Journal of Saudi Chemical Society (2017) 21, 545–557
King Saud University
Journal of Saudi Chemical Society www.ksu.edu.sa www.sciencedirect.com
ORIGINAL ARTICLE
Removal of caffeine from water by combining dielectric barrier discharge (DBD) plasma with goethite Jian Wang a, Yabing Sun a,*, Hao Jiang a, Jingwei Feng b a State Key laboratory of Pollution Control & Resources Reuse, School of the Environment, Nanjing University, Nanjing 210046, PR China b School of Civil and Hydraulic Engineering, Hefei University of Technology, Hefei 230009, PR China
Received 23 June 2016; revised 31 July 2016; accepted 14 August 2016 Available online 22 August 2016
KEYWORDS Dielectric barrier discharge; Plasma; Caffeine; Goethite
Abstracts In this research, dielectric barrier discharge plasma was developed to cooperate with goethite for removing caffeine in aqueous solution. Goethite was characterized by X-ray diffraction and scanning electron microscopy. The effects of input power, initial concentration and catalysts concentration on the removal efficiency of caffeine were evaluated. Furthermore, the degradation pathways of caffeine were also discussed preliminarily. In the case of caffeine concentration at 50 mg L1, the degradation efficiency of caffeine was improved from 41% to 94% after 24 min on the conditions of input power of 75 W by combining goethite catalysts (2.5 g L1), while the energy efficiency could be enhanced 1.6–2.3 times compared to the single DBD reactor. The reaction mechanism experiments demonstrated that attack by hydroxyl radical and ozone was the main degradation process of caffeine in aqueous solution. These studies also provided a theoretical and practical basis for the application of DBD-goethite in treatment of caffeine from water. Ó 2016 King Saud University. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction Caffeine (C8H10N4O2) is one of the most consumed stimulating substances, being present in cola drinks, coffee, tea and energy drinks [1]. Caffeine has a high water solubility (Ks value * Corresponding author. E-mail address:
[email protected] (Y. Sun). Peer review under responsibility of King Saud University.
Production and hosting by Elsevier
is more than 10,000 mg L1) and low octanol–water partition coefficient, which make it easy to dissolve in aqueous environment [2,3]. Due to the low efficiency of conventional wastewater treatment process, caffeine has been detected in many surface water and ground water. Caffeine has been found at concentrations of respectively 0.03 to 1.28 nmol L1 in Switzerland [4], while concentrations of 0.05 to 0.5 nmol L1 were found in Germany [5]. Caffeine is also detected in groundwater with concentrations ranging from 0.05 to 1.2 nmol [6]. Many treatment technologies are introduced to remove caffeine from wastewater, including activated carbon adsorption, biodegradation, ozonation, photo-Fenton, photo-oxidation, electrochemical oxidation, and photo-
http://dx.doi.org/10.1016/j.jscs.2016.08.002 1319-6103 Ó 2016 King Saud University. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
546 electrolysis [7–9]. However, the low efficiency, incomplete degradation, and high energy consumption limit the wide application of these technologies. Non-thermal plasma (NTP) has been reported as efficient technique for the removal of pollutants in air, but this technique was not applied until now for the removal of water pollutants on a large scale [10,11]. Reddy et al. [12] combined CeO2, Fe2O3/CeO2 and ZrO2/CeO2 catalysts with nonthermal plasma reactor for the degradation of phenol in water. The results confirmed that mineralization of phenol was enhanced by the catalyst addition up to 43.7%. The similar results were obtained on the removal of a model pharmaceutical compound sulfamethoxazole (SMX) [13], it had been observed that ZrO2/CeO2 promoted the degradation and mineralization of SMX. They [14] also designed a dielectric barrier discharge reactor for the degradation of crystal violet and investigated effect of several operating parameters on the removal efficiency. Among these techniques, dielectric barrier discharge plasma was based on the use of a discharge electrode plate covered with a dielectric barrier of quartz [15]. Electric discharges generated in water induced different physical and chemical effects like high electric fields, UV irradiation, overpressure shock waves, and the formation of active species like atomic oxygen O, hydroxyl radical OH, ozone radical ions, hydrogen peroxide, hydroperoxyl radicals, and ozone are generated, which are capable of oxidizing organic pollutants [16– 18]. However, some disadvantages of single DBD plasma, high energy consumption and low mineralization efficiency, limit the wide spread use of this technology. Goethite is one of the most thermodynamically stable iron oxide [19] and formed close-packed array of O2 and OH anions with Fe3+. During last decades, iron minerals have drawn significant interest for their potential application in the field of wastewater treatment due to their excellent absorption and catalytic capacities [20,21]. Recently, Lin et al. [22] found that more than 99% of 2,4,4-trichlorobiphenyl was degraded in a goethite-catalyzed Fenton-like system due to the increase inOH radical concentration. Zhou and coworkers [23] designed a novel heterogeneous sonophotolytic goethite/oxalate Fenton-like system (SP-FL) for removal of sulfamethazine. Compared with Fenton-like system, they found that SP-FL system could achieve synergistic degradation of antibiotic sulfamethazine. However, no relevant information (to our best knowledge) was provided regarding degrading of PPCPs combined with dielectric barrier discharge plasma with goethite catalysts. In the present study, the degradation of aqueous caffeine was conducted using a DBD plasma with goethite catalysts for the first time. In order to provide a theoretical and technical reference for the application of caffeine degradation by catalysts-plasma, the effect of goethite on pollutants degradation was investigated and its degradation mechanism was discussed. Furthermore, the degradation pathways of caffeine were also discussed preliminarily. 2. Experimental 2.1. Materials Caffeine (C8H10N4O2) was purchased from Sigma–Aldrich, goethite (FeOOH) was kindly provided by Zhenjiang Chemical
