Chemical Engineering Journal 210 (2012) 9–17
Contents lists available at SciVerse ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Sonochemical degradation of estradiols: Incidence of ultrasonic frequency Mauro Capocelli a,⇑, Eadaoin Joyce b, Amedeo Lancia c, Timothy J. Mason b, Dino Musmarra d, Marina Prisciandaro a a
Dipartimento di Ingegneria Industriale, dell’Informazione e di Economia, University of L’Aquila, Viale Giovanni Gronchi 18, 67100 L’Aquila, Italy Sonochemistry Centre, Coventry University, Priory Street, CV1 5FB Coventry, United Kingdom c Department of Chemical Engineering, University ‘‘Federico II’’ of Napoli, Piazzale V. Tecchio 80, 80125 Napoli, Italy d Department of Civil Engineering, Seconda Università di Napoli, Real Casa dell’Annunziata, Via Roma 29, 81031 Aversa (CE), Italy b
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
" Higher ultrasonic frequencies
increase the removal efficiency of E2, EE2 and PNP. " The efficiency is mainly related to hydroxyl radical production. " The optimal frequency depends also on the bubble collapse intensity and duration. " We model a total amount of hydroxyl radicals which agrees with experiments.
a r t i c l e
i n f o
Article history: Received 17 April 2012 Received in revised form 10 August 2012 Accepted 23 August 2012 Available online 31 August 2012 Keywords: Sonochemistry Estradiol Numerical simulation Advanced oxidation process Ultrasound frequency
a b s t r a c t This paper presents an experimental study of the ultrasonic degradation of organic pollutants in terms of the effect of ultrasonic frequency (40–380–850–1000 kHz). The removal efficiency of two endocrine disrupting compounds [17b-estradiol (E2) and 17a-ethinylestradiol (EE2)] is investigated using laboratory scale ultrasonic baths at low power intensities. Higher ultrasonic frequencies were found to be more effective for pollutant degradation with 850 kHz the best: 9.0 101 mg/kW h for E2 and 6.8 101 mg/kW h for EE2 at initial concentrations of 1 ppm. Additionally, the removal of p-nitrophenol was investigated under the same conditions, as a dosimetry reaction for estimating the hydroxyl radical production, key component in organic pollutant removal. In order to describe the overall phenomena occurring inside the reactor and to predict the apparent hydroxyl radical production, a simulation algorithm is proposed. It incorporates the solution of ODE systems that embodies bubble dynamics, heat and mass transfer through the bubble wall and chemical reactions in the gas–vapor phase for an initial bubble nuclei population. The experimental degradation measures of p-nitrophenol over the range of frequencies studied were found to be comparable with the model results of hydroxyl radicals produced during the sonication treatment. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction In recent years a number of new and complex chemical compounds have appeared in the effluent streams of wastewater processing plants as direct result of human activities; among the compounds causing most emerging concern are a group of organic ⇑ Corresponding author. Tel.: +39 0817685942. E-mail address:
[email protected] (M. Capocelli). 1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2012.08.084
pollutants such as estradiol, known as endocrine disruptor chemical (EDC). The discharge of EDCs (particularly estrogenic hormones) into the environment affects water quality and also impacts on the health of wildlife and humans [1–3]; moreover, several researchers attribute some changes observed in reproductive and developmental effects to the presence of EDCs in water [4–6]. Effluents from treatment plants are a major source of bestradiol (E2) and 17a-ethinylestradiol (EE2) across the globe because conventional wastewater treatment is not effective in
10
M. Capocelli et al. / Chemical Engineering Journal 210 (2012) 9–17
Nomenclature c C cp D EC f h I k kr K m N n p POH pOH PUS Peg Pev r R Re R t tUS tc T
velocity of sound in the medium (m/s) concentration (g/m3) specific heat (J/kg K) diffusivity (cm2/s) energy consumption (kW h/L) frequency of ultrasounds (Hz) average heat transfer coefficient (m1) intensity of ultrasounds irradiation (W/m2) thermal conductivity (W/m K) first-order kinetic constant (min1) average mass transfer coefficient (m1) mass (kg) bubble size distribution (n. bubbles/m4) concentration of chemical species in the bubble (mol/m3) pressure (Pa) total hydroxyl radical production (mol/m3) specific hydroxyl radical production (mol/n. bubbles) ultrasonic power (W) Peclet number for thermal diffusion () Peclet number for mass diffusion () reaction rate (mol/m3 s) bubble radius (m) Reynolds number (–) dimensionless gas constant (–) time (s) sonication treatment time (s) collapse period (s) temperature (K)
degrading these chemicals to concentrations at which they can no longer cause adverse effects [7–9]. It is for these reasons that new processes are urgently needed in more efficient wastewater treatment plants of the future: new technologies must be found that are capable of degrading such complex molecules into simpler ones in order to avoid the current deterioration in water quality [10]. Ultrasound has been recognized to be effective for the treatment of chemical and biological contamination in several kinds of wastewater and to be a consolidated technique in environmental protection [11–17]. The physical and chemical effects of acoustic cavitation have been well documented over the last few decades, particularly the release of free radicals attributed to the pyrolysis of water molecules during bubble collapse, a process that generates very high temperatures and pressures [18]. Due to the formation of highly reactive free molecules, in particular OH radicals and other reactive species such as hydrogen peroxide, ultrasonic cavitation has been employed as an advanced oxidation process (AOP) for wastewater treatment [10,18]. Recent investigations are suggesting a high efficiency in degrading sex hormones through ultrasonic cavitation [19,20]; in these works the effect of different parameters has been investigated in order to evaluate the guidelines in the optimization of sonochemical reactions design as an application in hormone degradation. In order to optimize chemical reactions, number of parameters, such as frequency of ultrasound, intensity of irradiation and bulk temperature of liquid, which affect the cavitational yields, need to be optimized [21]. In particular, the effect of frequency on the rates of sonochemical reactions has given rise to controversy because many experimental difficulties arise in minimizing the effect of other factors such as the reactor geometry and the source of ultrasonic field [22]. In most cases, even though with different
