Abstract

The removal of polycyclic aromatic hydrocarbons (PAHs) in activated sludge was evaluated using two lab-scale bioreactors, coupled or not with a disintegration system (sonication). Mass balances performed on each system underlined that PAHs removal was significantly improved after sludge disintegration, especially for the higher molecular weight PAHs studied, which tended to adsorb to suspended matter. A model was developed in order to study the effect of sludge disintegration on the content of dissolved and colloidal matter (DCM), and to predict the potential impacts on PAHs availability and degradation. Results showed that this new model was efficient for capturing apparent degradation improvement trends and for discriminating between the involved mechanisms. This study showed that DCM content increased after sludge disintegration, and proved to be the main driver for improving PAHs apparent degradation.

INTRODUCTION

The presence of micropollutants in aquatic environments is an issue of growing concern. In the last decades, intensive work has been focusing on the occurrence and removal of these pollutants over a large range of municipal wastewater treatment plants (Choubert et al. 2011; Martin Ruel et al. 2011, 2012; Grandclément et al. 2017; Liu et al. 2017; Qiao et al. 2017). Removal of PAH (and micropollutants in general) within activated sludge can be related to three main biotic or abiotic phenomena: (1) volatilization by stripping (governed by the physical properties of the micropollutant through Henry's law constant), (2) sorption to the sludge due to hydrophobic interactions between pollutants and suspended solids that lead to removal through sludge wastage and (3) biotic transformation, including both complete biodegradation or transformation into byproducts, lead to the effective elimination of organic compounds, but can also induce an apparent elimination via the formation of bound residues, rendering them inaccessible to chemical extraction, as has been demonstrated on soils (Kästner et al. 1999). Abiotic degradation processes such as photolysis or cleavage of chemical bonds are not expected to be significant in WWTP (Margot et al. 2015). Modelling the fate of these hazardous substances in wastewater effluents has now become an important topic to study (Plósz et al. 2013). Modelling represents a useful tool for assessing the design and control strategies that aim to improve the removal of micropollutants. However, extending the existing Activated Sludge Models (ASM) by including micropollutants is still seldom done. The current status and challenges regarding the modelling of micropollutants in biological wastewater treatment processes have been reviewed (Plósz et al. 2013; Pomiès et al. 2013). Pomiès et al. (2013) provided a detailed insight into removal mechanisms (volatilization, sorption, biodegradation), mathematical equations and parameter values, thereby revealing the need for more confrontation between simulation and experimental data in order to better calibrate and validate the models (Pomiès et al. 2013).

The biodegradation kinetics of polycyclic aromatic hydrocarbons (PAHs) have been studied, but due to a large variety of bacterial biomasses (either pure or mixed culture), different treatment systems, and a wide range of different PAHs concentrations and wastewater mixtures, the comparison between studies was rather difficult. For instance, Zielińska et al. (2012) determined the removal rate of phenanthrene by a nitrifying biomass and by membrane bioreactor (MBR) and Mozo et al. (2011) determined the removal rates of acenaphthylene, acenaphthene, phenanthrene, anthracene and fluoranthene by MBR and the biodegradation rate of PAHs decreases with increasing molecular weight (Mozo et al. 2011). Indeed, high molecular weight PAHs are more slowly biodegraded (Kanaly & Harayama 2010), and this is partly explained by the fact that such molecules are likely to adsorb to suspended matter, which reduces its bioavailability.

Degradation of these hydrophobic compounds is driven by their interaction with other organic compounds, either soluble, colloidal, or particulate. The effect of adsorption on particulate matter reduces the bioavailability and hence the biodegradation rate. Stringfellow & Alvarez-Cohen (1999) evaluated the relationship between adsorption of PAHs to bacterial biomass and its biodegradation. Barret et al. (2010a) assessed the overall removal of 13 PAHs during anaerobic digestion and highlighted the need to consider adsorption on particulate matter, but also on soluble and colloidal organic matter, which can on the other hand increase the bioavailability. A three compartment model was suggested, and successfully applied in an anaerobic digestion model by Barret et al. (2010b), by considering that the PAHs fraction present in the liquid phase, possibly adsorbed to dissolved and colloidal matter (DCM), was bioavailable. Indeed, Johnsen & Karlson (2004) also showed that bioavailability of PAHs (fluorene, phenanthrene, fluoranthene and pyrene) was increased with the presence of dissolved organic matter excreted in the liquid phase, such as bacterial biosurfactants or extracellular polymeric substances.

Modelling PAHs biodegradation kinetics in aerobic activated sludge (AS) systems is relatively seldom done. Guha et al. (1999) estimated the Monod kinetic parameters and yield coefficients for naphthalene, phenanthrene and pyrene in aerobic batch liquid systems inoculated with a mixed bacterial culture and Moscoso et al. (2012) estimated the specific growth rate and the specific degradation rate for phenanthrene, pyrene, and benzo(a)pyrene by a consortium of bacteria. However, kinetics are generally described by a simplified first-order model (Tang et al. 2005; Manoli & Samara 2008). The acceleration of the PAHs biodegradation kinetics due to the presence of DCM has already been reported in the literature (Holman et al. 2002; Smith et al. 2009; Ke et al. 2018). As a consequence, sludge disintegration or solubilization, from particle to colloidal forms, is expected to improve the bioavailability and biodegradation of PAHs. Moreover, Mozo et al. (2011) already showed that high shear stress could release PAHs initially adsorbed to suspended solids and hence increase the concentration of PAHs in the liquid phase (free or adsorbed to DCM). This study showed that residual PAHs measured in the wasted sludge was reduced and hence confirmed the positive effect of sludge disintegration on the bioavailability of hydrophobic micropollutants in the MBR. In this perspective, modifications of the actual ASM were suggested for modelling sludge reduction processes either by Cannibal™ process (Johnson et al. 2008) or by thermal treatment (Laurent et al. 2011).

The main goal of this present study is to extend the existing ASM1 in order to define the degradation of 16 PAHs stated as hazardous in the EU legislation. In addition, the effect of sludge disintegration and DCM content were both included and validated by confronting the model with experimental data obtained in two activated sludge (AS) systems run in parallel: one simple AS and one AS coupled with a sludge disintegration module (sonication). Experiments were here designed in order to test the ability of the new model to capture the effect of sludge disintegration on the removal of PAHs.

