UV analysis

Journal of Environmental Management 167 (2016) 206e213

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Research article

Removal of Mefenamic acid from aqueous solutions by oxidative process: Optimization through experimental design and HPLC/UV analysis Renata Colombo a, b, *, Tanare C.R. Ferreira b, Renato A. Ferreira b, Marcos R.V. Lanza b a b

~o Paulo, 03828-000, Sa ~o Paulo, SP, Brazil Escola de Artes, Ci^ encias e Humanidades, Universidade de Sa ~o Carlos, Universidade de Sa ~o Paulo, Caixa Postal 780, 13560-970, Sa ~o Carlos, SP, Brazil Instituto de Química de Sa

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 August 2015 Received in revised form 9 November 2015 Accepted 13 November 2015

Mefenamic acid (MEF) is a non-steroidal anti-inflammatory drug indicated for relief of mild to moderate pain, and for the treatment of primary dysmenorrhea. The presence of MEF in raw and sewage waters has been detected worldwide at concentrations exceeding the predicted no-effect concentration. In this study, using experimental designs, different oxidative processes (H2O2, H2O2/UV, fenton and Photofenton) were simultaneously evaluated for MEF degradation efficiency. The influence and interaction effects of the most important variables in the oxidative process (concentration and addition mode of hydrogen peroxide, concentration and type of catalyst, pH, reaction period and presence/absence of light) were investigated. The parameters were determined based on the maximum efficiency to save time and minimize the consumption of reagents. According to the results, the photo-Fenton process is the best procedure to remove the drug from water. A reaction mixture containing 1.005 mmol L
1 of ferrioxalate and 17.5 mmol L
1 of hydrogen peroxide, added at the initial reaction period, pH of 6.1 and 60 min of degradation indicated the most efficient degradation, promoting 95% of MEF removal. The development and validation of a rapid and efficient qualitative and quantitative HPLC/UV methodology for detecting this pollutant in aqueous solution is also reported. The method can be applied in water quality control that is generated and/or treated in municipal or industrial wastewater treatment plants. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Advanced oxidative processes Photo-fenton reaction Experimental designs Mefenamic acid Anti-inflammatory

1. Introduction There are numerous potentially hazardous chemicals and new substances constantly being developed and released into the environment. In recent years, research on water pollution has increasingly focused on a new group of pollutants known as emerging contaminants (Bolong et al., 2009). Pharmaceutically active substances are an important group of emerging contaminants. Their extended use and the incomplete elimination in wastewater treatment plants have resulted in these contaminants, in their native form or as metabolites, continuously introduced into sewage waters, mainly through excreta, disposal of unused or expired drugs, or directly from pharmaceutical discharges (Gros et al., 2007; Bueno et al., 2009 and Verlicchi et al., 2012). This

^ncias e Humanidades, Universidade * Corresponding author. Escola de Artes, Cie ~o Paulo, 03828-000, S~ de Sa ao Paulo, SP, Brazil. E-mail address: [email protected] (R. Colombo). http://dx.doi.org/10.1016/j.jenvman.2015.11.029 0301-4797/© 2015 Elsevier Ltd. All rights reserved.

fact, together with the reported evidence of harmful effects produced in the natural ecosystems (Bringolf et al., 2010 and Fent et al., 2006), has recently prompted calls for some pharmaceuticals to be included in the European Union's priority list of pollutants (European Commission, 2013). Mefenamic acid, IUPAC name 2-(2,3-dimethylphenyl)aminobenzoic acid, is a non-steroidal anti-inflammatory drug (NSAID), indicated for relief of mild to moderate pain, and for the treatment of primary dysmenorrhea (Rawlinson and Davis, 1983, Patient, 2014). The presence of mefenamic acid has been detected in sewage effluents as well as in receiving waters in different countries (Gros et al., 2007; Araujo et al., 2011; Tauxe-Wuersch et al., 2005; Jone et al., 2006 and Soulet et al., 2002). In treated effluent, mefenamic acid has been detected at concentrations exceeding the predicted no-effect concentration of 0.43 mg/L (Fent et al., 2006, Werner et al., 2005). Studies have pointed out that mefenamic acid was persistent in wastewater effluents after municipal wastewater treatment (Tauxe-Wuersch et al., 2005). The Advanced Oxidation Processes (AOP) have shown to be promising for

