Experimental design methodology

Comparing with biological treatment, the major limitation of AOPs is their relatively high operational costs for complete oxidation of organic compounds [34,35].

32

In order to reduce the operational cost and achieve high performance of AOPs, the experimental conditions, i.e. pH, temperature and dosages of reactants as shown in Table 2-6, must be optimized. The traditional approach of changing one variable at a time while fixing all other variables constant to investigate the effect of parameter on the response is a time consuming method especially while multifactor and several responses considered. Moreover, this approach has limitation in assuming that various parameters don’t have interacted effect between variables. To overcome these drawbacks, the statistical based optimization methodology which consider interaction effect between variables, such as the CCD, is appropriate to be applied in a multifactor system with minimum number of well-chosen experiments [31,36].

The CCD is a modern experimental design approach, which has been widely used in several applications [13,37-39] to fit the experimental data and develop a statistically significant second-order polynomial equation. The CCD is a star design including factorial design, axial or star point and central point (three to five replicates). The total number of experiments (N) required for CCD is determined as per the Eq. 2-18 [39,90,91], where K represents the number of independent variables and n_c is the central point. Due to the varying units of the different factors (actual values), independent variables are normalized in the form of dimensionless coded values (- , -1, 0, 1, ), which is also useful to obtain more accurate estimate of the regression coefficient and reduce the interrelationship between linear and quadratic terms [87,92]. The value of was calculated by Eq. 2-19. The transformation of natural variable into coded value (-1 and +1) is made according to the Eq. 2-20 [76].

(2-18)

33

Table 2-6. The independent and dependent variables of experimental design selected in Fenton processes.

Method

34

Table 2-6. The independent and dependent variables of experimental design selected in Fenton processes (contined)

Method

35

(2-19) (2-20)

where and are the natural and coded values of independent variables.

and are the maximum and minimum values of independent variables, respectively.

RSM is used along with different type of experiment designs, i.e. factorial design and CCD, for optimization by least squares technique to assess the conditions that could yield the most desirable response [16,31,37]. A full second-order polynomial equation and its corresponding regression coefficients for two factors are as shown in Eq. 2-21.

(2-21)

where Y represents predicted response; X₁ and X₂ represents independent variables; the set of regression coefficients consist the intercept ( ), linear ( ), interaction ( ) and quadratic coefficients ( ) [37].

For the purpose of applying effectively AOPs in the wastewater treatment, the optimization of operational parameters plays an important role in target compounds degradation. For such a goal, experimental design along with RSM has been used extensively in several applications to generate a quadratic model and yield the most desirable response that considers the synergistic and antagonist effects between the variables [13,31,93]. Besides, the response surface plot can be constructed to locate the optimum point of the multifactor system [16].

36

Chapter 3

Materials and methods

3.1 Chemical reagents

Carbofuran was obtained from Shida Chemical Industries (Taoyuan, Taiwan) and was used as received (HPLC grade ＞ 98% purity). Titanium sulfate (TiSO₄, 5% w/w) was purchased from Nacalai Tesque (Japan) and H₂O₂ (33%, w/w) was supplied by Panreac Chemicals (Spain). Stock solution of Fe³⁺ (1000 mg L^-1) was prepared by dissolving ferric sulfate (Yakuri Pure Chemicals, Japan) in double distilled water. The HPLC grade methanol was used in carbofuran analysis. All other chemicals were reagent grade and the solutions were prepared using double distilled water.

3.2 Experimental apparatus

Fig. 3-1 shows the schematic diagram of the experimental setup used for carbofuran degradation. A 1.6 L double-walled reactor was used in all experiments.

Several ports were provided on the reactor for feeding the reactants and sampling the solution. Moreover, pH and temperature probes were fixed in the reactor. A Teflon-coated stirrer was installed in the reactor to mix the solution at 175 rpm. During the experiment, H2O2 was added continuously into reactor at a flow rate of 1 mL min^-1

37

` `

Thermal probe

pH probe

UV lamp

Stirrer

Sampling port

Syringe pump

H2O2

Thermostate

Fig. 3-1. Schematic diagram of the experimental setup.

