Desalination 248 (2009) 836–842 Monitoring of the quality of winery influents/effluents and polishing of partially treated winery flows by homogeneous Fe(II) photo-oxidation Natasa Anastasioua,b, Maria Monoua, Dionissios Mantzavinosb, Despo Kassinosa* a Department of Civil and Environmental Engineering, University of Cyprus, 75 Kallipoleos, 1678 Nicosia, Cyprus b Department of Environmental Engineering, Technical University of Crete, Polytechneioupolis, GR-73100 Chania, Greece Tel: 003522892275; Fax: 0035722892295; email: dfatta@ucy.ac.cy Received 2 October 2008; accepted 12 November 2008 Abstract Winery wastewaters contain high concentrations of organic compounds including phytotoxic and recalcitrant compounds like phenols. Its treatment by conventional processes is difficult due to the variability of the characteristics of the liquid waste. The main objectives of this work were to (1) monitor onsite the quality of winery wastes prior to and following sequential physical and biological treatment and (2) assess the efficiency of coupling physical and biological treatment to photo-Fenton oxidation serving as the final polishing step. A partially treated effluent with chemical oxygen demand (COD) and biochemical oxygen demand (BOD5) values of 1060 and 210 mg/L respectively was subject to photo-Fenton oxidation at H2O2 and Fe2+ concentrations between 34–175 and 0.5–2 mM respectively, solution pH0 = 2.5, under continuous UV-A irradiation provided by a 125 W lamp. In general, organic matter degradation increased with increasing treatment time reaching values of COD or BOD removal as high as 80% after 4 h of reaction. Regarding the effect of initial iron and hydrogen peroxide concentrations, there appears to be an optimum dosage for both, above which treatment performance deteriorated. Hence, the combined biological + photo-Fenton oxidation resulted in 95% COD removal. Keywords: Advanced oxidation; BOD removal; COD removal; Photo-Fenton; Winery wastewaters 1. Introduction The wine industry generates large volumes of wastewaters originating from various washing steps during the crushing and pressing of grapes *Corresponding author. Presented at the Conference on Protection and Restoration of the Environment IX, Kefalonia Greece, June 30–July 3, 2008 as well as the rinsing of fermentation tanks, barrels and other items of equipment. According to Vlyssides et al. [1], the total production of wastewater from a winery is about 1.2 times greater than the production of wine. Winery wastewater is characterized by high organic content, seasonal production, unpleasant odours and variable composition, which is associated with the winemaking technologies employed (i.e., 0011-9164/09/$– See front matter © 2009 Published by Elsevier B.V. doi:10.1016/j.desal.2008.11.006 N. Anastasiou et al. / Desalination 248 (2009) 836–842 for red, white or special wines) and the working period (i.e., peak wastewater generation occurs during the ‘crush’, in other words, when grapes are actively being processed into juice for fermentation); all these make winery effluents difficult to treat fully by conventional biological methods. In addition, its treatment by conventional treatment processes is difficult due to the fact that these methods do not provide a comprehensive solution because of the need to dispose of sludge or other by-products derived from such processes. Wine process wastewaters are typically treated onsite in large aerobic ponds to biologically degrade the BOD at about 60–90 days detention time. Several treatment alternatives have been proposed by many researchers through experiments on both pilot scale and full scale with the aim to find efficient technologies characterized by low cost and easy management. Conventional systems include activated sludge reactors, sequencing batch reactors (SBR) and aerobic biofilm systems, such as rotating biological contactors (RBC). Some advanced treatment systems have been also applied, operating under aerobic conditions, such as moving bed biofilm reactors or anaerobic conditions