J. Wang et al. Reagent Co. Ltd, all other chemicals and solvents were of analytic grade and without any purification (Aladdin Inc.), and all solutions were prepared in ultrapure water (R = 18.25 MX,) which was purified by a Milli-Q Gradient system (Millipore, Molsheim, France). 2.2. Experimental reactor and method The schematic of plasma-catalyst system (Fig. 1(a)) was described by our previous literature [24]. The plasma system consisted of a high voltage alternating current (AC) power source (CTP-2000 K, Nanjing Suman Electronics Co., Ltd., China) and a plasma reactor. The electrodes were in air and it was a filamentary discharge. The actual discharge phenomenon is shown in Fig. 1(b). The discharge voltage and current were measured by an oscilloscope (TDS3032, Tektronix), which was displayed in Fig. S1. The equivalent voltage and current per delivered could be integrated by the area of corresponding state variables, and input power was calculated by the following equation. P¼
1 T
Z
T
UðtÞ IðtÞ dt
ð1Þ
0
The main device used to treat the caffeine solution was DBD reactor containing a quartz reaction still and electrodes (Fig. 1(c)). The distance between catalysts and electric barrier was about 1 cm. The reaction still was placed in the center of two aluminum electrodes. The goethite catalysts were placed into the quartz reaction still and suspended in the solution. The goethite concentration was adjusted by adding different quantity of goethite. The liquid circulation system was driven by two peristaltic pumps with a flow rate of 50 mL min1. In this study, 100 mL caffeine solution with the initial concentration of 50 mg L1 was pumped into the plasma reactor and was made to flow through goethite in the reaction still with reaction duration of 24 min. The residence time (RT) of plasma reactor was about 18 s. Each sample (1.5 mL) was taken every 4 min. Actually, most oxidants (e.g. OH, H) produced in the discharge plasma always exit for an extremely short time except for O3 and H2O2. In order to avoid the effect of residual oxidants on the toxicity test, 2 mg Na2S2O3 was added into 100 mL reaction solution to quench the oxidants (e.g. O3, H2O2). Prior to the experiment, the reaction was still placed in dark for 60 min to ensure that the adsorption–desorption equilibrium of caffeine on the surface of goethite. 2.3. Characterization Crystal phase of samples were obtained by X-ray diffraction (XRD, D/Max-rB, Japan Rigaku Corporation) with Cu Ka radiation source of wavelength 0.1541 nm in the range of 3o to 70o with a scanning speed of 4o min1. The specific surface area, the pore volume and the pore diameter of the catalysts and other samples were determined by the N2 adsorption–desorption isotherms (NOVA 3000e) at the liquid-nitrogen temperature. The morphologies of the goethite was analyzed by a scanning electron microscopy (Hitachi S - 3400N II).
Removal of caffeine from water
547
Figure 1
Schematic diagram of the experimental setup.
2.4. Analytical method The caffeine concentration was monitored by a high performance liquid chromatography system (HPLC, Dionex UltiMate 3000) with a C18 analytical column (4.6 150 mm, 5 lm, Akzo Nobel, Netherlands) and UV detection wavelength of 273 nm. The mobile phase consisted of 60% methanol and 40% water and the flow rate was 1.0 mL min1 with injection volume of 20 lL. Furthermore, the UV–Vis absorbance spectrum of caffeine was recorded by Shimadzu UV1750. Ozone concentration was detected according to CJ/ T3028.2–94 (Chinese Standard). For H2O2 measurement is based on the method of previous literature [25]. The acute toxicity assays were carried out according to the standard toxicity measurement protocol [26]. The identification of caffeine and its degradation intermediates were performed by a Thermo Finnigan Surveyor Modular HPLC system with a Finnigan LCQ Advantage MAX ion trap mass spectrometer (LC–MS) and separated by a C18 HPLC column (150 2.1 mm, 3 lm Thermo, USA). The mobile phase consisted of 60% methanol and 40% water with a steady flow rate of 0.2 mL min1. An electrospray interface (ESI) was used with the following conditions: The injection volume, 10 lL; the spray voltage, 4.5 kV; the ion-transfer capillary temperature, 200 °C, respectively. The nebulizer and sheath gas flow rate were 35 arbitrary unit loaded by nitrogen and the collision gas pressure was 0.25 MPa by argon. The spectra were obtained both in negative and positive ion scan mode with the m/z range from 60 to 600. 3. Results and discussions
confirmed that crystal structure and morphology of goethite had no change after their use in the DBD plasma reactor. The surface area of goethite before and after were similar (Table S1), indicated that plasma electric discharge process had no effect on the pore structure of goethite. 3.2. Synergetic degradation 3.2.1. Effect of goethite concentration The effect of goethite concentration on the degradation efficiency in the reaction system was shown in Fig. 3(a). The input power was 75 W and the initial concentration of caffeine was 50 mg L1. The removal efficiency of caffeine was only 41% in single DBD plasma after a 24 min treatment. However, when goethite concentrations in solution increased from 0.8 to 3.2 g L1, the removal efficiency of caffeine was improved from 67% to 98%. In addition, the goethite alone had no effect on the degradation of caffeine without DBD plasma (Fig. 3a). Based on the above results, we could conclude that goethite improved the removal efficiency of caffeine of DBD plasma system. In the plate type DBD system under air atmosphere, active species O3 and H2O2 were produced as the following reactions. The bond energy of oxygen was 5.12 eV [27]. After acquiring energy from electric field, O was produced by the direct electron impact dissociation [28]. This latter interacted then with oxygen molecules in the bulk gas to form ozone as reaction 3 [15]. The ozone and O dissolved in water could generate OH which was believed to play a key role in the degradation process of most organic pollutants [24,29]. As a result, caffeine might be decomposed in a series of reactions with active species.
3.1. XRD and SEM analyses of goethite SEM images (Fig. 2(a) and (b)) revealed the individual cells of goethite fibers with a length of approximately 1–2 lm. XRD patterns are shown in Fig. 2(d), the main diffraction peaks could be found at (1 1 0), (1 3 0), (0 2 1), (1 1 1), (1 4 0), (2 2 0), (1 4 1), (1 5 1) and (0 0 2) respectively, which was consistent with the spacing planes of goethite (JCPDS 29–0713). The XRD patterns and SEM images of used goethite catalysts were the same with that of fresh one (Fig. 2(c) and (d)). These results
O2 þ e ! O
ð2Þ
O þ O2 ! O3
ð3Þ
O þ H2 O ! 2 OH
ð4Þ
O3 þ OH ! OH2 þ O2
ð5Þ
2 OH2 ! H2 O2 þ O2
ð6Þ
548
J. Wang et al.
a
b
c
d
(110)
before reaction after reaction
(111)
(130) (021)
10
20
30
(141)
(220) (140)
40
(151)
50
60
(002)
70
2 Theta (degree) Figure 2
(a) and (b) SEM images of goethite before reaction, (c) SEM image of goethite after reaction and (d) XRD pattern of goethite.