R. Colombo et al. / Journal of Environmental Management 167 (2016) 206e213

degrading this pharmaceutical compound. The ozonation and O3/ UV degradation processes are the most commonly studied evaluation techniques (Gimeno et al., 2010; Gerrity and Snyder, 2011; Chang et al., 2012). The combination of ozone and UV radiation led to the best results, which exhibited MEF removal percentage of 60e80%. The use of UV/H2O2, Fenton and photo-Fenton in MEF degradation is also described, focusing on the degradation of this compound in municipal wastewater treatment plant effluents and in the presence of over 31 compounds (De la Cruz et al., 2012). In this study univariate analysis were carried out evaluating the H2O2 and catalyst concentration parameters and in a limited range (10, 25 and 50 mg L
1 of H2O2 and 5 mg L
1 added or 1.48 mg L
1 of total iron already present). Due to the fact that an oxidative process has many potentially important variables for its effectiveness, the evaluation of all these parameters using a univariate analysis requires many experiments, and the lack of correlation between all the variables can result in inaccurate results. Optimization strategies that use chemometric tools are generally more efficient and economical than the univariate analysis, since they allow studying the interactions between two or more variables in order to assess to what degree this interaction effects all other parameters in the system (Bruns et al., 2006; Rozaca et al., 2010 and Arslan-Alaton et al., 2010). This evaluation is important to define a process that is efficient, faster, simpler and at lower cost, enabling its application in municipal water and wastewater treatment. Multivariate analysis evaluating the influence and interaction effects of the most important variables in the oxidative process not been reported yet for MEF and it was applied in this study in this context. The fractional factorial design was proposed to simultaneously evaluate the different types of AOPs, as well as the influence and interaction effects of the seven most important variables in the oxidative process. The response surface methodology was developed to establish the best concentrations of hydrogen peroxide and catalyst (ferrous ions) and to provide an optimal region for more efficient mefenamic acid degradation. The fractional factorial designs were chosen because these types of designs are widely used in experiments involving several factors. This is also usually performed early in a response surface study, since it allows to identify, with a limited number of experiments, which factors have a significant effect and which can be considered as having little or no effect on the response (Montgomery, 1984; Torrades et al., 2003; San Miguel et al., 2014 and Zakrzewska et al., 2014).

2.2. Experimental procedures All experiments were carried out using a recirculation flowthrough UV photoreactor (Fig. 1), coupled to a thermostatically  ~o Paulo, SP, Brazil) and irradiated controlled bath (Nova Etica, Sa using a Philips (Amsterdam, The Netherlands) 15 W UVC lamp (l max 254 nm; maximum incident photon flux 3.33.10
13 photons.s
1). In each experiment, the reactor was filled with 2 L of solution containing 10 mg L
1 mefenamic acid (prepared using the generic medicament) and operated at a flow rate of 150 L h
1 ~o [determined using a rotameter model AP300SS from Applitech (Sa Paulo, SP, Brazil)].

2.3. Hydrolysis and photolysis assay Based on the literature data which reports that mefenamic acid is photodegraded when exposed to natural sunlight for long periods of time (Werner et al., 2005), preliminary experiments were performed in order to establish the influence of hydrolysis and photolysis process on the degradation of mefenamic acid. Hydrolysis was performed in the absence of hydrogen peroxide, ferrous ions and light and at original pH of reaction mixture (pH 6.1). Photolysis experiments were performed in the absence of hydrogen peroxide and ferrous ions, irradiating the reaction mixture, at its original pH of 6.1, with UV light (l max 254 nm). The kinetics of both the hydrolytic and photolytic reactions were performed determining the amounts of mefenamic acid that remained in aliquots of the reaction mixture, which were collected at adequate reaction times with a reaction of 180 min and analyzed by HPLC.

2. Experimental 2.1. Chemicals Mefenamic acid (analytical standard;  99% pure) was obtained from SigmaeAldrich (St. Louis, MO, USA; product number 31058), while a generic medicament (Ponstan®; Medley, S~ ao Paulo, SP, Brazil) containing 500 mg of mefenamic acid was purchased from a local drug store. Purified water (resistivity 18.2 MU) was prepared using a Millipore (Eschborn, Germany) Milli-Q water purification system. All other reagents were of analytical grade unless otherwise stated. Methanol, acetonitrile and sulphuric acid were obtained from Mallinckrodt (Xalostoc, Edomex., Mexico), formic acid (reagent grade), sodium sulphite was from Merck (Darmstadt, Germany), sodium oxalate and a 30% (w/w) solution of hydrogen ~o Paulo, SP, Brazil), peroxide (reagent grade) were from Ecibra (Sa ferric nitrate nonahydrate was from Químis (Diadema, SP, Brazil), and ferrous sulphate heptahydrate was from Synthy (Diadema, SP, Brazil). Purified water (resistivity 18.2 MU cm) was prepared using a Millipore (Eschborn, Germany) Milli-Q water purification system.

207

Fig. 1. Scheme of recirculation flow-through UV photoreactor.