such as UASB reactors [2]. Biofilm systems offer an interesting advantage for the treatment of wastewater with high-rate soluble compounds diffusible into the biofilm. Andreottola et al. [3] demonstrated a full-scale, two-stage fixed bed biofilm reactor for the treatment of winery effluents capable of reaching about 90% COD removal on average over 12 months of operation. In other studies, high COD removal efficiencies (80–95%) were achieved in aerobic-activated sludge [4] and jet-loop [5] bioreactors as well as in anaerobic digesters [6]. The possibility of winery effluents’ valorization was also investigated [7] showing that most of the ethanol present in winery effluents (which accounts for about 80% of the effluent COD) could be recovered by stripping. Recently, electrocoagulation at 2 A current intensity (10 V of voltage) was proved to be an effective method 837 for the pre-treatment of winery wastewater since it can remove 28–42% of COD and 16–28% of BOD in 10 min batch treatment [8]. In addition to the incomplete mineralization of the organic load, the management of the sludge produced through the application of the aforementioned methods constitutes an important problem. In recent years, advanced oxidation processes have gained considerable attention for the treatment of industrial effluents, including, amongst others, agro-industrial wastes. Of the various processes involved, Fenton oxidation provides a simple and effective method of generating hydroxyl radicals which can subsequently oxidize a wide array of organic pollutants. Furthermore, process efficiency may be enhanced in the presence of light irradiation (hn) (i.e., photo-Fenton reactions) through the following redox cycle: Fe3þ þ H2 O þ hv ! Fe2þ þ Hþ þ OH ð1Þ Fe2þ þ H2 O2 ! Fe3þ þ OH þ OH ð2Þ In a recent work [9], the treatment of winery effluents (initial COD of 4000 mg/L) by heterogeneous photo-Fenton oxidation (i.e., iron contained in clays) was investigated and the results were compared to those obtained by a H2O2-assisted, TiO2 photocatalytic process. Depending on the conditions employed (i.e., oxidant concentration, photocatalyst loading, use of artificial or natural solar irradiation), COD removal values between 35% and 60% were achieved with the TiO2 photocatalytic system being more efficient than the photo-Fenton process. Photo-Fenton oxidation has been also applied for the treatment of various type of wastewaters including medical x-ray film developing wastewaters [10], textile effluents [11], cellulose bleaching effluents [12], wastewaters contaminated with diesel [13], petroleum refinery sourwater [14], hospital wastewater [15] and olive mill wastewater [16], demonstrating good removal efficiencies for all these types of 838 N. Anastasiou et al. / Desalination 248 (2009) 836–842 capacity of about 5000 tons of grapes and was first subject to sequential grid removal, biological oxidation in SBR and sand filtration to reduce both organic and solid contents. The flow diagram of the treatment process is shown in Fig. 1. Samples taken from the inlet and outlet streams of the treatment battery as well as from the aeration tank were analyzed with respect to their major physicochemical properties, which are summarized in Table 1. Sampling and analysis were performed at three different periods within a month of operation. Samples were collected in amber glass vials and kept at 28C until analysis. All parameters were measured according to standard methods [17]. The pH and conductivity were measured using a multi-parameter measurement probe (WTW InoLap Multilevel 3), BOD using OxiDirect WTW OxiTop IS6 and wastewater, which are difficult to treat with biological processes. This work deals with the performance of an onsite facility to treat a winery effluent. The facility consisting of various mechanical, physical and biological steps was monitored with respect to its efficiency in effluent treatment and was found capable of only partially treating the effluent through the biological process applied; in this view, homogeneous photo-Fenton was tested as a potential polishing stage to achieve effluent mineralization. 