Catalytic Ozonation Using Fe 2 + in the Treatment of Dairy Effluent in a Semi-Batch Process with Recycle

This work describes the treatment of dairy industry effluent using catalytic ozonation with Fe as its catalyst in a semi-batch process with recycle. A fractional factorial design 2 was used with a reduction percentage of total organic carbon (TOCred) as response. Optimal conditions were obtained by the reaction time of 30 min, ozonator power of 35 W, O2 flow rate of 0.125 L min , Fe concentration of 1.0 g L and pH 4.0 for a 2 L raw effluent. TOCred of 64.03% represents a decrease in concentration from 473.0 to 170.1 mg L, with an estimate cost of US$ 0.03 L effluent. The treatment performed was not sufficient to discharge it directly on surface water; however, significant reduction of TOCred, among physical and chemical characteristics makes a better product to be proceeded into a biological treatment.


Introduction
Due to high consumption of water and large volume of effluent produced, dairy industry is considered the most pollutant in comparison to other food industries. 1 Treatment is, therefore, required since it is composed of high molecular weight and aromatic compounds that are not easily degraded, considered as pollutants.
Dairy effluents are composed of milk and processed products derived from milk as well as disinfectants, detergents, spices and essences (butter and cheese production) that are diluted in the washing process of equipment, pipes, floor and other facilities of such industry. 2ndustries have currently used treatment alternatives based on physical and biological processes that although show restricted efficiency, have several limitations.These limitations generally include primary treatment for the removal of suspended solids and fat material, and a secondary biological treatment.Problems related to the secondary biological treatment are related to high foam production, low biodegradability of sludge, low resistance to load shock, difficulties in removing nutrients (nitrogen and phosphorus) and problems in the degradation of fat and other specific types of pollutants as dyes, apart from unpleasant odor. 3,4us, efficient and economically viable technologies are required to treat these types of effluents. 5In this sense, advanced oxidation processes (AOPs) are technologies extremely efficient based on the oxidizing power of several radicals formed that are used in the destruction of organic compounds of difficult degradation (partial or total). 6,7atalytic ozonation may be considered as a homogeneous process based on ozone activation by metal ions present in aqueous solution.These reactions are characterized by single-phase development or as a heterogeneous process in the presence of metal oxides. 8Also, the transition metals form low solubility products under certain pH conditions.
As pH is related to concentration of hydroxyl ions (OH -), it directly influences the decomposition of molecular ozone.In situations where a small concentration of hydroxyl ion is present (pH < 4), the reaction takes place predominantly by electrophilic addition of molecular ozone (direct reaction). 9In an alkaline medium, hydroxyl ions in high concentration may start the decomposition of molecular ozone leading to the formation of hydroxyl radical (indirect reaction). 10he reaction system becomes heterogeneous in an alkaline medium with Fe 2+ ions in the catalytic ozonation processes producing a resultant precipitation of these ions.However, this precipitation does not prevent to work in this pH. 11The removal rates of the organic load in alkaline medium are not always satisfactory.Due to the complexity of environmental matrix, numerous compounds considered sequestrants of hydroxyl radical, such as HCO 3 -, CO 3 2-, CH 3 COO -and humic substances might be present resulting in decrease of the efficiency of the process. 9he reaction between the hydroxide anion and ozone leads to the formation of superoxide radical anions (O 2 •-) and hydroperoxide radicals (HO 2 • ), as shown in equation 1. 12 O 3 + OH -→ O 2 •-+ HO 2 Hydroperoxide radical shows acid-base equilibrium (equation 2). 12 The radical anion ozonide (O 3 •-) formed in the reaction between ozone and the superoxide radical anion (O 2