MATERIALS AND METHODS

Experimental set-up and operational conditions

Two lab-scale activated sludge processes were operated in parallel, fed with the same synthetic wastewater, during 96 days: a conventional activated sludge system (AS, reactor 1) and an activated sludge system coupled to an ultrasound loop (US-AS, reactor 2). Cylindrical bioreactors (20 cm inner diameter, 11 L volume) were equipped with fine bubble diffuser, Rushton impellers (100 rpm) and water jacket (temperature was controlled at 20 °C). The settlers had a volume of 2.5 L. Reactors, contactors and settlers were made out of glass and the piping, as other elements, were made in PTFE. Peristaltic pumps were used for feeding and recirculation (100% recirculation). Temperature and dissolved oxygen (DO) were monitored (optical probe, Hamilton) as well as pH (WTW probe). pH was naturally stabilized at 7.5 ± 0.5. Air flow rate was imposed to maintain DO at 3.5 ± 0.5 mg·L−1. On the second reactor (US-AS), an ultrasonic device (SONOPULS HD 2070, Bandelin Electronic) was combined with the activated sludge process.

The bioreactors were seeded with secondary sludge which originated from an activated sludge process receiving petrochemical effluent (Petrol refinery), in which bacteria acclimatised to PAHs were already present. After 20 days of similar operation, the ultrasonication system (US) was used in the second reactor. In this system, a volume of the mixed liquor was regularly (3 cycles of 21 minutes per day) pumped to a 300 mL glass side stream vessel, where the tip of a TT13 ultrasonic probe (operating frequency of 20 kHz and supplied power of 70 W) was immerged. The provided power density was 0.23 W·mL−1. The sonicated sludge returned directly to the bioreactor after sonication. The treatment frequency by ultrasounds (defined as the ratio between the amount of sludge treated per day over the amount of sludge in the system) was 0.08 d−1 and a sludge recycle ratio (volume of sludge sonicated per day over the volume of wastewater treated per day) of 0.08 L·L−1 was applied, leading to 30% of suspended solids solubilisation.

The synthetic wastewater was made with: (1) a solution containing sodium acetate and ethanol stored at 4 °C and prepared with tap water, (2) a solution containing carbonate, ammonium, salts, buffer solution and oligo-elements prepared with tap water and (3) the PAHs solution prepared in methanol. The PAHs concentrations were chosen in order to mimic a real industrial wastewater comparable to different literature references (Sponza & Oztekin 2010; Sponza & Gok 2011). The inlet flow rates of each solution were 2.1 L·d−1, 8.9 L·d−1 and 0.00415 L·d−1 respectively (total effluent flow rate was 11 L·d−1 corresponding to 24 hours' hydraulic retention time). PAHs represented 0.17% of the total inlet chemical oxygen demand (COD) (1,325 mg·L−1) and the organic loading rate was 1.06 kgCOD·m−3·d−1. The PAHs solution was stored and injected with a specific device (Dosimat) for accurate control of very low flow rate. The resulting concentrations in the reactors feed are summarized in Table 1.

Table 1

PAHs concentrations in the influent

Compound
 
Conc. (μg·L−1Compound   Conc. (μg·L−1
NAP Naphthalene 389 B(a)Ant Benzo(a)anthracene 0.84 
ACY Acenaphthylene 65.6 CHRYS Chrysene 0.84 
ACE Acenaphthene 99.1 B(b)Fluo Benzo(b)fluoranthene 0.67 
FLUOREN Fluorene 72.4 B(k)Fluo Benzo(k)fluoranthene 0.74 
PHEN Phenanthrene 22.9 B(a)Pyr Benzo(a)pyrene 0.74 
ANT Anthracene 22.9 DiB(ah)Ant Dibenzo(ah)anthracene 0.58 
FLUORAN Fluoranthene 22.1 B(ghi)Per Benzo(ghi)perylene 0.66 
PYR Pyrene 20.6 I(123cd)Pyr Indeno(123cd)pyrene 0.04 
Compound
 
Conc. (μg·L−1Compound   Conc. (μg·L−1
NAP Naphthalene 389 B(a)Ant Benzo(a)anthracene 0.84 
ACY Acenaphthylene 65.6 CHRYS Chrysene 0.84 
ACE Acenaphthene 99.1 B(b)Fluo Benzo(b)fluoranthene 0.67 
FLUOREN Fluorene 72.4 B(k)Fluo Benzo(k)fluoranthene 0.74 
PHEN Phenanthrene 22.9 B(a)Pyr Benzo(a)pyrene 0.74 
ANT Anthracene 22.9 DiB(ah)Ant Dibenzo(ah)anthracene 0.58 
FLUORAN Fluoranthene 22.1 B(ghi)Per Benzo(ghi)perylene 0.66 
PYR Pyrene 20.6 I(123cd)Pyr Indeno(123cd)pyrene 0.04 

Analytical methods

Sludge samples were separated by centrifugation (8,200 g, 20 min, 20 °C) followed by filtration through a glass fibre membrane (1.2 μm pore size). Two fractions were hence defined: a particulate fraction (solids after centrifugation and the residual particulate matter in the filter) and a liquid fraction containing DCM. Total solids, volatile solids and COD in total sludge, particulate phase and liquid phase were determined using the standard method (Afnor 1994) following the NFT 90101 and NFT 90105 for TSS/VSS and COD measures respectively. The particle size distribution of aggregates was analysed using a laser diffraction technique (Mastersizer, Malvern 2000, UK) with a particle size ranging from 0.02 μm to 2,000 μm.