O3 þ 3H2 O ! 3H2 O2
ð7Þ
Caffeine þ active species ð OH;O ; O3 ; etcÞ ! products
ð8Þ
strated that the weak surface FeO–H bonds were favorable sites for producing OH from aqueous ozone. During our experiment, the pH values of reaction solution was decreased with increasing reaction time (insert figure of Fig. 3(c)). The reason was that air discharge would certainly produce NOx which caused the acidification of the solution. Briefly, the NOx might be produced via excited electronic states, whose life was short [27,32]: P N2 þ e ! N2 ðA3 Þ þ e ð10Þ
As demonstrated by above experimental results that addition of goethite catalysts could significantly enhance the removal efficiency of caffeine in the solution. There were several factors that could be accounted for this phenomenon. The concentration of ozone of aqueous solution was measured as shown in Fig. 3(b). Both in the reaction solutions with or without goethite catalysts, the trend of ozone generation was the same that O3 concentration increased with increasing discharge time, and the ozone concentration in the solution with goethite catalysts was lower than that without goethite (Fig. 3(b)). In the absence of goethite, the ozone concentration reached a maximum and remained at constant of 9.8 mg L1 after discharge time of 4 min. These results indicate that the decomposition rate of O3 in the solution was accelerated by adding goethite catalysts to generate secondary oxidizing agent, such as OH, and reacted with caffeine molecules and its intermediates. Muruganandham et al. [30] found that aFeOOH was a stable and efficient catalyst for decomposition of ozone in aqueous medium and did not alter the goethite morphology and its composition.
Then, the interception of NOx by water molecules generated nitrous acid (HONO) and nitric acid (HNO3) [33]. In this case, the charge state of surface hydroxyl groups were changed. Protonated hydroxyl groups predominated when the solution pH was far below the surface equilibrium constants of goethite (pKa1 = 6.4) [30].
FeOOH þ O3 ! OH
BFeOH þ Hþ ! BFeOHþ 2
ð9Þ
Zhang et al. [31] reported that the catalytic activity of synthetic a-FeOOH in promoting OH generating from aqueous ozone was found to be reversely related to the IR stretching frequencies of surface hydroxyl groups. The results demon-
N2 ðA3
P
Þ þ O ! NO þ N ð2 DÞ
ð11Þ
NO þ O ! NO2
ð12Þ
N ð2 DÞ þ NO ! N2 O
ð13Þ
ð14Þ
In addition, the effect of pH on the removal efficiency of caffeine was investigated in Fig. 3(c) by changing the initial pH of solution. The results indicated that goethite-DBD plasma system displayed a better performance on the degradation of caf-
Removal of caffeine from water
Removal efficiency (%)
100
549
0 g L-1 0.8 g L-1 1.6 g L-1 2.5 g L-1 -1 3.2 g L no plasma
80
60
a
40
20
0 0
4
8
12
16
20
24
Time (min) 12
b
Goethite þ hm ! e þ hþ
ð15Þ
H2 O þ hþ ! Hþ þ OH
ð16Þ
O2 þ e ! O 2
ð17Þ
-1
Concentration (mg L )
10
promoted the decomposition of ozone to form OH radical [35]. The acidification and alkalization of the aqueous solution weakens the nucleophilicity of the O atoms of surface hydroxyl groups, which prohibited the interaction of hydroxyl groups with O3 [30,31]. Finally, the removal efficiency of caffeine was decreased with altering the initial pH of reaction solution. Second, the band gap of goethite was 2.1 eV corresponding to absorption wavelength of 590 nm [36]. Previous researches had demonstrated that wavelength of UV light emitted from plasma was in the range of 315–380 nm, and intensity of UV light was less than 4000 a.u. Clearly, even though the UV intensity generated by plasma was small, it could simulate the photocatalytic activity of goethite and produced some reactive oxidative species like Eqs. (15)–(17) [37–39]. These reactive species would attack caffeine molecules and make the contribution to the overall removal process.
8 O3 without goethite
6
O3 with goethite H2O2 without goethite
4
H2O2 with goethite
2 0 0
5
10
15
20
25
Time (min) 100
Removal efficiency (%)
c 90
80
0
5
10
15
20
25
30 8
Time (min) 6
70
4
pH
2
60
0
5
6
7
8
9
10
pH values Figure 3 (a) Effect of goethite concentration on the removal efficiency, (b) Changes of O3 and H2O2 concentration in the solution without caffeine and (c) effect of pH on the removal efficiency (insert figure was the change of pH during experiment).
feine at neutral condition. Resonance structure of ozone gave ozone both an electrophilic and nucleophilic character [34]. Therefore, ozone could combine with H atom and O atom of surface hydroxyl groups of goethite. This combination
Moreover, in the DBD system, active species, H2O2, was produced, which played a significant role in the degradation of the target contaminants [40]. The concentration of H2O2 increased to 2.8 mg L1 within 4 min (Fig. 3(b)), caused by the discharging process and formation of plasma active species. After 4 min, the H2O2 concentration slightly decreased, implying that hydrogen peroxide was decomposed as the reaction process. In another way, the concentration of H2O2 in caffeine solution without goethite was also higher than that with goethite (Fig. 3(b)). The heterogeneous Fenton degradation of caffeine was carried out at this conditions on the surface of goethite in the suspensions. Previous research [41,42] had suggested the following chain reactions in iron mineral catalyzed oxidation system. The symbol () represented the solid surface of goethite. BFe3þ þ H2 O2 ! BFeðOOHÞ2þ þ Hþ
ð18Þ
BFeðOOHÞ2þ ! BFe2þ þ HO2
ð19Þ
BFe2þ þ H2 O2 þ Hþ ! BFe3þ þ H2 O þ OH
ð20Þ
As a direct product of Fenton reaction, hydroxyl radical would be continuously produced, which made the contribution to the degradation of caffeine. According to the above results, goethite had an important effect on the improvement of caffeine degradation efficiency in DBD plasma reactor. However, the exact contribution of each mechanism of goethite improving the elimination efficiency of caffeine in DBD plasma system was not clear due to the limitation of detecting technology. 3.2.2. Effect of initial caffeine concentration The effect of initial concentration on the removal of caffeine was investigated. The input power was 75 W and the concentration of goethite was 2.5 g L1. As illustrated in Fig. 4(a), the degradation efficiency of caffeine was directly related to initial concentration. With the decrease in initial concentrations from 100 to 25 mg L1, the degradation efficiency of caffeine increased from 39% to 61% in single DBD reactor, while the degradation efficiency of caffeine increased from 91% to
550
J. Wang et al.
Removal efficiency (%)
100
constant at the same conditions, which led to a more intensive competition for reactive species and the decline of degradation efficiency.
