 170,000U mL�1, Fluka).

Sulphuric acid (98%, Frutarom) was used for the acidification offeed OMW prior to UF process. Calcium hydroxide (lime) (reagentgrade, Romical) was used for the adjustment of pH of the NFpermeate to neutral values after the end of membrane treatment.For the chromatographic analysis of phenolic compounds thefollowing were used: tyrosol (>99.5% (GC), analytical standard,Fluka), hydroxytyrosol (�98% (HPLC), Sigma), vanilic acid (�97%(HPLC), Sigma), syringic acid (98%, Sigma), p-coumaric acid (�98%(HPLC), Sigma), gallic acid (97.5e102.5% (titration), Sigma) andcaffeic acid (�98% (HPLC), Sigma) and trans-cinnamic acid (>99.9%,Sigma).

The fertilizer diammonium phosphate (DAP) with the grade of18-46-0 (18%N - 46%P2O5 - 0%K2O) from the Jordanian Phoaphate

Mining Company (JPMC), was used and added to tap water for thegreenhouse phytotoxicity experiments.

2.2. OMW characteristics

OMW samples were supplied by three-phase olive mills locatedin each country participating in this study (i.e. Cyprus, Israel, Jordanand Portugal) during the milling campaign of 2013e2014. Thequalitative characteristics of the OMW samples used in this study,determined according to Standard Methods (APHA, 1998), areshown in Table 1.

2.3. Experimental set-up and procedures

The technologies tested were: (i) biological treatment throughthe jet-loop reactor with ultrafiltration membrane system (Jac-to.MBR) (Portugal), (ii) coagulation/flocculation followed by solarphoto-Fenton oxidation (Cyprus) and (iii) integrated membranefiltration and separation processes (UF/NF) (Israel).

2.3.1. Jacto.MBR reactorJacto reactor is a jet aeration system that combines efficient

oxygen transfer with high turbulent mixing to be used for biolog-ical treatment of wastewater. Some of the advantages of this systeminclude its simple operation, the effectiveness of the treatment, theproduction of very lowamounts of sludge and the energy efficiency.In this study a new type of Jacto reactor was developed, using theprevious design described in Eus�ebio et al. (2007), but the settlerwas removed and a cross-flow UF membrane (cut-off 300 kDa) wasincorporated in the system (Jacto.MBR). It was made by AISI 304stainless steel to ensure good mechanical strength, no corrosionand easy maintenance. The reactor (capacity: 110 L), has a cylin-drical configuration, a plane bottom, and a cover to minimize theeffect of foaming. A foam breaker was placed on the top of thereactor. The liquid recirculates through the nozzle and the flow isdriven by a high-velocity liquid jet. The air is inhaled through theinlet of the Venturi ejector at the top, dispersed by the jet energyand bubbles are distributed throughout the reactor volume, while aminimum power input is required (Zehner and Kraume, 2000). Themass transfer coefficient increases with the increasing liquid flowrate because the turbulence generated by the liquid reduced thesize of gas bubbles and increased the interfacial area (Fadavi andChisti, 2005). The aerobic biological treatment of OMW was car-ried out under good conditions of aeration rate (i.e. 0.33 vvm) andturbulent mixing. Oxilyser and pHlyser II probes (S:CAN, Austria)were used in order to daily monitor the temperature and the dis-solved oxygen by fluorescence and the pH, respectively. All dataacquisition and registration of signals were done on-line using adata logger with a control terminal Conlyte4 (S:CAN, Austria).

2.3.2. Coagulation/flocculation followed by solar photo-Fentonoxidation

Solar photo-Fenton is a solar-driven chemical oxidation process,in which ferrous iron acts as the catalyst, while hydrogen peroxideis the oxidant. The OMW pre-treatment prior to the solar photo-Fenton process was performed in a tank specifically designed forthe coagulation/flocculation procedure. The tank (90 L) is con-structed of stainless steel and it has a cone shape. The OMW pre-treatment consisted at first of coagulation with FeSO4$7H2O(6.67 g L�1), followed by flocculation using the anionic poly-electrolyte FLOCAN 23 (0.287 g L�1), which were the optimumconcentrations found in a previous study of our group (Michaelet al., 2014). The photocatalytic experiments were carried out in asolar compound parabolic collector (CPC) pilot plant comprising sixborosilicate tubes mounted on a fixed platform tilted at the local

Table 1Main qualitative characteristics of raw OMW samples (mean values).

Parameters Units Raw OMW used for the experiments of:

Jacto.MBR Solar photo-Fenton oxidation UF/NF Greenhouse phytotoxicity

pH e 4.8 5.2 4.0 4.7EC mS/cm 14.1 12.5 6.9 6.0COD g/L 83.4 25.0 47.0 118.8BOD5 g/L e 5.1 e e

DOC g/L e 6.0 e e

TSS g/L 36.3 24 e 362.0TP g/L 0.13 4.2 e 0.3TN g/L 3.2 e e 96.8TPh g/L 7.9 4.2 4.0 2.94

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e134

latitude (35�), operated in batch mode. The overall volume capacityof the reactor VT is 100 L and the total irradiated volume Vi (tubesvolume) is 22.4 L. The experimental set-up and procedures of bothprocesses are described in detail elsewhere (Ioannou et al., 2014;Michael et al., 2014).

2.3.3. Integrated UF and NF treatmentA pilot-scale UF system (feed tank capacity: 40 L), with a GE-

Zenon UF ZW-10 module (PVDF hollow fiber, 2 mm OD, 0.04 mmpores and 0.91 m2 surface area) was operated in modified batchmode (topping up the feed tank as permeate was removed). TheZW-10 membrane was backwashed periodically with the optimumcycle being 20 s backwash every 3 min. Prior to its feeding to the UFsystem, the raw OMW was acidified to pH 2, in order to promotethe separation of water insoluble polyphenols from water solublepolyphenols. The acidified OMW pumped into the feed tank wasexposed to bubbling from the bottom by nitrogen, in order toprevent the oxidation of phenolic compounds and to protect theimmersed UF module from severe fouling.

The pilot-scale NF system equipped with an NF270 spiral-wound module (2.6 m2 active area, permeate flux: 28 L/m2 h,molecular weight cut-off of 150e300 Dalton, temperature range21e38 �C) was used as post-treatment of the UF process. Theavailable OMW sample volumes and the membrane unit tank deadvolumes lead to limited volume concentration factors (VCF) in boththe UF and NF to ~3e4. To get higher VCF values during the NFprocess, NF flat-sheet membranes were also investigated (i.e. (i)NF270: 120 cm2, VCF: 4e6, experimental temperature range25e27 �C, feed pressure: up to 50 bar, and (ii) GE DK5: molecularweight cut-off of 150e300 Dalton, VCF: 4e6, experimental tem-perature range 25e27 �C, feed pressure: ~40 bar). These wereinstalled in a flat-sheet flow cell with feed spacers that generatesimilar hydrodynamics to that of the spiral-wound elements(Sterlitech, USA). The sterilitech unit was equipped with a plate andframe heat exchanger that allowed maintaining constant temper-atures within ±1e2 �C. The minimal volume that could be main-tained during operation was 2 L, whereas the feed tank was10e20 L.

According to the scientific literature, processes of phenoliccompounds recovery involve typically a condensing step (i.e.thermal concentration, ultrafiltration or lyophilization) prior to thecarrying out of sequential extraction steps with organic solvents(e.g. methanol, ethanol or hydro-alcoholic solutions); while otherpractices include the application of resin chromatography, selectiveconcentration by liquid membranes or supercritical fluid extraction(De Leonardis et al., 2007; Rahmanian et al., 2014). In this study, theisolation of the high-added value phenolic compounds from the NFconcentrate was performed using the Membrane Aromatic Recov-ery System (MARS) - type process, which is a membrane contactorincluding a stripper with high pH solution. More specifically, in

MARS there is a membrane contactor wherein the extraction of thearomatic phenolic molecules into a high pH aqueous stripping so-lution involves conversion of phenols into ionized phenolates onthe strip side after diffusion of the non-ionized phenols across themembrane from the acidic feed stream. The difference in pH acrossthe membrane maintains the concentration driving force for phe-nols, since the effective phenol concentration on the strip side isclose to zero. The selective composite membrane, which wasdeveloped based on well-known methods (Strathmann, 1975;Ward et al., 1976; Baker et al., 1987; Kimmerle et al., 1988) andused in the contactor (membrane area: 21 cm2 and receiving vol-ume: 0.5 L) is stable at extreme pH values and allows permeation ofthe neutral organic species from the acidic feed through thepermeate, where it is ionized and becomes membrane imperme-able. Therefore, the ionized phenolate in the stripping solutioncannot diffuse back into the wastewater, resulting in a phenolconcentrate in the permeate stream. The membrane preventsmixing of the OMW and the stripping solution. The MARS mem-branes demonstrate a stable selective permeability of hydroxytyr-osol and tyrosol, with very low passage of other OMW components- such as acids, carboxylate, and non-aromatic carbon chains -which was achieved by developing a selective barrier by the use ofspecific modification of the selective barriers by formation of spe-cific channels.