•-
) quickly decomposes to form hydroxyl radicals.This formation of hydroxyl radicals may occur both in acid and alkaline media (equations 3 to 7). 13,14 Some studies [15][16][17] show that Fe 2+ catalyzes decomposition of O 3 to generate hydroxyl radicals following a distinct mechanism.In this case, the catalytic system O 3 /Fe 2+ involves direct reaction of Fe 2+ with ozone.Equations 8 to 14 illustrate proposed reaction mechanisms of Fe 2+ with ozone. 18Béltran et al. 19 proposed reaction mechanisms between Fe III and ozone with subsequent formation of hydroxyl radical, as shown in equation 15.
As shown in equations 8 to 15, the reaction mechanism is the transfer of electrons from the reduced metal to ozone, forming Fe 3+ ion and radical ion O 3 •-and so the hydroxyl radical.In the presence of excess Fe 2+ , hydroxyl radical can oxidize a second Fe 2+ providing a stoichiometric ratio of 0.5 moles of ozone per mole of ferrous ion. 18rslan 20 studied coagulation, ozonation and Fe 2+ ions catalyzed ozonation (FeSO 4 •7H 2 O) in commercial textile dyes at different pH (3 to 13).The response variable was the percentage of color and chemical oxygen demand (COD) reductions.In the coagulation procedure, a range of Fe 2+ concentrations (FeSO 4 •7H 2 O) from 0.25 to 5 g L -1 , at different pH values (3 to 13) was used.Largest reductions were obtained at pH 11 using 2.5 g L -1 of FeSO 4 •7H 2 O, yielding 96.9% reduction in dye and 54% reduction of COD.At pH 3, color reduction was 72.9% and COD 6.5%, which is attributed to the fact that dyes tend to aggregate in alkaline medium, causing reduction in solubility and ionization.In the process of ozonation and catalytic ozonation, 3.5 g h -1 of ozone mass flow was applied.After 10 min of ozonation and at different pH (3, 5, 8 and  11), in contradiction with coagulation experiments, the highest color and COD reductions were obtained at pH 3, with 77% color and 11% COD reduction, respectively.It should be pointed out that in an alkaline media, there is the decomposition ozone, occurring the formation of • OH radicals, waiting for an increase in the oxidation of pollutants.However, organic dyes, having polyaromatic structures and double bonds in their chromophoric groups, can react both in molecular pathway (acidic medium) as a radical • OH (alkaline medium).Considering the presence of acetic acid in the dye formulation, this will possibly compete with • OH alkaline solution, meaning that less • OH is available for dye oxidation.In the same study, ozonation tests were also conducted in the presence of Fe 2+ ions as FeSO 4 •7H 2 O, applied in concentrations ranging from 0.025 to 1.0 g L -1 at different pH (3 to 11).The color and COD removal efficiency was slightly higher in the presence than in the absence of Fe 2+ , showing the tendency to decrease with increasing pH, too.Increasing the concentration of Fe 2+ to 0.5 and 1.0 g L -1 at pH 3, showed an increase in the percentage of color and COD reductions.For 0.5 g L -1 Fe 2+ concentration, 76.2 and 25.2% color and COD reductions were observed, respectively.Increasing the concentration of Fe 2+ to 1.0 g L -1 , 94.6 and 47.8% in color and COD reductions were achieved, respectively.However, it possibly works in alkaline medium using Fe 2+ ions.COD and total organic carbon (TOC) removal rates do not depend only on catalyst, but also on the matrix to be treated.
Freire et al. 21applied the ozonation process for treating paper mill effluent by direct reaction (molecular ozone) and by indirect reaction (hydroxyl radical).These two processes were compared concerning TOC removal efficiency, total phenols and effluent color.Sevimli and Kinaci 22 compared the efficiency of Fenton process with ozonation and Fenton process concerning textile dying degradation.Balcioglu and Ötker 23 compared pH and H 2 O 2 concentration effect over the ozonation process, concerning COD efficiency removing of pharmaceutical effluents.Some references found in literature compare the efficiency of the ozonation process in the degradation of different substrates via direct and indirect reaction.For this work, both acidic media (pH 4.0) and alkaline media (pH 10.0) were used.
This work aims an experimental design to evaluate the efficiency of a dairy effluent treatment system from a dairy industry, Cia. de Alimentos Glória, in the city of Guaratinguetá, state of São Paulo, Brazil.The technique used advanced oxidation processes with the use of catalytic ozonation, using Fe 2+ as a catalyst in a recycling semi-batch process, besides a new reactor configuration due to the foam generated during the process.
This work aims a technological innovation since there is no scientific reference to scientific work about homogeneous catalytic ozonation in a semi-batch process with recycling for this type of effluent.