V v We y Y
volume of solution (L) bubble volume (m3) Weber number (-) radial coordinate in the bubble (m) cavitational yield (g/kW h)
Greek symbols a stoichiometric weight () c ratio of specific heats () l viscosity (Pa s) q density (kg/m3) s dimensionless time r surface tension of water (N/m) x0 bubble natural frequency (Hz) Superscripts 0 dimensionless value – average value Subscripts 0 initial value atm atmospheric value f final value l liquid value mix gas mixture value w bubble wall value v vapor value 1 in the bulk liquid value
experimental conditions, the existence of an optimal ultrasonic frequency in pollutant degradation has been presented in literature [21,22]. Experimental results in emerging compound degradation can be usefully coupled with a theoretical study in order to understand phenomenology issues and to allow a qualitative and quantitative optimization, as some remarkable works [23,24] demonstrated recently. For a true analysis of sonochemistry effects, complete models have to consider the complex phenomena associated with cavitation such as water condensation and vaporization, gas diffusion and chemical kinetics; this approach is the basis of several numerical simulations that gave computational efficiency and physically realistic views of the phenomena associated with cavitation bubbles [23–26]. The present work focuses on an experimental and theoretical study of sonochemistry used as an advanced oxidation process. Experimentally, the efficiency of ultrasound in the degradation of estradiol from water was evaluated at different ultrasonic frequencies, a key parameter that has not been investigated yet for this kind of pollutant. Fixing the reactor configuration and the source of ultrasound allowed the isolation of the investigated parameter’s effect. Experimental results have been discussed by comparing them with the degree of 4-nitrophenol (PNP) degradation obtained under the same experimental conditions. An algorithm has been used to solve the model equations of bubble dynamics and chemical reactions for a wide range of initial nuclei sizes. This approach has been used to estimate a total amount of hydroxyl radical production and to discuss the experimental results. The calculations were performed in Matlab™, without the help of auxiliary software. Despite the simplifying assumptions, simulation results of hydroxyl radical production correctly describe the reaction rates experimentally observed.
M. Capocelli et al. / Chemical Engineering Journal 210 (2012) 9–17
2. Experimental apparatus and procedures Two types of ultrasonic apparatus were used in this work. At 40 kHz, an ultrasonic cleaning bath (Langford Ultrasonic, Model 475TT) with the solution to be treated in a conical flask of 300 ml was used. For sonication at higher frequencies (380–850– 1000 kHz), the equipment used is outlined schematically in Fig. 1. It consists of a glass batch reactor with an inner diameter of 0.05 m equipped with a cooling jacket. An ultrasonic transducer of the same section is fitted into the base and can be coupled to different ultrasonic generators (Meinhardt Ultrasonics multi-frequency, with adjustable amplitude levels). For all experiments, the liquid volume was 250 ml and the temperature was maintained in the range 15–20 °C. A thermocouple was submerged into the reaction solution to measure the temperature. The actual ultrasonic power PUS, absorbed in the reactor, was estimated by employing the calorimetry method [27] which assumes that the sonicated medium is perfectly mixed, the acoustic intensity is uniform throughout the solution and that the thermal capacitance of the reactor can be neglected in relation to that of the liquid. The energy output of the generators was adjusted to dissipate the same ultrasonic power in the medium (finding a power intensity I = 0.3 W/cm2) at different frequencies. This approach makes the ultrasonic frequency to be the only operation parameter that affects reaction yields. The degradation of EDCs was measured using the model solutes E2 and EE2 (Sigma Aldrich, Gillingham, UK), all solvents used were of HPLC grade (Fisher Scientific, Loughborough, UK). Aqueous solutions of E2 and EE2 (1 ppm, pH 5) were prepared by dissolving the required amounts in de-ionized water. The EDCs were analyzed using a Shimadzu prominence series HPLC. The limits of detection and quantification based on signal to noise were 0.05 mg l1 and 0.03 mg l1 for E2 and 0.16 mg l1 and 0.12 mg l1 for EE2. The quantification of PNP and 4-nitrocatechol (4-NC) was achieved by measuring their respective absorbance peaks (at 401 nm and 512 nm) in alkaline solution following the method proposed by Kotronarou and coworkers [28]. PNP initial concentration was 100 lM; 1.5 ml samples were withdrawn at t = 0, 15, 30, 60 min and analyzed using UV–VIS spectrometry. In some cases additional samples were withdrawn at 45 min. Acidity of tested
11
solutions was measured and controlled at pH 5. All the experiments, previously described, were repeated more than three times and reproducible results were successfully achieved. 3. Simulation model In wastewater treatment modeling, any single cavitating bubble can be seen as a microreactor, source of radical species [21,22]. The purpose of the modeling work is to simulate bubble dynamics and chemical reactions for an initial nuclei distribution, assuming that homogeneous nucleation takes place in the liquid phase. Mathematical model mainly consists of the single bubble dynamics (SBD) and chemical reaction model, separately discussed in 3.1– 3.2, and is solved by using the algorithm sketched in Fig. 2, which is developed in Matlab™ code. The algorithm allows for the calculation of the total hydroxyl radical production by simulating the collapse of a single bubble with an initial radius R0. The time variation of bubble radius R, temperature T, pressure p and water content mv are obtained numerically thanks to the SBD for an initial size of bubble nucleus with the input parameters of water temperature Tw and pressure p1(t). The latter is given by a simple time-varying sinusoidal function (1), where the intensity I is considered to be uniform in the liquid bulk; c is the sound velocity in the liquid, f is the ultrasound frequency and patm is the atmospheric pressure.
