Photocatalytic degradation of acesulfame K

Process Safety and Environmental Protection 1 1 3 ( 2 0 1 8 ) 10–21

Contents lists available at ScienceDirect

Process Safety and Environmental Protection journal homepage: www.elsevier.com/locate/psep

Photocatalytic degradation of acesulfame K: Optimization using the Box–Behnken design (BBD) Seong-Nam Nam a,c , Hyekyung Cho b,d , Jonghun Han c , Namguk Her c , Jaekyung Yoon b,∗ a

Department of Civil and Environmental Engineering, Chung-Ang University, 84 Heukseok-ro, Dongjak-gu, Seoul 06974, Republic of Korea b New and Renewable Energy Research Division, Hydrogen Laboratory, Korea Institute of Energy Research, 71-2 Jang-dong, Yuseong-gu, Daejeon 305-343, Republic of Korea c Department of Civil and Environmental Science, Korea Army Academy at Yeongcheon, 495 Hoguk-ro, Yeongcheon-si, Republic of Korea d Department of Chemical and Biomolecular Engineering, Yonsei University, 50 Yonsei-ro, Seodaenum-gu, Seoul 03722, Republic of Korea

a r t i c l e

i n f o

a b s t r a c t

Article history:

In this research, photocatalytic degradation of acesulfame K, one of the most popular arti-

Received 27 March 2017

ficial sweeteners, has been carried out under variations of the initial concentration, pH,

Received in revised form 1

concentration of persulfate, and amount of natural organic matter (NOM). The removal

September 2017

efficiencies for 30-min, 60-min and 180-min reaction time have been applied to response

Accepted 4 September 2017

surface methodology using the experimental responses obtained by a four-factor-three-level

Available online 12 September 2017

Box–Behnken design (BBD). This provided 29 experimental data for the initial concentration of acesulfame K ranging from 300 to 900 ␮g/L, pH of solution ranging from 4 to 10, persul-

Keywords:

fate concentration ranging from 0 to 10 mg/L, and amount of natural organic matter (NOM)

Acesulfame K

ranging from 0 to 5 mg/L, which were consecutively coded as A, B, C, and D at three levels

Box–Behnken design (BBD)

(−1, 0, and 1). The analysis of variance (ANOVA) tests with 95% confidence limits deter-

Optimization

mined the significance of independent variables and their interactions consisting of the

Photocatalysis

polynomial regression equation. The optimum values of the selected variables were deter-

Persulfate

mined by numerical optimization, and the experimental conditions were found to reach

Response surface methodology

complete mineralization for 30 min and thereafter, at initial concentration of 887.2 ␮g/L; pH

(RSM)

of 4; persulfate concentration of 9 mg/L, and NOM concentration of 5 mg/L. © 2017 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1.

Introduction

Artificial sweeteners, e.g., acesulfame, aspartame, cyclamate, saccharin, and sucralose, have been used as sugar substitutes in considerable

(Mayer and Kemper, 1991), and are discharged into aquatic environments through wastewater treatment plants (WWTPs), because some of these compounds are not efficiently degraded through WWTPs. For example, removal efficiencies of acesulfame and sucralose were

amounts in beverages, food, health and dietary products, pharmaceuti-

reported to be less than 20% (Scheurer et al., 2009; Buerge et al., 2009),

cals, and animal feed (Kroger et al., 2006). These sweeteners have no or

consequently, these are detected to ng/L to mg/L levels in wastewater (Loos et al., 2013), surface waters (Muller et al., 2011; Perkola and

low calories, and are a few hundreds to thousands times stronger than sugar; thus their uses have been substantially increased by commercial industries. Following ingestion, artificial sweeteners are excreted from human bodies mostly unchanged at rates of >90% of the ingested dose

∗

Sainio, 2014), groundwater (van Stempvoort et al., 2011), and drinking water (Scheurer et al., 2010). Of artificial sweeteners, acesulfame is particularly resistant to degradation in WWTPs; thus, it was found

Corresponding author. E-mail addresses: [email protected], [email protected] (J. Yoon). http://dx.doi.org/10.1016/j.psep.2017.09.002 0957-5820/© 2017 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

11


Table 1 – Chemical information of acesulfame K.