Sampling and preservation
The effluent used was from the dairy industry Cia. de Alimentos Glória in the city of Guaratinguetá, São Paulo, Brazil.Samples were collected directly from the homogenizer tank of such industry in quantity enough to perform all of the experimental design planned; they were stored in plastic recipients and frozen down in order to minimize alteration in the physicochemical characteristics of the effluent.Analysis and experiments were conducted at Escola de Engenharia de Lorena (EEL-USP) in the Laboratories of Basic and Environmental Sciences Department (LOB) and Chemical Engineering Department (LOQ).

Methodology
A volume of 2 L of dairy effluent was used in each experiment, conditioned naturally to room temperature (23 ± 2 °C) before experiments.In order to minimize any intrinsic physicochemical alteration to this type of effluent, only the corresponding volume of effluent was thawed to be used in possible experiments of the day.After homogenizing the effluent, the volume was measured in a flask and transferred to a Becker.pH was adjusted according to the experimental condition with commercial solutions of 98% m/m H 2 SO 4 that were measured using a benchtop pHmeter.After the adjustment, the effluent was introduced to the reactor.After setting up all modules and adaptation of pumps and tubes, O 2 flow rate (99.50% of purity) for cylinder was adjusted using a valve according to each experiment.During this phase, the whole system was verified using only injection of O 2 .Afterwards and concomitantly, the ozonator was started up at the specific power and the addition of FeSO 4 •7H 2 O began.This solution was previously solubilized with commercial 98% m/m H 2 SO 4 , being introduced to the reactor with a peristaltic pump during 20 min of the total reaction time according to design of experiment levels.Concomitantly, the aeration of the system was initiated for foam formation reduction.
In general, homogeneous catalytic ozonation application should be associated with some technique for the removal of metallic ions from the treated effluent.Toxic character of some metals may restrict the application of this type of process.Furthermore, the presence of ions like iron, although less dangerous from a toxicological point of view, may be associated with undesirable effects, such as the arising of blemishes in clothes, as well as an unpleasant taste and odor in water. 11iquid foam was drained into a reservoir that could be recycled to the oxidative process through a pulse peristaltic pump (started up at the beginning of the process), without compromising the initial volume of the reactor.With a gradual reduction of foam formation during the rest of the reaction, the air pump was shut down at the last 10 min of the process.After 30 min of reaction, an aliquot (30 mL) was taken to perform TOC analysis.It is imperative to explain that prior to TOC analysis, considering the heterogeneous catalysis, the final product (obtained after oxidation process) was filtered and pH was conditioned in an acidic media.Considering the homogeneous catalysis, the final product was conditioned in alkaline solution for Fe precipitation, and then, pH was conditioned in an acidic media.
Organic carbon determinations were carried out using a total organic carbon Shimadzu analyzer, model TOC-VCPH, based on the catalytic oxidation at high temperatures.
In the determination of TOC, the analytical curve of non-purgeable organic carbon (NPOC) was prepared using a standard solution of potassium biphthalate, linear range 0-1000 mg L -1 .To the TOC determinations, samples were diluted with deionized water according to their concentrations.Samples were prepared between pH 2.9 to 3.1, without any precipitate and after homogenization they were injected in the combustion chamber of the equipment of TOC determination.

Quantification of O 3 mass flow
Iodometric titration was used to quantify the concentration of O 3 that was bubbled in a KI solution in sulfuric medium that consists in an indirect titration process.The end point for titration was obtained using amido gum as an indicator. 24The formation of tri-iodine ion (I 3 -), stable and soluble reactional product, which possesses the same number of equivalents than ozone, was obtained as shown in equations 16 to 18.
The concentration of tri-iodine was dosed by using a standard solution of sodium thiosulfate as shown in equation 19.
It is relevant to emphasize that the combination among ozonator power and O 2 flow rate did not show linearity, and thus, the O 3 mass flow rate (mg min -1 ) was quantified in all experiments of the experimental matrix.It was quantified the O 3 that was injected in the reactor, before the contact with the effluent.So, the products of the effluent do not interfere with this quantification.