-1
25 mg L 50 mg L-1 100 mg L-1 25 mg L-1 + goethite 50 mg L-1 + goethite 100 mg L-1 + goethite
80
60
a
3.2.3. Effect of input power
40
20
0 0
4
8
12
16
20
24
Time (min)
Removal efficiency (%)
100
60 W 75 W 90 W 60 W + goethite 75 W + goethite 90 W + goethite
80
60
3.2.4. The effect of radical scavengers
b
40
20
0 0
4
8
12
16
20
24
Time (min)
Removal efficiency (%)
100
Plasma generated in electrical discharges in liquid or at the gas–liquid interface led to the formation of several oxidizing species: radicals (H and OH) and molecules (H2O2, O3, etc.). These active species would react with caffeine. In order to identify the role of these radicals, a series of diagnostic experiments were conducted by adding different radical scavengers during the degradation process. The initial concentration of caffeine was 50 mg L1 and the concentration of goethite was 2.5 g L1 with an input power of 75 W. As shown in Fig. 4(c), with no addition into DBD-goethite system, the removal efficiency of caffeine was 94% in 24 min. In the presence of 2.50 mL L1 iso-propyl alcohol, an efficient OH quencher [44], the degradation efficiency of caffeine dropped to 76%. These results confirmed that reaction with OH was a major pathway of caffeine degradation. ðCH3 Þ2 CHOH þ OH ! ðCH3 Þ2 C OH þ H2 O
c
80
Fig. 4(b) showed the effect of input power on the degradation of caffeine. The initial concentration of caffeine was 50 mg L1 and the concentration of goethite was 2.5 g L1. It was clear that increased input power improved caffeine degradation. The removal efficiency of caffeine was only 24% at input power of 60 W, while removal efficiency of caffeine reached 51% at input power of 90 W in DBD system. Higher input power promoted the electron participation rates [43], which provided a stronger electric field and enhanced the intensity of active particles collision. Then more active species were generated which led to a higher removal efficiency of aqueous caffeine.
ð21Þ
As a hydrogen radical ( H) scavenger, CCl4 could quench H in solution (reaction 22) [45]. As shown in Fig. 4(c), the removal efficiency of caffeine reached 99% by adding 0.1 mL CCl4 into DBD-goethite reaction system, which was 5.5% higher than that with on addition. The decrease inH concentration inhibited the frequency of recombination between hydrogen atom with hydroxyl radical. As a result, the degradation rate of caffeine increased.
60
40 no addition EDTA isopropanol CCl4
20
0 0
4
8
12
16
20
24
Time (min) Figure 4 Effect of (a) initial concentration of caffeine, (b) input power and (c) different radical scavengers on the removal efficiency of caffeine.
CCl4 þ H ! CCl3 þ HCl
ð22Þ
ð23Þ
H þ OH ! H2 O 1
200 mg L EDTA was added into reaction solution, which was a commonly holes acceptor [46]. The removal efficiency of caffeine decreased from 94% to 82%, suggesting that holes promoted the degradation of caffeine in an aqueous solution. The degradation kinetics of caffeine under several factors is shown in Fig. S2. 3.2.5. Energy yield
99% in DBD-goethite system after a 24 min treatment. A higher initial concentration of caffeine meant more caffeine molecules and its intermediates existed in solution while the amount of reactive species generated in reaction system was
The energy yield Y (g kW1 h1) was used to compare the results at different initial concentrations of caffeine, input power, goethite concentration and radical quenchers. The energy yield Y could be calculated by the following equation [24].
Removal of caffeine from water Yðg kW1 h1 Þ ¼
551
C0 Vg Pt
ð24Þ
where C0 was the caffeine initial concentration (g L1), V was the volume of reaction solution (L), g was the removal rate of caffeine (%), P was the input power (kW) and t was the reaction time (h). In general, the energy yield decreased with an increasing removal efficiency of caffeine (Fig. 5). This could be explained that less caffeine and more intermediates existed in solution for continuous degradation, which led to less collision opportunities between caffeine molecules and active radicals, finally resulting in the decrease in energy efficiency. For the same removal efficiency of caffeine, the energy yield increased with increasing goethite concentration (Fig. 5(a)). These phenomena confirmed that adding goethite would dramatically promote the energy efficiency of DBD reaction system. With the increasing initial concentration from 25 to 100 mg L1, energy yield was improved from 0.051 to 0.131 g kW1 h1 in sole DBD plasma, while from 0.083 to 0.303 g kW1 h1 in DBD-goethite system (Fig. 5(c)). Obviously, a higher initial concentration resulted in a greater energy efficiency, which implied that DBD plasma exhibited a better energy performance on the removal of a higher concentration caffeine. Fig. 5(b) showed the value of energy
yield almost kept a constant of 0.06 g kW1 h1 in DBD plasma reactor. However, the energy efficiency increased with the increasing input power in composite DBD-goethite system. Meanwhile, the energy yield showed the same trend as the removal efficiency when different quenchers were added into reaction system (Fig. 5(d)). Finally, it was concluded that a higher initial concentration, input power, catalysts concentration and lower removal efficiency always resulted in a higher energy yield, and the addition of goethite could significantly improve the performance of energy efficiency. These results implied that the addition of goethite catalysts could promote the removal efficiency of caffeine and energy yield through several mechanisms. Compared to our previous literature [24] for removing triclocarban by DBD plasma combined with TiO2/activated carbon fibers, the energy yield were similar, where expensive raw materials were required, goethite catalysts might provide a practical and low cost approach for elimination of caffeine from water. 3.3. Effect of ozone To further investigate the role of ozone in the experiment, pure ozone was purged into the solution containing 50 mg L1 0.5
0.6 -1
0gL -1 0.8 g L -1 1.6 g L -1 2.5 g L -1 3.2 g L
(b) -1
0.4 60 W 75 W 90 W 60 W + goethite 75 W + goethite 90 W + goethite
-1
0.4
(a) Energy yield (g kW h )
-1
-1
Energy yield (g kW h )
0.5
0.3 0.2 0.1 0.0
0.3
0.2
0.1
0.0 0
20
40
60
80
100
0
Removal efficiency of caffeine (%)
25 mg L -1 -1 50 mg L -1 100 mg L -1 25 mg L + goethite -1 50 mg L + goethite 100 mg L-1 + goethite
60
80
0.6
(c)
100
(d) 0.4
-1
-1
Energy yield (g kW h )
0.8
-1
-1
40
Removal efficiency of caffeine (%) 0.5
1.0
Energy yield (g KW h )
20
0.4
0.2
0.0
0.3
0.2 no addition EDTA isopropanol CCl4
0.1
0.0 0
20
40
60
80
Removal efficiency of caffeine (%)
100
0
20
40
60
80
100
Removal efficiency of caffeine (%)
Figure 5 Effect of (a) initial concentration of caffeine, (b) input power, (c) goethite concentration and (d) different radical scavengers on the energy yield.
