Advanced systems for the enhancement of the environmental performance of WINEries in Cyprus (WINEC)

Transcription

1 Advanced systems for the enhancement of the environmental performance of WINEries in Cyprus (WINEC) Deliverable 8 Engineering and Technological Parameters Development First Edition Nicosia, April 2011

2 Table of Contents Table of Contents Summary Winery wastewater Advanced Oxidation Processes (AOPs) Solar Fenton process Heterogeneous Fenton-like process Experimental part Chemicals Winery effluent Experimental set up and procedure Analytical equipment and methods Results and discussion Solar-Fenton process Effect of ferrous dosage Effect of H 2 O 2 dosage Effect of ph Effect of Temperature Effect of the solar irradiation Total phenols and BOD 5 removal Degradation kinetics Color removal Average oxidation state (AOS) Solar heterogeneous Fenton-like process Blank experiments Effect of heterogeneous catalyst Fe 2 O 3 /SBA Effect of H 2 O 2 concentration Effect of ph Effect of Temperature Total phenols and BOD 5 removal Degradation kinetics Conclusions References

3 Index of Figures/Schematics Schematic 1: Sequence of reactions occurring in the homogeneous photo-fenton system Schematic 2: Bench scale experimental set up of solar-fenton experiments Figure 1: Effect of initial ferrous concentration on the COD removal of winery wastewater effluent; [Fe 2+ ] 0 =5-20 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. 25 Figure 2: Effect of initial hydrogen peroxide concentration on the COD removal of winery wastewater effluent; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 = mg L -1, ph= and T=25±0.1 o C Figure 3: Effect of initial solution ph on the COD removal of winery wastewater; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C Figure 4: Effect of temperature on the COD removal of winery wastewater; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=15-40 o C Figure 5: COD removal of winery wastewater effluents under different processes; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C Figure 6: Concentration of Total phenols (a) and % Total phenols removal (b) of winery wastewater effluent during solar-fenton process; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C Figure 7: Measurement of BOD 5 values for the winery effluent before and after 120 min of solar-fenton treatment; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C Figure 8: Photocatalytic (COD (a) and DOC (b)) decomposition kinetics of winery wastewater effluent after solar-fenton oxidation; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C Figure 9: Color removal of winery wastewater effluents after solar-fenton treatment; Inset graph shows the color removal of winery wastewater after min of Fenton treatment; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. 38 Figure 10: % Color removal of winery wastewater after solar-fenton treatment; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C Figure 11: COD and DOC removal of winery wastewater after Fenton treatment; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C Figure 12: Changes in the average oxidation state, AOS; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C

4 Figure 13: Blank experiments depicting the effect of the solar light in the presence of the oxidant and/or the catalyst and the contribution of dark reactions to the overall efficiency of the process; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]=100 mg L -1, ph o =8 and T=23 C Figure 14: Trend of DOC removal at different concentrations of Fe 2 O 3 /SBA-15 catalyst during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [H 2 O 2 ] =100 mg L -1, ph o =8 (inherent) and T=23 о C Figure 15: Trend of DOC removal at different oxidant concentrations during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, ph o =8 (inherent) and T=23 о C Figure 16: Trend of DOC removal at different initial ph values during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1 and T=23 о C Figure 17: Trend of (a) DOC and (b) COD removal at different solution temperatures during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1 and ph o = Figure 18: Total phenols removal during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1 and ph o = Figure 19: Measurement of BOD 5 values for the winery effluent after 180 min of heterogeneous solar treatment (HSF) under different experimental conditions that result in different DOC removal percentages. Green: 20% DOC removal; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1, ph o =8 and T=23 C. Purple: 5% DOC removal; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1 and ph o =8 and T=40 C Figure 20: First order kinetics of the mineralization of the winery effluent

5 Abbreviations AOPs BOD 5 COD DOC HSF SBR TPh U.S. EPA Advanced Oxidation Processes Biological Oxygen Demand Chemical oxygen Demand Dissolved Organic Carbon Heterogeneous solar Fenton-like Sequenced Batch Reactor Total phenolic content United States Environmental Protection Agency 5

6 1. Summary Winery wastewater contains 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 winery wastewater. The aim of this study was to assess the use of Fenton and heterogeneous Fenton-like processes in winery wastewater treatment. In specific, the main reactions and the operating parameters (e.g. ferrous and hydrogen peroxide concentrations, ph, temperature) that affect these processes are reported. The advantages and drawbacks of these methods are highlighted, while the degradation kinetics is reported for both processes. A partially biologically treated effluent with chemical oxygen demand (COD) and biochemical oxygen demand (BOD 5 ) values of 270 mg L -1 and 112 mg L -1 respectively, was subjected to both homogeneous solar-fenton and heterogeneous Fenton-like oxidation. The experimental conditions during the study of homogeneous solar Fenton are summarized as follows: The H 2 O 2 and Fe 2+ concentrations range between and 5-20 mg L -1 respectively, solution ph 0 = and continuous solar irradiation was provided by a 1000 W lamp. During the study of heterogeneous Fenton-like the experimental conditions varied between 0 and 300 mg O 2 L -1 for H 2 O 2 and in the case of the iron catalyst between 50 and 200 mg L -1. The initial solution ph was studied at two different values i.e. 2.8 which correspond to the optimum ph value for homogeneous Fenton reactions and 8 which is the inherent ph of the biotreated winery effluent. Simulated solar irradiation was provided by a 150 W xenon lamp. 6

7 In general, organic matter degradation increased with increasing treatment time reaching values of COD or BOD 5 removal as high as 69% and 85% after 120 min of solar-fenton reaction. Regarding the effect of initial iron and hydrogen peroxide concentrations, there appears to be an optimum dosage for both. Total phenols removal was 71%, while 53% of the color removed after 120 min of solar-fenton treatment. Similarly, heterogeneous Fenton-like oxidation of the same effluent resulted in 20% COD removal, 20% DOC removal and 80% total phenols removal after 120 min of treatment. Despite the average results reported, the heterogeneous solar-fenton process led to increased biodegradability of the effluent (e.g. the BOD/COD ratio increased roughly from 0.2 to over 0.5). As in the case of the homogeneous treatment, there is an optimum dosage in both the amount of the heterogeneous iron catalyst (i.e. 100 mg Fe SBA-15 ) and the concentration of hydrogen peroxide (i.e. 100 mg O 2 L -1 ) for the reaction system under study. 7