2.4. Analytical methods

Chemical Oxygen Demand (COD), Biochemical Oxygen Demand(BOD5), Total Suspended Solids (TSS), Total Nitrogen (TN), TotalPhosphorous (TP) and Electrochemical Conductivity (EC) weremeasured according to Standard Methods (APHA, 1998). The con-centration of TPh was assessed by the Folin-Ciocalteu method(Singleton and Rossi, 1965; Dejmkova et al., 2009). Dissolvedorganic carbon (DOC) was monitored by direct injection of thefiltered samples (0.22 mm, Millipore) into an Aurora 1030 W TOCanalyzer. In the phytotoxicity experiments, phosphorous (P) wasdetermined using Vanadate-Molybdate-Yellow method, potassium(K) and sodium (Na) by flame photometry, and the various metals(i.e. Zn, Cu, Pb, Cd, Fe, Mg and Mn) by atomic absorption spec-troscopy (Chapman and Pratt, 1961).

High-performance liquid chromatograph (HPLC) Varian Pro Star,equipped with Thermo Hypersil Varian C18 reverse-phase column(250 � 4.6 mm), with 5 mm particle size and composed of a pumpmodule 240, PDA-module detector and Varian ProStar module 410Auto sampler, was used to analyze phenolic compounds in OMW.Separation was achieved by gradient elution using an initialcomposition of water with 0.1% acetic acid (A) and 70% acetonitrileinwater (B) for a total running time of 45 min. The concentration ofB was increased to 50% in 45 min. The flow rate was 1 mL min�1.Chromatograms were obtained at 254 nm and different phenolic

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e13 5

compounds were identified by comparing retention times withthose of commercial standards.

2.5. Lab-scale toxicity assessment for the solar photo-Fentontreated flow

According to the results of previous studies of our group(Papaphilippou et al., 2013; Michael et al., 2014) it was noted thatduring the solar photo-Fenton treatment of OMW, the intermediateoxidation products produced, depending on the oxidation time,may be as toxic as, or even more toxic, than the initial effluent. Forthis reason, the investigation of the toxicity of OMW prior to andafter the solar photo-Fenton oxidation was of significant impor-tance for this study, since this would enable a more comprehensiveevaluation of the environmental safety of the applied technology.

Toxicity measurements were carried out in samples taken atvarious times during the solar photo-Fenton treatment, with: (i)the Daphtoxkit FTM magna toxicity test and (ii) the Phytotestkitmicrobiotest (MicroBioTests Inc.). Toxicity tests were conductedaccording to the standard operating protocols for D. magna (ISO6341, 1996), and for three plant species (S. saccharatum,L. sativum, S. alba), respectively. Catalase (Micrococcus lysodeikticus,Fluka Biochemika) was used to remove the excess of H2O2 used inthe experiments for the ecotoxicity and phytotoxicity measure-ments, because it is non-toxic to microorganisms.

2.6. Greenhouse phytotoxicity assessment

In this study, a calcareous soil with a low organic matter contentclassified as fine-loamy, mixed, thermic, calcic paleargid (Khresatet al., 1998) was used for the phytotoxicity experiments per-formed in Jordan. The soil was air-dried and sieved through a 2 mmscreen. The main characteristics of the soil used was as follows: pH:8.18; EC: 0.61 dS/m; organic matter content: 0.72%; CaCO3: 13.38%;P: 7.10 mg/kg; Fe: 3.56 mg/kg; K: 452 mg/kg; Mn: 3.58 mg/kg; Zn:1.88 mg/kg; Cu: 1.22 mg/kg; Pb: 0.68 mg/kg and Cd: 0.06 mg/kg.

The qualitative characteristics of all OMW samples (treated anduntreated) that were used for irrigation in the phytotoxicity ex-periments are presented in Table 2. It should be mentioned that thetreated OMW samples were acidified to prevent any microbial ac-tivities and were shipped to Jordan for the phytotoxicity experi-ments. The pH of the tested OMW samples was neutralized to pH6.0 prior each phytotoxicity experiment. The greenhouse potexperiment was conducted in a randomized complete block designwith four replications. Each pot was filled with 5 kg air dry soil.Three seeds of Hybrid maize (Belaqziz et al., 2008) (variety: Merkurfrom Seminis Company - Hungary) were seeded per pot and thenthe pots were watered periodically to maintain water content atapproximate field capacity. The amount of solution used for

Table 2Chemical composition of the raw and treated OMW samples provided for the greenhous

Parameters Units Tap water Tap water þ fertilizer O

R

pH initial e 7.80 7.8 4pH adjusted e 7.80 7.8 6EC mS/cm 0.56 0.56 7TSS mg/L 10.00 10 1COD g/L n.d. n.d. 1TP mg/L 0.98 e 1TN mg/L 11.70 11.7 9P2O5 mg/L 34.30 41.8 3K2O mg/L 10.90 10.9 2TPh mg/L n.d. n.d. 2

n.d.: not detected.

watering the pots was calculated by the differences in the weight ofthe pot beforewatering and theweight of the pot when the soil wasat field capacity. After germination, seedlings were thinned to keeptwo homogenous plants per pot. At the end of the growing period(12 weeks), the above ground biomass (shoot) was harvested fromeach pot and the shoot fresh weight was recorded. Then the plantwas oven-dried at 70 �C for 48 h, and the over dry weight (dwt) wasrecorded. Plant samples were ground to a fine powder using alaboratory mill with 0.5 mm sieve.

At the end of the experiment, representative soil sample wastaken from each pot after thoroughly mixing and was analyzed forthe same parameters mentioned above. Soil samples were sievedthrough 2mm sieve for chemical and biological analyses. Moreover,at the end of the experiment, analysis of variance (ANOVA) wasused to determine the treatment effects. When the F ratio wassignificant, a multiple-means comparison was performed usingFisher's Least Significance Test (0.05 probability level). Statisticalanalyses were performed with SYSTAT Statistical Program(Wilkinson, 1990).

3. Results and discussion

3.1. Jacto.MBR reactor

The aeration system of Jacto.MBR combined with the turbulentmixing makes both soluble organic compounds and dissolved ox-ygen (DO) accessible for microorganisms. The DO was measuredaround 70% of the saturation, showing thus the good aerationtransference ability of the reactor. It should be noted that duringthe Jacto.MBR reactor operation, the acidic pH of OMW increased tovalues higher than 6.0 and this was probably due to the degradationof organic acids. During OMW treatment with continuous feeding,temperature values measured inside the reactor were self-maintained in the range between 30 and 40 �C. The Food toMicroorganism ratio (F/M) is an important factor controlling theCOD removal by microbial treatment. F/M ratio represents theloaded mass of organic compounds (substrate), in terms ofconsumed COD, in relation to the mass of microorganisms withinthe reactor, in terms of volatile suspended solids (VSS). If the sys-tem is submitted to heavy washing water loads that are low insoluble COD, the microbial population may be reduced because ofthe lack of food. Additionally, wash-out of microorganisms canoccur if the hydraulic loading is greater than the outflow rate. Withthe incorporation of an UFmembrane in the recirculation system ofthe Jacto.MBR reactor, the wash-out of microorganisms was avoi-ded, being possible to operate with high VSS (biomass concentra-tion, around 5e10 g/L), allowing thus for a high F/M ratio (1.27e1.67d�1).

Three values of hydraulic retention times (HRT) (i.e. 3, 6 and 9

e phytotoxicity experiments.

MW used for the experiments of:

aw OMW Jacto.MBR Solar photo-Fenton oxidation UF/NF

.70 6.19 2.80 2.02

.00 6.00 6.00 6.00

.60 5.30 1.50 11.40236.00 362.00 310.00 378.0018.80 12.10 0.83 10.87666.7 573.50 6.20 343.106.80 68.60 10.00 33.6069.50 300.80 152.00 152.00441.80 343.30 45.40 175.20.94 10.00 n.d. 21.00

0

20

40

60

80

100

3 6.67 10

X re

mov

al (%

)

FeSO4.7H2O (g/L)

CODDOCTSS

Fig. 2. Coagulation/flocculation pre-treatment of OMW samples (Experimental con-ditions: 3e10 g/L FeSO4$7H2O and 0.287 g/L FLOCAN 23).