p1 ðtÞ ¼ patm
pffiffiffiffiffiffiffiffiffiffiffi 2Iql c sinð2pftÞ
ð1Þ
Chemical reaction kinetics are solved with the assistance of the numerical results of SBD in order to simulate the production of hydroxyl radicals for each initial bubble radius and each collapse cycle. With the input of the main variables at the final stage of collapse, the chemical reaction system (Section 3.2) is solved in order to find bubble composition as a function of the run time. The main parameter used for describing experimental results is the OH radical production pOH (relative to a bubble with initial size R0). The calculations described above are repeated for a wide range of initial bubble radius. Since actual experiments reflect a distribution of bubble, it is useful to integrate the specific radical production pOH over the size distribution N(R0) as suggested in few literature works [29,30], in order to obtain the total amount of production POH per acoustic cycle:
PðOHÞ ¼
Z
R2
pOH NðR0 ÞdR0
ð2Þ
R1
where N(R0) is the initial distribution of air nuclei in water in the form proposed by Liu and Brennen [31] with nuclei concentration set to 100 nuclei/cm3, R1 and R2 are the limit of the nuclei distribution. The lower limit is the minimum bubble dimension necessary to undergo collapse, in our case R1 = 1 lm (Blake critical radius); the maximum bubble size which contributes to the transient cavitation depends on ultrasound frequency, in our case it varies from R2 = 40 lm at 1000 kHz and R2 = 800 lm at 40 kHz. Nonetheless the contribution of bigger bubbles to cavitational activity is very low (usually sonochemistry phenomena are associated to the size of resonance at a defined frequency [32]) and it is possible to reduce the total range of integration to R1 = 1 lm–R2 = 50 lm. Additionally, this limit is acceptable according to literature investigations on nuclei size distribution [31–33]: the contribution, in number per volume, given by bigger bubbles becomes negligible considering that the steady-state bubble size reflects a distribution function of the form N(R0) / 1/R30 . 3.1. Single bubble dynamics Fig. 1. Representation of experimental apparatus. R: reactor; TH: thermocouple; TR: transducer; CJ: cooling jacket; IN: cooling water inlet; OUT: cooling water outlet; G: ultrasounds generator; A: amplitude adjustment wheel.
The physical situation consists of a single spherical bubble in water, irradiated by an ultrasonic wave. Fragmentation and coales-
12
M. Capocelli et al. / Chemical Engineering Journal 210 (2012) 9–17
Fig. 2. Flow chart of the algorithm.
cence phenomena are neglected and bubble remains spherically symmetric during collapse. Pressure p is considered to be uniform both inside the bubble and in the liquid bulk. Because of the large heat capacity of water, temperature Tw of the bubble wall is constant and equal to the bulk liquid temperature that in turn is considered to be constant (293 K) and spatially uniform. The contents of the bubble are water vapor and non-condensable (gas diffusion is negligible if compared to the water vapor mass transfer for short-time bubble simulations). The reduced-order model of spherical bubble dynamics and the diffusive processes for the interior of a gas–vapor bubble consists of Eqs. (3)–(5) according to a dimensionless formulation of Preston and coworkers [34] who developed the global model by using average heat and mass transfer coefficients (6) and (7). The first Eq. (3) represents bubble wall acceleration while the second and third (4) and (5) are the time derivative of pressure and mass of water vapor inside the bubble, respectively. Parameters for non-dimensionalization are initial bubble radius R0, bubble natural frequency x0, water density ql and specific heat capacity of water cpl [34]. According to this choice, temperature is made dimensionless dividing by R20 x20 =cpl and pressure dividing by R20 x20 ql . Subscripts w denotes the bubble-wall value of variables, a hyphen denotes dimensionless variables. A detailed description of all variables is reported in the Nomenclature section.
(
@T 0 0 h ðT 0 T 0w Þ 0 @y y0¼1 @C 0 K 0 ðC 0 C 0w Þ @y0 y0¼1
ð3Þ ð4Þ ð5Þ
ð6Þ ð7Þ
The aim of this approach is to reduce a time consuming PDEs integration (which requires the solution of radial diffusion equations) in a simpler ODEs system that is still able to describe properly the bubble dynamics. Although the less numerical effort, the constant transfer model has shown a very good agreement with the full computations [34]. The method of Preston and coworkers allows to determine both heat and mass transfer coefficients by calculating the Peclet numbers [34]. Reynolds number Re, Weber number We, Peclet number for thermal diffusion Peg and for mass diffusion of air–vapor mixture Pev, are written as in Eqs. (8)-(11) according to previous considerations.
ql R20 x0 ll ql R30 x20 We ¼ r qmix cp R20 x0
Re ¼
)
4 R_ 0 ðtÞ 2 € 0 ðtÞ ¼ 1 R p00 ½p0 ðtÞ p01 ðtÞ 1:5R_ 0 ðtÞ2 0 Re R0 ðtÞ WeR0 ðtÞ R ðtÞ " # 0 _ 0v ðtÞ 3c0 0 _ 0 kw @T 0 T 00 m 0 _ ¼ 0 pðtÞ R T p ðtÞR ðtÞ v w R ðtÞ Peg R0 ðtÞ @y0 y¼1 p00 4pR0 ðtÞ2 1 p00 p0 ðtÞ 1 1 @C 0 _ 0v ¼ m Pev T 00 Rw T 0w 1 C w R0 ðtÞ @y0 y¼1
Eq. (6) shows the approximation of heat gradients: T represents the average temperature, h is the average heat transfer coefficient and y is the radial distance. A similar approximation has been is the average concentramade in Eq. (7) for mass transfer where C tion and K the average mass transfer coefficient.