Acesulfame K a b c d

Formula

Molecular weight (g/mol)

Solubilitya (g/L)

pKa b

Log Kow c

Log Koc d

C4 H4 KNO4 S

201.24

9.1 × 102

3.02

−1.33

1.302

Structure

Based on WSKOW v1.41. Based on ChemAxon. Based on KOWWIN v1.67 estimate. Based on PCKOCWIN v1.66.

at similar to concentrations in a septic tank and a ∼15-year old septic plume (Robertson et al., 2013). Due to its persistence in the environment, acesulfame has been used as a suitable wastewater tracer in groundwater (Buerge et al., 2009; Scheurer et al., 2010; Muller et al., 2011; van Stempvoort et al., 2011; Loos et al., 2013; Robertson et al., 2013; Perkola and Sainio, 2014), as well as surface waters (Spoelstra et al. 2013). Photocatalysis, which is one of the advanced oxidation processes (AOPs), has been extensively applied for degradation of a variety of pharmaceuticals and recalcitrant organic compounds that are known to be less effective to conventional water and wastewater treatment processes (Romero et al., 2011; Dimitrakopoulou et al., 2012; Kanakaraju et al., 2014). Many metal oxides and sulfides such as TiO2 , ZnO, WO3 , ZrO2 , Fe2 O3 , CdS and ZnS have been used for the photocatalytic degradation since those can respond to visible light (Cao et al., 2015; Chong et al., 2010; Baran et al., 2015; Villa et al., 2015). In TiO2 -based photocatalysis, several active oxygen species (e.g., hydroxyl radical, superoxide radical, and hydrogen peroxide) are formed by the reactions with generated electrons and holes. Of them, hydroxyl radical (• OH) is frequently considered as the major reactant responsible for the photocatalytic degradation of organic chemicals (Hoffmann et al., 1995). On the other hand, the addition of persulfate ion (S2 O8 2− ) to the photocatalytic process is reported to significantly increase the removal efficiency (Wang and Hong, 1999; Subramonian and Wu, 2014; Ahmadi et al., 2015). S2 O8 2− can generate the sulfate radical (SO4 • − ) in aqueous solution (Eq. (1)) (Kolthoff and Miller, 1951; Waldemer et al., 2007;

The present study presents an experimental approach in combination with RSM. Thus, the purpose of this study was to investigate the photocatalytic degradation of acesulfame using the immobilized nano-TiO2 /UV/PS system and to optimize experimental variables using a Box–Behnken design, one of the experimental designs for RSM. Many researches under the similar reaction system to this study have performed to investigate kinetics and mechanisms along with varying parameters (Moradi et al., 2016; Ahn et al., 2017). Thus, based on the understanding of known discoveries, this study does not repeatedly cover similar mechanistic studies, but the optimization of the degradation conditions using the lab-generated immobilized TiO2 nanotube system.

2.

Materials and methods

2.1.

Reagents and chemicals

Acesulfame potassium and ammonium persulfate were purchased from Sigma-Aldrich, Korea. The Suwannee River natural organic matter (SRNOM) used as a representative NOM was obtained from the International Humic Substance Society (IHSS, Atlanta, Georgia, USA). Water (HPLC grade) and acetonitrile (HPLC grade) were purchased from J.T. Baker.

Matzek and Carter, 2016). SO4 • − reacts with H2 O to produce • OH (Eq. (2)) (Hayon et al., 1972; Ghanbari and Moradi, 2017).

2.2.

2S2 O8 2− → 2SO4 • −

(1)

SO4 •− + H2 O ↔ • OH + SO4 2− + H+

(2)

Persulfate can also be used as an electron acceptor in the photocatalysis process, producing sulfate radical that has a higher redox potential (2.5–3.1 V) with a longer lifetime compared to • OH through the following reaction (Eq. (3)) (Anipsitakis and Dionysiou, 2003). S2 O8 2− + e− ↔ SO4 2− +SO4 • − CB

(3)

Therefore, the persulfate is supposed to accelerate the photocatalytic degradation. Assuming that the photocatalytic process is applied to wastewater containing artificial sweeteners, the presence of organic matter would be one of the important parameters to be considered because natural organic matter (NOM) is present in almost every type of water, and plays a role as a hydroxyl radical scavenger, supposing that they are highly reactive with each other (kC,OH • = 3.8 ± 1.9 × 108 MC −1 s−1 (Arakaki et al., 2013), kNOM,OH • = 1.02 ± 0.10 × 108 MC −1 s−1 (Donham et al., 2014)). In assessing the effects of the multivariate results, response surface methodology (RSM) has been proven to be useful for developing the regression model. In photocatalytic process, using RSM, multiple parameters can be optimized by systematic variation of all variables in a well-designed experiment with a minimum, but the required number of experiments.