2. Materials and methods 2.1. Winery effluent The effluent was taken from a winery located in the district of Paphos, Cyprus, with a seasonal Pr et r eat m en t Sec o n d ar y Tr e a t m en t Flow Sec o n d ar y Tr e a t m en t Air Blower PH Wastewater M Pumps’ Filter Screening L /T L/T Relief Valve L/T Filter washings L /T Relief Valve M M Pumps M Sand removal Equalization Tank M Stirrer Stirrer M Polymer photo-Fenton M M From Secondary Treatment M M Filter Pumps Sand filter Filter washings Fig. 1. Flow diagram of the treatment steps applied. Sludge Pump M Bioreactor Tertiary Treatment Coagulation Al 2(SO3)4 M M M Bioreactor Relief Valve Sludge Pump Sludge Pump M Excess Sludge Removal Stirrer Sludge ThickeningStorage Tank M Bioreactor Chlorination Tank 839 N. Anastasiou et al. / Desalination 248 (2009) 836–842 Table 1 Major physicochemical properties of winery effluents before and after treatment Sampling time: t0 (d) Inlet Aeration tank Outlet pH () Conductivity (mS/cm) Total solids, TS (mg/L) Total suspended solids, TSS (mg/L) COD (mg/L) BOD5 (mg/L) BOD5/COD () Cu (mg/L) Ni (mg/L) Cr (mg/L) Cd (mg/L) Zn (mg/L) 10.6 3.3 15086 1259 16250 3250 0.2 0.5 0.1 0.12 BDL 1 8 5.3 8461 596 7500 1875 0.25 0.06 0.05 0.007 BDL 0.5 7 4 5785 216 6250 1250 0.2 BDL BDL BDL BDL 0.01 Sampling time: t0 + 10 (d) Inlet Aeration tank Outlet pH () Conductivity (mS/cm) Total solids, TS (mg/L) Total suspended solids, TSS (mg/L) COD (mg/L) BOD5 (mg/L) BOD5/COD () Cu (mg/L) Ni (mg/L) Cr (mg/L) Cd (mg/L) Zn (mg/L) 10.4 3.6 10127 1240 3780 756 0.2 0.6 0.08 0.12 BDL 1 7.5 5.2 7983 410 1890 440 0.23 0.07 0.03 0.005 BDL 0.5 7 4.4 5355 202 800 160 0.2 0.05 BDL BDL BDL 0.08 Sampling time: t0 + 20 (d) Inlet Aeration tank Outlet pH () Conductivity (mS/cm) Total solids, TS (mg/L) Total suspended solids, TSS (mg/L) COD (mg/L) BOD5 (mg/L) BOD5/COD () Cu (mg/L) Ni (mg/L) Cr (mg/L) Cd (mg/L) Zn (mg/L) 8 4.6 9579 1050 4840 989 0.2 0.8 0.1 0.15 BDL 0.9 7.6 4.5 7813 240 2420 560 0.23 0.1 0.03 0.007 BDL 0.5 7.6 4.4 4093 140 1060 208 0.2 0.06 BDL BDL BDL 0.03 BDL, below detection limit. 840 N. Anastasiou et al. / Desalination 248 (2009) 836–842 COD through the application of the Open Reflux Method (5220B). Heavy metals were determined by AAS, Perkin Elmer-AAnalyst 200. 2.2. Photo-Fenton experiments Ferrous salt (FeSO47H2O, 98.0% purity) and hydrogen peroxide (35% solution) were purchased from Merck. Sulphuric acid (95–98%, extra pure, Merck) was used to adjust the pH of the system. Milli-Q water system (Millipore) was used for the preparation of solutions. Na2SO3 (> 98%, Fluka) was used to increase the pH of the solution and therefore facilitate the precipitation of ferrous and ferric ions, to deactivate H2O2 and cease the oxidation process.Merckoquant Peroxide-Test strips (0.5–25 mg/L H2O2) were used to monitor the elimination of unreacted hydrogen peroxide. These analytical test strips are used for the detection and semiquantitative determination of residual concentrations of hydrogen peroxide (colorimetric method). Batch photo-Fenton experiments were performed in duplicate, in a conical vessel, which was irradiated from the top by a 125 W UV-A lamp (Osram) with a black-glass bulb emitting in the range 315–400 nm. The lamp was switched on 30 min before the irradiation of the solution in order to secure a stabilized photon flow. Irradiation intensity was measured at 13 W/m2 with a Thorlabs HLAG PM100 radiometer. The entire system was enclosed in a chamber with its inner walls covered with aluminium foil in order to prevent any loss of the irradiation. In all cases, the treated effluent from the last sampling period was employed (initial concentration of 1060 mg/L COD and 210 mg/L BOD5) and 100 mL were loaded in the reactor. The appropriate volume of a 35% hydrogen peroxide solution was then added followed by the appropriate amount of FeSO4.7H2O to achieve H2O2 and Fe2+ initial concentrations in the range 34–175 and 0.5–2 mM respectively. The solution pH was adjusted to 2.5 adding appropri- ate volume of H2SO4. The vessel contents were continuously stirred, while the reaction temperature was kept below 408C through a water cooling system. At the end of each run, Fenton reactions were quenched adding the appropriate amount of sodium sulphite and samples were analyzed according to standard methods. 