8 2. Winery wastewater Wastewaters from the wine-production, are mainly originated from various washing operations during the crushing and pressing of grapes, as well as rinsing of fermentation tanks, barrels, bottles and other equipments or surfaces (Arnaiz et al., 2007). Volumes and pollution loads significantly change over the year, in relation to the working period (vintage, racking, bottling) and the winemaking technologies used (e.g. in the production of red, white and special wines (Rochard, 1995)). Wastewater from wine production presents a pollution risk due to the following: 1. Wine making generates significant volumes of wastewater that include winery wastewater and cleaning water. While this effluent is unlikely to contain toxic materials, it contains high levels of organic matter, which may reduce the presence of oxygen in watercourses. 2. High acidity or alkalinity, depending on the type of cleaning detergents used. 3. Other contaminants such as pesticides from the initial washing of fruit and biocides from cleaning (European Bank for Reconstruction and Development, 2009). 4. Polyphenols are responsible for strong inhibitory effects on microbial activity and must be removed during wastewater treatment, due to the environmental and public health risks they might pose (Melamane et al., 2007). Winery wastewater can cause eutrophication of water resources (natural streams, rivers, dams, ground water and wetlands). Furthermore, wastewaters can cause salinity, contamination with a wide range of chemicals, loss of soil structure and increased susceptibility to erosion. These impacts may be exacerbated by process 8

9 interruptions. Process interruptions may stem from power failure, fire, floods, storms, overloading/underloading of wastewater treatment systems, temporary unavailability of wastewater holding dam capacity and the absence of trained operators (Brown et al., 2009). The objectives of wastewater management are to be protective of the receiving environment and enhance water reuse opportunities. Reduction of organic strength, measured as biochemical oxygen demand (BOD 5 ), dissolved organic carbon (DOC) or chemical oxygen demand (COD), are the most important wastewater treatment parameters which need to be minimised prior to the effluent s release into the environment. The organic content of winery wastewater consists of highly soluble sugars, alcohols, acids and recalcitrant high-molecular-weight compounds (e.g., polyphenols, tannins and lignins) not easily removable by physical or chemical means alone. The concern regarding the presence of these organic compounds in industrial effluents and natural waters related to their unknown potential effects has provoked intense research on the development of methods capable of removing organic residues efficiently (Litter, 2005; Navarro et al., 2005; Anastasiou et al., 2009). Over the past 30 years, research and development concerning advanced oxidation processes (AOPs) has been immense as alternatives or as a complement to conventional wastewater treatment industrial effluents. The aim of the present work was to investigate the efficiency of solar-fenton and heterogeneous Fenton-like processes for the removal of the organic content of winery wastewater effluent and choose the more efficient of the two AOP systems. Bench scale experiments were carried out to evaluate the influence of various operational 9

10 parameters on the degradation and the efficiency of each one of the two processes. The work focused on the effect of ferrous and hydrogen peroxide concentrations, ph, and temperature, on the COD - DOC - BOD 5, color and total phenols removal and on the degradation kinetics for both processes. 2.1 Advanced Oxidation Processes (AOPs) The destruction of toxic pollutants as well as that of the biologically recalcitrant compounds demands the application of some non-biological technologies. These technologies consist mainly of conventional phase separation techniques (adsorption processes, stripping techniques) and methods which destroy the contaminants (chemical oxidation/reduction). Chemical oxidation aims at the mineralization of the contaminants to carbon dioxide, water and inorganics or, at least, at their transformation into harmless products. Obviously the methods based on chemical destruction, when properly developed, can provide a complete solution to the problem of pollutant abatement in contrast to those in which only a phase separation is realised with the consequent problem of the final disposal (Andreozzi et al., 1999). It has been frequently observed that pollutants like soluble sugars, alcohols, acids and recalcitrant high-molecular-weight compounds (e.g., polyphenols, tannins and lignins) not amenable to biological treatment may also be characterised by high chemical stability and/or by strong difficulty to be completely mineralized. In these cases, it is necessary to adopt reactive systems much more effective than those adopted in conventional treatment processes. 10

11 A lot of researches have addressed this objective in the last decade pointing out the prominent role of a special class of oxidation techniques defined as AOPs which usually operate at or near ambient temperature and pressure. All AOPs are characterised by a common chemical feature: the capability of exploiting the high reactivity of hydroxyl radicals in driving oxidation processes which are suitable for achieving the complete abatement and through mineralization of even less reactive pollutants. Hydroxyl radicals are extraordinarily reactive species and they attack most of the organic molecules with rate constants usually in the order of M -1 s -1 (Andreozzi et al., 1999; Malato et al., 2003; Klavarioti et al., 2009). They are also characterised as non-selective which is a useful attribute for an oxidant used in wastewater treatment and for solving pollution problems. The versatility of AOPs is also enhanced by the fact that they offer different possible ways for HO production thus allowing a better compliance with the specific treatment requirements. The most widely known AOPs include: heterogeneous photocatalytic oxidation, treatment with ozone (often combined with H 2 O 2, UVA, or both), H 2 O 2 /UV systems, Fenton, and photo-fenton type reactions. 2.2 Solar Fenton process The Fenton and photo-fenton systems have been widely applied in the treatment of industrial wastewater in the field of AOPs, and illustrated in Schematic 1 (Sabhi et al., 2001; Catastini et al., 2002; Will et al., 2004; Mosteo et al., 2008; Stasinakis, 2008; Malato et al., 2009; Anastasiou et al., 2009). 11

12 Oxidation with Fenton s reagent is based on ferrous ion and hydrogen peroxide, and exploits the reactivity of the hydroxyl radicals produced in acidic solution by the catalytic decomposition of H 2 O 2 : Fe 2 H O 2 2 Fe ph 3 3 HO HO (1) Hydroxyl radicals may be scavenged by reaction with another Fe 2+ : 2 3 HO Fe HO Fe (2) The rate of degradation of the organic pollutants by Fenton reaction could increase when an irradiation source (UV or visible light) is present. The positive effect of irradiation on the degradation rate is due to the photoreduction of Fe 3+ to Fe 2+ ions, a step that produces new HO radicals and regenerates Fe 2+ ions that can further react with more H 2 O 2 molecules. The photoreduction of Fe 3+ follows the equation (3) (Perez et al., 2002; Will et al., 2004): Fe 3 H O HO 2 Fe hv 2 H (3) The illumination leads not only to the formation of additional hydroxyl radicals but also to recycling of ferrous catalyst by reduction of Fe 3+. In this way, the concentration of Fe 2+ is increased and the overall reaction is accelerated (Tamimi et al., 2008). The advantages of the method are various: Fe 2+ is abundant and nontoxic, hydrogen peroxide can be easily handled and it is an environmentally friendly compound. No chlorinated compounds are formed as in other oxidative techniques, and there are no 12

13 mass transfer limitations because all of the reagents are in solution (Litter, 2005). Fenton reagent appears to be a very powerful oxidizing agent. Besides, the process is simple and non-expensive, taking place at low temperatures and at atmospheric pressure. The chemicals (ferrous ion and hydrogen peroxide) are readily available at moderate cost and there is no need for special equipment (Lucas and Peres, 2006). In the presence of Fenton reagent, photochemical reactions can be driven with photons of low energy, photons that belong to the visible part of the spectrum. Thus, photo- Fenton processes are a potential cost-reduced AOP that can be operated under solar irradiation (Perez et al., 2002; Garcia-Montano et al., 2006). The maximum catalytic activity of the Fe 2+ / Fe 3+ - H 2 O 2 system is at a ph of about At ph > 5, particulate Fe 3+ is generated and at a lower ph, the complexation of Fe 3+ with H 2 O 2 (equation (4)) is inhibited; therefore, the ph must be kept constant (Safarzadeh-Amiri et al., 1997). 3 Fe H2O2 Fe OOH2 H (4) 13