208

R. Colombo et al. / Journal of Environmental Management 167 (2016) 206e213

2.4. Oxidative processes evaluation Four different oxidative processes were evaluated in terms of efficiency of MEF degradation and mineralization of the organic compounds (present or eventually formed in the reaction mixture during the process). The processes chosen were those widely applied for the degradation of residual pharmaceuticals (H2O2, H2O2/UV, fenton and Photo-fenton) (Wols and Hofman-Caris, 2012). The influence of variables: concentration and addition mode of hydrogen peroxide, concentration and type of catalyst (iron ions), pH, reaction period and presence/absence of light were initially studied in the two extreme levels (low and high). Since there is no specific information in the literature concerning the behavior/interaction of these variable in oxidative processes of mefenamic acid, the concentration range of catalyst (0e2 mmol L
1 of iron ions) and hydrogen peroxide (0.1e10 mmol L
1) was chosen based on a published review of residual pharmaceuticals degradation by Advanced Oxidative Process (Melo et al., 2009) or/and concepts of AOP, such as: i) in the presence of excess oxidant, the hydroxyl radicals can form a radical which is less reactive than OH (Haber and Weiss, 1934 and Herney-Ramirez et al., 2010) or can to react with this less reactive radical lowering the reaction rate and ii) in the photo-Fenton reaction, the excess of catalyst form a large amounts of metal-containing sludge at the end of the process. This sludge does not only cause high environmental impacts, implying additional associated costs, but also represent the loss of significant quantities of catalytic metals (Haber and Weiss, 1934). Additionally, various competitive reactions are involved in the complex mechanism, thereby an optimized level of metal is very important. To obtain the hydrogen peroxide concentrations studied, aliquots of hydrogen peroxide solution (30 mg mL
1) were added in different modes. In the “total” addition mode the reaction was performed adding total hydrogen peroxide at the initial reaction period. In the “partial” addition mode the reaction was performed adding hydrogen peroxide gradually in the reaction mixture during the first 5 min of the reaction period (for experiments of 10 min reaction period) and in the first 30 min (for experiments of 60 min reaction time). The reaction mixture was studied at natural pH of 6.1 (to save time and reduce costs with process adjustments) and pH adjusted to 2.5, based on a literature report claiming that the optimum pH interval for the formation of hydroxyl radicals in Fenton processes is between 2.5 and 3.0 (Ying-Shiha et al., 2010; Nogueira and Guimar~ aes, 2000 and Andreozzi et al., 1999). The evaluation was carried out simultaneously using a 27
3 fractional factorial design, elaborated by Minitab® software (MinitabInc, USA), version 13. The fractional factorial design, composed by 16 experiments, was performed in triplicate and randomized to eliminate environmental variations. The degradation of mefenamic acid and mineralization of organic compounds present in the reaction solution were used as response to the experimental model. The variables and their respective levels used in each experiment are summarized in Table 1. After defining the most efficient oxidative process in the degradation of mefenamic acid and subsequently defining the most adequate parameters for this process, the two variables, catalyst concentrations (iron ions) and hydrogen peroxide, were more carefully evaluated using response surface methodology (performed by Minitab®, version 13). In this design two levels were used for each factor (5e30 mmol L
1 hydrogen peroxide and 0.01e2.0 mmol L
1 ferrioxalate). These levels were chosen according to the range specified in the previous 27
3 fractional factorial design (hydrogen peroxide concentration close to 10 mmol L
1 and catalyst concentration between 0 and 2 mmol L
1). Table 2 shows the experimental conditions of

response surface design. The results of the experiment (Table 2) were analyzed by the response surface regression (RSREG) procedure to fit a quadratic polynomial model (using the software MINITAB). The polynomial model for the MEF degradation was given in terms of coded factors, according to Equation (1).

y1 ¼ b0 þ b1 x1 þ b2 x2 þ b11 x21 þ b22 x22 þ b12 x1 x2

(1)