Configuration of semi-batch reactor with recycle to the catalytic ozonation process of dairy effluent
The reactor used in this work was developed after some exploratory experiments.From an operational point of view, it is worth to emphasize that the configuration developed in this work gives the possibility to use the reactor under different systems: batch, semi-continuous with recycle and continuous one.
The reactor used in this work was idealized aiming effluent treatment that, besides promoting oxidation of organic matter, would concomitantly eliminate the foam formed during the treatment, without compromising the process or influencing the kinetics of degradation.The reactor was projected in borosilicate glass that can be operated in a batch or a continuous process (Figure 1).It consists of three main modules: reactor, cone (foam separator) and reservoir.
The inferior part of the reactor is 0.42 m high and of 0.10 m internal diameter with volumetric capacity of 4.0 L supported by a wooden base.
The oxidation process of the organic matter is carried out in the reactor (1).It consists, in all its extension, of a cylinder shape and that is where the input of O 2 + O 3 (4) from the ozonator (6), catalytic solution (7a), effluent feed (7) and sampling (11) occurs.
The O 2 + O 3 input is done at the base of the reactor (4) by micro-bubbles, using an air-stone (4a) for better adsorption of O 2 + O 3 and homogenization of oxidative system.The oxygen of the O 2 cylinder ( 5) is converted into ozone by the method of electric discharge that occurs in the ozonator, taking into account that the input is done at the inferior part of the reactor allowing an ascending flow.
The entry of the catalytic solution ( 7) is done at the opposite side of the O 2 + O 3 flow.The catalytic solution (7a) is previously prepared in sulfuric medium and introduced to the reactor by peristaltic pump (8) that is activated 10 s after the ozone input and remains active for 20 min of the whole reaction time of 30 min.
The effluent recycle entry, after the disruption of foam, is processed by the entry (9) through a pulse pump (10).
Sampling (11) was carried out manually with a kind of an open/close glass valve that allows fast sampling and closure of the system, without interference.
Module ( 2) is in the shape of a cone and it is attached to the reactor by a metal clip strap (2a).This is composed of a glass tube (2b) curved at 180°.This tube is comprised of an internal capillary tube with a narrower end (2c), projected in favor of the flow, in which atmospheric air is introduced by a centrifuge pump with the function of disrupting superficial tension of the foam (2c).The curved end of the tube allows a perfect fit to a hose (2e) for the transportation of effluent to the third module (3) (reservoir) that is connected to a pulse pump (10).This projects the liquid back to the reactional system through the entry (9) as recycling, completing the cycle of configuration of the reactor.

Experimental design for the degradation of the dairy effluent
In order to optimize the parameters to be tested, a 2 4-1 fraction type of factorial statistical design was carried out, randomly, in duplicate with 3 central points, being the response variable the percentage of reduction of total organic carbon (TOCred) after 30 min of reaction.The independent variables (factors) proposed for this stage were ozonator power, O 2 flow rate of the cylinder, Fe 2+ concentrations and pH.Table 1 shows the variables with respective evaluated levels.
Estimate values of ozonator power were determined by experimental measurement using a power and energy consumption measuring equipment (ICEL, model ME-2500, 220 V and 60 Hz, conformity certifying number 201111011512).Feeding the ozone generator was carried out by using oxygen gas of high nominal purity of 99.50% and a valve attached to the cylinder with an adjustment from 0 to 2 L min -1 .Fe 2+ concentrations were 0.5, 0.75 and 1.0 g L -1 , corresponding to 5.03, 7.55 and 10.06 g of FeSO 4 •7H 2 O with 99.0% of purity for 2 L of effluent.
After evaluating better experimental conditions, the physically and chemically treated product was analyzed to verify if the quality was in accordance with discharge limits to hydric bodies or if it should undergo a biological treatment as activated sludge treatment.