14 Schematic 1: Sequence of reactions occurring in the homogeneous photo-fenton system. 2.3 Heterogeneous Fenton-like process The application of Fenton and photo-fenton processes based on homogeneous ferrous or ferric salts usually suffers two major drawbacks associated with (i) the narrow ph range of operation, typically between 2.5 and 3.5, to avoid the formation and subsequent precipitation of iron oxyhydroxides and (ii) the need to recover dissolved ions from the treated solution, thus requiring an additional treatment stage. In this respect, the immobilization of Fenton s catalyst on a heterogeneous matrix would enable its use under non-controlled ph conditions as well as its easy recovery from the treated effluent (Santos et al., 2007; Pariente et al., 2008; Malato et al., 2009). Several iron-containing materials have been studied for the degradation of model compounds in batch photochemical reactors, such as zeolitic Fe(III)-Y and Fe(II)-3X 14

15 materials (Enriquez et al., 2004), perfluorosulphonic Nafion polymers (Fernandez et al., 1999) and pillared clays (Azabou et al., 2007). The low exchange of iron species within the zeolite framework and its poor stability are the main drawbacks of these microporous materials. The presence of sulphonic acid groups in the polymeric matrix makes Nafion a good candidate for the anchoring of active iron ions in form of films or pellets. The main limitation of this catalytic system has been associated with the relative high cost of the perfluorosulphonic polymer. The interlamellar structure of pillared clays by the intercalation of large polyoxocations of iron that can be thermally transformed in oxides grafted to the clays layers have revealed a low deactivation due to iron leaching accompanied with remarkable results in terms of the pollutant abatement. Iron-containing mesoporous materials based on supporting Fe 2 O 3 particles over hexagonally pore channels of mesostructured SBA-15 silica support have resulted in a very promising catalyst as compared to other photo-fenton catalysts supported over different types of silica like amorphous xerogels and microporous zeolite materials (Martinez et al., 2007). It was determined that the physicochemical properties of Fe 2 O 3 /SBA-15 materials, with an extended surface area and pore distributions within the mesoporous range, were responsible for high degradation rates of phenol in photo-assisted Fenton-like processes. In fact, it has been demonstrated that the photo-activity of a catalyst can be strongly influenced by its surface area, crystal structure, particle size distribution or surface hydroxyl group density (Bahnemann et al., 1991). So far, few authors have studied the degradation of the synthetic or actual winery wastewaters using photo-fenton processes. Mosteo et al. (2006), studied the heterogeneous solar Fenton-like treatment of a synthetic winery effluent by applying 15

16 an experimental design methodology, while other studies investigated the homogeneous photocatalytic treatment of a winery wastewater (Ormad et al., 2006; Anastasiou et al., 2009). Up to now, there are no reports on the application of heterogeneous solar Fenton-like process as a post-treatment of an actual winery wastewater already subjected to biological pre-treatment. 16

17 3. Experimental part 3.1 Chemicals Solar-Fenton experiments were performed using iron sulphate (FeSO 4. 7H 2 O), reagentgrade hydrogen peroxide (35%, w/v, Merck) and sulphuric acid for ph adjustment (around , 95-97%, Merck). In the case of the solar-fenton process, the residual hydrogen peroxide was removed from the treated samples with MnO 2 or Na 2 SO 3 (Sigma-Aldrich). The respective solar-fenton experiments in heterogeneous phase were performed by employing a powder iron-containing SBA-15 mesostructured material as heterogeneous catalyst, which was synthesized by co-condensation of iron (FeCl 3.6H 2 O, Aldrich) and silica (tetraethoxysilicate, 98%, Aldrich) sources templated with Pluronic 123 (MW = 5800, BASF) under acidic conditions (Galleja et al., 2005). The resulting catalyst consists of crystalline hematite particles embedded into a mesostructured silica support with ca % of iron content in the bulk sample. The silica matrix is characterized by typical hexagonal arrangement of mesostructured SBA-15 materials, leading to highly extended surface areas (SBET ca. 500 m 2 g -1 ) with a narrow pore diameter within the mesoporous range (6-7nm). More details about preparation and characterization of iron-containing SBA-15 mesostructured material are described elsewhere (Martinez et al., 2007). The reagent 1,10 phenantroline (C 12 H 8 N 2. H 2 O) and ascorbic acid (C 6 H 8 O 6 ) used for the measurement of the total dissolved iron (both ferrous and ferric ions) 17

18 concentration of ferrous, were supplied by Sigma Aldrich while the ammonium acetate (CH 3 COONH 4 ) was supplied by Fluka. 3.2 Winery effluent The samples used as the matrix for the bench scale experiments, were collected from a winery wastewater treatment plant of Paphos, Cyprus, after the biological treatment (Sequential Batch Reactor - SBR) winery wastewater and were characterized before used. The Table of characterization of wastewater is given below. Table 1: Characteristic parameters of winery wastewater effluent after SBR treatment Parameter Value ph (20 o C) Total Solids (mg L -1 ) Total Volatile Solids (mg L -1 ) Suspended Solids (mg L -1 ) Suspended Volatile Solids (mg L -1 ) Total phenols (mg L -1 ) Total Nitrogen (mg L -1 ) COD (mg L -1 ) Soluble BOD 5 (mg L -1 ) Total Phosphorous (mg L -1 ) Fats and oils (mg L -1 ) < 4 Cu (mg L -1 ) Cd (mg L -1 ) Fe (mg L -1 ) Na + (mg L -1 ) K + (mg L -1 )

19 3.3 Experimental set up and procedure Work completed at University of Cyprus For solar-fenton experiments irradiation was provided by a 1 kw lamp (Xe-OP) in a solar simulator (Newport model 91193). The irradiation intensity of the simulator was determined using a radiometer Newport type and it was found W/m 2. Solar experiments were conducted in a photochemical batch reactor constructed by Pyrex (inner diameter 7 cm; height 11 cm) with a maximum capacity of 300 ml. The vessel contents were continuously stirred (500 rpm), while the reaction temperature was kept below 25±0.1 o C through a water cooling system. Samples were periodically withdrawn from the reactor and filtered through a 0.22 μm filter (Milli-pore) to remove the MnO 2 particles, which used in order to avoid further reactions due to the presence of H 2 O 2, and were analyzed with Merck Spectroquant kits for the COD measurement, TOC analyzer (Schimadzu TOC-V CPH/CPN ), UV/Vis spectrometer (Jasco V-530) and BOD measurement system ( Oxidirect), illustrated in Schematic 2. Photocatalytic experiments were performed in triplicate and, in some cases, in quadruplicate and mean values are quoted as results. The standard deviation of the experiments never exceeded 5%. 19