where y1 is the response (% MEF degradation), x1 and x2 are the coded process values for hydrogen peroxide and iron ions concentrations, respectively, and b is the constant coefficients. The mathematical models were validated by analysis of variance (ANOVA) at the 95% confidence level. After obtaining the statistical analysis, response surfaces graphs were constructed in order to establish the conditions required for the most desired factors. In all experiments, two design sample aliquots of the reaction mixture were collected at appropriate times (initial and final) and sodium sulphite (0.1 g) was added immediately to each sample to quench the peroxide and stop the reaction. Quenched samples were filtered through 0.45 mm cartridges and the levels of mefenamic acid analyzed by HPLC and mineralization determined by TOC. 2.5. Analytical methods The quantitative determination of mefenamic acid was carried out by High-Performance Liquid Chromatography (HPLC-UV/DAD) technique. A specific methodology was developed using a Shimadzu (Kyoto, Japan) Prominence LC 20 AT modular system comprising two CBM-20 A pumps, a CTO-10AS oven, a SIL 20A autosampler, a SPD-M20A Diode Array Detector and a LC-10 Workstation Class data processor. Separations were performed on a Pursuit (Varian, Walnut Creek, USA) C-18 column (250  4.6 mm i.d.; 5 mm), protected by a Pursuit C-18 guard column (20  4.6 mm i.d.; 5 mm). The elution was performed with mixtures of 0.5% formic acid in water (solvent A) and acetonitrile (solvent B) according to the program: 0e15 min: 45e100% B (linear gradient). The oven temperature was 25  C, the flow rate was 1.0 mL min
1, the injection volume was 20 mL (Rheodyne loop); and UV detection at 280 nm. The method developed was validated following the parameters specificity, linearity, range, accuracy, precision, limit of detection (LOD), limit of quantification (LOQ) and robustness, according to the ICH guideline (ICH, 2014). Mineralization of organic compounds was analyzed using a TOC analyzer (Shimadzu TOC-VCPN analyzer). 3. Results and discussion 3.1. Analytical parameters HPLC methodology was developed and validated for the degradation of mefenamic acid and determination of the efficiency of the oxidative processes studied. The specificity was confirmed using the data derived from the Diode Array Detector (DAD) in association with the software resource provided by the workstation. For the linearity and range assay the external standard method was utilized and an analytical curve was constructed to analyze, in triplicate, eight different concentrations of mefenamic acid in the range of 0.100e100.98 mg L
1. The curve exhibited linearity in the range chosen with a regression coefficient (R) of 0.9999805 and regression equation y ¼ 44371.25x þ 4933.091. The accuracy was evaluated by comparing the results obtained from the analysis of generic medicament solutions containing three different concentrations of mefenamic acid. The different concentrations were within the range, the analysis was performed in triplicate and the average recoveries of mefenamic acid of each of the three

R. Colombo et al. / Journal of Environmental Management 167 (2016) 206e213

209

Table 1 27
3 fractional factorial design array indicating the real values of the variables and the response functions MEF degradation and TOC removal. Experimental Oxidation run process

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Fenton H2O2/UV H2O2/UV H2O2/UV H2O2 photo-Fenton (Feþ2) photo-Fenton (Feþ3) Fenton H2O2 H2O2/UV H2O2 Fenton H2O2/UV H2O2 photo-Fenton (Feþ3) photo-Fenton (Feþ2)

Variables (factor and levels)

Response

H2O2 addition H2O2 concentration mode (mmol L
1)

Iron ions concentration (mmol L
1)

Iron ions type

pH Reaction period

Light UV Degradation of MEF (%)

Removal of TOC (%)

total total partial total total total

0.1 10.0 10.0 0.1 10.0 0.1

2 0 0 0 0 2

Feþ3 e e e e Feþ2

2.5 2.5 6.0 2.5 6.0 6.0

60 60 10 10 10 10

absence presence presence presence absence presence

14.24 41.47 16.43 14.91 16.76 11.14

16.09 14.01 4.38 2.29 0.73 1.00

partial

10.0

2

Feþ3

2.5 10

presence

9.55

29.48

þ2

absence absence absence absence absence presence absence presence

30.19 11.02 0.00 35.51 22.35 22.09 4.48 54.61

1.70 0.10 0.64 15.27 0.86 15.68 2.31 47.87

presence 18.51

31.29

partial total partial partial total partial partial total

10.0 0.1 0.1 10.0 10.0 0.1 0.1 10.0

2 0 2 0 2 0 0 2

Fe e Feþ3 e Feþ2 e e Feþ3

6.0 6.0 6.0 2.5 2.5 6.0 2.5 6.0

60 60 10 60 10 60 10 60

partial

0.1

2

Feþ2

2.5 60

Table 2 Surface response design array indicating the real values of the variables and the response function MEF degradation. [Experimental conditions: photo-Fenton process; mefenamic acid 10 mg L
1; addition of hydrogen peroxide at total mode; ferric ion as catalyst; pH 6.1; presence of light (l max 254 nm) and 60 min of reaction time]. Experimental run

Levels of each factor

Response
1

Hydrogen peroxide (mmol L 1 2 3 4 5 6 7 8 9 10 11 12 13

)

17.5000 17.5000 17.5000 5.0000 17.5000 30.0000 0.1777 17.5000 17.5000 5.0000 30.0000 35.1777 17.5000

concentrations analyzed were 99.62, 99.89 and 99.58%. The precision was evaluated in terms of repeatability and intermediate precision, using a generic medicament solution prepared at the concentration within the range and analyzed several times. The repeatability was performed under the same operating conditions over a short interval of time. Intermediate precision was performed under varying conditions e different days and different analysts. The coefficients of variation (CV) of the results were 0.80% for repeatability and 0.06% for intermediate precision. The limit of detection (LOD) and limit of quantification (LOQ) were calculated using the following expressions: LOD ¼ 3s/b and LOQ ¼ 10s/b, where s is the standard deviation of twenty measures of the blank and b is the slope of the analytical curves. The limit of detection and quantification determined for the method were 0.001 mg L
1 and 0.003 mg L
1. These limits were used to determined the initial concentration of mefenamic acid (10 mg L
1) in the reaction solution used in the oxidative study. The robustness of the analytical procedure was evaluated by small, but deliberate variations in the method parameters, hence indicating it is reliability during normal usage.