38

with a syringe pump. Two 8-W monochromatic UV lamps of 312 nm (with emission range between 280 nm and 360 nm) were placed axially in the reactor and kept in place with a quartz sleeve; the UV intensity of one 8-W UV lamp at 312 nm is 60 µW cm^-2. The reaction temperature was maintained at 25±1^oC during all the experiments by using a water bath.

3.3 Experimental overview

The present study was divided into three stages as shown in Fig. 3-2. In stage 1, the effects of operating parameters including irradiation time, carbofuran concentration, pH, H2O2 dosage rate and Fe³⁺ dosage were investigated. The carbofuran removal was fitted using the pseudo-first order reaction model to determine the kinetic parameters.

Furthermore, a number of control experiments including H2O2, Fe³⁺, UV, Fe³⁺ + H2O2, UV + Fe³⁺and UV + H2O2 were carried out to evaluate the efficiency of each process in carbofuran degradation and mineralization.

The CCD along with RSM was employed to find out the favorable reagent dosages for carbofuran degradation under the photo-Fenton process in the stage 2. The effects of H2O2 dosage rate (1-10 mg L^-1 min^-1) and Fe³⁺ dosage (1-100 mg L^-1) were evaluated and the second-order polynomial equations in terms of carbofuran and DOC removals and BOD₅/DOC ratio with different reaction times were developed.

Stage 3 was intended to evaluate the potential of the photo-Fenton process as a pretreatment of carbofuran under favorable reagent dosages obtained in stage 2. BOD₅, BOD5/COD ratio, average oxidation state (AOS), carbon oxidation state (COS) and Microtox^® test were used to investigate the variations of toxicity and biodegradability during the degradation and mineralization of carbofuran and to assess the appropriate

39

Fig. 3-2. Flowchart of the research.

40

time for coupling the photo-Fenton process with biological treatment. Moreover, the intermediates of carbofuran produced during the photo-Fenton treatment were identified by GC/MS.

3.4 Experimental procedure

Stock solution of carbofuran was prepared by dissolving 200 mg of carbofuran in 1 L double distilled water. Exactly 1 L of diluted carbofuran solution corresponding to a required initial concentration was added into the reactor. The initial pH was adjusted to a pre-determined level using 0.1 N H₂SO₄. Subsequently, a designed quantity of Fe³⁺

was pumped into the reactor and the contents were mixed thoroughly. The UV lamps were turned on for 15 min before experiment then mark the starting point of the experiment. The H2O2 was simultaneously added into the reactor at a constant flow rate.

At regular time intervals, 8 mL of sample was withdrawn from the reactor and filtered through a 0.45 µm membrane filter paper. Finally, the samples were analyzed for residual carbofuran, BOD₅, COD, DOC, H₂O₂ and carbofuran intermediates. The experimental conditions for the carbofuran degradation are shown in Table 3-1.

3.5 Experimental design and data analysis

In order to correlate the independent variables, i.e. H2O2 dosage rate and Fe³⁺

dosage, and dependent variables, i.e. carbofuran and DOC removals and BOD5/DOC.

Predetermined ranges of independent variables, i.e. H₂O₂ dosage rate (1, 2.3, 5.5, 8.7 and 10 mg L^-1 min^-1) and Fe³⁺ dosage (1, 15, 51, 86 and 100 mg L^-1), were selected for CCD. Three replications of central points (runs 4, 5, 6) were selected for CCD to check the reproducibility of data and evaluate the experimental error of the results and total 11

41

Table 3-1. Experimental conditions for carbofuran degradation

Experiments

42

photo-Fenton experiments shown in Table 3-2 were randomized by the statistical software to minimize systematic errors. The coded and natural levels of independent variables for the experimental design are shown in Table 3-1. Due to the varying units of the different factors (actual values), H₂O₂ dosage rate and Fe³⁺ dosage were normalized in the form of dimensionless coded values from - to , which is also useful to obtain more accurate estimate of the regression coefficients and to reduce the interrelationship between linear and quadratic terms [92]. In this study, the coded value of a two factor CCD was 1.414. A second-order polynomial equation and its corresponding regression coefficients were obtained from the experimental data using the MINITAB^® 14 statistical software.