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e136

days) and various organic loading rates (OLR: 6e18 kg/m3 d) wereexamined, in order to obtain the optimum operating conditions (asshown in Fig. 1); while the OMW feeding rates applied were 0.01,0.02 and 0.04 m3/d for HRT 9, 6 and 3 d, respectively. It wasobserved that under the higher examined retention time (i.e. HRT:9 d) and at OLR equal to 6.7 kg COD/m3 d, a high removal efficiencywas achieved, and more specifically 85% of COD and 80% of TPhwere removed. When decreasing the HRT to 6 d, the best efficiencyof the process was achieved at the higher feeding rate applied (i.e.OLR: 13.9 kg COD/m3 d), which was equal to 90% COD and 80% TPhremoval. On the other hand, when reduction of the HRT to 3 d wasapplied, the performance of the reactor was negatively affected (i.e.decreasing to 70% COD removal and 52% TPh removal), which wasmainly attributed to the low food/microorganism (F/M) ratio (i.e.high biomass content inside the reactor and low substrate avail-ability). It should be mentioned that during OMW biologicaltreatment with continuous feeding at HRT 9 d, the pH increased tovalues higher than 6.0; while after changing the feeding regimen toHRT 3 d, a greater fluctuation was observed in measured pH valuesbut with tendency to values around 6.0. According to the above, theoptimum operating conditions of this system were achieved attested HRT of 6 days and OLR equal to 13.9 kg COD/m3 d.

Twomain advantages of this process are the fact (i) that there isno need for chemicals, in order to adjust the effluents' pH prior totheir disposal in the environment, since the pH of the UF-permeateis already neutral, and (ii) that it is suitable for both small- andlarge-scale olive mills.

3.2. Solar photo-Fenton oxidation

OMWwas pre-treated with coagulation/flocculation, in order toreduce the solid fraction of the effluent, improving thus the effi-ciency of the solar photo-Fenton oxidation. The OMW pre-treatment consisted at first of coagulation with FeSO4$7H2O fol-lowed by flocculation using the anionic polyelectrolyte FLOCAN 23.Here it is noted that FeSO4$7H2O and FLOCAN 23 have beenselected as the coagulant and flocculant agent, respectively, ac-cording to previous findings of our group (Papaphilippou et al.,2013). The experiments were performed at constant FLOCAN 23concentration (0.287 g/L), whereas FeSO4$7H2O concentration wasaltered from 3 to 10 g/L, as shown in Fig. 2. The treatment of OMWwith 6.67 g/L of FeSO4$7H2O and 0.287 g/L of FLOCAN 23, whichwere the optimum reagents' concentrations, achieved 90% TSSremoval, while the COD and DOC were removed by 40% and 11%,respectively. Hence, coagulation/flocculation pre-treatment wasable to significantly reduce the solid fraction of OMW (expressed asTSS). However, the organic load of these effluents remained at high

Fig. 1. Effect of organic loading and hydraulic retention time on the COD and totalphenols removal efficiency in OMW treatment using the JACTO.MBR reactor.

values. As a consequence, solar photo-Fenton oxidationwas furtherapplied as a post-treatment, in order to reduce both OMWs' organicload and their ecotoxicity, as well.

The solar photo-Fenton experiments were carried out using pre-treated OMW (by coagulation/flocculation) that had been diluted30 times with secondary-treated urban wastewater effluent. Itshould be mentioned that the dilution of the OMW sample prior tothe solar photo-Fenton treatment was considered necessary mainlyto reduce the inner filter effects into the photoreactor. This dilution(30�) was selected as the optimum one, after a series of bench-scale experiments that were performed in a previous work of ourteam (Papaphilippou et al., 2013).

A series of pilot-scale experiments was conducted with severalcombinations of ferrous (0.02e0.1 g/L) and H2O2 concentrations(0.5e2.0 g/L), in order to obtain the optimum operating conditions(Fig. 3 (a) and (b)). As shown in Fig. 3 (a), the reduction of COD inOMW was found to be always faster in the early stages of the solaroxidation reaction (i.e. during the first 30 min of treatment) than inthe later stages. The maximum COD removal was observed at0.08 g/L Fe2þ within 240 min of irradiation, while an excess offerrous concentration (i.e. [Fe2þ] ¼ 0.1 g/L) in the system produceda decrease in the COD removal. It is well known, that when the ironconcentration is very low, H2O2 is consumed by less desirable re-actions, without the generation of hydroxyl radicals; while withhigher iron doses, the process is accelerated due to the regenerationof ferrous from ferric iron, resulting to the rapid generation ofadditional radicals (Zapata et al., 2009). On the other hand, too highconcentrations of iron can generate dark zones in the photoreactor,reducing thus the process efficiency (Malato et al., 2009; Michaelet al., 2014). Moreover, regarding the influence of oxidant concen-tration on the reaction rate (Fig. 3 (b)), it was observed that neithertoo low oxidant concentration (leading to a rate reduction of theFenton reaction), nor too high concentration (H2O2 act as scavengerof the hydroxyl radicals produced) may be applied. Usually how-ever, there is a rather broad concentration interval between bothextremes, where none of these two phenomena occurs. As aconsequence, the maximum COD removal, which was equal to 94%,was observedwith catalyst concentration of 0.08 g/L and H2O2 doseof 1.0 g/L after 240 min of solar treatment. Under the optimumexperimental conditions of solar photo-Fenton oxidation, BOD5,DOC, TSS and TPh were removed by 86%, 43%, 96% and 99.8%,respectively, as shown in Table 3.

It is noted that OMW is a complexmixture containing numerousorganic/inorganic compounds whose contribution to toxicity andplant phytotoxicity is not generally known yet. In addition, duringthe advanced chemical treatment of OMW, the oxidation productsproduced may be even more toxic, than the initial effluent, asmentioned before. Thus, the toxicity of OMW prior to and after thesolar photo-Fenton oxidation towards D. magna and the three plant

Fig. 3. Effect of (a) initial ferrous concentration (Experimental conditions: [H2O2]0 ¼ 1.5 g/L; pH ¼ 2.8e2.9) and (b) initial H2O2 concentration (Experimental conditions:[Fe2þ]0 ¼ 0.08 g/L; pH ¼ 2.8e2.9) on the COD removal of pre-treated OMW.

Table 3Removal efficiencies of the treatment processes applied.

Parameters % Removal Efficiency

Jacto.MBR Coagulation/flocculation þ solar photo-Fenton UF/NF

COD 90.0% 94.0% 92.0%DOC not measured 43.0% not measuredBOD5 not measured 86.0% not measuredEC not measured not measured 90.0%TSS 99.0% 96.0% not measuredTPh 80.0% 99.8% 95.0%

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e13 7

species was further investigated.Firstly, in order to determine the toxicity of the raw diluted

OMWused as a feed flow of solar treatment, a set of control toxicitytests was performed by exposing D. magna to these samples. Thecontrol tests showed 50% and 85% immobilization of D. magna after

24 h and 48 h of exposure time, respectively. With regard to theefficiency of the coagulation/flocculation pre-treatment inremoving OMW toxicity, it should be mentioned that it wasnegligible. The results of the solar treatment showed that thetoxicity towards D. magna usually increased during the first

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e138

minutes of the oxidation (i.e. 30 min of treatment), due to theformation of toxic oxidation products, while after a suitable treat-ment time (usually over 90min) the toxicity was gradually reduced,indicating thus the potential oxidation of these products. Finally,after 240 min of treatment the immobilization of daphnids after24 h of exposure was lower than 15%, while after 48 h of exposurethe respective value was lower than 25%. This was in accordancewith the fact that at the end of the solar photo-Fenton oxidation,the fraction of toxic phenolic compounds of OMW was completelyeliminated (TPh removal ~ 99.8%).

Secondly, the phytotoxicity of OMW towards three plant species(i.e. S. alba, S. saccharatum and L. sativum) was evaluated comparingthe mean number of germinated seeds and the mean root andshoot length in the control (i.e. tap water) and in each examinedOMW sample. Regarding the raw undiluted OMW, it should benoted that no germination of seeds was observed, mainly due to itshigh phenolic content (i.e. 4.2 g/L). On the other hand, the OMWpre-treated with coagulation/flocculation and the raw dilutedOMW sample (30�) provoked a negligible phytotoxic effect on theseed germination (germination inhibition (GI) < 5%), which waseliminated after 240 min of solar photo-Fenton oxidation.Furthermore, the phytotoxicity in the treated samples expressed asroot inhibition (RI) and shoot inhibition (SI), varied differentlycompared to the GI. The growth of the roots and shoots wasaffected by the treated samples, potentially reflecting their ownspecific mechanisms for growth and adaption to the oxidationproducts. Interestingly however, was the fact that the SI and RI ofthe treated samples were considerably higher compared to the GI.Nevertheless, it is important to note that the RI and SI of all plantspecies at the end of the solar photo-Fenton treatment were muchlower than that of the raw diluted OMW effluent (i.e. SI: 57e72%and RI: 45e62% for the three plant species), indicating thus thehigh efficiency of the applied treatment technology in removingphytotoxicity. More specifically, the SI and RI of the solar photo-Fenton treated OMW effluents were between 27-41% and18e32%, respectively, for the three plant species examined.