552
J. Wang et al.
caffeine without DBD plasma system. As shown in Fig. 6(a), the relative concentration (Ct/C0) decreased, reaching 0.054 and 0.276 for DBD-goethite plasma system and ozone treatment alone after 24 min, respectively. The results suggested that the contribution of ozone on the total oxidation process of caffeine was about 73% in plasma reactor. Our experiment confirmed that ozone played an important role in the degradation of caffeine by DBD-goethite plasma system.
3.5. Recyclability of goethite catalysts To evaluate the recyclability performance of goethite for wastewater treatment, the experiment was studied with the following conditions: the input power of 75 W with 50 mg L1caffeine concentration (concentration of goethite was 2.5 g L1). The reuse experiment of catalysts was repeated five times without any treatment. As shown in Fig. 6(c), the degradation efficiency of caffeine had a minor decrease and remained about 89% after five catalytic runs, which was 48% higher than that in sole DBD plasma. Experimental results demonstrated goethite had an excellent and stable performance on recyclability for caffeine removal.
3.4. Mineralization The mineralization of caffeine (50 mg L1) in DBD-goethite plasma reactor was evaluated by monitoring the changes of total organic (TOC) in the solution with 2.5 g L1 goethite catalysts. The results are shown in Fig. 6(b), more than 20% caffeine was oxidized to CO2 indicating that DBD plasma combined with goethite which had a high performance on the mineralization efficiency of caffeine than DBD plasma alone and were more likely to oxidize caffeine to nontoxic products.
3.6. Toxicity evaluation The change of acute toxicity of caffeine solution (50 mg L1) was evaluated by monitoring evolution of natural emission of luminescent bacterium P. phosphoreum T3 spp in samples after different reaction times (0, 4, 8, 12, 16, 20 and 24 min).
25
1.0
TOC removal efficiency (%)
a 0.8 ozone treament DBD-goethite treatment
Ct /C0
0.6
0.4
0.2
15
10
5
0
0.0 0
4
8
12
16
20
0
24
4
8
Time (min)
12
16
20
24
Time (min) 50
100
40
c
no goethite with goethite
30 90
Inhibition rate (%)
Removal efficiency (%)
b
DBD-goethite plasma DBD plasma
20
80
70
d
20 10 0 -10
0
4
8
12
16
20
24
-20 -30 -40 -50
60 1th
2nd
3th
4th
5th
Time (min)
Recyle numbers Figure 6 (a) Comparison of plasma treatment and ozone treatment, (b) Evolution of TOC during the course of degradation of caffeine, (c) The recyclability of goethite in DBD plasma for caffeine removal (100 mL, initial caffeine concentration of 50 mg L1, input power of 75 W) and (d) acute toxicity evolution of degraded caffeine solution: light inhibition of the bacterium P. phosphoreum T3 spp.
Removal of caffeine from water Table 1
553
Summary of identified intermediates determined by LC–MS upon degradation of caffeine in DBD-goethite reaction system.
Compound
Molecular formula
Structure
RT (min)
Molecular ions (m/z)
4.396
195.00 ([M+H])
3.074
195.70 ([M-H])
2.801
211.70 ([M-H])
3.085
243.7 ([M-H]+)
4.163
264.92 ([M+H]+)
2.793
165.72 ([M-H]+)
2.780
61.90 ([M-H]+)
O
2.813
225.71 ([M-H]+)
CHO
2.801
211.70 ([M-H])
3.085
243.7 ([M-H]+)
CHO
2.813
225.71 ([M-H]+)
COOH
2.825
241.70 ([M-H]+)
O
Caffeine
N
N
C8H10N4O2
N
N
O
CHO
C1
N
N
C8H12N4O2 O
N
N
COOH
C2
N
N
C8H12N4O3
N
N
O
COOH N
N
C3
O
C8H12N4O5
N
N
O
OH
COOH N
N
C4
COOH
C8H14N4O6 O
NH
N
OH O
C5
N
HN
C6H6N4O2
N
N H
O
H 2N
C6
C1H6N2O
HN OH
O N
N
C7
C8H10N4O4 O
N
N
OH O
C8
N
N
C8H12N4O3 O
N O
C9
NH2
O N
N
C8H12N4O5
CHO N
O
NH2
O
O
C10
N
N
C8H10N4O4 O
N
NO
O
C11
N
N
C8H10N4O5 O
N
NO
(continued on next page)