20 Schematic 2: Bench scale experimental set up of solar-fenton experiments For solar-fenton experiments, 300 ml of winery wastewater after filtration with 1μm was fed into the reactor and the ph was adjusted to by adding the required amount of 2M H 2 SO 4 aqueous solution. Then, the appropriate amount of Fe 2+ was added from an aqueous solution of FeSO. 4 7H 2 O. The iron was very well mixed with the winery wastewater solution before the addition of the appropriate volume of H 2 O 2 (35% w/w). The time at which the solar lamp was turned on was considered time zero or the beginning of the experiment which was taking place simultaneously with the addition of hydrogen peroxide. Samples were withdrawn at specified intervals from the reactor and were transferred in vials containing some mg of MnO 2 in order to avoid further reactions due to the presence of H 2 O 2 and after 10 min that the elimination of unreacted hydrogen peroxide has completed, the samples filtered through a 0.22 μm filter (Milli-pore) to remove the MnO 2 particles. 20

21 Work completed at Technical University of Crete Heterogeneous photocatalytic experiments were performed at Technical University of Crete using a laboratory scale solar simulator (Oriel, model 96000) equipped with an ozone free xenon lamp 150 W. The output irradiation intensity was calculated through chemical actinometry at Einstein/sec All experiments were conducted in a photochemical batch reactor made of borosilicate glass equipped with double wall for temperature control. In a typical experimental run 310 ml of pre-filtered (30 μm, PALL) winery wastewater were loaded in the reaction vessel along with an appropriate amount of SBA-15 catalyst. The suspension was left in the dark under continuous magnetic stirring for 30 min to reach equilibrium state. Samples were withdrawn right after ph adjustment and/or after equilibrium state between catalyst and substrate had been reached. A pre-determined volume of 30% H 2 O 2 (Merck) was inserted in the reaction medium and the lamp was turned on. This was considered the time zero of the reaction. The initial irradiated volume was 300 ml at all times. Samples were withdrawn at frequent time intervals and were filtered (0.45μm, RC, PALL) prior to each analysis. In order to eliminate any residual hydrogen peroxide remaining in the sample aliquots a small amount of MnO 2 particles or 120 μl of Na 2 SO 3 solution 0.13 mm were added. In order to verify the comparative results produced during the solar-fenton study in the University of Cyprus and the heterogeneous solar Fenton-like study in Technical University of Crete, a small number of experiments on homogeneous solar-fenton were also performed with the experimental apparatus available in Technical University of Crete 21

22 3.4 Analytical equipment and methods For the solar-fenton experiments, mineralisation was monitored by measuring the chemical oxygen demand (COD) with Merck Spectroquant kits and dissolved organic carbon (DOC) by direct injection of filtered samples into a Shimadzu TOC- VCPH/CPN, TOC analyser calibrated with standard solutions of potassium hydrogen phthalate. The ph was measured using a multi-parameter measurement probe (WTW InoLap Multilevel 3) and the BOD using OxiDirect. The % total phenols removal carried out (Folin-Ciocalteau method) by measuring the absorbance of the solution at 765nm with a UV/Vis Jasco V-530 spectrophotometer. Colorimetric determination of total iron concentration with 1,10-phenantroline was used according to ISO Hydrogen peroxide was analysed by a fast, simple spectrophotometric method using ammonium metavanadate, which allows the H 2 O 2 concentration to be determined immediately based on a red-orange peroxovanadium cation formed during the reaction of H 2 O 2 with metavanadate, with a maximum absorption at 450 nm (Klamerth et al., 2009). Quantofix Peroxide-Test sticks (0-100 mg L -1 H 2 O 2 ) 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). The analytical determinations employed at Technical University of Crete for the heterogeneous photocatalytic experiments are as follows: In order to assess the extent of mineralization, dissolved organic carbon (DOC) measurements were carried out by 22

23 injection of filtered samples (RC 0.45mm, PALL) into a Shimadzu-5050A TOC analyzer equipped with an NDIR. Each sample prior to DOC analysis was acidified with H 2 SO 4 solution 1N and was subjected to air purging for 10 min to eliminate the inorganic carbon content of the effluent. ph was determined with a Crison GLP 21 ph meter. COD was determined colorimetrically using a DR/2010 spectrophotometer (Hach Company, USA) according to the EPA approved reactor digestion method (Hach, 1992) in order to monitor the COD reduction of the solution during degradation. The total phenol content (TPh) was determined colorimetrically at 765 nm on a Shimadzu UV 1240 spectrophotometer using the Folin Ciocalteau reagent according to a modified method (Singleton et al., 1999). Gallic acid monohydrate was used as standard to quantify the concentration of total phenols in the winery effluent. Dissolved iron concentration was determined using the 1,10-phenanthroline spectrophotometric method (ISO 6332:1982). Semiquantitative determination of the residual concentration of hydrogen peroxide (colorimetric method) was performed with the use of Merckoquant peroxide test strips (Merck). The determination of the Biochemical Oxygen Demand (BOD) was performed with an OxiTop measuring system. 23

24 4. Results and discussion 4.1 Solar-Fenton process Preliminary experiments were carried out to find the suitable ferrous salt and hydrogen peroxide doses needed to oxidize the examined compound. To obtain the optimal initial Fe 2+ and H 2 O 2 concentrations, experiments were conducted with several combinations of catalyst (5-20 mg L -1 ) and oxidant ( mg L -1 ). Experiments were also carried out to find the optimum ph value (2-4) and temperature (15-40 o C) Effect of ferrous dosage Amount of ferrous ion is one of the main parameters to influence the Fenton and photo-fenton processes. In this study, to obtain the optimal initial Fe 2+ concentration, the investigation was carried out in the range of (5-20 mg L -1 ) at ph= , while the initial concentration of H 2 O 2 was kept constant at 500 mg L -1. The results for solar Fenton process are shown in Fig. 1. As seen the use of extra ferrous salt had an insignificant beneficiary effect (i.e. within experimental error) on the reduction of COD of the winery wastewater effluents and concretely after 120 min of reaction the reduction of COD ranged from 69-71% at 5-20 mg L -1 Fe 2+. It was known that the Fe 2+ had a catalytic decomposition effect on H 2 O 2. When Fe 2+ concentration increased, the catalytic effect also accordingly increased. When the concentration of Fe 2+ was higher, decomposition by Fe 2+ was easy to exit in the form of Fe(OH) 2+ in acidic environment. Because Fe(OH) 2+ had a strong absorption for UV 24