From the above, it can be concluded that the combination of thecoagulation/flocculation and solar photo-Fenton oxidation seemsto be a promising process for the effective purification of OMWeffluents, as well as for the minimization of their toxicity, and canbemainly applied to small-scale olive mills in countries, like Cyprusor other Mediterranean countries, with plenty of sunshine.

3.3. Integrated UF/NF process

Raw OMW samples were first acidified to a pH of 2.0 with sul-phuric acid (2.8 mL sulphuric acid/L OMW) to promote the sepa-ration of water insoluble polyphenols from water solublepolyphenols, as mentioned before. During the acidification, thesolid fraction of the effluent was separated and one portion settledto the bottom, while an oil rich fraction (4.5e9.8% oil) floated to thetop. The permeate rate in the UF system was set to ~20 L/h whichgave an initial flux of 21 L/m2 h, while the transmembrane pressure(TMP) was monitored over time and a permeate backwash wasapplied every 3min for 20 s as an optimal setting based on previousexperiments (data not shown). NF was used to concentrate thephenolic compounds from the UF permeate and get a reduced levelof these compounds in the NF permeate intended for irrigation. Itshould be mentioned that since monovalent ions pass through NFmembranes, and as consequence they do not contribute to theosmotic pressure, the NF systems usually operate at feed pressuresbelow those of RO systems. However, in this case the sulphate ionsproduced during the acidification step of OMW effluents werecontributed to the osmotic pressure, and as a result pressures above20 bar were needed to be applied in the NF step of this study.

Firstly, a quite noteworthy reduction of COD was observedduring the acidification step of OMW effluents with an average of30% COD removal, while a slight reduction of TPh was observed.Subsequently, during the first membrane filtration process applied(i.e. UF), both the COD and TPh removals were quite low (lowerthan 20%). On the other hand, the treatment efficiency of the NFstep using the spiral-wound modules was high. More specifically,the COD removal ranged from 76 to 92%, while the TPh removalachieved was up to 84%. The results are in line with those of thestudies of Coskun et al. (2010) and Ochando-Pulido et al. (2014),where up to 83% COD removal was achieved after the NF process, aswell as El-Abbassi et al. (2011), where 74% TPh recovery was ach-ieved after the UF process.

However, in case of using the NF270 spiral-wound membranemodules, the combination of significant dead volume and thelimited sample volumes (100e160 L), showed that it was notpossible to achieve a volume recovery exceeding 70% (i.e. themaximum volume concentration factor (VCF) achieved was equalto 3). Therefore, flat-sheet membrane modules were used, in orderto investigate their ability to achieve higher VCF values (i.e. VCF4e6). In comparing the NF270 membranes in spiral-wound andflat-sheet modules, it was shown that the overall performance ofthe flat-sheet modules was significant higher than that of spiralwound (i.e. in the case of flat-sheet modules the COD removalranged from 86 to 92% and TPh removal was up to 95%). When theNF systemwas operated in VCF values of 2e3, the fluxes of the flat-sheet membranes were higher (i.e. 30 and 20 LMH for DK5 and 38and 17 LMH for NF270, respectively) compared to those obtainedwith the spiral-wound NF270 membrane (i.e. 25 and 8 LMH). In thecase of applying higher VCF values (i.e. VCF 4 to 6) the fluxes of bothNF flat-sheets modules were kept above 10 LMH. The reduction ofthe specific flux (i.e. flux/applied TMP) by increasing the VCF valuesreflected both the increase of (i) the osmotic pressure of theconcentrate and (ii) the hydraulic resistance of foulants on thesurface of the membrane. The increase in the TPh and COD removalin the case of the flat-sheet membranes compared to spiral woundcan be attributed to the higher fluxes and the less fouling of theformer compared to the latter module configuration. In the case ofusing the flat-sheet membrane modules, the retentions of thevarious individual phenolic compounds examined in this study (i.e.tyrosol, hydroxytyrosol, vanilic acid, syringic acid, p-coumaric acid,gallic acid and caffeic acid), were found to be much higher for DK5NF membrane (~100% for all the examined phenolic compounds)compared to those of NF270 flat-sheet membrane (i.e. 50e95%).

Regarding the MARS process used for the recovery of phenoliccompounds concentrated in the NF concentrate stream (TPh:~4e7 g/L, ~4% polyphenols), it should be mentioned that themembranes that were developed and used in this study were foundto be selective mainly for tyrosol and hydroxytyrosol transportcompared to the other phenolic compounds, since their mass bal-ances were found to be close to 100%. More specifically, the passageof other OMW components, such as acids, carboxylate, polyphenolsand non-aromatic carbon chains, through these membranes wasvery low. As a consequence, a final effluent which was rich in thesetwo high-added value compounds (i.e. 15þ% polyphenol solution)was generated through MARS process. As a result, the MARS pro-cess appeared to be promising for real application, due to its highrecovery efficiency and simple operation, while the value of this by-product makes the entire process cost effective. In addition, thereare other processes achieving also significant separation ofphenolic compounds from OMW. For example, in the study of El-Abbassi et al. (2012), the direct contact membrane distillation(DCMD) process with polytetrafluoroethylene (PTFE) membraneswas able to separate polyphenols from OMW by 100% after oper-ating DCMD for 8 h. Moreover, Garcia-Castello et al. (2010) used a

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e13 9

system including MF, NF, osmotic distillation and vacuum mem-brane distillation (VMD), which was able to recover 78% of theinitial content of OMW' polyphenols in the permeate stream.

It is important to highlight that this combined membranetreatment technology is suitable for both small- and large-scaleolive mills, since a mobile OMW membrane treatment systemwas developed, with a treatment capacity up to 180 m3 per day (i.e.UF membrane area: 400 m2 and NF membrane area: 300 m2). Thismobile system is able to reuse the NF permeate mainly for irriga-tion purposes, whereas the UF and NF concentrates are able to besent to a central processing plant for isolation of the valuablephenolic compounds.

3.4. Greenhouse phytotoxicity experiments

The following samples were used as water sources for irrigatingmaize grown in pots to determine their phytotoxic effects on plantgrowth and nutrients uptake by maize: (i) tap water (control), (ii)tap water with the addition of recommended rate of fertilizer(80 kg DAP ha�1), (iii) raw OMW, (iv) Jacto.MBR treated effluent, (v)solar photo-Fenton treated effluent and (vi) UF/NF treated effluent.

First of all, the pH of the raw and treated OMW samples wasadjusted to a near neutral value (i.e. pH: 6), since their initial pH -especially of solar photo-Fenton and UF/NF treated effluents - wastoo acidic to be used for soil application, due to its inhibition to theplant growth. As shown in Table 2, according to the electricalconductivity (EC) values, the raw OMW, Jacto.MBR and UF/NFtreated effluents were strongly saline, while solar photo-Fentontreated effluent was moderately saline. In addition, the TSS in theraw OMWwere high (1236 mg/L), followed by the Jacto.MBR, solarphoto-Fenton and UF/NF treated effluents. Moreover, the COD ofthe raw OMW was 118.8 g/L, while the organic load of the treatedeffluents was significantly lower, ranging from 0.83 to 12.1 g/L COD.With regard to the levels of nutrients (i.e. N, P and K), these werehigher for the raw OMW, followed by the Jacto.MBR treatedeffluent, while considerably lower values were observed in thecases of UF/NF and solar photo-Fenton treated effluents.

It should be highlighted that the amounts of both tap water withand without fertilizer, and treated/untreated OMW effluents thatwere applied to the soil for the phytotoxicity experiments, variedfrom 5.41 L in the case of raw OMW to 12.71 L for the controlexperiment (i.e. tap water), as shown in Table 4. The variation in thewater amount applied to the soil was attributed to the fact that theapplied amounts were calculated based on the amount needed toreplenish soil field capacity water content. This was a function ofthe soil moisture lost through evaporation and transpiration, whichwas significantly affected by the type of the water sample applied.More specifically, the amount of the evapotranspiration in the potsreceiving the raw OMW effluent was found to be the lowest andtherefore, the added amount to replenish the soil field capacity

Table 4Comparison of the plant growth of maize after irrigation with the raw OMW and the tre

Effluent Water added(L/pot)*

dWt (g/plant)* dWt relaticontrol (%

Tap water (control) 12.71a 67.70b 100bTap water þ fertilizer 12.33a 90.01a 133aRaw OMW 5.41c 14.18e 21eJACTO.MBR 7.35b 28.86d 43dSolar photo-Fenton

oxidation8.18b 48.71c 72c

UF/NF 7.97b 31.28d 46d

dwt: dry weight.* Different letters within each column indicate significant differences using the LSD testData adopted from a previous work of our group (Rusan et al., 2016).

water content in the pot was the lowest (i.e. 5.41 L).Comparing all the water sources applied in the phytotoxicity

experiments, the plant dry weight (dwt) was the highest with thesoil application of tap water with fertilizer, followed by the controlsample (i.e. where tap water alonewas used), as shown in Table 4. Itshould be mentioned that the plant dwt obtained with the appli-cation of raw OMW was significantly reduced to 21% compared tothe control sample. On the other hand, the application in the soil ofthe treated OMW effluents of the different technologies tested,resulted in different effects on plant dwt. More specifically, solarphoto-Fenton treated effluents resulted in plant dry weight of 72%compared to the control experiment, while Jacto.MBR and UF/NFtreated effluents exhibited lower dwt, approx. 45% of the controlsample. From the above results, it can be concluded that the rawOMW was found to be highly phytotoxic to maize, as indicated bythe lower plant weights compared to those of the treated OMWeffluents, and especially the solar photo-Fenton treated effluent,which caused the lowest plant growth inhibition (i.e. 28% inhibi-tion). As a consequence, OMWeffluents' phytotoxicity was found todecrease to a different extent upon their treatment with the tech-nologies examined herein, mainly due to the sufficient removal ofTPh and other phytotoxic compounds from the raw effluents.Similar results were obtained in another study, where the sametreatment technologies were examined regarding their ability toreduce phytotoxicity of OMW effluents on seed germination ofbarley plant (Rusan et al., 2015).