3.6 Analytical measurements

The analysis included measurements of pH, carbofuran concentration, H₂O₂ concentration, DOC, BOD5, COD, Microtox^® test and intermediates. The description of each measurement was summarized as follows:

3.6.1 Analysis of carbofuran concentration

Carbofuran concentration in the samples was analyzed by the high performance liquid chromatography (HPLC) (Hitachi Co., Japan) equipped with a Hitachi L-2420 UV detector and a RP-18 GP 250 separation column (250 mm × 4.6 mm i.d., Kanto Chemicals, Japan). Exactly 20 µL of sample was injected manually and analyzed at 280nm. The mobile phase, composed of methanol and water (50:50, v/v), was pumped at a flow rate of 1 mL min^-1. Under these separation conditions, the retention time of

43

Table 3-2. Experimental design along with coded and natural levels for two independent variables.

Experimental run

x₁¹ x₂²

Coded value

Natural value (mg L^-1 min^-1)

Coded value

Natural value (mg L^-1)

1 -1 2.3 -1 15

2 -1 2.3 1 86

3 0 5.5 1.414 100

4 0 5.5 0 51

5 0 5.5 0 51

6 0 5.5 0 51

7 0 5.5 -1.414 1

8 1 8.7 1 86

9 1.414 10.0 0 51

10 -1.414 1.0 0 51

11 1 8.7 -1 15

1x1, H2O2 dosage rate (mg L^-1 min^-1)

2x₂, Fe³⁺ dosage (mg L^-1)

44

carbofuran was observed around 12 min. The method detection limit (MDL) for analysis of carbofuran is 0.16 mg L^-1. The percentage carbofuran removed was expressed as Eq. 3-1, where C0 is the initial concentration of carbofuran (mg L^-1) and Ct

is the residual concentration of carbofuran at reaction time t (mg L^-1).

Carbofuran removal (%) = (1 – Ct/C0) × 100% (3-1)

3.6.2 Analysis of H₂O₂ concentration

For determining the residual H₂O₂ concentration, the samples were mixed with 5%

titanium sulfate solution (the volume ratio of H2O2 sample to titanium sulfate solution was 10:1 v/v) and analyzed using a spectrophotometer (Hitachi U-3010, Japan) at 412 nm [94]. The residual H2O2 was measured for two purposes: (1) to make sure that the H2O2 supplied in the system is sufficient and (2) to compare the residual H2O2 with the variation of carbofuran concentration and DOC removal. The MDL for analysis of H₂O₂ is 0.04 mg L^-1.

3.6.3 Analysis of DOC concentration

Carbofuran mineralization was estimated in term of DOC concentrations. A TOC analyzer (O.I. Analytical Model 1030) was used for measuring the DOC of the samples.

For the DOC measurements, potassium phthalate solutions (5-80 mg L^-1) were used as the calibration standard. The MDL for analysis of DOC is 0.06 mg L^-1. The DOC removal (%) of carbofuran was calculated by using Eq. 3-2, where DOC0 is the initial DOC value of carbofuran (mg L^-1) and DOCt is the residual DOC value of carbofuran at reaction time t (mg L^-1).

45

DOC removal (%) = (1 – DOC_t/DOC₀) × 100% (3-2)

3.6.4 Microtox^® test

Acute toxicity of initial carbofuran solution and the samples collected at different time periods of the photo-Fenton reaction were measured by the Microtox^® test using Vibrio fischeri strain. Microtox® test was performed using a model M500 analyzer and

standard procedures recommended by the Microbics Corporation, USA. Toxicity is generally expressed as EC50 value, i.e. the concentration of sample that causes a 50%

reduction of the bioluminescence (Vibrio fischeri). In this paper, toxicity was evaluated at 5 and 15 min from the time of mixing at various dilutions with the Vibrio fischeri.