Photocatalysis experiments

Photocatalytic experiments were implemented using the selfrotating photocatalytic system, which is made up of the UV light source, tubular reactor vessel, and self-rotating body (∼90 rpm of rotating speed) with four-blade impellers and TiO2 nanotubes on Ti meshes (each mesh: 2 cm × 5 cm in W:H). The physicochemical properties (e.g., structural morphology, crystalline phase, size, etc.) of anodized TiO2 nanotubes on Ti meshes as well as its preparation and fabrication processes can be found in our previous paper (Kim et al., 2013). During the experiments, 1-L of synthesized solution was continuously circulated by 4-blade impeller incorporating with a peristatic pump to keep the solution homogeneous. Temperature (20 ± 1 ◦ C) and dissolved oxygen concentration (7.6 ± 0.9 mg L−1 ) were maintained by recirculating water through a closed and double jacketed reactor vessel. The UV irradiation was provided by a 1000 W xenon lamp (Oriel, USA), in which the light emission was above the 300 nm and the irradiated light intensity was measured to be ca. 68 mW cm−2 . More details of the system are well described in the previous paper (Kim et al., 2013), and readers are encouraged to refer it for further information. During the irradiation experiments, 2 mL of samples for acesulfame K were withdrawn every at 30 min time interval for 180 min (Table 1).

12


Table 2 – Instrumental conditions for acesulfame K analysis. Separation conditions Instrumentation Column Mobile phase A Mobile phase B Gradient

Agilent 6410 Triple Quadrupole LC–MS/MS System Agilent ZORBAX Eclipse XDB-C18, particle size 5 ␮m, 50 × 2.1 mm Water + 0.1% formic acid Acetonitrile + 0.1% formic acid Time (min) %A %B 100 0 0 35 65 1 80 20 8 90 10 10 100 0 11.5 0.4 mL/min 10 ␮L

Flow rate Injection volume MS conditions Ionization condition Spray voltage (V) Capillary temp (◦ C) Collision pressure (mTorr) Compound Acesulfame K

2.3.

ESI mode 4000 300 1.5 Precursor ion (m/z) 162

Product ion (m/z) 82.1

Collision energy (eV) 13

Ionization mode Negative

Analytical methods

Residual acesulfame K was analyzed using the liquid chromatography coupled with mass spectrometry (LC–MS/MS, Agilent 6410 triple-quadrupole). Sample separation was performed on ZORBAX Eclipse XDB-C18 column (Agilent, particle size 5 ␮m, 50 × 2.1 mm) with a gradient elution program using a mobile phase consisting of a mixture of 0.1% formic acid in water and 0.1% formic acid in acetonitrile. Table 2 summarizes the detailed analytical conditions. The removal efficiency of acesulfame K calculated using the following equation:

Removal (%) = 1 −

Ct Co

× 100

(4)

Fig. 1 – Decrease profiles of acesulfame K by adsorption only and photocatalysis only.

Here, Co and Ct are the initial and specific reaction time t concentrations of acesulfame K, respectively.

2.4.