554
J. Wang et al.
Table 1 (continued) Compound
Molecular formula
Structure O
C12
CHO
O
C13
O
O
257.7 ([M-H]+)
2.842
303.7 ([M-H]+)
2.784
103.7 ([M-H]+)
NO COOH
O N
N
C8H8N4O9
2.831
N
N
O
Molecular ions (m/z)
O
N
C8H10N4O6
RT (min)
COOH N
O
NO
COOH
C14
HN
C2H4N2O3 O
NH2
Before toxicity test, 2 mg Na2S2O3 was added to 100 mL reaction solution to quench the oxidant (e.g. O3, H2O2) and each sample was adjusted to 3% salinity and neutral. Fig. 6(d) showed the relative inhibition rate (I%) of natural emission in collected samples. It was observed that the original caffeine solution had a low relative inhibition rate of 10%. After a 4 min reaction, the relative inhibition rate of samples in both DBD and DBD-goethite systems slightly increased and reached 17% and 22%, respectively. It should be noted that the values of I in both reaction systems after 8 and 12 min degradation were minus. This phenomenon might be explained that a lower concentration of caffeine solution and some short chain compounds including carbon and nitrogen elements were obtained for the continuous degradation process, and these materials could be used as energy source or stimulant for bacterium. However, the relative inhibition rate gradually reached a steady state in both systems with further reaction, which indicated that the caffeine solution after a 24 min degradation in DBD and DBD-goethite system nearly had no effect on the bacterium either in inhibition or stimulation state. The difference was that the range ability of influence in DBDgoethite system was smaller than that in DBD system. 3.7. Proposed degradation pathways of caffeine The degradation of caffeine and the formation or decomposition of intermediates in DBD-goethite system were monitored by LC–MS analysis, which could be used to propose elemental compositions for the protonated or sodiated adducts of the detected components as well as their characteristic molecule ions. In general, 14 intermediates were identified and 4 degradation pathways were proposed in this process, in which C1, C5 and C13 had never been reported before as well as pathways 3 and 4. The molecular formulas, structure retention times and molecular ions are summarized in Table 1, and four possible degradation pathways are shown in Fig. 7. Caffeine molecule was identified with a positive mode (m/z at 195). The compound marked as C7 with a negative mode (m/z at 225.71) was likely the intermediate of radical oxidation according to a previous study [47], which contained two additional oxygen atoms over caffeine molecule. This compound was
generated by a ring opening due to the attack of OH to the C = N double bond of caffeine molecule. Moreover, the compounds C10, C11 and C13 might be the succeeding products of C7 via further oxidation. C11 was the product of C10 for the further oxidation of –CHO to –COOH. The compound C13 was likely originated from the breaking of 1C = 2C bond and oxidation of methyl (–CH3) of compound C11. The above pathway was defined as pathway 1. Ion mass chromatographs of intermediates could be found in Fig. S3. In pathway 2, C8 was detected in a negative mode with a m/ z of 211.70. The compound C8 was originated from C7 for that the N-CO bond was replaced by a hydrogen atom. Since the further attack of active species to C = C bond of C8, two carbonyl groups were generated on the carbon bonding with nitrogen, which resulted in the formation of C9. C12 was the degraded produce of C9 with the oxidation of –NH2 to – NO. Moreover, the oxidation of methyl and –CHO of C12 led to the generation of C13. However, pathways 1 and 2 were not mutual independence. C8 could be decomposed into C10 for the oxidation of –NH2 to –NO, and C10 might be degraded to C12 as the breakdown of C = C double bond. The detected compounds C1 (m/z at 195.70), C5 (m/z at 165.72) and C13 (m/z at 303.70) in a negative mode had never been reported before. In pathway 3, the compound C1 was originated from caffeine molecule via the breakdown of 2C3C bond. C2 was the consequent degraded product of C1 for the further oxidation of –CHO to –COOH. Moreover, C3 and C4 were generated from C2 for the continuous oxidation of C = N double bond. In pathway 4, the formation of C5 (m/z at 165.72 in a negative mode) was due to the demethylation of caffeine molecule. C6 was detected in a negative mode with a m/z of 61.9, which was likely the degradation product for the fracture of 1C-11 N and 2C-12 N. However, the precursor of C6 and the byproduct were not detected in either a positive or negative mode. Among all intermediates, C14 (m/z at 103.7 in a negative mode) was a highly oxidized product, which could be originated from C4, C5 and C13 for the breakdown of 1C-6 N and 2C-3C bonds and the demethylation when compounds got the attack of radical species. Interestingly, the process from C4, C5 and C13 to C14 was not a one-step procedure, and
Removal of caffeine from water
555 O
O N
N O
N
N
O
N
N
N H
O
C1
N O
N
N
N
O
N
N
H N
HN O
N H
NH2
O
O O N
C9
C2
COOH
O
O
O
N
N
HN
O HN
OH O
C3
N H
N
NO
O N
N
H2 N
COOH
C11
O
N
N
O
NH2
O
N
N
CHO N
N
NO
C10
N
N
CHO
O
C8
O
N
N
CHO N
C5
COOH
O
N
HN
O
N
1
N
N
O
O
C7
y
CHO
pa
wa th
4
2 ay thw
pa
OH
ath wa y
3 ay
O N
N
O
Caffeinep
hw pat
N
N
CHO N
O
NO
C12
OH
C6 O
COOH
O N
N
COOH
COOH
O
O
N
N
COOH N
N
O
NO
C13
NH OH
C4
COOH HN O
CO2, H 2O, NO3-1 NH2
C14 Figure 7
Proposed degradation pathway of caffeine in DBD-goethite system.