3. Results and discussion As seen in Table 1, effluent quality fluctuates considerably depending on the sampling period and this is due to the fact that different winemaking technologies (i.e., variety of grapes and types of wines) are involved. In all cases, the untreated effluent is alkaline due to washing of various items of equipment with KOH and its pH is adjusted to near-neutral values prior to biological treatment. Biological oxidation results in about 50% COD and BOD5 reduction, while the overall process results in 60–80% COD reduction; this is accompanied by about 85% TSS reduction. Interestingly, the effluent’s aerobic biodegradability (as assessed by the BOD5/COD ratio) is always low (about 0.2), regardless of the sampling point and period. The presence of trace quantities of various heavy metals is attributed to residual pesticides and antibacterial agents. The trend of changing the concentration of Fenton’s reagents on the reduction of COD (initial COD: 1060 mg/L) and BOD5 (initial BOD5: 210 mg/L) is shown in Figs 2 and 3 respectively. In general, organic matter degradation increases with increasing treatment time reaching values of COD or BOD5 removal as high as 80% after 4 h of reaction. The effluent pH as well as the concentrations of iron and hydrogen peroxide are the most important variables affecting the performance of photo-Fenton reactions at a fixed organic carbon concentration and illumination source. It is well-documented [18] that the optimum pH range is between 2 and 4 favouring the catalytic decomposition of oxidant to hydroxyl radicals as well as the decomposition of the organic N. Anastasiou et al. / Desalination 248 (2009) 836–842 60 Likewise, increased catalyst concentrations may be responsible for reduced degradation as the catalyst may compete with the organic material and scavenge radicals [20], that is, 40 Fe2þ þ HO2  ! Fe3þ þ HO 2 100 COD - 2 h - 1mM Fe2+ COD - 4 h - 1mM Fe2+ BOD - 2 h - 1mM Fe2+ BOD - 4 h - 1mM Fe2+ COD - 2 h - 1.3 mM Fe2+ COD - 4 h - 1.3 mM Fe2+ Removal, % 80 20 0 0 30 60 90 120 H 2O2 concentration, mM 150 180 Fig. 2. Effect of H2O2 concentration on COD and BOD5 reduction. matter; this is why photo-Fenton runs were carried out at pH0 = 2.5. Hydrogen peroxide is usually the limiting reactant in photo-Fenton processes. However, an excess of this reagent does not mean a continuous increase in the mineralization and degradation rates of the treated solution [19]. In fact, there appears to be an optimum dosage above which treatment performance deteriorates. This is so because excessive H2O2 eventually acts as radical scavenger, thus affecting treatment adversely, that is, 100 H2 O2 þ OH ! H2 O þ HO2  ð3Þ HO2  þ  OH ! H2 O þ O2 ð4Þ COD - 2 h - 103 mM H2O2 COD - 4 h - 103 mM H2O2 BOD - 2 h - 103 mM H2O2 BOD - 4 h - 103 mM H2O2 COD - 2 h - 137 mM H2O2 COD - 4 h - 137 mM H2O2 80 Removal, % 841 60 40 20 0 0 0.5 2+ Fig. 3. Effect of Fe reduction. 1 1.5 ð5Þ Moreover, high iron concentrations are not desirable since they are likely to increase iron precipitation [19]. A major concern typically associated with homogeneous Fenton reactions is the production of iron-containing sludge at the end of the process. However, at the experimental conditions employed in this study the amount of this sludge is far less than that produced during the physical–biological effluent treatment; therefore, it is envisaged that the iron-containing sludge could be managed alongside the biological sludge. Best results were obtained after 4 h of treatment at iron and hydrogen peroxide concentrations of 1 and 137 mM respectively yielding an effluent with 235 mg/L residual COD. Given that the original effluent (i.e., prior to any treatment) had