Experimental design and statistical analysis

Experimental design, mathematical modeling, and optimization were performed using Design-Expert 7.0 software. Box Behnken design (BBD) is one of the most common designs used in the RSM, along with central composite design (CCD), and is a class of rotatable or nearly-rotatable second-order designs based on three-level incomplete factorial designs. BBD is often considered to be a relatively efficient and ideal alternative to CCD because CCD requires more experiments, more time and higher costs to build the model equation (Ferreira et al., 2007). The experimental data designed by BBD were analyzed by the response surface regression (RSREG) procedure to fit the following second-order polynomial model (Eq. (5))

Y = ˇ0 +

k i=1

ˇi Xi +

k i=1

ˇii Xi2 +

k k

ˇij Xi Xj + e0

(5)

cross-product coefficient and ˇii is the quadratic coefficient, which refer to the effects of the interaction among independent variables. The multiple regression analysis can be applied to obtain the coefficient, and the equation can be used to predict the response. The coded values of the parameters can be determined from the following equation (Eq. (6)) (Montgomery, 2008):

xi =

Xi − X0 ıX

(6)

Here, X0 is the real value of the independent variable at the center point, Xi is the real value of the independent variable, and ıX is the step change values between low (−1) and high (+1) levels (Table 3).

3.

Results and discussion

3.1.

Photocatalytic degradation of acesulfame K

i=1 j=1

Here, Y is the predicted response (% removal efficiency of acesulfame K); Xi and Xj are variables; ˇ0 is the constant coefficient; and ˇi is the coefficient that determines the influence of parameter i in the response (linear term), ˇij is the

The photocatalytic degradation of acesulfame K was monitored over 180-min and the residual concentration was checked at every 30 min interval. The disappearance was fol-

13


Table 3 – Experimental ranges and levels of the independent test variables. Variables

Initial concentration (Co ) pH Persulfate (PS) NOM

Symbol

Unit

␮g/L – mg/L mg/L

A B C D

lowing the first order decay as expressed using Eq. (7) and Fig. 1. dC = −kC orCt = Co e−kt dt

(7)

where Co and Ct are the concentrations of acesulfame K (mg/L) at time 0 and t, respectively; t is the reaction time (minute); dC/dt is the first order decay rate, and k is the reaction rate constant (min−1 ). In order to investigate the adsorption effect of acesulfame K on TiO2 -coated meshes, control experiments were performed with Co = 600 ␮g L−1 at pH 4, 7, and 10 for 120 min. Fig. 1 shows that adsorptions of acesulfame K were negligible over the experimental range of pH. In addition, as preliminary tests without additional oxidant (i.e., persulfate) and radical scavenger (i.e., NOM), photocatalytic degradation was examined at Co = 600 ␮g L−1 at pH 7 for 180 min, and the apparent rate constant was found out to be 1.36 × 10−2 min−1 .

3.2.

The fitting of models

A four-factor three level Box–Behnken design consisting of 29 experimental runs was adopted to optimize the experimental data, including five replications at the center point. Table 4 summarizes the complete experimental design matrix that uses four factors as independent variables and the response based on experimental runs, and the results as response variables were expressed as empirical second order polynomial equations as shown in Eq. (8)–(10). Y1 (30-min removal) = 34.83 − 0.17A − 21.86B + 13.12C − 11.90D − 2.91AB + 4.90AC − 0.93AD − 12.12BC − 6.03BD − 1.06CD + 5.60A2 + 14.71B2 − 13.53C2 + 13.06D2

(8)

Y2 (60-min removal) = 53.46 − 0.27A − 16.95B + 20.17C − 14.25D − 0.80AB + 3.21AC − 1.23AD − 5.38BC − 8.49BD + 0.15CD + 5.74A2 + 10.51B2 − 8.23C2 + 7.82D2

(9)

Y3 (180-min removal) = 90.59 + 0.40A − 6.40B + 1162C − 9.91D − 0.76AB + 2.71AC + 0.83AD + 1.72BC − 7.10BD + 7.68CD − 0.87A2 + 1.72B2 − 6.90C2 − 2.57D2

(10)

Coded variable level −1

0

1

300 4 0 0

600 7 5 2.5

900 10 10 5

where, Y1 , Y2 and Y3 represent the % removals at 30-min, 60min and 180-min reaction, respectively. A, B, C and D are the coded values of the initial concentration of acesulfame K, pH, concentration of persulfate, and the amount of natural organic matter (NOM), respectively. The quality of the models was statistically evaluated based on the coefficient of determination (R2 ) and the analysis of variance (ANOVA) results as shown in Table 5. The ANOVA results of the second order quadratic regressions show that the models were highly significant, because the F-values (90.18, 52.86 and 62.62 for the three selected reaction times) are greater than 0.001. The p-values are