Peg ¼
Pev ¼
kmix R20
x0
Dmix
ð8Þ ð9Þ ð10Þ ð11Þ
13
M. Capocelli et al. / Chemical Engineering Journal 210 (2012) 9–17
System of Eqs. (3)-(5) can be written as a system of four first or_ der ODEs, assuming RðtÞ as an additional variable and equal to the time derivative of bubble radius. It can be solved by using a Runge– Kutta fourth order adaptive step-size method to get the radius history of the bubble along with the instantaneous number of water molecules trapped in the bubble, pressure and the temperature of the bubble contents with initial conditions (12)-(15) and by calculating the values of physical and chemical properties [35].
Rð0Þ ¼ 1
ð12Þ
_ Rð0Þ ¼0
ð13Þ
pð0Þ ¼ 1
ð14Þ
mv ð0Þ ¼ 0
ð15Þ
4. Results and discussion 4.1. Estradiol degradation Fig. 3a and b report the experimental results obtained during ultrasonic treatment for estradiol degradation, in terms of removal efficiency (written as the logarithm of the dimensionless concentration C/C0) as a function of the time of treatment, for E2 (Fig. 3a) and for EE2 (Fig. 3b), respectively. These figures clearly show that ultrasound is effective in degrading estradiol and that the operating frequency influences the efficiency of sonication. The slope of straight lines fitting experimental points allows the determination of a first-order kinetic constant kr as suggested in recent works [19,20]. Table 2 summarizes main treatment results and parameters: the third and fourth rows report cavitational yield Y (Eq. (17)), a useful parameter in the evaluation of energy
3.2. Chemical reactions Starting from output values of SBD variables, radical production is obtained by solving the chemical reaction system of ODE. The chemical rate of change for each species is given by the sum of all elementary reaction rates with their corresponding stoichiometric weight.
0 -1e-1
where v is the bubble volume, ai,j is the stoichiometric weight for the i-th compound in the j-th reaction; n is concentration of the i-th specie. The bubble collapse stage is simulated as an adiabatic compression in which the species are trapped in the bubble core, thus the diffusive contribution to the change in species concentration is neglected. This simplified approach is widely used in the literature [29,30,32] with the aim of lightening the computational weight of the model. The most interesting output for this section of the model is the specific OH radical production pOH. The parameters for the Arrhenius formulation come from a consolidated reaction model [36] and are listed in Table 1. Thermodynamic equilibrium prevails until the point of minimum radius during collapse, according to the comparison of time scale of bubble collapse with the time scales of reactions [36]. Additionally, the kinetic approach of Kamath and coworkers [37] confirms that the OH production time scale is around the 1% of the sound period and is therefore comparable with the duration of the elevated gas temperature. According to above considerations, temperature for the Arrhenius formulations (Table 1) is calculated from the SBD (see Section 3.1) as the maximum value for each collapse cycle. Moreover, only a few ‘‘generation’’ of hydroxyl radicals are produced per collapse event, hence the computed number of radicals instantaneously present in the bubble at the collapse stage coincides (at least in order of magnitude) with the number of radicals actually produced during each collapse [37].
-3e-1 -4e-1 -5e-1
(a) E2
-6e-1 0 -1e-1 -2e-1
ln(C/C0)
ð16Þ
ln(C/C0)
-2e-1
n_ i ¼ v Rj ai;j r i
-3e-1 -4e-1 -5e-1
(b) EE2
-6e-1 0
10
20
30
40
50
60
Time, min Fig. 3. Removal efficiency of (a) E2 and (b) EE2 using four different ultrasonic devices at different frequencies (d) 40 kHz, (s) 380 kHz, (.) 850 kHz, (D) 1000 kHz. Initial concentration in water C0: 1 ppm.
Table 1 Arrhenius parameters for considered radical reactions: frequency factors A are given in cm3 (mol/s)1 for the two-body reactions (III, IV, VII, VIII) and in cm6 (mol2/s)1 for the three-body reactions (I, II, V, VI), Ea/k is the activation temperature in K [34]. No.
I II III IV V VI VII VIII
Reaction
O + O + M M O2 + M O + H + M M OH + M O + H2 M H + OH H + O2 M O + OH H + H + M M H2 + M H + OH + M M H2O + M OH + H2 M H + H2O OH + OH M O + H20
Forward reaction
Backward reaction
A
c
Ea/k
A
c
Ea/k
1.20 1017 5.00 1017 3.87 104 2.65 1016 1.00 1018 2.20 1022 2.16 108 3.57 104
1 1 2.7 0.7 1 2 1.5 2.4
0 0 3150 8576 0 0 1726 1062
3.16 1019 3.54 1017 1.79 104 9.00 1013 7.46 1017 3.67 1023 5.20 109 1.74 106
1.3 0.9 2.7 0.3 0.8 2 1.3 2.2
59893 51217 2200 83 52177 59980 9529 7693
14
M. Capocelli et al. / Chemical Engineering Journal 210 (2012) 9–17
Table 2 Reactor volume, cavitational yield Y for E2 and EE2 degradation (mg kW h1), first-order degradation constants (min1) and regression coefficients (R2) at investigated frequencies. Comparison with previous work [19] at 20 kHz.