25 % COD removal light from 290 to 400 nm, the strength of UV light would decrease (Tamimi et al., 2008). This parameter is crucial for a large-scale wastewater treatment plant since it can influence not only the application costs as it determines the size of the photoreactor but also the operating costs as shorter reaction times are required. However, if iron concentration is high, the problem of the resulting iron-separation step arises at the end of the photocatalytic process. Consequently, it is preferable to select a smaller concentration by which it would be possible to achieve as short reaction time as possible without further treatment to be required for iron removal (Gernjak et al., 2003; Evgenidou et al., 2007). For all these reasons, the selected as optimum ferrous concentration, was the 5 mg L -1 Fe mg Fe(II)/L 10 mg Fe(II)/L 15 mg Fe(II)/L 20 mg Fe(II)/L time (min) Figure 1: Effect of initial ferrous concentration on the COD removal of winery wastewater effluent; [Fe 2+ ] 0 =5-20 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. 25

26 Effect of H 2 O 2 dosage In order to examine the effect of hydrogen peroxide concentration on the COD removal of the winery wastewater, experiments were carried out at different initial concentrations of oxidant (H 2 O 2 ). The influence of oxidant concentration on the kinetics was investigated by several studies and the main findings can be reduced to the fact that neither too low hydrogen peroxide concentration (leading to a rate reduction of the Fenton reaction) nor a too high concentration may be applied (H 2 O 2 competes successfully for hydroxyl radicals and becomes decomposed with oxidizing the pollutant). Usually, there is a rather broad concentration interval in between both extremes, where none of both phenomena occurs (Malato et al., 2009). Figure 2 shows the effect of initial concentration of H 2 O 2 which was studied in the range of mg L -1 at ph= and at constant ferrous concentration of 5 mg L -1. For the photo-fenton process, the addition of H 2 O 2 from 50 to 500 mg L -1 increases the reduction of COD from 58 to 69%, respectively. This can be explained by the effect of the additionally produced hydroxyl radicals. Further increases from 500 to 1000 mg L -1 cause no significant change in COD reduction, while a small decrease is only achieved, as Fig. 2 shown. This little decrease is due to the fact that at a higher H 2 O 2 concentration scavenging of HO radicals will occur as well as recombination of hydroxyl radicals, which can be expressed by the Eqs. (5) to (7) (Fernandez et al., 1999; Lucas and Perez, 2006; Tamimi et al., 2008): H2O2 H2O HO2 HO (5) HO 2 HO H2O O2 (6) HO HO H O 2 2 (7) 26

27 % COD removal It can be postulated that H 2 O 2 should be added at an optimum concentration to achieve the best COD removal. Hence, 500 mg L -1 of H 2 O 2 appears as an optimal dosage for solar-fenton process mg H2O2/L 100 mg H2O2/L 500 mg H2O2/L 750 mg H2O2/L 1000 mg H2O2/L time (min) Figure 2: Effect of initial hydrogen peroxide concentration on the COD removal of winery wastewater effluent; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 = mg L -1, ph= and T=25±0.1 o C Effect of ph The ph value affects the oxidation of organic substances both directly and indirectly. The Fenton and photo-fenton reactions are strongly ph dependent. The ph value influences the generation of hydroxyl radicals and thus the oxidation efficiency (Lucas and Peres, 2006; Tamimi et al., 2008; Malato et al., 2009). The effect of ph on the reduction of COD of the winery wastewater effluent by solar-fenton process is shown in Fig

28 % COD removal The experiments were carried out at ph between 2 and 4. At ph=2 the COD removal percentage is 45% and for ph=4 is 39% at 120 min. At ph=2.9 we can obtain the highest COD removal, which was 69% ph=2 ph=2.9 ph= Time (min) Figure 3: Effect of initial solution ph on the COD removal of winery wastewater; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. The optimum ph was found to be about 3. For ph values below 3, the reaction of hydrogen peroxide with Fe 2+ is seriously affected causing the reduction in hydroxyl radical production. Inhibition of radical HO formation at ph below 3 seems also to be due to the small amount of soluble iron (Fe 3+ ), responsible for the continuity of the oxidation process, occurring in the formation of Fe(OH) 2+ and Fe(OH) + 2 (Peres et al., 2004). The low COD removal at ph=2 is also due to the hydroxyl radical scavenging of H + ions (Eq. (8)) (Lucas and Peres, 2006). HO H e H2O (8) 28

29 The COD removal decreased at ph values higher than 4, because iron precipitated as hydroxide, which resulted in a reduction in the transmission of the radiation (photo- Fenton). Additionally, the oxidation potential of hydroxyl radical is known to decrease with increasing ph (Tamimi et al., 2008). Another reason for the inefficient degradation at ph > 3 is due to the dissociation and auto-decomposition of H 2 O 2 (Badawy et al., 2006). Thus, ph= is frequently postulated as an optimum ph for photo-fenton treatment (Safarzadeh-Amid et al., 1996), because at this ph precipitation does not take place yet and the dominant iron species in solution is Fe(OH) 2+ (Eq. (9)), the most photoactive ferric iron-water complex (Malato et al., 2009). Fe 2 2 ( OH) hv Fe HO (9) Effect of Temperature In order to investigate the role that the temperature plays at the reduction of COD, at the winery wastewater, experiments were conducted at different temperatures varying from 15 to 40 o C, which was controlled by a thermostated container. It has been reported that an elevated temperature can significantly increase the activity of the photo-fenton system (Gob et al., 2001; Lee and Yoon, 2004; Gernjak et al., 2006; Malato et al., 2009). Figure 4 shows the COD reduction of the winery wastewater at each temperature. The COD-values increased with increasing temperature and especially the COD removal ranged from 46 to 72% for 15 to 40 o C, respectively. 29

30 % COD removal This enhancement can be attributed to the increase of the Fe(OH) 2+ concentration and to the temperature dependence of the quantum yield of the photochemical reduction of Fe 3+ as well (Lee and Yoon, 2004). In particular, temperature turns into a key parameter when it is desirable to reduce reagents costs, or when high levels of iron are not allowed into the treated wastewaters (Torades et al., 2003) T=15oC T=25oC T=40oC Time (min) Figure 4: Effect of temperature on the COD removal of winery wastewater; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=15-40 o C Effect of the solar irradiation In order to evaluate the benefit of solar irradiation in the COD reduction of the winery wastewater, effluents after SBR treatment, the following experiments were carried out: (i) photolysis (only solar irradiation), (ii) solar irradiation + Fe 2+, (iii) solar 30