Before conducting the Microtox^® test, the pH value of the samples was adjusted to 7.

Moreover, the EC50 value used in this study was converted to toxicity unit (TU) using Eq. 3-3. Thus, TU is inversely proportional to EC₅₀ value, lower EC₅₀ value relates to a higher TU.

TU = 100/EC₅₀ (3-3)

3.6.5 Identification of intermediates

The major carbofuran degradation intermediates formed during the photo-Fenton reaction were identified using the GC/MS technique. A mixture of sample and dichloromethane (5:1 v/v) was shaken vigorously in a rotary shaker at 150 rpm for 30 min, and subsequently, analyzed in a Shimadzu GC/MS-QP2010 equipped with a HP-5 capillary column (30 m × 0.25 mm i.d., thickness of 0.25 μm). Helium was used as the carrier gas at a flow rate of 1.5 mL min^-1. The GC oven temperature was programmed as

46

follows: initially held at 80^oC for 2 min, increased to 210^oC at a rate of 10^oC min⁻¹ and held for 3 min, then raised from 210 to 310^oC at a rate of 30^oC min⁻¹ and finally held at 310^oC for 2 min. The injector and detector temperatures were maintained at 220 and 250^oC, respectively. The mass spectrometer was operated in the full-scan electron-impact (EI) mode at 70 eV.

3.6.6 Other analytical techniques

The pH and temperature were monitored continuously by a pH and temperature meter (Suntex TS-2), respectively. In the biodegradability test, variations in BOD₅ and COD at different stages of the photo-Fenton reaction were measured as per the Standard Methods [95]. The pH value of the samples was adjusted to 7 for the analysis of BOD₅ and sodium hydrogen sulphite was added into the samples to remove residual H2O2

before analysis of COD.

3.7 Carbofuran removal kinetics

The pseudo-first order reaction model as shown in Eq. 3-4 was used to study the carbofuran removal kinetics, where, C0 and Ct are the concentrations of carbofuran (mg L^-1) at reaction times zero and t, respectively. kapp is the apparent rate constant (min^-1) and t is the reaction time (min). The rate constants were estimated based on the linear plot of ln(C0/Ct) versus t.

ln(C₀/C_t) = k_appt (3-4)

47

Chapter 4

Results and discussions

There were numbers of well chosen experiments carried out in this study. Based on the results obtained, this chapter is divided into ten major sections listed as follows:

4.1 Effect of irradiation time on carbofuran degradation

4.2 Effect of carbofuran concentration on carbofuran degradation 4.3 Effect of initial pH on carbofuran degradation

4.4 Effect of H2O2 dosage rate on carbofuran degradation 4.5 Effect of Fe³⁺ dosage on carbofuran degradation 4.6 Chemical degradability of carbofuran

4.7 CCD of experiments for the photo-Fenton process 4.8 Degradation pathway

4.9 Toxicity and oxidation state assessment 4.10 Biodegradability assessment

Many studies have shown that the performance of the photo-Fenton process is highly related to the operating parameters. In order to achieve high removal efficiency, the optimum operation conditions must be investigated. Therefore, experiments were carried out to study the effects of operating parameters including irradiation time, carbofuran concentration, pH, H2O2 dosage rate and Fe³⁺ dosage on the degradation and mineralization of carbofuran and their results are described in sections 4.1-4.5.

48

Control experiments such as H₂O₂, Fe³⁺, UV, Fe³⁺ + H₂O₂, UV + Fe³⁺and UV + H₂O₂ were also carried out to evaluate the efficiency of each process and their results are summarized in section 4.6. The favorable reagent dosages obtained by CCD along with RSM were discussed and reported in section 4.7. The degradation pathway of carbofuran degradation under favorable reagent dosages is proposed and discussed in section 4.8 based on the intermediates identified. Finally, the appropriate time for coupling the photo-Fenton process with biological treatment has been discussed according to the variations of toxicity and biodegradability at favorable reagent dosages in sections 4.8 and 4.9, respectively.