Reactor volume (L) YE2 (mg/kW h) YEE2 (mg/kW h) kr (min1) [R2] for E2 kr (min1) [R2] for EE2
20 kHz bath [19] 0.6 kW
20 kHz bath [19] 2 kW
40 kHz
380 kHz
850 kHz
1000 kHz
0.200 1.55 1003 1.46 1003 6.49 1002 [0.9787] 6.47 1002 [0.9505]
0.600 1.07 1002 1.03 1002 6.48 1003 [0.8405] 6.22 1002 [0.9453]
0.250 1.60 1001 1.21 1002 1.35 1003 [0.9363] 0.97 1003 [0.9954]
0.250 4.75 1001 7.33 1002 7.19 1003 [0.9849] 5.99 1003 [0.9994]
0.250 9.00 1001 6.76 1001 8.39 1003 [0.9998] 8.89 1003 [0.9993]
0.250 1.42 1001 3.84 1001 2.19 1003 [0.9245] 3.19 1003 [0.9527]
EC, kWh/L
100
10-1
10-2 0
200
400
600
800
1000
Frequency, kHz Fig. 4. Electrical energy consumption for E2 (d) and EE2 (s). Comparison with literature data [HYPERLINK l ‘‘And 12’’ 20] for a 20 kHz commercial reactor (j) and a 20 kHz laboratory scale device (N).
efficiency of cavitational reactions [27,38], for E2 and EE2, respectively.
Y¼
m P US tUS
ð17Þ
where m is the degraded/produced mass of compounds, tUS is the sonication time and PUS is the ultrasonic power, which can be estimated by employing the calorimetry method [27]. This parameter allows the comparison with an experimental work [19] in the literature, which reports ultrasonic degradation of estrogens at different experimental conditions (20 kHz ultrasonic baths at higher levels of power density). The fifth and sixth rows in Table 2 report the estimation of a first-order kinetic constant and the comparison with the ones reported in literature. It is possible to observe an higher cavitational yield obtained at 380 and 850 kHz (second and third row) although with low first-order constant if compared with values at 20 kHz from literature [19]. This is due to a very much higher power density for the experiments reported by Suri and coworkers [19] who obtained higher degradation level with a higher power consumption. The effect of frequency, as Table 2 reports, can be analyzed in depth by looking at the electrical energy consumption (EC) of treatments, defined in Eq. (18) and reported in Fig. 4 for our work and for two literature experiments [20]:
EC ¼
PUS t US
V log
Ci Cf
ð18Þ
where PUS is the electrical power spent in the treatment time tUS in order to reach a final concentration Ci from an initial concentration Cf, in a batch reactor of volume V. In Fig. 4, continuous and dotted lines follow the results for E2 and EE2 respectively, while two iso-
lated points represent data from literature [20] for two apparatus (20 kHz commercial and laboratory scale ultrasonic devices) employed for hormones degradation. The latter seem to agree with the hypothetical tendency in Fig. 4, although they show lower energy consumption if compared with 40 kHz reactor. Even so, observing the global meaning of Fig. 4, our results underline the possibility of increasing the energy efficiency by choosing an optimal frequency range, as suggested in literature for similar compounds [38]. Direct experimental evidence shows that there is very little reaction between hydrogen peroxide and estradiol in the bulk solution, which was also confirmed by Belgiorno and coworkers [39] who established OH to be main oxidative species by performing experiments in the presence of radical scavengers. Moreover, the importance of hydroxylation becomes clear when analyzing the most probable degradation mechanism: E2 and EE2 are not degraded by pyrolysis in the bubble core due to their low fugacity, but the main reaction is hydroxyl-mediated and it occurs at the bubble–liquid interface (bulk liquid reactions are disfavored due to the hydrophobic character of the pollutant molecules). Latter considerations indicate that higher degradation rates at higher frequencies should be related to a better radical production. On the basis of assumptions hitherto mentioned, it is clear that the explanation of the frequency effect has to be found in the valuation (experimentally and numerically) of the hydroxyl radical production.
4.2. Numerical results and p-nitrophenol dosimetry Figs. 5a, b and 6a, b show the results of the SBD model at two different level of frequency in order to allow a first qualitative analysis of cavitation phenomena. Fig. 5a and b show dimensionless radius (R0 ) and pressure (p0 ) variations versus dimensionless time (s) for 300 and 600 kHz ultrasonic frequency f, respectively. It is possible to observe how the growing stage is smaller with higher pressure peaks at f = 600 kHz. Fig. 6a and b shows the dimensionless temperature (T0 ) and vapor mass (m0v ) history for f = 300 kHz and f = 600 kHz, respectively. These parameters are fundamental in the discussion of the experimental results because the extent of production of radicals through cavitation bubbles depends mainly on the amount of water vapor trapped in the bubble during transient collapse of the bubble and the temperature peak reached in the bubble during collapse [23,24]. The mass of water inside the bubble is higher at lower frequencies, due to the longer available time for the bubble to grow until the highest pressure is recovered; on the other hand temperature peaks are higher at 600 kHz (Fig. 6a and b): from a quantitative point of view, at the low frequency the number of water molecules trapped in the bubble at the instance of first collapse is 7.66 107 and the maximum temperature peak reached is 4016 K; at the high frequency the number of water molecules trapped in the bubble is 4.84 107 and the temperature peak is 6024 K. Hence, at higher frequencies, hydroxyl radical production should be enhanced due
15
M. Capocelli et al. / Chemical Engineering Journal 210 (2012) 9–17 1e+2
(a) 300 kHz
14
2e+0
1e+0
1e+0
5e-1
Pressure, p'
Radius, R'
1e+1 2e+0
Mass removed, %
12 10
15 min 30 min 45 min 60 min
8 6 4
1e-1 2 0
1e+2
(b) 600 kHz
2e+0
1e+0
1e+0
5e-1
Pressure, p'
Radius, R'
1e+1
1e-1
0
10
20
30
Time, τ Fig. 5. Dimensionless radius (continuous line) and pressure (broken line) inside a bubble of initial radius R0 = 1 lm, frequency f = 300 kHz (a) and 600 kHz (b); US intensity I = 0.5 W cm2.