1

Ferrioxalate (mmol L

)

1.00500 0.40214 1.00500 0.01000 2.41214 2.00000 1.00500 1.00500 1.00500 2.00000 0.01000 1.00500 1.00500

Degradation of mefenamic acid (%) 97.01 66.10 95.20 56.84 72.21 56.10 44.63 96.55 94.57 56.33 58.00 68.42 96.40

3.2. Evaluation for MEF degradation The kinetics study of hydrolytic and photolytic degradation reactions of the mefenamic acid demonstrated that mefenamic acid was not substantially degraded in the presence of water and light (l max 254 nm). After 180 min of reaction, in the hydrolysis process the degradation of the pharmaceutical stabilized at 7.21% and during the photolysis process at 17.03%. The responses obtained in fractional factorial design are summarized in Table 1 and represented in the Pareto (p ¼ 0.05) and main effects (p ¼ 0.05) charts, according to Figs. 2 and 3. The Pareto chart (Fig. 2) shows the results of mefenamic acid degradation, where the length of each bar on the chart and their signs showed that the variables reaction period and hydrogen peroxide concentration were significant at the 95% confidence level. The variables addition mode of hydrogen peroxide, pH, concentration and type of iron ions and presence/absence of light are not considered as significant by the Pareto chart, however the Mean Effect chart (Fig. 3) shows the favorable conditions to obtain the better degradation rate for mefenamic acid. The steep slope of the line that connects the mean responses to the low and high level

210

R. Colombo et al. / Journal of Environmental Management 167 (2016) 206e213

degradation rates. The reaction period was considered a relevant variable in the experimental design probably because of the high percentage difference in the mefenamic acid degradation in the extreme times of 10 min (where small degradation percentage occurred) and 60 min. Then, a kinetic study of mefenamic acid degradation was performed to establish the best time to be used in the next step. Using the best condition established by the fractional factorial design (experiment 15, Table 1) the degradation of mefenamic acid was evaluated at different times. The order of reaction was determined from the decay curve of MEF concentration (Fig. 4) by application of the Equation (2).

Fig. 2. Pareto chart for visualizing the effects of the chemical variables on the mefenamic acid degradation using 27
3 fractional factorial design.

n½MEF ¼ ln½MEF0
kt

(2) 2

On the basis of the observed linearity of the plots (R ¼ 0.98), it

Fig. 3. Mean Effect chart for visualizing the level of the chemical variables in the mefenamic acid degradation using 27
3 fractional factorial design.

was assumed that the reaction kinetics was pseudo-first order and that the slope represented the rate constant (k ¼ 0.043 min
1) for the removal of MEF. Considering that the goal of this study is the total degradation of mefenamic acid, even with the small

-3 -3.5 -4 -4.5

ln [Mef]

indicated the best rate by adding hydrogen peroxide in the total mode, in the presence of light (l max 254 nm) and using a longer reaction period (60 min). The significance of the hydrogen peroxide variable is in agreement with the literature-based evidence that the oxidant (hydrogen peroxide) addition approach is an important issue to consider when optimizing the organic degradation (Herney-Ramirez et al., 2010). However considering the influence and interaction effects of the variables using the multivariate optimization showed that adding hydrogen peroxide in the total mode helps the degradation rate for MEF, unlike the literature result which showed the need to add hydrogen peroxide divided into doses to prevent scavenging, which often occurs at excessive hydrogen peroxide doses. The variation in the natural pH from 6.1 to 2.5 does not considerably contribute to improve the degradation of MEF, showing better response when using natural pH. The gentle slope of the line that connects the mean responses to ferric and ferrous ions shows that both represent almost the same rate of degradation behavior, with a slight increase using ferric ions. Although the precipitation of ferric ions in the form of insoluble hydroxide occurs at pH values above 3, the use of ferrioxalate [a polycarboxylate complex of Fe (III)] stabilized the ferric ions enabling the photo-Fenton process in a wide pH range (Katsumata n, 2009; Me ndez-Arriaga et al., et al., 2010; Prato-Garcia and Buitro 2010 and Melo et al., 2009). The Mean Effect chart also shows that the use of high concentrations of hydrogen peroxide (close to 10 mmol L
1) and low concentrations of ferric ion provided higher

-5 -5.5 -6 -6.5 0

10

20

30

40

50

60

70

Time (min) Fig. 4. Decrease in mefenamic acid concentration during photo-Fenton process. [Reaction mixture: mefenamic acid 10 mg L
1; ferrioxalate 2 mmol L
1; hydrogen peroxide 10 mmol L
1 (attained by total addition); pH 6.0 and presence of light (l max 254 nm)].