On the other hand, it should be noted that the behavior of plantheight was quite different compared to the plant dry weightbehavior, with UF/NF treated effluent caused the highest inhibition(59%), followed by the raw OMW (39%), the Jacto.MBR treatedeffluent (26%), while the solar photo-Fenton treated effluent causedonce again the lowest plant height inhibition (8%) (Table 4). It is notclear why the plant height responded differently than the plantweight, but the unavoidable variation in the light intensity in thegreenhouse could have caused this variation in the plant height,since the plant tends to grow higher under lower light intensity.

Furthermore, the recovery of the irrigation water applied in thecase of applying raw OMW effluents was the lowest (2.74 gdWt/L)compared to the tap water with and without fertilizer (7.30 and5.32 gdWt/L, respectively) and the treated OMW samples (whichranged from 3.93 to 5.96 gdWt/L), as presented in Table 4, suggestingthat the loss of soil moisture was mainly due to the lower evapo-transpiration and not to the leaching and deep percolation (AbuAwwad, 1996), since the latter are controlled in a pot experiment.According to the study of Jameel et al. (2011), the presence of oiland grease contained in OMWeffluents leads to the formation of oillayer onwater surface, which causes significant pollution problems,such as reduction of light penetration, photosynthesis and oxygendiffusion. This surface oil layer plays the role of hydrophobic mulch,modifying thus the behavior of water in the soil profile and

ated effluents of the technologies examined.

ve to)*

Plant height (cm)* Plant height relativeto control (%)*

Recovery(gdwt/L)*

113.0b 100b 5.32b128.5a 114a 7.30a69.3e 61e 2.74d83.8d 74d 3.93c103.4c 92c 5.96b

45.8f 41f 3.93c

at the 0.05 probability level.

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e1310

reducing water loss through evaporation (Sahraoui and Mellouli,2010).

In addition, the percentage concentrations of the nutrients N, Pand K in the maize plant were found to be slightly higher both inthe case of using tap water with fertilizer and raw OMW, comparedto the treated OMW effluents and the control sample, as was ex-pected (Table 5). However, the uptake of these nutrients by theplant was found to be higher only in the case of applying tap waterwith fertilizer in the soil and not in the case of applying raw OMWeffluent. Similar trends were found for the uptake of Fe and Zn,while the differences in the uptake of Mn, Cu, Cd and Pb wereinsignificant. On the other hand, it should be highlighted that theplant uptake of K in case of applying tap water (control) to soil wasas high as that obtained with the addition of fertilizer in tap water,indicating thus the soil was originally rich in available potassium.However, the uptake of N and Pwas found to be lower in the controlsample compared to the tap water with the addition of fertilizer,suggesting thus that the soil was deficient in those two nutrients.Moreover, it should be mentioned that the uptake of these nutri-ents in case of applying in the soil treated OMW samples differswidely. As shown in Table 5, the highest nutrients' plant uptakewasobserved when applying solar photo-Fenton treated effluents,while Jacto.MBR and UF/NF treated effluents' uptake was quitesimilar. On the other hand, it should be mentioned that the uptakeof metals (i.e. Mn, Zn, Cu, Cd and Pb) in case of applying the rawOMW to the maize plant was approximately the samewith those ofapplying all the treated OMW effluents that were examined in thisstudy.

4. SWOT analysis (Strengths, Weaknesses, Opportunities andThreats) of the OMW treatment processes tested above

The SWOT analysis is a tool that identifies the strengths,weaknesses, opportunities and threats of a project, technology ormethod. Specifically, SWOT analysis is a basic, straightforwardmodel that assesses what a proposed new technology or methodcan add to the existing operation of a business, showing the posi-tive and negative areas, as well as its potential opportunities andthreats (Pickton and Wright, 1998). The method of SWOT analysistakes into consideration all the relevant information and separatesthem into internal (Strengths and Weaknesses) and external (Op-portunities and Threats) factors.

In this study, a SWOT analysis was performed, in order todetermine the strengths of the treatment technologies proposed,under specific circumstances, as well as the possible obstacles thatneed to overcome or minimize for achieving the desired levels ofOMW treatment, in terms of the quality of the treated effluents

Table 5Plant nutrients and heavy metals content of maize after irrigation with the raw OMW an

Parameters Tap water (control)* Tap water þ fertilizer* Raw

N (%) 1.46c 1.81a 1.69P (%) 0.14c 0.22a 0.21K (%) 2.04b 2.53a 2.36dwt (g/plant) 67.70b 90.01a 14.1N (g/plant) 0.99b 1.63a 0.24P (g/plant) 0.09b 0.20a 0.03K (g/plant) 1.38b 2.28a 0.32Fe (mg/plant) 3.68a 4.12a 1.74Mn (mg/plant) 1.89 1.96 1.44Zn (mg/plant) 0.39b 0.74a 0.27Cu (mg/plant) 0.02 0.16 0.03Cd (mg/plant) 0.06 0.03 0.03Pb (mg/plant) 0.18 0.18 0.08

* Different letters within each raw indicate significant differences using the LSD test at tData adopted from a previous work of our group (Rusan et al., 2016).

discharged into the environment. In order to evaluate the possibleimpact of each factor on each technology, the scoring system wasbased on the level of the impact on the treatment technology. Thescoring regarding the potential impact of each factor has beenclassified as: (i) low impact - monitor for changes; (ii) mediumimpact - focus on after the high impact and priority items havebeen resolved; and (iii) high impact - the main focus (i.e. thereforeit needs to be ensured that adequate resources exist to addressthese issues).

Based on the answers given by the partners of this study and byvarious relevant stakeholders with regard to the potential impact ofeach factor for each treatment technology (i.e. relative importance,data not shown), the average value of each factor was calculated asindicated in the Average Potential Impact Weight (APIW) inTables S1 and S2, for internal and external factors, respectively. Thisaverage value indicates the weight of each factor showing howsignificant the factor is to the successful implementation of theproposed technologies for the treatment of OMW effluents. Thesum of the APIW for both the strengths and the weaknesses (i.e.internal factors) equals to one (1) unit, regardless the number of thefactors under study. The same applies for the sum of the APIW ofthe opportunities and threats (i.e. external factors).

4.1. Internal factors (Strengths and Weaknesses)

The internal factors (Strengths and Weaknesses) are shown inTable S1. Based on the APIW results, the most important strengthsseem to be how well-tested a proposed technology is, whether it isan environmentally friendly technology and how much value eachtechnology adds to the society (high impact scoring, i.e. APIW: 0.1).Further to these factors, the cost of implementation (i.e. capitalcost), the lack of strict regulations, the operational/maintenancecost, the uniqueness of the technology, the local communitiessupport, the global economic crisis and the weak financial positionof the majority of the olive millers are considered to be of mediumimpact weaknesses (i.e. APIW: 0.05). Finally, the lowest impact isgiven to the suitability of the proposed technologies for large-scaleolivemills and the cost needed for the storage of the OMWeffluentsprior to the treatment (low impact scoring, i.e. APIW: 0.025).

4.2. External factors (Opportunities and Threats)

The external factors (Opportunities and Threats) are presentedin Table S2. Based on the results, it will be of great value if thetreated OMWwould be available for irrigating the olive trees and ingeneral reducing the existing water consumption cost for olivemills (high impact scoring, i.e. APIW: 0.1). The major threats are

d the treated effluents of the technologies examined.

OMW* Jacto.MBR* Solar photo-Fenton oxidation* UF/NF*

ab 1.36c 1.55bc 1.57bca 0.17bc 0.15bc 0.18bca 2.09b 2.05b 2.08b8c 28.86d 48.71c 31.28de 0.39bc 0.76bc 0.49bcd 0.04c 0.07b 0.06ce 0.60d 1.00c 0.65db 1.44b 1.58b 1.92b

1.62 1.81 1.91c 0.21c 0.36b 0.44b

0.03 0.03 0.020.03 0.04 0.030.12 0.08 0.09

he 0.05 probability level.