(a) 300 kHz
1e-4
1e+4
1e-4 1e+3
Temperature, T'
1e+5
2e-4
Water vapor mass, mv'
380
850
1000
Frequency, kHz
2e+0
5e-5 1e+2
0
(b) 600 kHz
1e-4
1e+4
1e-4 1e+3
Temperature, T'
1e+5
2e-4
Water vapor mass, mv'
40
5e-5 1e+2
0 0
10
20
30
Time, τ Fig. 6. Dimensionless vapor mass content (continuous line) and temperature (broken line) inside a bubble of initial radius R0 = 1 lm, frequency f = 300 kHz (a) and 600 kHz (b); US intensity I = 0.5 W cm2.
to more violent (higher pressure and temperature peaks) and numerous collapse cycles; additionally in such conditions OH radicals could be ejected more efficiently in solution before they have
Fig. 7. P-nitrophenol degradation as a function of frequency for different treatment times. Initial concentration in water C0: 100 lM.
time to combine in the bubble cavity thanks to more rapid pulsation and collapses [22,39]. The collapse time for the resonant bubble at different frequency is 5.1, 0.5, 0.22 and 0.19 ls for 40, 380, 850 and 1000 kHz respectively. However, with shorter acoustic period at higher frequency, the size of cavitating bubbles decreases; these phenomena reduce the amount of water molecules trapped in the bubble and, consequently, the number of radicals which could be ejected. The optimal frequency, experimentally found in pollutants degradation, is a direct consequence of these two antagonistic phenomena [22]: the effective oxidant production, which is higher at high frequency because of more numerous and intensive collapses, and the real availability of them, together with pollutant molecules, which is lower at higher frequencies. An extensively investigated pollutant [22,28,40,41], p-nitrophenol, has been chosen as an indicator of the apparent radical concentration (a dosimeter) in order to investigate the features of hydroxylation, which appears to be the main degradation mechanism for estradiol. Then, in order to understand the quantitative effect of frequency on the oxidative process, it is useful to look at the experimental results obtained for PNP degradation, shown in Fig. 7. It reports the mass removal (as a percentage) at different times for the investigated frequencies. Results with the ultrasonic bath at 850 kHz show the highest removal efficiency with cavitational yield (17) of 7.54 101 mg/kW h. Globally, the results indicate that the removal efficiency increases as frequency increases with a decrease in the range of highest frequencies. This is in agreement with the indication of SBD simulations and several results mentioned in literature [40–46]. Experimental results can be used to validate the outputs of the proposed mathematical algorithm in predicting hydroxyl radical production. High temperature reactions inside the cavitating bubble are discounted because of the low vapor pressure of PNP, the most plausible degradation mechanism, particularly at low solute concentration) involves the OH radical attack in the gas–liquid shell [28]. In this extremely rapid reaction (at pH 5) 70% of radicals are expected to attack PNP to produce 4-nitrocathecol; other side reactions lead to the formation of NO 2 , and to a lesser extent of p-BC and hydrochinone. The cited mechanism appears to be controlled by the transport of substrate into the interface (diffusioncontrolled) [28]. This observation suggests that PNP degradation as a function of run time, reflects the interface availability of oxidant and pollutant at the collapse stage and could be strongly influenced by the collapse time. This latter can influence the diffusion of PNP which is considered the rate-limiting step, while the reaction of substrate and radicals appears to occur very fast [22,28,40].
16
M. Capocelli et al. / Chemical Engineering Journal 210 (2012) 9–17
14
40 kHz
380 kHz
850 kHz
1000 kHz
Concentration, μM
12 10 8 6 4 2 0 14
Concentration, μM
12 10 8 6 4 2 0 0
10
20
30
40
50
60
0
10
20
Time, min
30
40
50
60
Time, min
Fig. 8. NC production (d), PNP degradation (s), numerical simulation of OH production (line) for different ultrasonic frequencies and US intensity I = 0.3 W cm2.
Fig. 8 shows the numerical prediction of hydroxyl radical production (straight line) as well as degradation of PNP (expressed as lM removed) and production of 4-NC (expressed as lM produced). The latter is an additional indicator of OH production, since it is the main PNP hydroxylation products. Looking at the global indications in Fig. 8, numerical and experimental results are in very good agreement: theoretical radical production is higher than 4-nitrocatechol production and is comparable with PNP degradation. The OH concentration has to be higher than 4-NC, mainly because catechol is only one among many different reaction byproducts and can be degraded as well [28]; this agrees with numerical prediction of hydroxyl radical production. Indeed, the model may give a maximum theoretical limit of 4-NC production, if one considers nitrocathecol as the indicator of collisions between hydroxyl radicals, migrating over the fragmenting bubbles into the bulk, and PNP molecules. According to the model predictions, the decreasing of oxidation rate, observed form 850 kHz to 1000 kHz, is not directly addressed to a lower theoretical radical production: as aforementioned and extensively pointed in literature [22,29,46], the degradation efficiency is additionally influenced by the radical ejection and transport from the bubble; the optimum, found near 850 kHz for the investigated compounds (estradiol as well as p-nitrophenol), is a consequence of two antagonistic phenomena, namely the rate of radical production and the duration of cavity collapse that affects their actual availability. This feature is similar for the degradation of tested EDCs, which as mentioned in Section 4.1 occurs through radical reactions in the vapor–liquid interface. The decrease in degradation at 1000 kHz for E2 and EE2 is even more accentuated, the behavior at this frequency is one of the worst, comparable with the 40 kHz bath. This is due to the low fugacity of estradiol (vapor pressure is ca. 108 Pa for E2 and 102 Pa for PNP at 30 °C), which
does not allow the migration of EDCs at interface during such fast collapses relative to 1000 kHz. Furthermore, and in conclusion, it should be noted that the described reaction mechanism is particular attractive to explain the degradation qualitatively, but it is a simplified model which does not take into account the contribution of pyrolysis in the hot core and bulk liquid reactions.