R. Colombo et al. / Journal of Environmental Management 167 (2016) 206e213

degradation percentage obtained after 40 min, an appropriate monitoring of the reaction up to 60 min of reaction was used.

3.3. Optimization of oxidant and catalyst concentrations to MEF degradation In the present study, after determining the best process (photoFenton) and some of its best parameters for the efficient degradation of mefenamic acid, the ideal hydrogen peroxide and ferric ion concentrations were carefully studied by response surface methodology. In this design 13 photo-Fenton experiments were performed with five combinations of ferrioxalate and hydrogen peroxide concentrations, including quintuplicate determinations of the central points (which correspond to 17.5 mmol L
1 of hydrogen peroxide and 1.005 mmol L
1 of ferrioxalate), for statistical validity. Table 2 shows the experimental conditions used. The Adjusted R Square value of 90e99% and p-value (Significance F) close to zero shows that the model is a good fit for the data and confirms the significance of the quadratic coefficient (Bruns et al., 2006). The very low p-value (0.009) obtained in the analysis of variance of the present data showed that the quadratic polynomial model has a statistical significance and represents the relationship between the response and the significant variables. The results also revealed that the concentrations of hydrogen peroxide (x1) and iron ions (x2) were important variables for the process (p < 0.05). Besides the analysis of variance, it was also determined how well the model fits the data through of R-squared (R2) statistic. The 0.95 value to determine the coefficient (R2) showed that 95.0% of the MEF degradation was attributed to the independent variables (95.0% of the model adequacy). The Rsquared adjusted for the number of predictors in the model (adjusted R-squared or Adj R2) was 0.915, which suggested that there were excellent correlations between the model and the independent variables. The entire relationships between reaction factors and response could be better understood by examining the contour plots (Fig. 5). In the Contour plot, it was observed that the best conditions (those that promote the degradation of about 95%) were located in the central point of the graph, demonstrating the adequate levels chosen. The results found in the contour graph also indicated that the most advantageous combination for the degradation of mefenamic acid is located at the center point, that is, at concentrations of 17.5 mmol L
1 hydrogen peroxide and 1.005 mmol L
1 of ferrioxalate ion.

Fig. 5. Contours of the estimated response surface for the mefenamic acid degradation under different hydrogen peroxide and ferrioxalate concentrations. Experimental conditions: mefenamic acid 10 mg L
1; addition of hydrogen peroxide at total mode; ferric ion as catalyst; pH 6.0; presence of light (l max 254 nm) and 60 min reaction time.

211

Once the experimental conditions of the center point were performed in quintuplicate (experiments 1, 3, 8, 9 and 13 of Table 2) the average amount of mefenamic acid degradation under these conditions were determined. The degradation percentages of mefenamic acid in each experiment were 97.01; 95.20; 96.55; 94.57 and 96.40%, respectively, generating an average of 95.95% degradation. 3.4. Oxidative processes evaluation for TOC mineralization In parallel to the investigation regarding the different oxidative processes and their variables for mefenamic acid degradation, the mineralization of organic compounds (initially present or formed during the reaction) in aqueous solution were also evaluated. Table 1 summarizes the responses obtained in each experiment. Fig. 6 shows the Pareto chart for the TOC where the variables presence/absence of light, reaction period, concentration and type of iron ions were considered as those that most influenced mineralization. The Mean Effect chart (Fig. 7) shows that the best mineralization of organic compounds occurs in presence of light and using ferric ions (photo-Fenton process). These conditions coincide with the best conditions for mefenamic acid degradation, however, the addition of more concentrations of ferric ions (close to 2 mmol L
1) have greater influence over the mineralization rate than the degradation of mefenamic acid. The variables addition of hydrogen peroxide and pH are not considered significant; however, the analysis of the Main Effects chart shows that the best conditions for the mineralization of organic compounds occur with partial addition of hydrogen peroxide at pH 2.5. These conditions are contrary to those found as the best conditions for mefenamic acid degradation, due to the presence of organic substances (excipients from generic medicament and/or degradation products formed during the oxidation process) in the reaction mixture. These products are usually polar compounds of lower molecular mass and consequently requires different conditions (high concentration of oxidative reagents and catalyst) to degraded it (Melo et al., 2014; Babic et al., 2013 and Wu et al., 2012). 4. Conclusion Mefenamic acid is efficiently degraded in water using a recirculation flow-through UV photoreactor and photo-Fenton process. The choice of the best oxidative process and its variable was based on the experimental design results (fractional factorial designs and response surface methodology) and this combination proved to be a powerful tool in the optimization of the oxidative process. The

Fig. 6. Pareto chart to view the effects of the chemical variables on the organic compounds mineralization using 27
3 fractional factorial design.