4.1 Effect of irradiation time on carbofuran degradation

The effect of irradiation time was evaluated under an initial carbofuran concentration of 50 mg L^-1 and pH 3 with the Fe³⁺ dosage at 5 mg L^-1 and H2O2 dosage rate at 0.1 mg L^-1 min^-1. The profiles of carbofuran and DOC removals are shown in Fig.

4-1. After 240 min, carbofuran and DOC removals were reached around 97% and 42%, respectively. These results indicate that even with low dosage of Fe³⁺ and dosage rate of H2O2, it is possible to obtain higher carbofuran degradation as well as its mineralization.

However, no significant increase in the carbofuran removal was observed between 120-240 min (less than 10%). The DOC removal profile indicates that carbofuran mineralization rate could be enhanced by increasing the UV irradiation time.

4.2 Effect of carbofuran concentration on carbofuran degradation

In order to investigate the effect of carbofuran concentration for the photo-Fenton degradation of carbofuran, the experiments were repeated at 21 and 80 mg L^-1

49

(a)

0 40 80 120 160 200 240

0 20 40 60 80 100

C ar bof ur an rem ov al ( % )

Time (min)

(b)

0 40 80 120 160 200 240

0 20 40 60 80 100

D O C r em ov al ( % )

Time (min)

Fig. 4-1. The profiles of (a) carbofuran (b) DOC removals with time (Fe³⁺ dosage at 5 mg L^-1, H2O2 dosage rate at 0.1 mg L^-1 min^-1 and pH 3).

50

carbofuran concentration with two different reagent dosages at pH 3, i.e. H₂O₂ dosage rate of 1.3 mg L^-1 min^-1, Fe³⁺ dosage of 10 mg L^-1 and H2O2 dosage rate of 1.3 mg L^-1 min^-1, Fe³⁺ dosage of 40 mg L^-1. Fig. 4-2 shows the trend of degradation and mineralization of carbofuran with the reaction time. It can be noticed in Figs. 4-2a and b that carbofuran concentration has great negative effect on carbofuran degradation and mineralization, regardless of the amount of H₂O₂ dosage rate and Fe³⁺ dosage employed.

The reduction of carbofuran removal accompanies with the increase of carbofuran concentration. In addition, carbofuran removal increased with increasing of limiting factor Fe³⁺ dosage at high carbofuran concentration (Fig. 4-2a). After 30 min of reaction, carbofuran were completely removed in cases of A and C. However, the carbofuran removal has reached only less than 74% in cases of B and D. Similar variations were also determined for DOC removal (Fig. 4-2b). Mineralization efficiencies obtained at low carbofuran concentration (A and C) were greater than those at high carbofuran concentration (B and D).

4.3 Effect of initial pH on carbofuran degradation

The photo-Fenton reaction is strongly pH dependent as solution pH will affect the generation of ‧OH, the oxidation performance and the species concentration of Fe³⁺

complex in aqueous solution. The experiments were conducted using 50 mg L^-1 carbofuran concentration at various pH values (2-4) with Fe³⁺ dosage of 35 mg L^-1 and H2O2 dosage rate of 4 mg L^-1 min^-1. The effect of initial pH on the carbofuran degradation in the photo-Fenton process is shown in Fig. 4-3 and Table 4-1. As shown in Fig. 4-3, the profiles demonstrate that the removals of carbofuran and DOC are greatly affected by the initial pH level. Almost 100% degradation of carbofuran was

51

Fig. 4-2. The profiles of (a) carbofuran (b) DOC removals with time at pH 3, using H2O2: 1.3 mg L^-1 min^-1, Fe³⁺: 10 mg L^-1 (A and B) and H2O2: 1.3 mg L^-1 min^-1, Fe³⁺:

40 mg L^-1 (C and D) (Solid points represent carbofuran 21 mg L^-1 and hollow points represent carbofuran 81 mg L^-1).

52

Fig. 4-3. The profiles of (a) carbofuran (b) DOC removals with time at various pH (Fe³⁺ dosage at 35 mg L^-1, H2O2 dosage rate at 4 mg L^-1 min^-1 and carbofuran 50 mg

L^-1).