5. Conclusions The paper studies the effectiveness of ultrasound in degrading aromatic compounds such as estradiol, addressing the effect of frequency and isolating it from the influence of other operating parameters. Additionally, hydroxyl radical production, which exerts a prevailing role in degradation, has been estimated by p-nitrophenol degradation and by numerical simulations. The experimental results show increasing sonication efficiency with increasing ultrasonic frequency for both estradiol and p-nitrophenol; this can be described in terms of higher OH production at higher frequencies. Additionally it has been shown the existence of an optimal frequency, for the maximum oxidation efficiency that depends on radical production as well as radical ejection, transport mechanisms of the target compounds and effective reaction zone. Moreover, this paper provides a physical insight into the sonochemical degradation of aromatics compounds such as estradiol and p-nitrophenol by proposing a mathematical model. The numerical simulation of the model gives an explanation of the degradation mechanism in conjunction with bubble dynamics behavior (qualitative) and with hydroxyl radical production (quantitative). The proposed approach, which takes into account both for the chemical mechanism and the dynamics of the
M. Capocelli et al. / Chemical Engineering Journal 210 (2012) 9–17
evolution of nuclei population, finds a direct correlation with experimental results and offers a benchmark for interpreting experimental results. References [1] E. Diamanti-Kandarakis, J. Bourguignon, L.C. Giudice, R. Hauser, G.S. Prins, A.M. Soto, R.T. Zoeller, A.C. Gore, Endocrine-disrupting chemicals: an endocrine society scientific statement, Endocr. Rev. 30 (2009) 293–342. [2] T.A. Hanselman, D.A. Graetz, A.C. Wilkie, Manure-borne estrogens as potential environmental contaminants: a review, Environ. Sci. Technol. 37 (2003) 5471– 5478. [3] G.G. Ying, R.S. Kookana, Y. Ru, Occurrence and fate of hormone steroids in the environment, Environ. Int. 28 (2002) 545–551. [4] D.G.J. Larsson, M. Adolfsson-Erici, J. Parkkonen, M. Pettersson, A.H. Berg, P.-E. Olsson, L. Forlin, Ethinylestradiol – an undesired fish contraceptive?, Aquat Toxicol. 45 (1999) 91–97. [5] E.F. Orlando, L.J. Guillette Jr., Sexual dimorphic responses in wildlife exposed to endocrine disrupting chemicals, Environ. Res. 104 (2007) 163–173. [6] L. Santos, Ecotoxicological aspects related to the presence of pharmaceuticals in the aquatic environment, J. Hazard. Mater. 175 (2010) 45–95. [7] M. Auriol, Y. Filali-Meknassi, R.D. Tyagi, C.D. Adams, R.Y. Surampalli, Endocrine disrupting compounds removal from wastewater, a new challenge, Process Biochem. 41 (2006) 525–539. [8] M. Esperanza, M.T. Suidan, R. Marfil-Vega, C. Gonzalez, G.A. Sorial, P. McCauley, R. Brenner, Fate of sex hormones in twopilot-scalemunicipal wastewater treatment plants: conventional treatment, Chemosphere 66 (2007) 1535– 1544. [9] R. Kanda, J. Churchley, Removal of endocrine disrupting chemicals during conventional wastewater treatment, Environ. Technol. 29 (2008) 323–325. [10] P.G. Gogate, A.B. Pandit, A review of imperative technologies for wastewater treatment I: oxidation technologies at ambient conditions, Adv. Environ. Res. 8 (2004) 501–551. [11] T.J. Mason, E. Joyce, S.S. Phull, J.P. Lorimer, Potential uses of ultrasound in the biological decontamination of water, Ultrason. Sonochem. 10 (2003) 319–323. [12] X. Wu, E.A. Joyce, T.J. Mason, The effects of ultrasound on cyanobacteria, Harmful Algae 10 (2011) 738–743. [13] T.J. Mason, Sonochemistry and the environment – providing a ‘‘green’’ link between chemistry, physics and the engineering, Ultrason. Sonochem. 14 (2007) 476–483. [14] E. Gonze, L. Fourel, Y. Gonthier, P. Boldo, A. Bernis, Wastewater pretreatment with ultrasonic irradiation to reduce toxicity, Chem. Eng. J. 73 (1999) 93–100. [15] K.P. Gopinath, K. Muthukumar, M. Velan, Sonochemical degradation of Congo red: optimization through response surface methodology, Chem. Eng. J. 157 (2010) 427–433. [16] S. Merouani, O. Hamdaoui, F. Saoudi, M. Chiha, Sonochemical degradation of rhodamine B in aqueous phase: effects of additives, Chem. Eng. J. 158 (2010) 550–557. [17] D.T. Sponza, R. Oztekin, Removals of PAHs and acute toxicity via sonication in a petrochemical industry wastewater, Chem. Eng. J. 162 (2010) 142–150. [18] T.J. Mason, J.P. Lorimer, Sonochemistry: Theory, Applications and Uses of Ultrasound in Chemistry, Ellis Horwood, New York, 1989. [19] R.P.S. Suri, M. Nayak, U. Devaiah, E. Helming, Ultrasound assisted destruction of estrogen hormones in aqueous solution: effect of power density, power intensity and reactor configuration, J. Hazard. Mater. 146 (2007) 472–478. [20] G. Andaluri, E.V. Rokhina, R.P.S. Suri, Evaluation of relative importance of ultrasound reactor parameters for the removal of estrogen hormones in water, Ultrason. Sonochem. (2012), http://dx.doi.org/10.1016/j.ultsonch.2011.12.005. [21] V.S. Sutkar, P.G. Gogate, Design aspects of sonochemical reactors: techniques for understanding cavitational activity distribution and effect of operating parameters, Chem. Eng. J. 155 (2009) 26–36. [22] C. Petrier, A. Francony, Ultrasonic waste-water treatment: incidence of ultrasonic frequency on the rate of phenol and carbon tetrachloride degradation, Ultrason. Sonochem. 4 (1997) 295–300.