Total PAHs concentrations and PAHs concentrations in aqueous phase were measured by GC-MS following the NF EN ISO 17993 (quantification limit Q.L. = 0.05 μgPAH·L−1). The PAHs concentration in the solid phase was measured on 1 g of solids after freeze-drying and applying an automatic solvent extraction method with dichloromethane (Soxtec HT2 Tecator, Foss). Analysis of each extracted sample was performed with a GC-MS analyser (Gas chromatograph VARIAN 3,900) equipped with a mass spectrometer detector (VARIAN 2,100 SaturnT) allowing the electronic impact and operating under ‘Full Scan’ acquisition mode. Liquid samples were analysed by the automatic headspace technique for volatile compounds (naphthalene, acenaphthylene, acenaphthene) and solid phase micro-extraction (SPME) with 7-μm polydimethylsiloxane (PDMS) SPME fibres for other PAHs. Calibration was performed with deutered PAHs semi-volatile internal standard mix reaching a detection limit of 0.1 μg·L−1

Model description

All simulations and parameter identifications were performed using AQUASIM software (Reichert 1994). The standard ASM1 model (Henze et al. 2000) was extended, as shown in Figure 1, in order to consider the fate of PAHs and sludge disintegration (see Table SI-1 for the stoichiometry and Table SI-2 for the kinetics). Each compound (16 PAHs) concentration was added as new variable (TPAH). For each compound, the total concentration (TPAH) was divided into three fractions: the free dissolved PAH (SPAH), the PAH adsorbed to particulate matter (XPAH) and the PAHs adsorbed to DCM (CPAH). Considering PAHs adsorption as rapid mechanisms, partition was predicted with a set of specific equilibrium constants for each molecule (KPART and KDCM) as proposed by Barret et al. (2010b). The different concentrations can be determined by the set of following equations: 
formula
(1)
 
formula
(2)
 
formula
(3)

Considering the literature studying the disintegration processes, it was also suggested to include a colloidal fraction for both biodegradable and unbiodegradable organic matter (Guellil et al. 2001; Laspidou & Rittmann 2002; Jiang et al. 2008). Hence biodegradable colloids and unbiodegradable colloids are considered. The model considers that are hydrolysed and transformed into particulate slowly biodegradable compounds and then into soluble readily biodegradable substrate . The hydrolysis rate is modelled by a first order kinetic (Laspidou & Rittmann 2002; Jiang et al. 2008), with a first order constant qHyd. The are rapidly re-adsorbed into endogenous residue particles following a first order kinetic with a rate constant qADS as proposed by Guellil et al. (2001).

New kinetic processes described the fate of PAHs, including biodegradation (processes 9–24), volatilization (25–41) and sludge disintegration by sonication (42–47). Processes (1) to (41) are referred to in the following text as model ASM1-PAH (for conventional AS, i.e. reactor 1) while model ASM1-PAH-DCM also includes processes (42) to (47) (for US-AS, i.e. reactor 2).

PAHs biodegradation

A first order kinetic expression was used for PAHs biodegradation. It was assumed that ordinary heterotrophic organisms can grow with both the easily biodegradable substrate and the specific substrate. PAHs concentration in the liquid phase, constituted by the free PAHs (SPAH) and the PAHs adsorbed to DCM (CPAH), was considered as the bioavailable fraction able to be metabolised by microorganisms (Delgadillo-Mirquez et al. 2011). Therefore, the fraction of PAHs adsorbed to particulate matter was not considered bioavailable unless this fraction was transferred into the liquid phase by applying sludge disintegration.

PAHs gas/liquid transfer

Assuming that the equilibrium between the gas and liquid phases was rapidly reached for these hydrocarbons, the volatilisation rate was directly related to the Henry's constant of each compound (HPAH), known from literature, the air flow rate and the soluble concentration of PAHs as stated in Mozo et al. (2012): 
formula
(4)

With QG the air flux (m3·d−1), HPAH the Henry's law constant (atm·m−3.mol−1), V the reactor volume (m3) and Vm the molar volume, 0.0224 m3·mol−1 for ideal gases at a temperature of 20 °C and at the pressure, Patm, of 1 atm.

Disintegration

Disintegration of both the endogenous residue and active biomass with release of colloids and soluble matter and biodegradable colloids was considered. The disintegration was assumed to follow a first order kinetic (Jones et al. 2008; Ramdani et al. 2010) with the same disintegration specific rate for both and .

Calibration and validation procedure

The overall approach was undertaken by performing two distinct steps: (1) calibrate the model with the data obtained from the AS reactor (reference) and (2) validate the model on data obtained from the US-AS reactor (sludge disintegration). Therefore, the exact same set of biological parameters values were used for both systems and only the physical description was different; that is, related to disintegration.

Most of the parameters (originated from ASM1) were provided from the literature (see Table 2). Regarding kinetic and stoichiometric parameters, only the heterotroph growth yield, YOHO, and the PAHs degradation rates (qPAH,Bio) were calibrated with the mass balance performed on AS process (reactor 1). YOHO was identified based on experimental VSS and excess sludge production in AS (reactor 1) was observed. Composite parameters and conversion factors for COD to VSS can be found in Table SI-3. Regarding the physical phenomena, the values of the partitioning coefficients KPART and KDCM for the 16 PAHs were determined in a previous study with dedicated experiments (Mozo et al. 2018); logarithmic values of these partitioning experiments are summarized in Table SI-4. The same values were used for both reactors 1 (AS) and 2 (US-AS). For the second reactor (US-AS), the specific disintegration rate () was calculated based on the degree of solubilisation performed in the ultrasound cell (30%) and suspended solids solubilised per day in reactor 2. The values of and were provided from the literature (Table 2). For the ASM1-PAH model, retention of suspended solids in the settler fret,X was estimated to predict TSSout, VSS, sludge retention time (SRT) and Yobs. For the ASM1-PAH-DCM model, the retention of suspended solids and colloids in the settler, fret,X and fret,C respectively, were estimated simultaneously to predict the same parameters but also the effluent partitioning (CODt,out, CODs,out) and the filtered COD in the mixed liquor (CODs,reactor).