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e13 11

considered to be the poor environmental awareness in combinationwith the weak regulations. The funding difficulties of the treatmenttechnologies are an additional threat lowering the motivation leveland the interest of the olive millers in adapting the proposedtechnologies (medium impact scoring, i.e. APIW: 0.05). On theother hand, the positive externalities on real estate prices and thepotential to benefit from the high-added value by-products (i.e.phenolic compounds) as an additional business income stream,constitute the main opportunities for the real application of thesetechnologies (medium impact scoring, i.e. APIW: 0.05). The lowestimpact is given to the difficulties for funding the treatment plant,the licensing procedures, the declining olive oil production, and thedifficulties to store and/or transfer the OMW prior to treatment(low impact scoring, i.e. APIW: 0.025).

From all the above, it can be concluded that the selection of anappropriate technology for a proper and effective management ofOMW produced by an individual olive mill or a group of olive millssituated in a specific area, is based on a variety of factors. Althoughthe above mentioned ranking is not absolute, it is indicative thatthis hierarchy should serve as the basis for the decision makingprocess. Therefore, combining the abovementioned results with thetreatment efficiency and the capital/operational/maintenance costof each treatment technology, the possible profit and the positiveexternalities, the decision makers can have a strong indicationconcerning the optimum strategy.

5. Economic evaluation of the OMW processes tested above

One of the main objectives of this study was to estimate theoverall cost of the three processes applied needed for the treatmentof 1 m3 of OMW. The pilot-scale approach regarding OMW treat-ment served as the basis for carrying out an assessment of theapplication of the tested technologies from an economic, social andenvironmental point of view, taking into consideration a number ofassumptions and projections. This is without doubt one of the mostimportant outcomes of this study, since it provides clear andvaluable information for olive mill operators/managers, who mayrealize the economic potential of the treated OMWeffluents, investin the proposed treatment solutions and reduce the legal risksassociated to the environmental non-compliance of their effluents.

Olive mill owners/operators have to manage the cost neededevery season for the treatment of OMW, in order to be safely dis-charged into the environment, the OMW odor and the variouscomplaints they receive, in a way that is acceptable by the localauthorities. However, (i) the weak enforcement of the currentlegislation of each country, (ii) the fluctuation in their olive oilproduction level each year and the unstable financial income, (iii)the global financial crisis of the last several years that affected allsectors across the board and (iv) the lack of any tangible financialmotivation to invest in a new treatment technology for properlyprocessing their OMW, minimize significantly their interest tomove forward towards a more environmentally friendly system forthe treatment of their effluents.

It is important to underline that generally the olive mills aresmall enterprises, family businesses in the main, scattered aroundthe olive production areas, making individual on-site treatmentoptions unaffordable (Paraskeva and Diamadopoulos, 2006). Takingthis into consideration, a viable scenario would include the oper-ation of a treatment plant to serve a cluster of olivemills co-existingin the same area, reducing thus the overall investment and oper-ational cost for the treatment of OMW. Thus, this cost estimation isbased on a cluster of olive mills that are expected to produce anaverage of (i) 50 m3 of OMW per day and operating for 365 d/yr inthe case of Jacto.MBR, taking into account that the effluent will bestored in a tank in order to continuously feeding the reactor

throughout the year, (ii) 50 m3/d and operating for 365 d/yr in thecase solar photo-Fenton oxidation (i.e. a storage tankwill be used inorder to store OMWeffluents during the period that olive mills didnot operate, in order to be able the treatment plant to be fed andoperate all year round without interruption), and (iii) 180 m3/dayoperating for 42 days/season in the case of UF/NF technology, for atotal period of ten years.

The cost benefit analysis of the three treatment technologiesapplied in this study for the treatment of OMW is divided in twomain financial areas: (i) the capital cost and (ii) the operational andmaintenance cost.

5.1. Capital cost

In this study, the investment (capital) cost is considered toinclude civil and mechanical works, buildings, engineering designsand supervision of on-site infrastructure, start-up costs and work-ing capital. The capital cost of the Jacto.MBR reactor and the solarphoto-Fenton technology were 80,000 V and 83,000 V, respec-tively, for the treatment of 50 m3 OMW/d, while the capital cost ofthe integrated UF/NF technology was 238,000 V for a treatmentcapacity of 180 m3 OMW/d, based on data taken from themanufacturing companies of these treatment technologies.

5.2. Operation and maintenance cost

It should be mentioned that the operation/maintenance cost ofeach technology was estimated for a period of ten years, while forthe calculation of the operating costs the seasonal variations in theflow of OMW have been taken into consideration. The analysis ofthe operation and maintenance expenses mainly includes thefollowing: (i) personnel salaries for maintenance and operation, (ii)cost of electrical energy based on the electricity prices of eachcountry (i.e. Jacto.MBR _ Portugal (0.15 V/kWh in average), solarphoto-Fenton oxidation _ Cyprus (0.25 V/kWh) and UF/NF _ Israel(0.06 V/kWh)), and (iii) chemicals' consumption and consumables.In addition, maintenance costs include both regular repairs (e.g.mechanical, electrical, electronic and civil parts, etc.) and minor ormajor replacements (e.g. small or large parts for pumps, blowers,motors, etc.). This cost is usually estimated as a percentage of thetotal investment cost, and in the technologies examined herein,approx. 1.5% of the initial investment cost per annum can beconsidered as representative. Finally, the chemicals' cost considersthe reagents needed for the advanced chemical oxidation process(i.e. 1.10 V/Kg H2O2, 1.30 V/Kg FeSO4$7H2O, 2 V/Kg H2SO4 and0.55 V/Kg NaOH), as well as the chemicals used for the pHadjustment and cleaning of themembranes in the case of the UF/NFprocess (i.e. 1.6 V/m3 of treated OMW).

According to the above, the operating cost of the Jacto.MBRreactor reached on an average annual level the amount of 9700 V

for the treatment of 50 m3 OMW per day, the solar photo-Fentonprocess reached 21,200 V/year for 50 m3/d and the integrated UF/NF technology 55,300 V/year for the treatment of 180 m3/d.

From all the above, it can be concluded that the total ownershipcost (capital and operational cost) for a 10-year period of the abovementioned technologies is as follows: (i) Jacto.MBR reactor(177,000 V for 50 m3/d and operation period: 365 d/yr), (ii) solarphoto-Fenton oxidation (295,000 V for 50 m3/d and operationperiod: 365 d/yr) and (iii) integrated UF/NF processes (751,000 V

for 180 m3/d and operation period: 42 days/season). More specif-ically, the above overall costs are corresponded to 0.97 V/m3 forJacto.MBR process, 1.6 V/m3 for solar photo-Fenton oxidation, and9.94 V/m3 for UF/NF and membrane contactor technology.

It should be mentioned, that while the overall cost for treatingOMW with the combined membrane process is higher than the

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e1312

other processes examined herein, it actually can have a significantoperating profit by recovering and isolating the high-added valuephenolic compounds present in it (i.e. hydroxytyrosol and tyrosol).According to the study of Nunes et al. (2005), the prices of purifiedhydroxytyrosol extract products can range from 14,900 to20,900V/kg for pure hydroxytyrosol. According to the results of thesame study, the price of a dilute solution of 1.2% hydroxytyrosolextract was 250 V/kg (Nunes et al., 2005). Assuming that the priceof the raw unpurified hydroxytyrosol is 10% of the price of thepurified compound, the price of the raw unpurified compoundcould correspond to 1500e2000 V/kg. According to the experi-mental results of our study, the hydroxytyrosol content of the NFconcentrate fed to the contactor of MARS system was around2.0e2.5 kg/m3 of OMW. Since the volume of the NF concentratewas17% of the feed volume (i.e. 83% volume recovery through NF step)and up to 90% of unpurified hydroxytyrosol could be recoveredfrom the concentrate in the contactor (i.e. MARS system), therecould be a potential profit of 370e610 V/m3 of OMW, by selling inthe market this high-added value compound; however, furtherstudy is required to confirm such a profit.

6. Conclusions

This study provides clear and valuable information regardingthe three tested technologies applied for the purification of OMW,taking into consideration the treatment efficiency, the effluentquality and phytotoxicity, the environmental/social impacts andthe overall cost estimation, that can be used in olive mills accordingto their specific needs. Based on the technological and economicanalysis performed, this study offers technological solutions for themanagement and treatment of OMW effluents for a cluster of olivemills with various capacities and characteristics. Specifically, olivemills with high annual production of OMW can apply membrane-based technologies for both OMW purification (i.e. organic con-tent removal) and recovery of valuable by-products, as well asbiological treatment technologies, while olive mills with lowercapacity can utilize a 'destruction' technology (i.e. solar photo-Fenton oxidation and Jacto.MBR) for the removal of the organiccontent and reduction of phytotoxicity present in these effluents, aswell.