212

R. Colombo et al. / Journal of Environmental Management 167 (2016) 206e213

Fig. 7. Mean Effect chart to view the level of the chemical variables in the organic compound mineralization using 27
3 fractional factorial design.

experimental design indicated that the most efficient degradation of mefenamic acid occurred using photo-Fenton process with a reaction mixture containing 1.005 mmol L
1 of ferrioxalate and 17.5 mmol L
1 of hydrogen peroxide, added totally at the initial reaction period and with pH 6.1. 95.95% degradation of mefenamic acid occurs under these conditions in 60 min. Despite the efficiency of the photo-Fenton reaction in mefenamic acid degradation, TOC removal (mineralization) was not complete because the best conditions for the degradation of mefenamic acid are not the same for the mineralization of other compounds initially present or formed during the process. The use of experimental design for optimization of the photo-Fenton degradation can make a significant contribution to the reduction of the experiment numbers and environmental impact of the process. Acknowledgments The authors wish to thank the Brazilian Funding Institutions CNPq (grant numbers 150772/2010-6 and 473629/2011-9), FAPESP (grant number 2011/10264-0) for financial support and for the provision of fellowships. References Andreozzi, R., Caprio, V., Insola, A., Marotta, R., 1999. Advanced oxidation processes (AOP) for water purification and recovery. Catal. Today 53, 51e59. Araujo, L., Villa, N., Camargo, N., Bustos, M., García, T., Prieto, A.J., 2011. Persistence of gemfibrozil, naproxen and mefenamic acid in natural waters. Environ. Chem. Lett. 9, 13e18. Arslan-Alaton, I., Yalabik, A.B., Olmez-Hanci, T., 2010. Development of experimental design models to predict photo-Fenton oxidation of a commercially important naphthalene sulfonate and its organic carbon content. Chem. Eng. J. 165, 597e606. Babic, S., Perisa, M., Skoric, I., 2013. Photolytic degradation of norfloxacin, enrofloxacin and ciprofloxacin in various aqueous media. Chemosphere 91, 1635e1642. Bolong, N., Ismail, A.F., Salim, M.R., Matsuura, T., 2009. A review of the effects of emerging contaminants in wastewater and options for their removal. Desalination 239, 229e246. Bringolf, R.B., Heltsley, R.M., Newton, T.J., Eads, C.B., Fraley, S.J., Shea, D., Cope, W.G., 2010. Environmental occurrence and reproductive effects of the pharmaceutical fluoxetine in native freshwater mussels. Environ. Toxicol. Chem. 29, 1311e1318. Bruns, R.E., Scarminio, I.S., Barros, N.B., 2006. Statistical Design e Chemometrics. Elsevier, The Netherlands. mez, M.J., Ferna ndez-Alba, A.R., 2009. Bueno, M.J.M., Aguera, A., Hernando, M.D., Go Evaluation of various liquid chromatography-quadruplo-linear ion trap-mass spectrometry operation modes applied to the analysis of organic pollutants in wastewater. J. Chromatogr. A 1216, 5995e6002.