Table 2

Parameters used in the model

Symbol Definition Value Units Source 
 Decay rate for XANO 0.15 d−1 Henze et al. (2000)  
 Decay rate for XOHO 0.62 d−1 Henze et al. (2000)  
 Fraction of XOHO transformed into CUE by sonication 0.08 gCUE·gXOHO−1 Assumed equal to 
 Fraction of XU generated in biomass decay 0.08 gXU·gXOHO−1 Henze et al. (2000)  
 N content of XBio 0.086 gN·gXBio−1 Henze et al. (2000)  
 N content of XU,E 0.06 gN·gXUE−1 Sperandio et al. (2012)  
 Conversion factor XU in TSS 0.1 gISS·gVSS−1 Sperandio et al. (2012)  
 Conversion factor XB in TSS 0.1 gISS·gVSS−1 Henze et al. (2000)  
 Oxygen transfer coefficient 194 d−1 Measured 
 Oxygen half-saturation coefficient for XOHO 0.2 gO2·m−3 Henze et al. (2000)  
 Half-saturation coefficient for XOHO 20 gO2·m−3 Henze et al. (2000)  
 Half-saturation coefficient for hydrolysis of XB 0.03 gO2·m−3 Henze et al. (2000)  
 Maximum specific growth rate for XOHO d−1 Henze et al. (2000)  
 Maximum specific adsorption rate of CUE on XUE 331 d−1 Guellil et al. (2001)  
 Maximum specific hydrolysis rate of CB on XB 7.41·10−7 d−1 Jiang et al. (2008)  
 Rate constant for PAHs uptake rate Table 4 m3·gO2−1·d−1 Estimated 
 Rate constant for PAHs stripping Equation (4) d−1 Calculated 
 Disintegration specific rate 0.0243 d−1 Calculated 
 Maximum specific hydrolysis rate d−1 Henze et al. (2000)  
 Growth yield coefficient for XOHO 0.525 gCOD-XOHO·gCOD−1 Estimated 
Symbol Definition Value Units Source 
 Decay rate for XANO 0.15 d−1 Henze et al. (2000)  
 Decay rate for XOHO 0.62 d−1 Henze et al. (2000)  
 Fraction of XOHO transformed into CUE by sonication 0.08 gCUE·gXOHO−1 Assumed equal to 
 Fraction of XU generated in biomass decay 0.08 gXU·gXOHO−1 Henze et al. (2000)  
 N content of XBio 0.086 gN·gXBio−1 Henze et al. (2000)  
 N content of XU,E 0.06 gN·gXUE−1 Sperandio et al. (2012)  
 Conversion factor XU in TSS 0.1 gISS·gVSS−1 Sperandio et al. (2012)  
 Conversion factor XB in TSS 0.1 gISS·gVSS−1 Henze et al. (2000)  
 Oxygen transfer coefficient 194 d−1 Measured 
 Oxygen half-saturation coefficient for XOHO 0.2 gO2·m−3 Henze et al. (2000)  
 Half-saturation coefficient for XOHO 20 gO2·m−3 Henze et al. (2000)  
 Half-saturation coefficient for hydrolysis of XB 0.03 gO2·m−3 Henze et al. (2000)  
 Maximum specific growth rate for XOHO d−1 Henze et al. (2000)  
 Maximum specific adsorption rate of CUE on XUE 331 d−1 Guellil et al. (2001)  
 Maximum specific hydrolysis rate of CB on XB 7.41·10−7 d−1 Jiang et al. (2008)  
 Rate constant for PAHs uptake rate Table 4 m3·gO2−1·d−1 Estimated 
 Rate constant for PAHs stripping Equation (4) d−1 Calculated 
 Disintegration specific rate 0.0243 d−1 Calculated 
 Maximum specific hydrolysis rate d−1 Henze et al. (2000)  
 Growth yield coefficient for XOHO 0.525 gCOD-XOHO·gCOD−1 Estimated 

RESULTS

Modelling global parameters and performances

Evolution of total and filtered outlet COD and COD removal (shown in Figure SI-1) underlines that similar behavior was observed in both reactors before applying sonication in reactor US-AS (day 21), confirming a satisfying reproducibility under the same operating conditions. In the stabilized period, results showed that residual COD (total and soluble) in the effluent was higher in the conventional activated sludge system compared to the system coupled to sonication. Total outlet COD concentration was 157 ± 33 and 97 ± 25 mgO2·L−1; meanwhile, filtered COD was 51 ± 10 and 34 ± 12 mgO2·L−1 in AS and US-AS respectively. COD removal efficiency was good in both systems but significantly higher in US-AS (84.8%) compared to AS (75.5%). As shown in Table 3, experimental COD removal and mixed liquor volatile suspended solids (MLVSS) content values obtained in both bioreactors were correctly predicted by the model. Results revealed that by combining sonication with the AS system, suspended particulate matter disintegrated into colloidal and soluble matter. Simultaneous visual observations reported that the settleability of the sludge also improved (results not shown), resulting in lower TSS in the outlet of the US-AS process (89 ± 20 g·m−3) compared to the AS system (107 ± 20 g·m−3). This settleability improvement has already been reported in the literature: Gonze et al. (2003) underlined that by disrupting the flocs and releasing them into water, ultrasound reduces the size of particles but also increases their cohesion and density. This increase in density causes the increase of settling velocity even if the volume of particles is reduced. Wunsch et al. (2002) also showed that ultrasound, by creating high local shear stresses capable of breaking the filamentous structure of the sludge, allows the tackling of bulking problems and thus improves solids retention. Another hypothesis is that ultrasound, by the release of polymeric substances, induced the aggregation of small particles. Indeed, the particle number size distribution was centered on 2.2 μm in the AS reactor against 3.6 μm in the US-AS reactor, while the two effluents had a similar size distribution (see Figure SI-2) indicating that intermediate size particles were retained in the US-AS system. As a consequence, the suspended solids retention was better in the US-AS reactor (fret,X = 0.985 in US-AS) compared to the simple AS reactor (fret,X = 0.969 for AS) and this contributed to a significant increase of both the SRT and MLVSS content in the US-AS reactor. For both these reasons (SRT increase and suspended solids disintegration) the observed sludge production yield was slightly lower in the US-AS reactor compared to the AS reactor alone. This response in production yield was captured by the model, although the exact same intrinsic yield (YOHO) was used for modelling both systems. Indeed, YOHO was calibrated to fit VSS, TSS, COD and sludge production. The obtained value (0.525 gCOD·gCOD−1) was slightly lower than the common value used for ASM1 modelling domestic wastewater (0.67 gXOHO·gSB−COD−1). This is rational regarding the fact that synthetic carbon sources were used in this study and the obtained sludge yield is comparable to the thermodynamic growth yield calculated for acetate, ethanol or methanol (Tchobanoglous et al. 2003).