31 irradiation + H 2 O 2 (photo-bleaching), (iv) Fe 2+ + H 2 O 2 (dark Fenton), and (v) solar irradiation + Fe 2+ + H 2 O 2 (solar-fenton). Figure 5 shows the COD removal of all these five processes, for the same concentrations of catalyst and oxidant. Only the solar irradiation (direct photolysis) showed no significant COD reduction on the winery wastewater effluent, which was around 6.5% after 120 min of irradiate time. The combination of solar irradiation and catalyst (Fe 2+ ) leading to a very small COD reduction of 8.6% after 120 min under solar irradiation, a value that is compared to the COD reduction due to the photolysis. With regard to the efficiency of H 2 O 2 (oxidant) to the COD removal of winery wastewater, experiments were performed with the optimum concentration of the oxidant, which was the 500 mg L -1, and the solar irradiation duration 120 min. The results showed that the COD removal due to photo-bleaching was 38%, a reduction which it is owed to the formation of some hydroxyl radicals form due to the irradiation of H 2 O 2. Dark Fenton experiments (in the absence of solar light) were showed that 45% of the COD reduction is due to the Fenton reagent. Finally the solar- Fenton showed the optimum COD reduction which was 69% after 120 min of the photocatalytic treatment. 31

32 % COD removal Photolysis Solar irradiation + Fe(II) Solar irradiation + H2O2 Dark Fenton Solar Fenton Time (min) Figure 5: COD removal of winery wastewater effluents under different processes; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. As the above figure shows, the order from the most efficient to the least efficient process is the following: solar-fenton (69%) > dark-fenton (45%) > photo-bleaching (38%) > solar irradiation + Fe 2+ (8.6%) photolysis (6.5%) Total phenols and BOD 5 removal Phenols are on the United States Environmental Protection Agency (U.S. EPA) priority pollutants list. However, phenols are often used by various industries and found in various consumer goods. Therefore, contamination of soil and groundwater by phenols are common in the world (Yang and Yong, 1999). Figure 6 shows the % total phenols removal of winery wastewater after 120 min of solar-fenton process which was 71% after 120 min of solar irradiation. 32

33 % Total phenols removal Total phenols (mg/l) (a) time (min) (b) Time (min) Figure 6: Concentration of Total phenols (a) and % Total phenols removal (b) of winery wastewater effluent during solar-fenton process; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. The BOD test measures the molecular oxygen utilized during a specified incubation period for the biochemical degradation of organic material (carbonaceous demand) and the oxygen used to oxidize inorganic material such as sulfides and ferrous iron. It may also measure the amount of oxygen used to oxidize reduced forms of nitrogen (nitrogenous demand) unless their oxidation is prevented by an inhibitor. The seeding and dilution procedures provide an estimate of the BOD at ph 6.5 to 7.5. Measurements of oxygen consumed were done in a 5-days test period (BOD 5 ). The BOD test after 120 min of solar-fenton coagulation was performed using the OxiDirect BOD measurement system. The results of BOD 5 measurements are shown in Fig.7. It can be seen that after the 120 min of solar-fenton treatment the 33

34 BOD 5 (mg/l) BOD reduces from 112 mg L -1 to 17 mg L -1 (i.e. on Day 5). This figure the sum of the amount of BOD produced from COD after treatment and the original BOD minus the amount oxidised during treatment (Eq. (10)). This corresponds to a BOD 5 removal as high as 87% after 120 min of reaction. As for the BOD/COD ratio before the treatment was equal to 1.1 while after 120 min of treatment significant decrease of BOD/COD ratio appeared equal to 0.53, perhaps due to the mineralization of part of biodegradable organic matters by hydroxyl radicals. According Metcalf and Eddy (1991), if the BOD/COD ratio for untreated wastewater is 0.5 or greater the waste is considered to be easily treatable by biological means. BOD FinalMeasured BOD (10) fromcod ( BODoriginal BODoxidised) Day 1 Day 2 Day 3 Day 4 Day 5 Before Fenton treatment After Fenton treatment Figure 7: Measurement of BOD 5 values for the winery effluent before and after 120 min of solar-fenton treatment; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. 34

35 4.1.7 Degradation kinetics Kinetic studies of photo-fenton processes can be performed assuming that the reaction between hydroxyl radicals and the pollutant is the rate determining step as has been reported by other authors (Malato et al., 2001; Evgenidou et al., 2007; Nunez et al., 2007; Pariente et al., 2008). Under the assumption that HO radicals rapidly achieve a constant steady-state concentration in the solution, the winery wastewater COD and TOC removal may be described by pseudo-first order kinetic expression Eq. (11): dc r dt k C HO k C appt C kapp C C Co e ln kapp t (11) C o where C is the concentration of the COD and DOC of winery wastewater, k is the reaction rate constant and k app is a pseudo-first order constant. The experimental data fit well in the linear kinetic equation and consequently the winery wastewater oxidation follows the pseudo first order law like shown in Fig.8 ((a) and (b)). The degradation rate constant, k (min -1 ), was determined from the slop of the - ln(c/c o )=f(t) plot, where C o and C are the concentrations of the substrate at times 0 and t (min) respectively and was min -1 and min -1 for COD and DOC respectively. It should be noted that linear fitting was done for those reaction times necessary to achieve up to 51% COD and 25% DOC removal of winery wastewater treatment. This criterion was set assuming a steady state production of hydroxyl radicals from hydrogen peroxide decomposition which is likely to occur mainly during the early stages of the reaction. 35

36 -ln (C/Co) -ln (C/Co) 1.2 (a) 1 y = x R 2 = Time (min) 0.35 (b) y = x R 2 = Time (min) Figure 8: Photocatalytic (COD (a) and DOC (b)) decomposition kinetics of winery wastewater effluent after solar-fenton oxidation; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C Color removal Wastewaters originated from the wine industry contain various pollutants including a high content of organic matter and color (depending on the wine production). There is 36

37 a considerable need for the removal of color from wastewater effluents. This work investigates its removal from the partially biologically treated winery wastewater effluent by the solar-fenton technology. Prior to color measurements, the wastewater samples were filtered using 0.22 μm filters by Millipore, to prevent turbidity. The color measurement followed the ADMI Tristimulus Filter Method (Method 2120E in Standard Methods). To determine the color in ADMI unit, a light scan from 350 to 700 nm was performed using a UV-Vis Jasco V-530 spectrometer (Kang et al., 2000; Kang et al., 2002; Pala and Tokat, 2002). The results show that color was removed effectively by Fenton oxidation. The color removal is markedly related with the amount of HO formed (Kang et al., 2000). The color removal reached a maximum of 53% at a reaction time of 120 min, under 500 mg L -1 H 2 O 2 and 5 mg L -1 Fe 2+, while 44% of the color was removed after the first 5 min of solar-fenton treatment (Fig. 9 and Fig. 10). Color content was determined by measuring the absorbance at three wavelengths (390, 400, 410), and taking the sum of these absorbencies (Pala et al., 2002). 37

38 % Color removal Abs (cm-1) Std 5 min 10 min 15 min 20 min 25 min 30 min 60 min 90 min 120 min Wavelenght (nm) Figure 9: Color removal of winery wastewater effluents after solar-fenton treatment; Inset graph shows the color removal of winery wastewater after min of Fenton treatment; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C time (min) Figure 10: % Color removal of winery wastewater after solar-fenton treatment; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. 38