Analytic characterization of effluent in natura
Table 2 shows the results from the physicochemical characterization of the effluent in natura of the studied industry, along with data from pertinent legislation of discharge in hydric bodies: Article 18 of Environmental Control Agency of São Paulo State (CETESB), 25 an environmental agency in the State of São Paulo, Brazil, Resolution 357/05 of National Council of Environment (CONAMA) 26 and CONAMA Resolution 430/11 that complements and changes 357/05. 27ccording to Table 2, there is a great variation in the analyzed parameters, which can be explained by different cleaning processes, physical and chemical characteristics of raw materials and processing type of products.The variation is related to the collections made during the experimental period, not only in one day.
The dairy effluent is characterized as a white color type of effluent due to the physicochemical characteristics.The cleaning system "clean in place" (CIP), in which acidic and alkaline solutions are added, carries this cleaning water to the treatment plant resulting in a pH that can range from 2.0 to 12.0 in the homogenizer tank. 30ases produced by the fermentation of anaerobic bacteria cause the odor; suspended solids and/or emulsifiers cause turbidity.Suspended solids are derived from milk coagulants.The loss of fat from milk that corresponds to 90% of total fat material of the effluent, 31 also contributes to the high concentration of solids.
The presence of nitrogen is related to high content of proteins but since milk is diluted in the cleaning process, this value is found under the limits of legislation concerning discharge in hydric bodies.
3][34] High values of COD are due to substances present in the milk; spills, leaks, deficient operations of equipment, overload of tanks and loss in the process also contribute to the increase of concentration of the organic matter of the effluent. 353][34] As in BOD, only the biodegradable fraction can be measured, the higher this value is close to COD, the easier the effluent can be biodegrated. 36,37aw liquid effluents of dairy industries show mean values of BOD 5 /COD in the range of 0.50 to 0.70. 4 The ratio BOD 5 /COD has been used by several researchers to express biodegradability of effluents of environmental relevance, being used as a parameter in the selection of type of effluent treatment.According to results from raw effluents, the mean ratio of BOD 5 /COD was 0.85.After the treatment by catalytic ozonation, this ratio reached the value of 1.27.The increase in BOD 5 /COD obtained suggests that the chemical process may be used as a pre-treatment in order to improve biodegrability of the effluent.According to Balcioglu and Ötker 23 the most biodegradable fraction of an effluent may be increased by ozonation, since it leads to formation of low molecular weight by-products that are readily biodegraded.
Thus, the use of biological processes for this type of effluent can be justified, something that had already been taken into account by the company.Although activated sludge process is widely used, it can be observed that in several visits to the company there was a strong and characteristic odor, high sludge production, consumption of nutrients and hydraulic retention time in the effluent treatment.Besides all of the above mentioned, biological processes have some limitations, as a narrow range of optimal conditions (pH, temperature, nutrient concentration, residence time).Among several other restrictions, it can be cited the inability of biological processes to metabolize recalcitrant compounds. 38everal studies have been conducted in order to develop technologies that minimize industrial effluents volume and toxicity, aiming to allow not only the removal of contaminant substances, but also the complete mineralization.Toxicity associated with industrial effluents may be intimately associated with the presence of recalcitrant or refractory compounds. 39calcitrant or refractory compounds are not biodegraded by organisms that normally exist in treatment of biological systems, considering hydraulic retention times usually applied.However, they are frequently disposed on aquatic bodies. 11n this situation, chemical processes appear as a promising alternative for degradation of these pollutant chemical species. 40or the parameter oils and fat, it can be observed that its concentration is high due to the content of fat present in the raw effluent from processed raw material and the process adopted by industry. 31