Table 3

Comparison of overall experimental and modelled performances for both activated sludge processes (retention factor fret,X = 0.969 for AS; fret,X = 0.9853 and frect,C = 0.86 for US-AS)

 
 
AS process
 
US-AS process
 
Parameter Unit Exp. Model Exp. Model 
COD removal (based on soluble) 92.1 95.2 94.7 99.7 
COD removal (based on total) 75.5 82.4 84.8 88.0 
MLVSS g·m−3 1,485 ± 150 1,533 2,280 ± 440 2,352 
Sludge production yield gVSS·gCOD−1 0.123 0.119 0.119 0.086 
SRT 12.6 ± 1.1 12.8 24.1 ± 4.5 25.7 
 
 
AS process
 
US-AS process
 
Parameter Unit Exp. Model Exp. Model 
COD removal (based on soluble) 92.1 95.2 94.7 99.7 
COD removal (based on total) 75.5 82.4 84.8 88.0 
MLVSS g·m−3 1,485 ± 150 1,533 2,280 ± 440 2,352 
Sludge production yield gVSS·gCOD−1 0.123 0.119 0.119 0.086 
SRT 12.6 ± 1.1 12.8 24.1 ± 4.5 25.7 

Estimation of PAHs biodegradation rates

Mass balances were calculated with average values over the last 40 days of the running experiment, including PAHs concentrations measured in both the mixed liquor and effluent from AS and US-AS systems (Tables SI-5 and SI-6). The mass balance was not performed for indeno(1,2,3-c,d)pyrene because the analytical error in concentration assessment was of comparable order to the inlet concentration. Regarding the mixed liquor, the total amount of low molecular weight PAHs (between 128.17 and 202.25 g·mol−1) was six-fold higher in the sludge obtained in the AS reactor compared to the US-AS reactor, whilst the total amount of higher molecular weight PAHs (between 228.29 and 278.35 g·mol−1) was 2-fold higher in the US-AS reactor compared to the AS reactor. Total concentrations of low molecular weight as well as higher molecular weight PAHs in the effluent were more than three times lower in the US-AS reactor compared to the AS reactor and this was also true when considering the total amount of the 15 US-EPA PAHs with concentrations of 4.8 and 1.7 μgPAH·L−1 in the AS and US-AS reactors respectively. Filtered effluent samples presented relatively low PAHs concentrations in both reactors; in most cases, values were below the quantification limit (0.05 μgPAH·L−1) especially for the higher molecular weight PAHs. At last, results show that, for all the PAHs compounds whose concentration levels were quantifiable, concentrations measured in the total and filtered effluent were systematically lower in the reactor combined with sonication compared to the AS reactor alone. The decrease of concentration in US-AS compared to AS systems varied between 54% and 93%, see Tables SI-5 and SI-6.

Table 4 shows the specific PAHs removal rates (qPAH,bio) estimated for the 15 PAHs and the volumetric removal rates predicted and measured in both reactors. The range of specific PAHs removal rates were estimated based on results obtained in the AS reactor by using a non-linear least square method implemented in AQUASIM for fitting simulated values and experimental data and validate with the comparison to the US-AS process. The simulation was confronted to the content of PAHs adsorbed to wasted sludge, the flux of PAHs in the total effluent and the flux of PAHs associated to the particulate as well as the liquid fractions of the effluent. As expected, removal rates were higher for low molecular weight compounds compared to more complex and higher molecular weight PAHs. The volumetric removal rate was systematically higher in the US-AS reactor compared to the AS reactor alone, in particular for high molecular weight compounds, with up to 28% improvement for B(b)Fluo. This improvement was well reproduced by the model even though the same specific kinetic rates were used, while including the disintegration processes and adapting the solid retention factors. The volumetric removal rate, including the specific rates, the volumetric load but also the active biomass content and the availability of PAH, the satisfying prediction given by the model underlines that solubilisation of PAHs and the increase of biomass retention time (SRT) are both liable to explain this improvement.

Table 4

Estimated (on reactor AS) biodegradation specific rate qPAH,bio, for PAHs, and volumetric removal rate predicted and experimental in AS and in US-AS

  Model specific rate qPAH, BIO AS (reactor 1) volumetric removal rate mg·m−3·d−1
 
AS-US (reactor 2) volumetric removal rate mg·m−3·d−1
 
Molecule (m3·gDCO−1·day−1Experimental Model Experimental Model 
NAP 6.65 382.54 ± 0.05 382 382.53 ± 0.12 382 
ACY 0.0926 63.86 ± 1.07 63.83 64.15 ± 0.77 63.92 
ACE 0.0367 95.03 ± 3.23 95.21 96.46 ± 1.82 95.54 
FLUOREN 0.0854 70.48 ± 1.97 70.36 71.04 ± 0.74 70.50 
PHEN 0.551 22.43 ± 0.02 22.43 22.43 + 0.01 22.45 
ANT 0.235 22.30 ± 0.28 22.30 22.43 ± 0.06 22.40 
FLUORAN 0.443 21.64 ± 0.12 21.61 21.65 ± 0.02 21.67 
PYR 0.797 20.17 ± 0.04 20.17 20.17 ± 0.02 20.21 
B(a)Ant 0.0450 0.764 ± 0.046 0.764 0.774 ± 0.002 0.815 
CHRYS 0.0303 0.734 ± 0.026 0.735 0.774 ± 0.029 0.805 
B(b)Fluo 0.0067 0.438 ± 0.246 0.442 0.556 ± 0.103 0.588 
B(k)Fluo 0.0120 0.564 ± 0.141 0.568 0.688 ± 0.069 0.687 
BaPyr 0.0208 0.605 ± 0.125 0.607 0.698 ± 0.055 0.697 
DiB(ah)Ant 0.0093 0.418 ± 0.172 0.421 0.528 ± 0.077 0.510 
B(ghi)Peryl 0.0099 0.488 ± 0.300 0.490 0.589 ± 0.123 0.578 
  Model specific rate qPAH, BIO AS (reactor 1) volumetric removal rate mg·m−3·d−1
 