According to the results obtained, all treatment technologiesexamined herein, showed a high performance for the purificationof OMW effluents. More specifically, the Jacto.MBR process wasfound to be highly efficient in removing the organic content of theraw OMW, with a quite low operational cost. However, it was notfound capable of removing OMW phytotoxicity, which should beconsidered for its real application. Solar photo-Fenton oxidationprocess was found to be the most efficient process in removingphytotoxicity from OMW samples, while it also showed a high ef-ficiency in removing OMWorganic load and completely eliminatingits toxic phenolic fraction. This advanced technology, taking intoaccount its reasonable overall cost, can be applied inMediterraneancountries, with plenty of sunshine. On the other hand, the necessityof a pre-treatment step, as well as the extra dilution of influentOMW sample, in order to reduce the solid fraction and enhance thetreatment efficiency, are the major issues that should be taken intoaccount for the real application of this technology. The integratedmembrane processes applied were found efficient in removing theCOD and TPh from the permeate stream, while the high-addedvalue phenolic compounds were significantly concentrated in theconcentrate flow, being capable for recovering and further valori-zation in various industries. Despite the significant higher overallcost of this combined membrane technology, a significant oper-ating profit could be potentially achieved from the valorization ofthe high-added value phenolic compounds. However, further

studies are deemed necessary in order to estimate with greateraccuracy the potential profit from the sale and valorization of theseby-products recovered by this process. Moreover, the lower effi-ciency comparing to the other processes in removing OMWphytotoxicity could be highlighted as the major limitation of thefeasibility of this technology in real application.

Finally, a SWOT analysis was performed, in order to determinethe most important strengths/opportunities of the treatmenttechnologies proposed, as well as the possible obstacles/weak-nesses that the olive millers need to overcome or minimize forachieving the desired levels of OMW treatment. Specifically, themost important strengths/opportunities identifiedwere: howwell-tested a proposed technology is, whether it is an environmentallyfriendly technology and the potential use of treated effluents forirrigating the olive trees, reducing thus the existing water con-sumption cost for olive mills; while the possible weaknesses foundwere: the overall cost of these technologies, the weak financialposition of the olive millers, the poor environmental awareness andthe lack of legislations.

Acknowledgments

This work was prepared in the framework of the project“Mediterranean Cooperation in the Treatment and Valorization ofOlive Mill Wastewater (MEDOLICO, I-B/2.1/090)”which was fundedby the European Union under the “ENPI Cross-Border CooperationMediterranean Sea Basin Programme”. MEDOLICO total budget was1.9 million Euro and it was co-financed through the EuropeanNeighborhood and Partnership Instrument (90%) and nationalfunds of the countries participating in the project (10%).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.watres.2017.02.020.

References

Abu Awwad, A., 1996. Irrigation water management for onion trickle irrigated withsaline drainage water. Dirasat 23 (1), 46e55.

Amor, C., Lucas, M.S., García, J., Dominguez, J.R., De Heredia, J.B., Peres, J.A., 2015.Combined treatment of olive mill wastewater by Fenton's reagent and anaer-obic biological process. J. Environ. Sci. Health, Part A 50 (2), 161e168.

APHA, 1998. Standard Methods for the Examination of Water and Wastewater,twentieth ed. American Public Health Association, Washington, DC, USA.

Azbar, N., Bayram, A., Filibeli, A., Muezzinoglu, A., Sengul, F., Ozer, A., 2004. A reviewof waste management options in olive oil production. Crit. Rev. Environ. Sci.Technol. 34 (3), 209e247.

Baker, R.W., Yoshioka, N., Mohr, J.M., Khan, A.J., 1987. Separation of organic vaporsfrom air. J. Membr. Sci. 31 (2e3), 259e271.

Belaid, C., Khadraoui, M., Mseddi, S., Kallel, M., Elleuch, B., Fauvarque, J.F., 2013.Electrochemical treatment of olive mill wastewater: treatment extent andeffluent phenolic compounds monitoring using some uncommon analyticaltools. J. Environ. Sci. 25, 220e230.

Belaqziz, M., Lakhal, E.K., Mbouobda, H.D., El Hadrami, I., 2008. Land spreading ofolive mill wastewater: effect on maize (Zea maize) crop. J. Agron. 7 (4),297e305.

Ca~nizares, P., Paz, R., S�aez, C., Rodrigo, M.A., 2009. Costs of the electrochemicaloxidation of wastewaters: a comparison with ozonation and Fenton oxidationprocesses. J. Environ. Manag. 90 (1), 410e420.

Ca~nizares, P., Lobato, J., Paz, R., Rodrigo, M., Saez, C., 2007. Advanced oxidationprocesses for the treatment of olive-oil mills wastewater. Chemosphere 67,832e838.

Cassano, A., Conidi, C., Drioli, E., 2011. Comparison of the performance of UFmembranes in olive mill wastewaters treatment. Water Res. 45 (10),3197e3204.

Chapman, H.D., Pratt, P.F., 1961. Methods of Analysis for Soils, Plants and Water.University of California. Division of Agricultural Sciences, Riverside, CA,pp. 169e170.

Chatzisymeon, E., Diamadopoulos, E., Mantzavinos, D., 2009. Effect of key operatingparameters on the non-catalytic wet oxidation of olive mill wastewaters. WaterSci. Technol. 59, 2509e2518.

Chiavola, A., Farabegoli, G., Antonetti, F., 2014. Biological treatment of olive mill

L. Ioannou-Ttofa et al. / Water Research 114 (2017) 1e13 13

wastewater in a sequencing batch reactor. Biochem. Eng. J. 85, 71e78.Coskun, T., Debik, E., Demir, N.M., 2010. Treatment of olive mill wastewaters by

nanofiltration and reverse osmosis membranes. Desalination 259, 65e70.De Leonardis, A., Macciola, V., Lembo, G., Aretini, A., Nag, A., 2007. Studies on

oxidative stabilisation of lard by natural antioxidants recovered from olive-oilmill wastewater. Food Chem. 100 (3), 998e1004.

Dejmkova, H., Scampicchio, M., Zima, J., Barek, J., Mannino, S., 2009. Determinationof total phenols in foods by boron doped diamond electrode. Electroanalysis 21(9), 1014e1018.

El-Abbassi, A., Khayet, M., Hafidi, A., 2011. Micellar enhanced ultrafiltration processfor the treatment of olive mill wastewater. Water Res. 45, 4522e4530.

El-Abbassi, A., Kiai, H., Hafidi, A., Garcıa-Payo, M., Khayet, M., 2012. Treatment ofolive mill wastewater by membrane distillationusing polytetrafluoroethylenemembranes. Sep. Purif. Technol. 98, 55e61.

El Hajjouji, H., El Fels, L., Pinelli, E., Barje, F., El Asli, A., Merlina, G., Hafidi, M., 2014.Evaluation of an aerobic treatment for olive mill wastewater detoxification.Environ. Technol. 35 (24), 3052e3059.

Eus�ebio, A., Mateus, M., Baeta-Hall, L., S�a�agua, M.C., Tenreiro, R., Almeida-Vara, E.,Duarte, J.C., 2007. Characterization of the microbial communities in jet-loop(JACTO) reactors during aerobic olive oil wastewater treatment. Int. Bio-deterior. Biodegrad. 59 (3), 226e233.

Fadavi, A., Chisti, Y., 2005. Gas-liquid mass transfer in a novel forced circulation loopreactor. Chem. Eng. J. 112, 73e80.

Ferreira, F.C., Livingston, A., Han, S., Boam, A., Zhang, S., 2003. Membrane aromaticrecovery system (MARS)-A new process for recovering phenols and aromaticamines from aqueous streams. Membr. Sci. Technol. 8, 165e181.

Garcia-Castello, E., Cassano, A., Criscuoli, A., Conidi, C., Drioli, E., 2010. Recovery andconcentration of polyphenols from olive mill wastewaters by integratedmembrane system. Water Res. 44, 3883e3892.

Gernjak, W., Maldonado, M., Malato, S., Caceres, J., Krutzler, T., Glaser, A., Bauer, R.,2004. Pilot-plant treatment of olive mill wastewater (OMW) by solar TiO2photocatalysis and solar photo-Fenton. Sol. Energy 77, 567e572.

Hamza, M., Sayadi, S., 2015. Valorisation of olive mill wastewater by enhancementof natural hydroxytyrosol recovery. Int. J. Food Sci. Technol. 50 (3), 826e833.

Han, S., Ferreira, F.C., Livingston, A., 2001. Membrane aromatic recovery system(MARS) - a new membrane process for the recovery of phenols from waste-waters. J. Membr. Sci. 188 (2), 219e233.

Hodaifa, G., Ochando-Pulido, J.M., Rodriguez-Vives, S., Martinez-Ferez, A., 2013.Optimization of continuous reactor at pilot scale for olive-oil mill wastewatertreatment by Fenton-like process. Chem. Eng. J. 220, 117e124.