some other intermediates must exist in between, but none of them have been reported before. 4. Conclusions DBD plasma combined with goethite was investigated for elimination of caffeine from water. The removal efficiency of
caffeine (50 mg L1) was markedly improved from 41% to 94% when goethite catalysts (2.5 g L1) were added into DBD plasma system with an input power of 75 W after 24 min. Meanwhile, the energy efficiency was enhanced by 1.6 to 2.3 times. DBD-goethite system exhibited an excellent performance for caffeine removal on degradation and energy efficiency. Reaction with hydroxyl radical was the main
556 degradation pathway of caffeine solution. Product analysis indicated that there were four transformation pathways of caffeine during the reaction. Acknowledgements This work is supported by the National Science and Technology Major Project on Water Pollution Control and Treatment (No. 2014ZX07204-008). The authors also thank for funding supports from the Project of State Key Laboratory of Pollution Control and Resource Reuse (No. PCRRF11014). The authors also would like to express our heartfelt gratitude to anonymous reviewers for their constructive advices. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jscs.2016. 08.002. References [1] A. Ashour, M.A. Hegazy, M. Abdel-Kawy, M.B. Elzeiny, Simultaneous spectrophotometric determination of overlapping spectra of paracetamol and caffeine in laboratory prepared mixtures and pharmaceutical preparations using continuous wavelet and derivative transform, J. Saudi Chem. Soc. 19 (2015) 186–192. [2] T.R. Dias, M.G. Alves, R.L. Bernardino, A.D. Martins, A.C. Moreira, J. Silva, A. Barros, M. Sousa, B.M. Silva, P.F. Oliveira, Dose-dependent effects of caffeine in human Sertoli cells metabolism and oxidative profile: relevance for male fertility, Toxicol. 328 (2015) 12–20. [3] J.L. Sotelo, G. Ovejero, A. Rodrı´ guez, S. A´lvarez, J. Gala´n, J. Garcı´ a, Competitive adsorption studies of caffeine and diclofenac aqueous solutions by activated carbon, Chem. Eng. J. 240 (2014) 443–453. [4] A. Bernabeu, R.F. Vercher, L. Santos-Juanes, P.J. Simo´n, C. Lardı´ n, M.A. Martı´ nez, J.A. Vicente, R. Gonza´lez, C. Llosa´, A. Arques, A.M. Amat, Solar photocatalysis as a tertiary treatment to remove emerging pollutants from wastewater treatment plant effluents, Catal. Today 161 (2011) 235–240. [5] G. Strauch, M. Mo¨der, R. Wennrich, K. Osenbru¨ck, H.R. Gla¨ser, T. Schladitz, C. Mu¨ller, K. Schirmer, F. Reinstorf, M. Schirmer, Indicators for assessing anthropogenic impact on urban surface and groundwater, J. Soil. Sediment 8 (2007) 23– 33. [6] R.L. Seiler, S.D. Zaugg, J.M. Thomas, D.L. Howcroft, Caffeine and pharmaceuticals as indicators of waste water contamination in wells, Ground Water 37 (1997) 405–410. [7] J.G. Yu, X.H. Zhao, H. Yang, X.H. Chen, Q. Yang, L.Y. Yu, J. H. Jiang, X.Q. Chen, Aqueous adsorption and removal of organic contaminants by carbon nanotubes, Sci. Total Environ. 482–483 (2014) 241–251. [8] A.Y. Lin, C.A. Lin, H.H. Tung, N.S. Chary, Potential for biodegradation and sorption of acetaminophen, caffeine, propranolol and acebutolol in lab-scale aqueous environments, J. Hazard. Mater. 183 (2010) 242–250. [9] W.J. Kim, J.D. Kim, J. Kim, S.G. Oh, Y.W. Lee, Selective caffeine removal from green tea using supercritical carbon dioxide extraction, J. Food. Eng. 89 (2008) 303–309. [10] P.M.K. Reddy, B.R. Raju, J. Karuppiah, E.L. Reddy, C. Subrahmanyam, Degradation and mineralization of methylene blue by dielectric barrier discharge non-thermal plasma reactor, Chem. Eng. J. 217 (2013) 41–47.
J. Wang et al. [11] P.M.K. Reddy, B. Ramaraju, C. Subrahmanyam, Degradation of malachite green by dielectric barrier discharge plasma, Water Sci. Technol. 67 (2013) 1097–1104. [12] P.M.K. Reddy, A. Dayamani, S. Mahammadunnisa, C. Subrahmanyam, Mineralization of phenol in water by catalytic non-thermal plasma reactor-an eco-friendly approach for wastewater treatment, Plasma. Process. Polym. 10 (2013) 1010–1017. [13] P.M.K. Reddy, C. Subrahmanyam, Catalytic plasma reactor for degradation and mineralization of pharmaceuticals and personal care products, J. Adv. Oxid. Technol. 18 (2015) 161–166. [14] P.M.K. Reddy, C. Subrahmanyam, Green approach for wastewater treatment-degradation and mineralization of aqueous organic pollutants by discharge plasma, Ind. Eng. Chem. Res. 51 (2012) 11097–11105. [15] P.M.K. Reddy, S. Mahammadunnisa, C. Subrahmanyam, Catalytic non-thermal plasma reactor for mineralization of endosulfan in aqueous medium: a green approach for the treatment of pesticide contaminated water, Chem. Eng. J. 238 (2014) 157–163. [16] P.M.K. Reddy, K. Krushnamurty, S.K. Mahammadunnisa, A. Dayamani, C. Subrahmanyam, Preparation of activated carbons from bio-waste: effect of surface functional groups on methylene blue adsorption, Int. J. Environ. Sci. Technol. 