39 % DOC and COD removal Average oxidation state (AOS) Figure 11 shows the removal of COD and DOC during the photocatalytic treatment of the winery wastewater. COD and DOC decreased during the photocatalytic process. The initial COD of the treated solution is equivalent to 103 mg L -1. By the end of the process the COD reached approximately 32 mg L -1. The initial DOC was equivalent to 46 mg L -1 and the final after 120 min of oxidation was 24 mg L -1. The DOC removal after 120 minutes of treatment (48%) is lower than the COD removal (69%, at the same time), which indicates that the oxidized organic matter is not completely mineralized to carbon dioxide (Navarro et al., 2005) % DOC removal % COD removal Time (min) Figure 11: COD and DOC removal of winery wastewater after Fenton treatment; [Fe 2+ ] 0 =5 mg L -1 and [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. The efficiency of the oxidative process is more clearly shown by the AOS parameter (average oxidation state), which can be calculated by Eq. (12), in which DOC and 39

40 COD are expressed in moles of C L -1 and of O 2 L -1, respectively, at the sampling time (Sirtori et al., 2009). ( DOC COD) AOS 4 (12) DOC AOS constitutes a potentially valuable parameter that can be used to estimate the degree of oxidation in a complex solution consisting of the initial compound and its oxidation by-products (Velegraki et al., 2010). AOS is between +4 for CO 2, the most oxidized state of C, and -4 for CH 4, the most reduced state of C (Sarria et al., 2002). AOS usually increases with treatment time until almost reaching a plateau, as shown in Fig.12. These results suggest that more oxidized organic intermediates are formed at the beginning of the photocatalytic process, and after a certain time, the chemical nature of most of them no longer varies substantially (Sarria et al., 2002), even if the photo-fenton treatment continues. Formation of more oxidized intermediates indirectly demonstrates that the treatment can improve biodegradability. At the moment that AOS stabilizes, the chemical treatment is only mineralizing organic contaminants, but with no partial oxidation (Sirtori et al., 2009). 40

41 AOS time (min) Figure 12: Changes in the average oxidation state, AOS; [Fe 2+ ] 0 =5 mg L -1 [H 2 O 2 ] 0 =500 mg L -1, ph= and T=25±0.1 o C. and However, its use does not provide information concerning the degree of partial and total oxidation that has occurred since the beginning of the reaction, but simply reports on the oxidation state at a given time without following the progress of partial oxidation reactions (Mantzavinos et al., 2000). COD removal by partial oxidation (COD partial) at different reaction times can be expressed as follows (Eq. (13)): COD partial DOC COD COD DOC 0 (13) 0 Eq. (14) can be expanded to give additional information with respect to the relative contribution of partial and total oxidation reactions that have occurred; the index ε represents the efficiency of COD removal through partial oxidation and is defined as the ratio of COD removal by partial oxidation to the overall COD removal (Mantzavinos et al., 2000): 41

42 COD partial (14) ( COD0 COD) where ε takes values from 0 to 1 corresponding to total and partial oxidation, respectively. 42

43 4.2 Solar heterogeneous Fenton-like process Blank experiments In order to assess the range of the factors influencing the efficiency of the present photocatalytic reaction system, various blank tests were performed. The reaction system was irradiated with solar light i) in the absence of both catalyst and oxidant (which allows to evaluate the extent of direct photolysis), ii) just in the presence of oxidant (i.e. H 2 O 2 ), iii) just in the presence of catalyst Fe 2 O 3 /SBA-15 (to investigate the contribution of surface reactive species generated from the illumination of the catalyst particles). The contribution of the latter phenomena is negligible (< ca. 5% after 180 min) as shown in Fig. 13. In order to assess the contribution of dark-fenton reactions to the overall catalytic degradation of photo-fenton systems additional experiments using 100 mg L -1 Fe SBA-15 under dark reaction conditions were carried out: (i) without hydrogen peroxide (dark adsorption) and (ii) in the presence of 100 mg L -1 hydrogen peroxide (dark Fenton). Dark adsorption onto the catalyst surface was negligible (< ca. 3% after 180 min). The contribution of dark Fenton reactions was also negligible leading to 5% DOC removal, after 180 min; the respective values for the photo-fenton run were almost 20%, thus confirming the strong influence of solar radiation on Fenton s reactivity for the mineralization of the winery effluent. 43

44 Figure 13: Blank experiments depicting the effect of the solar light in the presence of the oxidant and/or the catalyst and the contribution of dark reactions to the overall efficiency of the process; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]=100 mg L -1, ph o =8 and T=23 C Effect of heterogeneous catalyst Fe 2 O 3 /SBA-15 In order to examine the effect of the heterogeneous catalyst Fe 2 O 3 /SBA-15 on the efficiency of the photocatalytic process, experiments were carried out at different initial concentrations of the iron containing catalyst (e.g. [Fe] SBA-15 = 50 mg L -1, [Fe] SBA-15 = 100 and [Fe] SBA-15 = 200 mg L -1 ). The concentrations chosen correspond to the iron content of the Fe 2 O 3 /SBA-15 catalyst i.e %. The results are presented in Fig. 14 and it is depicted that all the tested catalyst concentrations lead to similar mineralization degrees (i.e. between 5 and 20%) after 180 min of solar photocatalytic treatment. However, it is obvious that the utilization of [Fe]SBA-15 = 200 mg L -1 inhibits the efficiency of the reaction, both in terms of mineralization extent (i.e. 7%) and mineralization rate. This effect could be due to the 44

45 fact that higher amounts of catalyst particles are suspended in the reaction mixture, thus preventing the photons from reaching the active sites of the catalyst and initiating solar Fenton reactions. In view of this, the study focused on lower catalyst concentrations and it was shown that the addition of 100 mg L -1 Fe SBA-15 leads to marginally higher mineralization (i.e. 18%) even from the first 30 min of the treatment. Figure 14: Trend of DOC removal at different concentrations of Fe 2 O 3 /SBA-15 catalyst during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [H 2 O 2 ] =100 mg L -1, ph o =8 (inherent) and T=23 о C. It is interesting to note that the DOC reduction commences even from the beginning of the process, which means that complete oxidation reactions may also occur during the heterogeneous photocatalytic treatment of a biologically pre-treated winery effluent. Nonetheless, practically no further mineralization occurs after the first 30 min of the reaction. Taking into account that hydrogen peroxide is still available in the reaction medium after 180 min, thus eliminating the possibility of the complete 45