Ioannou, L., Michael, C., Kyriakou, S., Fatta-Kassinos, D., 2014. Solar Fenton: frompilot to industrial scale application for polishing winery wastewater pretreatedby MBR. J. Chem. Technol. Biotechnol. 89 (7), 1067e1076.

Ioannou, L.A., Michael, C., Vakondios, N., Drosou, K., Xekoukoulotakis, N.P.,Diamadopoulos, E., Fatta-Kassinos, D., 2013. Winery wastewater purification byreverse osmosis and oxidation of the concentrate by solar photo-Fenton. Sep.Purif. Technol. 118, 659e669.

ISO 6341, 1996. Water Quality - Determination of the Inhibition of the Mobility ofDaphnia Magna Straus (Cladocera, Crustacea) - Acute Toxicity Test.

Jameel, A.T., Muyubi, S.A., Karim, M.I.A., Alam, M.Z., 2011. Removal of oil and greaseas emerging pollutants of concern (EPC) in wastewater stream. IIUM Eng. J. 12(4), 161e169.

Kallel, M., Belaid, C., Mechichi, T., Ksibi, M., Elleuch, B., 2009. Removal of organicload and phenolic compounds from olive mill wastewater by Fenton oxidationwith zero-valent iron. Chem. Eng. J. 150 (2), 391e395.

Khoufi, S., Aloui, F., Sayadi, S., 2009. Pilot scale hybrid process for olive millwastewater treatment and reuse. Chem. Eng. Process. Process Intensif. 48 (2),643e650.

Khresat, S.A., Rawajfeh, Z., Mohammad, M., 1998. Morphological, physical andchemical properties of selected soils in the arid and semi-arid region in north-western Jordan. J. Arid Environ. 40 (1), 15e25.

Kimmerle, K., Bell, C.M., Gudernatsch, W., Chmiel, H., 1988. Solvent recovery fromair. J. Membr. Sci. 36, 477e488.

Kontos, S.S., Iakovides, I.C., Koutsoukos, P.G., Paraskeva, C.A., 2016. Isolation of pu-rified high added value products from olive mill wastewater streams throughthe implementation of membrane technology and cooling crystallization pro-cess. Transactions 47, 337e342.

Malato, S., Fernandez-Ibanez, P., Maldonado, M., Blanco, J., Gernjak, W., 2009.Decontamination and disinfection of water by solar photocatalysis: recentoverview and trends. Catal. Today 147, 1e59.

Marques, I.P., 2001. Anaerobic digestion treatment of olive mill wastewater foreffluent re-use in irrigation. Desalination 137, 233e239.

Mechri, B., Echbili, A., Issaoui, M., Braham, M., Elhadj, S.B., Hammami, M., 2007.Short-term effects in soil microbial community following agronomic

application of olive mill wastewaters in a field of olive trees. Appl. Soil Ecol. 36(2), 216e223.

Mekki, A., Dhouib, A., Sayadi, S., 2013. Review: effects of olive mill wastewaterapplication on soil properties and plants growth. Int. J. Recycl. Org. Waste Agric.2 (1), 1e7.

Mekki, A., Dhouib, A., Sayadi, S., 2006. Changes in microbial and soil propertiesfollowing amendment with treated and untreated olive mill wastewater.Microbiol. Res. 161 (2), 93e101.

Michael, I., Panagi, A., Ioannou, L.A., Frontistis, Z., Fatta-Kassinos, D., 2014. Utilizingsolar energy for the purification of olive mill wastewater using a pilot-scalephotocatalytic reactor after coagulation-flocculation. Water Res. 60, 28e40.

Michailides, M., Panagopoulos, P., Akratos, C.S., Tekerlekopoulou, A.G., Vayenas, D.V.,2011. A full-scale system for aerobic biological treatment of olive mill waste-water. J. Chem. Technol. Biotechnol. 86 (6), 888e892.

Nunes, A., Duarte, A.R., Sousa, A.R., Martis, A., Dias, F., Angelo, G., Gois, P.,Goncalves, R., Duarte, V., 2 June 2005. In: Olidrox - a Natural Extract from Ol-ives, Presentation at the Cohitec Innovation Conference. COTEC, Portugal.

Obied, H.K., Allen, M.S., Bedgood, D.R., Prenzler, P.D., Robards, K., Stockmann, R.,2005. Bioactivity and analysis of biophenols recovered from olive mill waste.J. Agric. Food Chem. 53 (4), 823e837.

Ochando-Pulido, J.M., Stoller, M., Di Palma, L., Martinez-Ferez, A., 2014. Thresholdperformance of a spiral-wound reverse osmosis membrane in the treatment ofolive mill effluents from two-phase and three-phase extraction processes.Chem. Eng. Process. Process Intensif. 83, 64e70.

Papaphilippou, P.C., Yiannapas, C., Politi, M., Daskalaki, V.M., Michael, C.,Kalogerakis, N., Mantzavinos, D., Fatta-Kassinos, D., 2013. Sequentialcoagulation-flocculation, solvent extraction and photo-Fenton oxidation for thevalorization and treatment of olive mill effluent. Chem. Eng. J. 224, 82e88.

Paraskeva, P., Diamadopoulos, E., 2006. Technologies for olive mill wastewater(OMW) treatment: a review. J. Chem. Technol. Biotechnol. 81 (9), 1475e1485.

Pickton, D.W., Wright, S., 1998. What's SWOT in strategic analysis? Strateg. Change 7(2), 101e109.

Piotrowska, A., Iamarino, G., Rao, M.A., Gianfreda, L., 2006. Short-term effects ofolive mill waste water (OMW) on chemical and biochemical properties of asemiarid Mediterranean soil. Soil Biol. Biochem. 38 (3), 600e610.

Rahmanian, N., Jafari, S.M., Galanakis, C.M., 2014. Recovery and removal of phenoliccompounds from olive mill wastewater. J. Am. Oil Chem. Soc. 91 (1), 1e18.

Rusan, M.J., Albalasmeh, A., Zuraiqi, S., Bashabsheh, M., 2015. Evaluation of phyto-toxicity effect of olive mill wastewater treated by different technologies on seedgermination of barley (Hordeum vulgare L.). Environ. Sci. Pollut. Res. 22,9127e9135.

Rusan, M.J., Albalasmeh, A.A., Malkawi, H.I., 2016. Treated olive mill wastewatereffects on soil properties and plant growth. Water, Air & Soil Pollut. 227 (5),1e10.

Sahraoui, H., Mellouli, H.J., 2010. Effects of olive mill wastewater spreading on thephysical characteristics of soil. EQA-Int. J. Environ. Qual. 5 (5), 1e6.

Sanches, S., Fraga, M.C., Silva, N.A., Nunes, P., Crespo, J.G., Pereira, V.J., 2016. Pilotscale nanofiltration treatment of olive mill wastewater: a technical andeconomical evaluation. Environ. Sci. Pollut. Res. 1e13.

Singleton, V.L., Rossi, J.A., 1965. Colorimetry of total phenolics withphosphomolybdic-phosphotungstic acid reagents. Am. J. Enol. Vitic. 16 (3),144e158.

Strathmann, H., 1975. Water treatment by reverse and forward osmosis. Desalina-tion 16, 179e186.

Ward, W.J., Browall, W.R., Salemme, R.M., 1976. Membrane fabrication/manufacturing techniques. J. Membr. Sci. 1, 99e108.

Wilkinson, L., 1990. SYSTAT: the System for Statistics. SYSTAT, Inc, Evanston, IL.Yalılı Kılıç, M., Yonar, T., Kestio�glu, K., 2013. Pilot-scale treatment of olive oil mill

wastewater by physicochemical and advanced oxidation processes. Environ.Technol. 34 (12), 1521e1531.

Yay, A.S.E., Oral, H.V., Onay, T.T., Yenigün, O., 2012. A study on olive oil millwastewater management in Turkey: a questionnaire and experimentalapproach. Resour. Conserv. Recycl. 60, 64e71.

Zagklis, D.P., Vavouraki, A.I., Kornaros, M.E., Paraskeva, C.A., 2015. Purification ofolive mill wastewater phenols through membrane filtration and resin adsorp-tion/desorption. J. Hazard. Mater. 285, 69e76.

Zapata, A., Velegraki, T., Sanchez-Perez, J., Mantzavinos, D., Maldonado, M.,Malato, S., 2009. Solar photo-Fenton treatment of pesticides in water: effect ofiron concentration on degradation and assessment of ecotoxicity and biode-gradability. Appl. Catal. B Environ. 88, 448e454.

Zehner, P., Kraume, M., 2000. Bubble Columns. Ullmann's Encyclopedia of IndustrialChemistry. Wiley-VCH Verlag GmbH & Co. KGaA. http://dx.doi.org/10.1002/14356007.b04_275.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具