12 (2014) 1363–1372. [17] P.M.K. Reddy, S.K. Mahammadunnisa, B. Ramaraju, B. Sreedhar, C. Subrahmanyam, Low-cost adsorbents from biowaste for the removal of dyes from aqueous solution, Environ. Sci. Pollut. Res. 20 (2013) 4111–4124. [18] P.M.K. Reddy, P. Verma, C. Subrahmanyam, Bio-waste derived adsorbent material for methylene blue adsorption, J. Taiwan Inst. Chem. Eng. 58 (2016) 500–508. [19] H. Liu, T. Chen, R.L. Frost, An overview of the role of goethite surfaces in the environment, Chemosphere 103 (2014) 1–11. [20] X. Guo, H. Dong, C. Yang, Q. Zhan, C. Liao, F. Zha, L. Gao, Application of goethite modified biochar for tylosin removal from aqueous solution, Colloids Surf. A 502 (2016) 81–88. [21] J.X. Fan, Y.J. Wang, T.T. Fan, F. Dang, D.M. Zhou, Effect of aqueous Fe(II) on Sb(V) sorption on soil and goethite, Chemosphere 147 (2016) 44–51. [22] Z.R. Lin, X.H. Ma, L. Zhao, Y.H. Dong, Kinetics and products of PCB28 degradation through a goethite-catalyzed Fenton-like reaction, Chemosphere 101 (2014) 15–20. [23] Z. Tao, X. Wu, Y. Zhang, J. Li, T.T. Lim, Synergistic catalytic degradation of antibiotic sulfamethazine in a heterogeneous sonophotolytic goethite/oxalate Fenton-like system, Appl. Catal. B. Environ. 136–137 (2013) 294–301. [24] J. Wang, Y. Sun, J. Feng, L. Xin, J. Ma, Degradation of triclocarban in water by dielectric barrier dischargeplasma combined with TiO2/activated carbon fibers: effect of operatingparameters and byproducts identification, Chem. Eng. J. 300 (2016) 36–46. [25] R.F.P. Nogueira, M.C. Oliveira, W.C. Paterlini, Simple and fastspectrophotometric determination of H2O2 in photo-Fenton reactions usingmetavanadate, Talanta 66 (2005) 86–91. [26] L. Xin, Y. Sun, J. Feng, J. Wang, D. He, Degradation of triclosan in aqueous solution by dielectric barrier discharge plasma combined with activated carbon fibers, Chemosphere 144 (2015) 855–863. [27] R. Li, Y. Liu, W. Cheng, W. Zhang, G. Xue, S. Ognier, Study on remediation of phenanthrene contaminated soil by pulsed dielectric barrier discharge plasma: the role of active species, Chem. Eng. J. 296 (2016) 132–140. [28] U. Kogelschatz, B. Eliasson, M. Hirth, Ozone generation from oxygen and air discharge physics and reaction-mechanisms, Ozone Sci. Eng. 10 (1988) 367–377. [29] G.A.G. Sara, A.A. Amine, C. Guillerme, A. Bouzaza, D. Wolbert, Association of surface dielectric barrier discharge and
Removal of caffeine from water
[30]
[31]
[32]
[33] [34]
[35]
[36]
[37]
[38]
photocatalysis in continuous reactor at pilot scale: butyraldehyde oxidation, by-products identification and ozone valorization, Chem. Eng. J. 292 (2016) 276–283. M. Muruganandhama, J.J. Wu, Granular a-FeOOH – A stable and efficient catalyst for the decomposition of dissolved ozone in water, Catal. Commun. 8 (2007) 668–672. T. Zhang, C. Li, J. Ma, H. Tian, Z. Qin, Surface hydroxyl groups of synthetic a-FeOOH in promoting OH generation from aqueous ozone: property and activity relationship, Appl. Catal. B. Environ. 82 (82) (2008) 131–137. A. Ricard, E. Panousis, F. Clement, T. Sindzingre, J.F. Loiseau, Production of active species in a N-2 DBD plasma afterglow at atmospheric gas pressure, Eur. Phys. J. Appl. Phys. 42 (2008) 63–66. N. Wu, W. Chen, Two photon dissociation dynamics of NO2 and NO2 + H2O, J. Phys. Chem. A 116 (2012) 6894–6900. S. Yan, A. Jia, S. Merel, S.A. Snyder, K.E. O’Shea, D.D. Dionysiou, W. Song, Ozonation of cylindrospermopsin (cyanotoxin): degradation mechanisms and cytotoxicity assessments, Environ. Sci. Technol. 50 (2016) 1437–1446. M. Sui, L. Sheng, K. Lu, F. Tian, FeOOH catalytic ozonation of oxalic acid and the effect of phosphate binding on its catalytic activity, Appl. Catal. B. Environ. 96 (2010) 94–100. J. Ma, C. Zhu, J. Lu, H. Liu, L. Huang, T. Chen, D. Chen, Catalytic degradation ofgaseous benzene by using TiO2/goethite immobilized on palygorskite:preparation, characterization and mechanism, Solid. State. Sci. 49 (2015) 1–9. J. Ma, C. Zhu, Y. Xu, J. Lu, L. Huang, Z. Yang, Photocatalytic degradation of gaseous benzene with H3PW12O40/TiO2/ palygorskite composite catalyst, J. Saudi Chem. Soc. 21 (2017) 132–142. C. Zhu, X. Wang, Q. Huang, L. Huang, J. Xie, C. Qing, T. Chen, Removal of gaseous carbon bisulfide using dielectric
557
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
barrier discharge plasmas combined with TiO2 coated attapulgite catalyst, Chem. Eng. J. 225 (2013) 567–573. B. Gao, L. Liu, J. Liu, F. Yang, Photocatalytic degradation of or FeOOH doped 2,4,6-tribromophenol on Fe2O3 ZnIn2S4 heterostructure: Insight into degradation mechanism, Appl. Catal. B: Environ. 147 (2014) 929–939. D. Zhu, L. Jiang, R.L. Liu, P. Chen, L. Lang, J.W. Feng, S.J. Yuan, D.Y. Zhao, Wire–cylinder dielectric barrier discharge induced degradation of aqueous atrazine, Chemosphere 117 (2014) 506–514. A. Tiya-Djowe, S. Laminsi, G.L. Noupeyi, E.M. Gaigneaux, Non-thermal plasma synthesis of sea-urchin like a-FeOOH for the catalytic oxidation of Orange II in aqueous solution, Appl. Catal. B. 176–177 (2015) 99–106. E.G. Garrido-Ramı´ rez, B.K.G. Theng, M.L. Mora, Clays and oxide minerals as catalysts and nanocatalysts in Fenton-like reactions-a review, Appl. Clay. Sci. 47 (2010) 182–192. V.B. Avramenko, Some particularities of the formation and flow of the plasma jet of an ablation pulse plasma accelerator with cylindrical geometry, High. Temp. 51 (2013) 753–758. W. Song, W.J. Cooper, S.P. Mezyk, J. Greaves, B.M. Peake, Free radical destruction of b-blockers in aqueous solution, Environ. Sci. Technol. 42 (2008) 1256–1261. S. Rong, Y. Sun, Degradation of TAIC by water falling film dielectric barrierdischarge – Influence of radical scavengers, J. Hazard. Mater. 287 (2015) 317–324. N. Lu, Y. Lu, F. Liu, K. Zhao, X. Yuan, Y. Zhao, Y. Li, H. Qin, J. Zhu, H3PW12O40/TiO2catalyst-induced photodegradation of bisphenol A (BPA): kinetics, toxicity and degradation pathways, Chemosphere 91 (2013) 1266–1272. F. Qi, W. Chu, B. Xu, Catalytic degradation of caffeine in aqueous solutions by cobalt-MCM41 activation of peroxymonosulfate, Appl. Catal. B. 134–135 (2013) 324–332.