AS-US (reactor 2) volumetric removal rate mg·m−3·d−1
 
Molecule (m3·gDCO−1·day−1Experimental Model Experimental Model 
NAP 6.65 382.54 ± 0.05 382 382.53 ± 0.12 382 
ACY 0.0926 63.86 ± 1.07 63.83 64.15 ± 0.77 63.92 
ACE 0.0367 95.03 ± 3.23 95.21 96.46 ± 1.82 95.54 
FLUOREN 0.0854 70.48 ± 1.97 70.36 71.04 ± 0.74 70.50 
PHEN 0.551 22.43 ± 0.02 22.43 22.43 + 0.01 22.45 
ANT 0.235 22.30 ± 0.28 22.30 22.43 ± 0.06 22.40 
FLUORAN 0.443 21.64 ± 0.12 21.61 21.65 ± 0.02 21.67 
PYR 0.797 20.17 ± 0.04 20.17 20.17 ± 0.02 20.21 
B(a)Ant 0.0450 0.764 ± 0.046 0.764 0.774 ± 0.002 0.815 
CHRYS 0.0303 0.734 ± 0.026 0.735 0.774 ± 0.029 0.805 
B(b)Fluo 0.0067 0.438 ± 0.246 0.442 0.556 ± 0.103 0.588 
B(k)Fluo 0.0120 0.564 ± 0.141 0.568 0.688 ± 0.069 0.687 
BaPyr 0.0208 0.605 ± 0.125 0.607 0.698 ± 0.055 0.697 
DiB(ah)Ant 0.0093 0.418 ± 0.172 0.421 0.528 ± 0.077 0.510 
B(ghi)Peryl 0.0099 0.488 ± 0.300 0.490 0.589 ± 0.123 0.578 

Modelling mass balance for PAHs

Figure 2 shows the mass balance calculated for each PAHs compound and in each system, obtained experimentally and simulated. Volatilization was not significant for most of the PAHs except for acenaphthene with 10% and 5% in AS and US-AS respectively (results not shown). Results show that biodegradation was the most important contributing mechanism for the removal of all molecules, which indicates that the biomass was well acclimatized. PAHs which adsorbed to particles in the wasted sludge and in the effluent contributed for up to 30% of the most hydrophobic molecules (from benzo(a)anthracene to benzo(ghi)perylene). The new model allowed prediction of the improvement of PAHs biodegradation observed previously in the US-AS reactor and this for all molecules except for benzo(b)fluoranthene which was only slightly overestimated. Moreover, besides for the latter compound, the standard deviation between experimental and modelled removal percentages was below 4.6%. The improvement of PAHs biodegradation in the US-AS reactor was especially visible for the most hydrophobic compounds, for which initial biodegradation (as found in the AS reactor) was limited by the adsorption on particulate matter. Moreover, the residual PAHs measured in effluent and wasted sludge was lower in US-AS compared to the AS reactor, and this was also correctly captured by the model.

Figure 1

Diagram of ASM1-based models for the modelling of PAHs biodegradation.

Figure 1

Diagram of ASM1-based models for the modelling of PAHs biodegradation.

Figure 2

PAHs mass balance. Experimental and simulated contribution of removal + volatilisation (a), final residue in effluent (b), and adsorbed to wasted sludge (c) for both AS and US-AS reactors.

Figure 2

PAHs mass balance. Experimental and simulated contribution of removal + volatilisation (a), final residue in effluent (b), and adsorbed to wasted sludge (c) for both AS and US-AS reactors.

DISCUSSION

Experimental results obtained in this study show that sludge disintegration clearly impacted the removal of PAHs in AS systems, and that the newly developed model was able to correctly predict these results. Previous experiments showed that sonication (operated under the same conditions in terms of specific supplied energy) did not have any direct impact on PAHs besides their transfer from the particulate to the liquid phase (Mozo et al. 2018). Therefore, the improvement of PAHs removal rates observed in the US-AS reactor highlights that solid disintegration had a positive effect on the apparent biodegradation of PAHs. At least two main mechanisms can be assumed: one being the increase of SRT due to a better retention of solids (increased settleability) and the other being an increase of PAHs bioavailability due to the adsorption and transport of PAHs with the soluble and colloidal matter.

Of course the question of the fraction actually accessible to biodegradation can be raised. If it is generally admitted that a sorbed micropollutant is not available for biotransformation by microorganism (Rittmann & McCarty 2001; Artola-Garicano et al. 2003; Barret et al. 2010c; Delgadillo-Mirquez et al. 2011; Plósz et al. 2013) no consensus emerges from literature and studies assuming that the fraction sorbed to particles is degraded can be found (Cowan et al. 1993; Lee et al. 1998). This is largely due to the difficulty of designing specific experiments that could dissociate the mechanisms. In any case degradation of the sorbed-to-particles compounds is indirectly taken into account when we consider, as is the case in this study, the hydrolysis of entrapped organics that simultaneously induces a transfer of PAHs to the aqueous phase via equilibrium equations. As a result, modeling was seen here as a tool for discriminating between the two mechanisms considered.

In order to evaluate the effect of solids retention on the improvement of PAHs biodegradation observed in the US-AS reactor, a simulation was performed using the ASM1-PAH model (meaning without processes 42–47) while maintaining SRT and fret,X at the US-AS values. Results from this simulation are referred as ‘model ASM1-PAH – US-AS (case 2)’ in Figure 3. ‘ASM1-PAH – US-AS (case 2)’ model predicted only a slight increase of the removed percentage of PAHs and a limited diminution of the PAHs concentration in the outlet compared to ASM1-PAH-AS. This increase is explained by the increase in the concentration of the mixed liquor and increase of the SRT, due to a better retention of particles (increase of fret,X and resulting decrease of TSSout) in the US-AS reactor. However, when considering these effects alone, the simulation clearly underestimated the difference of PAHs removal values observed between the two systems.

Figure 3

Modelling only the effect of the retention on (a) PAHs removal (biodegradation + volatilisation) and (b) PAHs concentration in the effluent.

Figure 3

Modelling only the effect of the retention on (a) PAHs removal (biodegradation + volatilisation) and (b) PAHs concentration in the effluent.

Indeed, the predicted fractions of removed PAHs and PAHs concentrations in the outlet are far from the experimental values. This result shows that the change in solid retention alone cannot explain the improvement of PAHs biodegradation in the US-AS bioreactor and it is therefore necessary to take into account sludge disintegration processes in order to efficiently predict observations.