17
[23] T. Sivasankar, V.S. Moholkar, Physical insights into the sonochemical degradation of recalcitrant organic pollutants with cavitation bubble dynamics, Ultrason. Sonochem. 16 (2009) 769–781. [24] T. Sivasankar, V.S. Moholkar, Mechanistic approach to intensification of sonochemical degradation of phenol, Chem. Eng. J. 149 (2009) 57–69. [25] L. Csoka, S. Katekhaye, P.R. Gogate, Comparison of cavitational activity in different configurations of sonochemical reactors using model reaction supported with theoretical simulations, Chem. Eng. J. 178 (2011) 384–390. [26] R. Jamshidi, B. Pohl, U.A. Peuker, G. Brenner, Numerical investigation of sonochemical reactors considering the effect of inhomogeneous bubble clouds on ultrasonic wave propagation, Chem. Eng. J. 189–190 (2012) 364–375. [27] T.J. Mason, J.P. Lorimer, D.M. Bates, Quantifying sonochemistry: casting some light on a ‘black art’, Ultrasonics 30 (1992) 40–42. [28] A. Kotronarou, G. Mills, M.R. Hoffmann, Ultrasonic irradiation of p-nitrophenol in aqueous solution, J. Phys. Chem. 95 (1991) 3630–3638. [29] A.J. Colussi, L.K. Weavers, M.R. Hoffmann, Chemical bubble dynamics and quantitative sonochemistry, J. Phys. Chem. A 102 (1998) 6927–6934. [30] M. Capocelli, M. Prisciandaro, A. Lancia, D. Musmarra, Modeling of cavitation as an advanced wastewater treatment, Desalination Water Treat. (2012), http://dx.doi.org/10.1080/19443994.2012.705094. [31] Z. Liu, C.E. Brennen, Cavitation nuclei population and event rates, J. Fluid Eng. 120 (1998) 728–737. [32] C.E. Brennen, Cavitation and Bubble Dynamics, Oxford University Press, New York, 1995. [33] Y. Iida, M. Ashokkumar, T. Tuziuti, T. Kozuka, K. Yasui, A. Towata, J. Lee, Bubble population phenomena in sonochemical reactor: II. Estimation of bubble size distribution and its number density by simple coalescence model calculation, Ultrason. Sonochem. 17 (2010) 480–486. [34] T. Preston, T. Colonius, C.E. Brennen, A reduced-order model of diffusive effects on the dynamics of bubble, Phys. Fluids 19 (2007) (2007) 1–12. [35] D. Green, R. Perry, Perry’s Chemical Engineers’ Handbook, Eighth ed., McGrawHill, New York, 2007. [36] X. Lu, A. Prosperetti, R. Toegel, D. Lohse, Harmonic enhancement of single bubble sonoluminescence, Phys. Rev. E 67 (2003) 1–9. [37] V. Kamath, A. Prosperetti, F.N. Egolfopoulos, A theoretical study of sonoluminescence, J. Acoust. Soc. Am. 94 (1993) 248–260. [38] Y. Son, M. Lim, J. Khimc, L.K. Kim, M. Ashokkumar, B. Pandit, Comparison of calorimetric energy and cavitation energy for the removal of bisphenol-A: the effects of frequency and liquid height, Chem. Eng. J. 183 (2012) 39–45. [39] V. Belgiorno, L. Rizzo, D. Fatta, C. Della Rocca, G. Lofrano, A. Nikolau, V. Naddeo, S. Meric, Review on endocrine disrupting-emerging compounds in urban wastewater: occurrence and removal by photocatalysis and ultrasonic irradiation for wastewater reuse, Desalination 215 (2007) 166–176. [40] C. Petrier, M.F. Lamy, A. Francony, A. Benahcene, B. David, Sonochemical degradation of phenol in dilute aqueous solutions: comparison of the reactions rates at 20 and 487 kHz, J. Phys. Chem. 98 (1994) 10514–10520. [41] P. Chowdhury, T. Viraraghavan, Sonochemical degradation of chlorinated organic compounds, phenolic compounds and organic dyes – a review, Sci. Total Environ. 407 (2009) 2474–2492. [42] M.H. Entezari, P. Kruus, Effect of frequency on sonochemical reactions II. Temperature and intensity effect, Ultrason. Sonochem. 3 (1996) 19–24. [43] I. Hua, M.R. Hoffmann, Optimization of ultrasonic irradiation as an advanced oxidation technology, Environ. Sci. Technol. 31 (1997) 2237–2243. [44] M.H. Entezari, C. Petrier, P. Devidal, Sonochemical degradation of phenol in water: a comparison of classical equipment with a new cylindrical reactor, Ultrason. Sonochem. 10 (2003) 103–108. [45] D. Kobayashi, C. Honma, A. Suzuki, T. Takahashi, H. Matsumoto, C. Kuroda, K. Otake, A. Shono, Comparison of ultrasonic degradation rates constants of methylene blue at 22.8 kHz, 127 kHz, and 490 kHz, Ultrason. Sonochem. 19 (2012) 745–749. [46] R. Kidak, N.H. Ince, Effects of operating parameters on sonochemical decomposition of phenol, J. Hazard. Mater. B137 (2006) 1453–1457.