Evaluation of experimental matrix and TOCred in the design of semi-batch process with recycle
With the purpose of increasing efficiency in the ozonation reactions, especially in the removal of organic matter with ozone consumption inferior to the direct ozonation, the catalytic process was used.Some considerations have to be taken into account, however, as for the employment of catalytic process: (i) solubility of catalyst in the reactional media; (ii) difficulty in reusing the catalyst used; (iii) need of using ion removal techniques due to its toxic characteristics and/or unwanted adverse effects. 11errous ions were selected as catalysts for this work.Therefore, there is no definition of the type of catalysis of the process because in acidic pH, ferrous ions are soluble (homogeneous), but in pH 7 and 10, there is a formation of a product less soluble with Fe 2+ ion (heterogeneous).The maximum concentration of Fe 2+ in this work was 1.0 g L -1 , because in higher concentrations it was verified a little reduction on the percentage of TOC reduction (TOCred).
The maximum power of the ozonator in this work was 35 W.There was a notorious and characteristic odor of ozone when pure oxygen from the cylinder was used.This could be determined qualitatively by O 3 indicator paper that darkened quickly under power higher than 35 W. Maximum and minimum flow rates were inverted during the treatment, being the flow rate at the low level of 0.5 L min -1 and high level of 0.125 L min -1 , which does not compromise the statistical evaluation of the process.To this extent, atmospheric air was injected through a silicon hose attached to the tube containing a capillary tube, enabling the disruption of superficial tension of the foam.In all experiments and at the first 10 min of oxidation reaction, there was an intense formation of foam, possibly based on the fast degradation and respective mineralization of effluent to CO 2 .Some different causes may be considered to explain this foam formation.The wastewater may contain various substances used for cleaning equipment and factory items, such as alkaline phosphates, acids, surfactants and complexing compounds. 41The presence of large quantities of O 2 + O 3 could also enhance the foam formation.Furthermore, the foam formation could be favored by the fast degradation and respective mineralization of effluent to CO 2 .However, hypotheses investigation were not performed in this work.
Table 3 presents the percentage of TOCred and mass ratio of O 3 by oxidized carbon.
Overall, it was observed that replicates showed low deviation, which indicates good repeatability and good control of random errors.
An analysis based on results showed that the best average for TOCred (61.29%) was obtained with experiments in duplicate 1 and 19, suggesting that the ozonator power (A) must work at its highest level (35 W), O 2 flow rate (B) must be at 0.125 L min -1 , Fe 2+ concentration (C) in 0.5 g L -1 of effluent and pH (D) 4.0.
Due to the ozonator power and O 2 flow rate, the average ozone mass rate could be determined for the experimental design, being the O 3 mass rate of duplicates 1 and 19 (3.85 mg min -1 ).The best result for the O 3 mass relation by carbon degraded (TOC) was 0.09, which indicates that the ozonator power and O 2 flow rate selected were pertinent to the proposed process.Considering economical and safety aspects, it is important because O 3 in excess can compromise the process.It is not possible just through the analysis of the results of Table 3 to evaluate and optimize the best experimental conditions.However, all of the TOCred reduction responses were assessed for statistical analysis generated by Minitab 16.
First, the graphic of probabilities of percentage reduction of each output variable (TOCred) was generated, to verify the normality of experimental data of this planning.Figure 2 shows the graph of normal probability in terms of percentage of TOCred.The points in the figure correspond to the 19 experiments of the exploratory factorial matrix fractional 2 4-1 , conducted in duplicate with three central points.
It can be seen in Figure 2 that the graph does not follow a normal distribution with a p-value of 0.033.Even though data is aligned, 95% of data points are within the limit of experimental error for a 95% level, showing confidence to experiments.The various effects of variables may be seen in Figure 3 and Table 4 (analysis of variance, ANOVA) and in Figure 4 (Pareto chart).
Figure 3 shows the main effects of parameters for percentage of TOCred.It can be seen that for the highest percentage of TOCred, the best condition is found with factors A, B and C adjusted in the central point (33 W; 0.25 L min -1 and 0.75 g Fe 2+ L -1 respectively), while D factor was at low level (pH 4.0).The analysis of variance consists in a group of statistical models such as least squares for the evaluation of total sample variance that can be attributed to different factors of experimental error.Figure 3 shows that the best and highest percentage of TOCred is working with the power flow, and [Fe 2+ ] at the central point, or 33W power, the O 2 flow rate 0.25 L min -1 and [Fe 2+ ] 0.75 g L -1 effluent, obtaining reductions getting TOCred above 56%.
Table 4 presents factors of analysis of variance (ANOVA) involved in the catalytic ozonation process.
Through ANOVA (Table 4), it can be observed that all studied variables were significant on the studied process with p-value, which is approximately zero (since we showed only three digits on the table, it appears as being "0.000").A p-value ranges from 0 to 1.The p-value is calculated from the observed sample and represents the probability of incorrectly rejecting a null hypothesis that is true indeed (type I error).Minitab 16 automatically displays p-values for most hypothesis tests.These results were also confirmed by a Pareto diagram with 95% confidence, presented in Figure 4.
To the variable response TOCred, the factor power (A) was more significant than the interaction power-iron (AC) which indicated that the direct ozonation is more reoccurring.
Based on the design of experiment used, a statistical model was proposed in order to predict the behavior of the catalytic ozonation process.Equation 20 represents the model obtained (considering ANOVA in The mathematical model obtained does not demonstrate lack of adjustment and presents an R 2 of 99.73% of variations around the mean.Therefore, the optimization of variables can be found at power (35 W), O 2 flow rate (0.125 L min -1 ), Fe 2+ concentration (1.0 g L -1 ) and pH (4.0), with 64.26% of TOCred.
The best experiment with variables at the optimized level were carried out obtaining a mean value (triplicate) of 64.03% in the TOCred, which is close to the theoretical one, 64.26%.