Chang, E.E., Liu, T.Y., Huang, C.P., Liang, C.H., Chiang, P.C., 2012. Degradation of mefenamic acid from aqueous solutions by the ozonation and O3/UV processes. Sep. Purif. Technol. 98, 123e129. nez, J., Esplugas, S., Grandjean, D., Alencastro, L.F., Pulgarín, C., De la Cruz, N., Gime 2012. Degradation of 32 emergent contaminants by UV and neutral photofenton in domestic wastewater effluent previously treated by activated sludge. Water Res. 46, 1947e1957. European Commission, 2013. New Chemicals, Drugs Added to EU water pollution watch list. http://www.no-pills.eu/wp-content/uploads/2013/04/20130704_EUParliament_new-chemicals-pharmaceutic_EN.pdf. Accessed November 2015. Fent, K., Weston, A.A., Caminada, D., 2006. Ecotoxicology of human pharmaceuticals: review. Aquat. Toxicol. 76, 122e159. Gerrity, D., Snyder, S., 2011. Review of ozone for water reuse applications: toxicity, regulations, and trace organic contaminant oxidation. Ozone Sci. Eng. 33, 253e266. Gimeno, O., Rivas, J., Encinas, A., Beltran, F., 2010. Application of advanced oxidation processes to mefenamic acid elimination. World Acad. Sci. Eng. Technol. 4, 895e897. , D., 2007. Wastewater treatment plants as a pathway Gros, M., Petrovic, M., Barcelo for aquatic contamination by pharmaceuticals in the ebro river basin (Northeast Spain). Environ. Toxicol. Chem. 26, 1553e1562. Haber, F., Weiss, J., 1934. The catalytic decomposition of hydrogen peroxide by iron salts. Proc. Roy. Soc. A 134, 332e351. Herney-Ramirez, J., Vicente, M.A., Madeira, L.M., 2010. Heterogeneous photo-Fenton oxidation with pillared clay-based catalysts for wastewater treatment: a review. Appl. Catal. B Environ. 98, 10e26. International Conference on Harmonisation (ICH), 2014. Validation of analytical procedures: text and methodology Q2(R1). http://www.ich.org/products/ guidelines/quality/article/quality-guidelines.html. Accessed November 2015. Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2006. Partitioning behavior of five pharmaceutical compounds to activated sludge and river sediment. Arch. Environ. Contam. Toxicol. 50, 297e305. Katsumata, H., Okada, T., Kaneco, S., Suzuki, T., Ohta, K., 2010. Degradation of fenitrothion by ultrasound/ferrioxalate/UV system. Ultrason. Sonochem. 17, 200e206. , A.G., Bautitz, I.R., Nogueira, R.F.P., 2009. Degradaç~ rmacos Melo, S.A.S., Trovo ao de fa residuais por processos oxidativos avançados. Quim. Nova 32, 188e197. Melo, S.R.O., Homem-de-Mello, M., Silveira, D., Simeoni, L.A., 2014. Advice on degradation products in pharmaceuticals: a toxicological evaluation. J. Pharm. Sci. Technol. 68, 221e238. ndez-Arriaga, F., Esplugas, S., Gime nez, J., 2010. Degradation of the emerging Me contaminant ibuprofen in water by photo-Fenton. Water Res. 44, 589e595. Montgomery, D.D., 1984. Design and Analysis of Experiment, second ed. Willey, New York. ~es, J.R., 2000. Photodegradation of dichloroacetic acid and Nogueira, R.F.P., Guimara 2,4-dichlorophenol by ferrioxalate/H2O2 system. Water Res. 34, 895e901. Patient, 2014. Mefenamic acid for pain and inflammation e Ponstan. http://www. patient.co.uk/medicine/mefenamic-acid-for-pain-and-inflammation-ponstan. Accessed November 2015.  n, G., 2009. Solar photo assisted advanced oxidation process Prato-Garcia, D., Buitro of azo dyes. Water Sci. Technol. 59, 965e972. Rawlinson, J.A., Davis, P., 1983. Mefenamic Acid: Some Practical Aspects of the Treatment of Rheumatic Disease. Oxford University Press, Oxford. Rozasa, O., Contrerasa, D.M., Mondacab, A., Perez-Moyac, M., Mansillaa, H.D., 2010. Experimental design of Fenton and photo-Fenton reactions for the treatment of

R. Colombo et al. / Journal of Environmental Management 167 (2016) 206e213 ampicillin solutions. J. Hazard. Mater. 177, 1025e1030. San Miguel, E.R., Vital, X., de Gyves, J., 2014. Cr(VI) transport via a supported ionic liquid membrane containing CYPHOS IL101 as carrier: system analysis and optimization through experimental design strategies. J. Hazard. Mater 273, 253e262. Soulet, B., Tauxe, A., Tarradellas, J., 2002. Analysis of acidic drugs in swiss wastewaters. Int. J. Environ. An. Ch. 82, 659e667. Tauxe-Wuersch, A., De Alencastro, L.F., Grandjean, D., Tarradellas, J., 2005. Occurrence of several acidic drugs in sewage treatment plants in Switzerland and risk assessment. Water Res. 39, 1761e1772. rez, M., Mansilla, H.D., Peral, J., 2003. Experimental design of Fenton Torrades, F., Pe and photo-Fenton reactions for the treatment of cellulose bleaching effluents. Chemosphere 53, 1211e1220. Verlicchi, P., Al Aukidy, M., Zambello, E., 2012. Occurrence of pharmaceutical compounds in urban wastewater: removal, mass load and environmental risk after a secondary treatment: a review. Sci. Total Environ. 429, 123e155.

213

Werner, J.J., McNeill, K., Arnold, W.A., 2005. Environmental photodegradation of mefenamic acid. Chemosphere 58, 1339e1346. Wols, B.A., Hofman-Caris, C.H., 2012. Review of photochemical reaction constants of organic micropollutants required for UV advanced oxidation processes in water. Water Res. 46, 2815e2827. Wu, Q., Shi, h., Adams, C.D., Timmons, T., Ma, Y., 2012. Oxidative removal of selected endocrine-disruptors and pharmaceuticals in drinking water treatment systems, and identification of degradation products of triclosan. Sci. Total Environ. 439, 18e25. Ying-Shiha, M., Chi-Fanga, S., Jih-Gaw, L., 2010. Degradation of carbofuran in aqueous solution by ultrasound and Fenton processes: effect of system parameters and kinetic study. J. Hazard. Mater. 178, 320e325. Zakrzewska, G., Herdzik-Koniecko, I., Cojocaru, C., Chajduk, E., 2014. Experimental design and optimization of leaching process for recovery of valuable chemical elements (U, La, V, Mo and Yb and Th) from low-grade uranium ore. J. Hazard. Mater 275, 136e145.