46 consumption of oxidant, it could be assumed that some kind of catalyst deactivation occurs which leads to the actual termination of further oxidation reactions Effect of H 2 O 2 concentration The effect of the amount of oxidant was examined by employing different initial concentrations of hydrogen peroxide varying between 0 and 300 mg L -1 (Fig. 15). At oxidant concentrations as low as 50 mg L -1 H 2 O 2 no decrease in the DOC content of the winery effluent is observed, which is similar to the corresponding experiment without oxidant addition. However, increasing the H 2 O 2 up to 100 mg L -1 causes almost 20% mineralization within 60 min, which is due to the fact that higher amounts of hydroxyl radicals are produced (Muruganandham and Swaminathan, 2004). Higher treatment times do not lead to further improvement in the mineralization efficiency of the process. At higher peroxide concentration i.e. 300 mg L -1, the mineralization is completely supressed, impeding the performance of the process, which can attributed to the to the hydroxyl radical scavenging effect of H 2 O 2 (Muruganandham and Swaminathan, 2004). Since working at even higher H 2 O 2 concentrations would raise additional issues of concern such as the extra cost and the need to remove any unreacted peroxide from the effluent, it was deemed unnecessary to investigate the effect of higher amounts of oxidant. The corresponding COD measurements (data not shown) follow the same decrease as in the case of DOC removal, which indicates that the oxidative reactions taking place during the heterogeneous photocatalytic process lead to the complete elimination of the removed organic content through total oxidation reactions. 46

47 DOC / DOCo [H2O2] = 50 mg/l [H2O2] = 100 mg/l [H2O2] = 300 mg/l [H2O2] = 0 mg/l Time (min) Figure 15: Trend of DOC removal at different oxidant concentrations during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, ph o =8 (inherent) and T=23 о C Effect of ph The initial ph constitutes a very decisive factor in regards to the performance of the heterogeneous photocatalytic process. The ph value of the reaction solution may affect both the surface charge on the catalyst and the ionization state of the substrate and consequently the extent of the organic substrate adsorption onto the catalyst particles (Dhananjeyan et al., 1996; Bhatkhande et al., 2001). In order to evaluate the effect of the effluent s initial ph on the degradation efficiency, two different ph values were examined, i.e. 2.8 and 8. The acidic ph was considered because it is the optimum ph value for homogeneous photo-fenton and Fenton reaction systems (Feng et al., 2006) and also because it is necessary to test the stability of the heterogeneous Fe 2 O 3 /SBA-15 catalyst under conditions that generally promote the leaching of the 47

48 catalyst (Sum et al., 2005). The second ph was considered because it is the inherent ph of the winery effluent after it has passed through biological treatment DOC / DOCo ph = 2.8 ph = Time (min) Figure 16: Trend of DOC removal at different initial ph values during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1 and T=23 о C. It is obvious from Fig. 16 that similar results are obtained under both experimental conditions. Almost 20% of the organic content is completely mineralized after 180 min of treatment at initial ph 2.8, which is the same removal obtained at inherent ph (i.e. 8). In order to exclude the possibility of homogeneous photo Fenton reactions contributing to the mineralization observed at ph 2.8, the concentration of the total dissolved iron in the reaction solution was measured during treatment. The results showed no leaching of the iron from the catalyst to the solution; therefore, it is safe to assume that the degradation that occurred is solely due to photo-assisted heterogeneous Fenton reactions. Similar results were reported for the reduction in COD measurements, which reached ca. 20% after 180 min, thus leading to the 48

49 assumption that only complete oxidation takes place during the heterogeneous solar Fenton-like process. These results show that heterogeneous photocatalysis is a very flexible method for the treatment of non-biodegradable effluents with ph values ranging from 2.8 to 8. As the winery effluent of the present study has an inherent ph value around 8, no ph adjustment was deemed necessary for the experiments that follow Effect of Temperature The effect of the solution temperature on the treatment efficacy was evaluated by performing experiments at different temperature values i.e. 23 o C and 40 o C (Fig. 17). Higher temperature (i.e. 40 o C) seems to cause a decrease in the mineralization efficiency of the photocatalytic process (Fig. 17(a)) that could be attributed to partial thermal decomposition of H 2 O 2 towards H 2 O and O 2 (Yip et al., 2005). However, the peroxide measurements during treatment do not corroborate the assumption above, since no substantial peroxide elimination was recorded. Furthermore, the corresponding results in the COD measurements (Fig. 17(b)) show that both temperatures lead to similar degradation degrees (i.e. ca. 20% COD reduction after 180 min of reaction). This observation could indicate that the increase in temperature favours partial oxidation rather that total oxidation and/or that some kind of catalyst deactivation of catalyst occurs. 49

50 1.0 (a) DOC / DOCo T=23 C T=40 C Time (min) (b) Figure 17: Trend of (a) DOC and (b) COD removal at different solution temperatures during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1 and ph o = Total phenols and BOD 5 removal The effect of the heterogeneous solar Fenton-like on the abatement of the total phenols content (TPh) is presented in the following figure (Fig. 18). It can be seen that, under optimum experimental conditions, about 80% of the phenolic content is removed within 60 min of treatment. Beyond that point, no additional abatement is 50

51 recorded which comes to agreement with the corresponding elimination trend for the DOC content (Fig. 16) TPh / TPho Time (min) Figure 18: Total phenols removal during the heterogeneous solar Fenton-like treatment of a pre-treated winery effluent; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1 and ph o =8. With this in view, it could be assumed that part of the DOC that is removed is owed to the TPh content that is removed during photocatalytic treatment. Furthermore, taking into account the average results (i.e. ca. 5 to 20% DOC reduction) during heterogeneous photocatalysis, it was deemed necessary to make an assessment of the biodegradability of the winery effluent after treatment with heterogeneous solar Fenton-like (i.e. HSF) under different experimental conditions. The results of the BOD 5 measurements are shown in Fig. 19. It can be seen that after HSF treatment that led to merely 5% DOC removal the BOD increases up to a value of 30 mg L -1 (i.e. on Day 5). This corresponds to a BOD/COD ratio that equals to

52 BOD 5 (mg/l) which reveals low biodegradability of the resulting wastewater. In turn, after HSF treatment that led to almost 20% DOC removal the BOD seems to increase up to 155 mg L -1 (i.e. on Day 5). In the latter case, the corresponding BOD/COD ratio exceeds 0.5 which is indicative of a biodegradable effluent. These results indicate that the application of heterogeneous photocatalysis under mild experimental conditions can enhance significantly the biodegradability of a biorecalcitrant effluent, which has already undergone biological treatment that eliminated any biodegradable fraction present in the raw wastewater Day 1 Day 2 Day Day 4 80 Day % (HSF) 20 % (HSF) Figure 19: Measurement of BOD 5 values for the winery effluent after 180 min of heterogeneous solar treatment (HSF) under different experimental conditions that result in different DOC removal percentages. Green: 20% DOC removal; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1, ph o =8 and T=23 C. Purple: 5% DOC removal; [DOC] o =100 mg L -1, [Fe] SBA =100 mg L -1, [H 2 O 2 ]= 100 mg L -1 and ph o =8 and T=40 C. 52