Economic evaluation of semi-batch process with recycle
Estimate values of energetic consumption of equipment were determined by experimental measurement considering the time of 30 min.Table 5 shows the estimated cost of experiments for the experimental planning used.
Overall, Table 5 shows that the best result of the process is also the one with lower cost/TOCred ratio.As it can be observed in experiments 1 and 19 of design 2 4-1 , in which the mean was 61.29% of degradation of TOCred concentration and its cost/TOCred ratio (0.50), the ratio is the lowest of all the other experiments.
According to the statistical modelling, the optimization of variables shows 64.03% of TOCred with a cost of US$ 0.03644 L -1 of effluent, in which the change happens only with the Fe 2+ concentration factor.Even with the change in ferrous concentration, the estimate is relatively the same (US$ 0.03 L -1 of treated effluent), being the best cost/TOCred ratio of the catalytic ozonation process the one in which optimized variable conditions were ozonator power (35 W), O 2 flow rate (0.125 L min -1 ), Fe 2+ concentration (1.0 g L -1 ) and pH (4.0).
Loures et al. 44 carried out a study of the application of photo-Fenton process followed by a biological treatment in and TOCred were 91, 74 and 91%, respectively.The total estimate cost of photo-Fenton treatment was US$ 0.0893 L -1 of effluent with a cost/TOCred ratio of 1.62.Salazar and Izário Filho 45 studied the heterogeneous photocatalytic process to reduce the organic matter of dairy industry effluent.Titanium dioxide (TiO 2 ) was used as the semiconductor with solar radiation.Best conditions of the treatment step having percentage of reduction of COD were: time of reaction of 180 min, TiO 2 anatase and pH 5.0 obtaining 54.77% of COD reduction.
In addition, Samanamud et al. 46 studied the hetereogeneous photocatalytic process on dairy effluent treatment.In this study, ZnO was employed as the semiconductor using solar radiation.Degradation was measured in terms of percentage of TOCred, the highest efficiency was found at pH 8.0 with a TOCred of 14.23% of degradation with a reaction time of 180 min.The photocatalytic treatment had an estimate cost of US$ 0.0055 L -1 of effluent, but with a cost/TOCred ratio of 0.39.
It can be observed that in the heterogeneous catalysis with UV radiation, the reaction time to obtain a significant reduction of organic content is higher when compared to homogeneous catalytic processes.

Figure 3 .
Figure 3. Main effects for percentage of TOCred variation measurements.

Table 1 .
Factors and levels for the treatment of 2 L of the effluent during 30 minutes

Table 2 .
Physical and chemical characteristics of the effluent in natura -: Not specified; LOQ: limit of quantification; COD: chemical oxygen demand; BOD 5 : biochemical oxygen demand after five days incubation; TOC: total organic carbon; N-NH 3 : ammoniacal nitrogen; N-organic: organic nitrogen; ST: total solids; STF: fixed total solids; STV: volatile total solids; surfactant: extracted with chloroform and methylene blue in a separation funnel and determined by visible spectrophotometry at 655 nm; oil and grease: determined by solvent extraction method (Soxhlet method); SM: standard methods; ICP-OES: inductively coupled plasma optical emission spectrometry; MBAS: methylene blue active substances.Vol.26, No. 7, 2015

Table 4 )
, having the response factor percentage of TOCred.

Table 5 .
Values of energy consumption and reagents per liter of dairy effluent treated by catalytic ozonation process of duplicate and triplicate experiments dairy effluents.Best experimental conditions had percentage of TOCred as response variable and they were: temperature at 30 ºC; pH 5.0; concentration of peroxide of 0.255 mol L -1 ; Fe 2+ concentration of 0.0108 mol L -1 ; UV irradiation of 15 W and time of 60 min.Under these conditions, they obtained 54.96% of TOCred.Results showed high efficiency with biological treatment after the AOP where COD, BOD 5 US$ 1.00 = R$ 2.576;421 kWh = US$ 0.11837.43