The new model developed within this study captured the effect of sludge disintegration on the removal of PAHs by combining two concepts: the concept of PAHs adsorption on DCM and its bioavailability (a concept developed by Barret et al. (2010c) for anaerobic systems and validated by Mozo et al. (2018) for activated sludge), and the fact that sludge disintegration releases endogenous compounds and DCM (developed by Jiang et al. (2008)). The fact that these effects were successfully captured by the model supports the idea that DCM plays a major role in improving the bioavailability of PAHs. In the literature, two mechanisms have been reported for explaining the degradation of micropollutants adsorbed to DCM. The first mechanism is based on the assumption that low molecular weight DCM might cross microbial membranes and form a micropollutant-DCM complex. Indeed, molecules up to a few kDa have already been shown to cross bacterial membranes (Nikaido 2003). The second assumed mechanism considers that particles might transport adsorbed micropollutants in the vicinity of the cell surfaces (Smith et al. 2009), which is a preliminary step before diffusion through the cell membrane.

Concerning industrial perspectives, a techno-economic study could favour the implementation of this hybrid process. As an example, based on modelling, we estimate that an improvement in benzo(b)fluoranthene removal from 47% to 66% could be obtained with sonication applied in the sludge recycling loop, implying 3.2 kWh per m3 of treated wastewater, whereas doubling the reactor volume of activated sludge would lead to an increase from 47% to only 51%. Thus a complete economic evaluation will have to compare the energy costs associated with the implementation of ultrasound and the infrastructure costs that would be required to achieve similar performances with a conventional process.

Finally, DCM content is generally relatively low in activated sludge systems. But new treatment systems with growing quantities of DCM are being developed (Wang & Chen 2018). Indeed, the short retention time applied in HRAS (High Rate Activated Sludge) is likely to significantly modify the partitioning of PAHs compared to conventional activated sludge, and hence impact the resulting biodegradation of PAHs. Furthermore, wastewater treatment plants are increasingly considered as facilities to recover water resources (van der Hoek et al. 2016; Solon et al. 2019) and in this perspective the combination of several processes would be needed. In such facilities, the disintegration of the solid phase is involved, whether experienced during recirculation through a MBR or applied as an additional process to improve anaerobic digestion, for example. Recirculation between processes units may hence contribute to the transport of DCM-adsorbed-PAHs. The developed model will therefore be useful for such systems as it will be able to both determine the effect of disintegration on solubilization and predict the resulting fate of PAHs within the system.

CONCLUSION

  • Sludge disintegration (in our study, the use of sonication) can optimise conventional activated sludge systems by improving performances and reliability for the removal of PAHs in wastewater treatment plants.

  • A new model was developed in order to include the effect of sludge disintegration on DCM content and subsequent PAHs bioavailability.

  • The model predicted a better degradation of a large range of PAHs in an AS system coupled with sonication, a slight reduction of sludge production and a better effluent quality.

  • Combining specifically designed experiments and modelling proved to be efficient to study the role of sludge disintegration on PAHs removal: increase of apparent biodegradation was shown to be better explained by the improvement of PAHs availability compared to the increase of solid retention.

ACKNOWLEDGEMENTS

The authors thank Gerard Cancel and Delphine Delagnes for their technical contribution to this work and the ANRT for financial support (CIFRE n° 1261/2009).