a COD content of 4840 mg/L, a process train comprising conventional treatment followed by photo-Fenton post-treatment may result in 95% mineralization. Ormad et al. [21] studied the degradation of winery wastewaters by photo-Fenton reactions (using Fe(III) as the catalyst and synthetic wastewater) regarding the effect of catalyst, hydrogen peroxide and organic carbon concentrations as well as the reaction time on treatment performance. The present study is in agreement with the study performed by Ormad et al. [21] since both studies proved that the homogeneous photo-Fenton is capable of depredating up to 95% the organic load and the main influencing parameters are the catalyst and oxidant dosages. 2 Fe2+ concentration, mM 4. Conclusions concentration on COD and BOD5 In this study, it has been found that homogeneous Fenton photo-oxidation is an appropriate 842 N. Anastasiou et al. / Desalination 248 (2009) 836–842 process for the post-treatment of winery effluents. The results lead to two major conclusions as follows: 1. Organic matter degradation increases with increasing treatment time and can reach COD or BOD removal values as high as 80% after 4 h of reaction. 2. A process train consisting of conventional treatment followed by photo-Fenton posttreatment has the potential to result in nearly completely effluent mineralization since the application of the conventional treatment followed by the photo-Fenton post-treatment resulted in 95% organic carbon degradation. Given that Fenton systems are relatively inexpensive as well easy to handle and operate (i.e., compared to other advanced processes), these results pinpoint the potential of process combination to improve performance employing techniques that can easily be integrated in existing treatment schemes. References [1] A.G. Vlyssides, E.M. Barampouti, S. Mai. Water Sci. Technol. 51 (2005) 53–60. [2] G. Andreottola, P. Nardelli, F. Nardin, (1997) Demonstration plant experience of winery wastewater anaerobic treatment in a hybrid reactor. Proceedings of the 2nd International Specialized Conference on Winery Wastewaters. Bordeaux, France, 243–251. [3] G. Andreottola, P. Foladori, P. Nardelli, A. Denicolo. Water Sci. Technol. 51 (2005) 71–79. [4] M. Brucculeri, D. Bolzonella, P. Battistoni, F. Cecchi. Water Sci. Technol. 51 (2005) 89–98. [5] A. Eusebio, M. Mateus, L. Baeta-Hall, E. Almeida-Vara, J.C. Duarte. Water Sci. Technol. 51 (2005) 107–112. [6] R. Moletta. Water Sci. Technol. 51 (2005) 137–144. [7] T. Colin, A. Bories, Y. Sire, R. Perrin. Water Sci. Technol. 51 (2005) 99–106. [8] F. Kirzhner, Y. Zimmels, Y. Shraiber. Sep. Purif. Technol. 63 (2008), 38–44. [9] P. Navarro, J. Sarasa, D. Sierra, S. Esteban, J.L. Ovelleiro. Water Sci. Technol. 51 (2005) 113–120. [10] C. Stalikas, L. Lunar, A. Rubio, D. PerezeBendito. Wat. Res. 35 (2001) 3845–3856. [11] M. Perez, F. Torrades, X. Doménech, J. Peral. Wat. Res. 36 (2002) 2703–2710. [12] F. Torrades, M. Pérez, H. Mansilla, J. Peral. Chemosphere. 53 (2003) 1211– 1220. [13] S.A.O. Galvão, A.L.N. Mota, D.N. Silva, J.E.F. Moraes, C.A.O. Nascimento, O. Chiavone-Filho. Sci. Total Environ. 367 (2006) 42–49. [14] A. Coelho, A. Castro, M. Dezotti, G.L. Sant’ Anna Jr. J. Hazard. Mater. B137 (2006) 178–184. [15] P. Kajitvichyanukul, N. Suntronvipart. J. Hazard. Mater. B138 (2006) 384–391. [16] L. Rizzo, G. Lofrano, M. Grassi, V. Belgiorno. Sep. Purif. Technol. 63 (2008) 648–653. [17] L.S. Clesceri, A.E. Greenberg, A.D. Eaton, American Public Health Association/ American Water Works Association, water Environment Federation, (1998). Standard Methods for the Examination of Water and Wastewater, 20th edition, Washington DC, USA. [18] D. Mantzavinos. Process Saf. Environ. 81 (2003) 99–106. [19] M. Rodriguez, S. Malato, C. Pulgarin, S. Contreras, D. Curco, J. Gimenez, S. Esplugas, Sol. Energy 79 (2005) 360–368. [20] C. Berberidou, I. Poulios, N.P. Xekoukoulotakis, D. Mantzavinos. Appl. Catal. B-Environ. 74 (2007) 63–72. [21] M.P. Ormad, R. Mosteo, C. Ibarz, J. Ovelleiro. Appl. Catal. B-Environ. 66 (2006) 58–63.