53

Table 4-1. Effect of various pH on carbofuran and DOC removals (Fe³⁺ dosage at 35 mg L^-1, H2O2 dosage rate at 4 mg L^-1 min^-1 and carbofuran 50 mg L^-1).

pH

Carbofuran

removal (%)

DOC

removal (%)

Pseudo-first order carbofuran removal kinetics

30 min 120 min 30 min 120 min k_app (×10^-2 min^-1) R²

2 70 100 7 82 3.5 0.92

2.5 95 100 22 92 8.1 0.88

3 100 100 31 93 8.6 0.94

3.5 61 100 9 63 3.2 0.96

4 37 100 7 49 2.1 0.85

achieved in between 100-120 minutes of the photo-Fenton process under all initial pH values (Fig. 4-3a). At the end of 30 minutes, 37% degradation of carbofuran was observed at pH 4, whereas it was 61% at pH 3.5 and almost 100% at pH 3. This indicates that the lower pH values are better for the photo-Fenton decomposition of carbofuran. However, the decomposition of carbofuran was reduced to 95% and 70%

under pH 2.5 and 2, respectively at the end of 30 minutes of photo-Fenton reaction.

Similar trends can be seen in Fig. 4-3b, where the DOC removal at the end of 120 minutes increased from 49% at pH 4 to 93% at pH 3, and then decreased to 82% at pH 2.

It can also be seen in Table 4-1 that increasing solution pH from 2 to 3 enhances the removal of carbofuran at 30 min reaction from 70% to 100%; however, further

54

increase in pH (above 3) decreases the removal efficiency, and at pH 4 the removal efficiency decreased to 37%. The reduced efficiency at higher pH values (pH > 3) is due to the precipitation of Fe³⁺ as inactive Fe(OH)3 thus hindering the reaction between Fe²⁺

and H₂O₂. Furthermore, H₂O₂ can also be hydrolyzed to water and oxygen (Eq. 4-1) at higher pH values [14]. The decomposition performance under low pH condition (pH < 3) decreases since excess H⁺ will react with ‧OH via Eq. 2-17 and produce water [33].

The consumption of H2O2 and ‧OH can decrease the oxidation performance of target compounds. Moreover, the dominant Fe(OH)²⁺ reaches a maximum at pH around 3 [14].

The lower quantity of Fe(OH)²⁺ at pH below 3 causes lower amount of Fe²⁺ reduction and ‧OH production in Eq. 2-12 resulted in poor performance of carbofuran removal.

2H₂O₂ → 2H₂O + O2 (4-1)

The variations of DOC removal at 30 min and 120 min are similar with the results of carbofuran removal. At pH value of 3, the DOC removals at 30 min and 120 min were found to be 31% and 93%, respectively. Furthermore, the carbofuran removals were fitted well with the pseudo-first order reaction model and the R² values ranged between 0.85 to 0.96. The apparent rate constants of carbofuran removal were increased with the decrease in pH value from 4 to 3. Further decrease in pH (below 3) decreased the apparent rate constant of carbofuran removal. The highest apparent rate constant (8.6 × 10^-2 min^-1) was obtained at pH 3. Base on the results obtained, the optimal pH for the oxidation of carbofuran using photo-Fenton process is 3, which is in good

The variations of DOC removal at 30 min and 120 min are similar with the results of carbofuran removal. At pH value of 3, the DOC removals at 30 min and 120 min were found to be 31% and 93%, respectively. Furthermore, the carbofuran removals were fitted well with the pseudo-first order reaction model and the R² values ranged between 0.85 to 0.96. The apparent rate constants of carbofuran removal were increased with the decrease in pH value from 4 to 3. Further decrease in pH (below 3) decreased the apparent rate constant of carbofuran removal. The highest apparent rate constant (8.6 × 10^-2 min^-1) was obtained at pH 3. Base on the results obtained, the optimal pH for the oxidation of carbofuran using photo-Fenton process is 3, which is in good

在文檔中 利用光 Fenton 法降解與去除水中加保扶毒性之研究 (頁 46-0)