REFERENCES

REFERENCES
Afnor
1994
Qualité de l'eau (Water Quality). In Recueil de Normes, Afnor, ed. (Afnor Editions)
.
Barret
M.
,
Barcia
G. C.
,
Guillon
A.
,
Carrere
H.
&
Patureau
D.
2010a
Influence of feed characteristics on the removal of micropollutants during the anaerobic digestion of contaminated sludge
.
J. Hazard. Mater.
181
,
241
247
.
Choubert
J. M.
,
Martin Ruel
S.
,
Esperanza
M.
,
Budzinski
H.
,
Miège
C.
,
Lagarrigue
C.
&
Coquery
M.
2011
Limiting the emissions of micro-pollutants: what efficiency can we expect from wastewater treatment plants?
Water Sci. Technol. J. Int. Assoc. Water Pollut. Res.
63
,
57
65
.
Cowan
C. E.
,
Larson
R. J.
,
Feijtel
T. C. J.
&
Rapaport
R. A.
1993
An improved model for predicting the fate of consumer product chemicals in wastewater treatment plants
.
Water Res.
27
,
561
573
.
Delgadillo-Mirquez
L.
,
Lardon
L.
,
Steyer
J.-P.
&
Patureau
D.
2011
A new dynamic model for bioavailability and cometabolism of micropollutants during anaerobic digestion
.
Water Res.
45
,
4511
4521
.
Gonze
E.
,
Pillot
S.
,
Valette
E.
,
Gonthier
Y.
&
Bernis
A.
2003
Ultrasonic treatment of an aerobic activated sludge in batch reactor
.
Chem. Eng. Process.
42
,
965
975
.
Grandclément
C.
,
Seyssiecq
I.
,
Piram
A.
,
Wong-Wah-Chung
P.
,
Vanot
G.
,
Tiliacos
N.
,
Roche
N.
&
Doumenq
P.
2017
From the conventional biological wastewater treatment to hybrid processes, the evaluation of organic micropollutant removal: a review
.
Water Res.
111
,
297
317
.
Guellil
A.
,
Thomas
F.
,
Block
J.-C.
,
Bersillon
J.-L.
&
Ginestet
P.
2001
Transfer of organic matter between wastewater and activated sludge flocs
.
Water Res.
35
,
143
150
.
Henze
M.
,
Gujer
W.
,
Mino
T.
&
van Loosdrecht
M.
2000
Activated Sludge Models ASM1, ASM2, ASM2D and ASM3
.
IWA Publishing
,
London
,
UK
.
Holman
H.-Y. N.
,
Nieman
K.
,
Sorensen
D. L.
,
Miller
C. D.
,
Martin
M. C.
,
Borch
T.
,
Mckinney
W. R.
&
Sims
R. C.
2002
Catalysis of PAH biodegradation by humic acid shown in synchrotron infrared studies
.
Environ. Sci. Technol.
36
,
1276
1280
.
Jiang
T.
,
Myngheer
S.
,
De Pauw
D. J. W.
,
Spanjers
H.
,
Nopens
I.
,
Kennedy
M. D.
,
Amy
G.
&
Vanrolleghem
P. A.
2008
Modelling the production and degradation of soluble microbial products (SMP) in membrane bioreactors (MBR)
.
Water Res.
42
,
4955
4964
.
Johnson
B. R.
,
Daigger
G. T.
&
Novak
J. T.
2008
The use of ASM based models for the simulation of biological sludge reduction processes
.
Water Pract. Technol.
3
,
3
11
.
Jones
R.
,
Parker
W.
,
Khan
Z.
,
Murthy
S.
&
Rupke
M.
2008
Characterization of sludges for predicting anaerobic digester performance
.
Water Sci. Technol. J. Int. Assoc. Water Pollut. Res.
57
,
721
726
.
Kästner
M.
,
Streibich
S.
,
Beyrer
M.
,
Richnow
H. H.
&
Fritsche
W.
1999
Formation of bound residues during microbial degradation of [14C] anthracene in soil
.
Appl. Environ. Microbiol.
65
,
1834
1842
.
Ke
Y.
,
Ning
X.
,
Liang
J.
,
Zou
H.
,
Sun
J.
,
Cai
H.
,
Lin
M.
,
Li
R.
&
Zhang
Y.
2018
Sludge treatment by integrated ultrasound-Fenton process: characterization of sludge organic matter and its impact on PAHs removal
.
J. Hazard. Mater.
343
,
191
199
.
Laurent
J.
,
Guignard
R.
,
Casellas
M.
&
Dagot
C.
2011
Sludge minimization by thermal lysis during biological wastewater treatment – experimental study, model set up and calibration
. In
SIMULTECH
. pp.
360
365
.
Lee
K. C.
,
Rittmann
B.
,
Shi
J.
&
McAvoy
D.
1998
Advanced steady-state model for the fate of hydrophobic and volatile compounds in activated sludge
.
Water Environ. Res.
70
,
1118
1131
.
Liu
Z.
,
Li
Q.
,
Wu
Q.
,
Kuo
D. T. F.
,
Chen
S.
,
Hu
X.
,
Deng
M.
,
Zhang
H.
&
Luo
M.
2017
Removal efficiency and risk assessment of polycyclic aromatic hydrocarbons in a typical municipal wastewater treatment facility in Guangzhou, China
.
Int. J. Environ. Res. Public. Health
14
,
861
.
Margot
J.
,
Rossi
L.
,
Barry
D. A.
&
Holliger
C.
2015
A review of the fate of micropollutants in wastewater treatment plants
.
Wiley Interdiscip. Rev. Water
2
,
457
487
.
Martin Ruel
S.
,
Choubert
J. M.
,
Esperanza
M.
,
Miège
C.
,
Madrigal
P. N.
,
Budzinski
H.
,
Le Ménach
K.
,
Lazarova
V.
&
Coquery
M.
2011
On-site evaluation of the removal of 100 micro-pollutants through advanced wastewater treatment processes for reuse applications
.
Water Sci. Technol. J. Int. Assoc. Water Pollut. Res.
63
,
2486
2497
.
Martin Ruel
S.
,
Choubert
J.-M.
,
Budzinski
H.
,
Miège
C.
,
Esperanza
M.
&
Coquery
M.
2012
Occurrence and fate of relevant substances in wastewater treatment plants regarding Water Framework Directive and future legislations
.
Water Sci. Technol. J. Int. Assoc. Water Pollut. Res.
65
,
1179
1189
.
Moscoso
F.
,
Teijiz
I.
,
Deive
F. J.
&
Sanromán
M. A.
2012
Efficient PAHs biodegradation by a bacterial consortium at flask and bioreactor scale
.
Bioresour. Technol.
119
,
270
276
.
Mozo
I.
,
Stricot
M.
,
Lesage
N.
&
Sperandio
M.
2011
Fate of hazardous aromatic substances in membrane bioreactors
.
Water Res.
45
,
4551
4561
.
Mozo
I.
,
Lesage
G.
,
Yin
J.
,
Bessiere
Y.
,
Barna
L.
&
Sperandio
M.
2012
Dynamic modeling of biodegradation and volatilization of hazardous aromatic substances in aerobic bioreactor
.
Water Res.
46
,
5327
5342
.
Mozo
I.
,
Bounouba
M.
,
Mengelle
E.
,
Lesage
N.
,
Sperandio
M.
&
Bessiere
Y.
2018
Modelling PAH partitioning during sludge disintegration: the key role of dissolved and colloidal matter
.
Chemosphere
209
,
883
891
.
Nikaido
H.
2003
Molecular basis of bacterial outer membrane permeability revisited
.
Microbiol. Mol. Biol. Rev. MMBR
67
,
593
656
.
Plósz
B. G.
,
Benedetti
L.
,
Daigger
G. T.
,
Langford
K. H.
,
Larsen
H. F.
,
Monteith
H.
,
Ort
C.
,
Seth
R.
,
Steyer
J.-P.
&
Vanrolleghem
P. A.
2013
Modelling micro-pollutant fate in wastewater collection and treatment systems: status and challenges
.
Water Sci. Technol. J. Int. Assoc. Water Pollut. Res.
67
,
1
15
.
Pomiès
M.
,
Choubert
J.-M.
,
Wisniewski
C.
&
Coquery
M.
2013
Modelling of micropollutant removal in biological wastewater treatments: a review
.
Sci. Total Environ.
443
,
733
748
.
Ramdani
A.
,
Dold
P.
,
Déléris
S.
,
Lamarre
D.
,
Gadbois
A.
&
Comeau
Y.
2010
Biodegradation of the endogenous residue of activated sludge
.
Water Res.
44
,
2179
2188
.
Rittmann
B. E.
&
McCarty
P. L.
2001
Environmental Biotechnology: Principles and Applications
.
McGraw-Hill Education
,
New York, NY, USA
.
Smith
K. E. C.
,
Thullner
M.
,
Wick
L. Y.
&
Harms
H.
2009
Sorption to humic acids enhances polycyclic aromatic hydrocarbon biodegradation
.
Environ. Sci. Technol.
43
,
7205
7211
.
Solon
K.
,
Volcke
E. I. P.
,
Spérandio
M.
&
Loosdrecht
M. C. M. v.
2019
Resource recovery and wastewater treatment modelling
.
Environ. Sci. Water Res. Technol.
5
,
631
642
.
Sperandio
M.
,
Paul
E.
,
Bessiere
Y.
&
Liu
Y.
2012
Sludge production: quantification and prediction for urban treatment plants and assessments of strategies for sludge reduction
. In:
Biological Sludge Minimization and Biomaterials/Bioenergy Recovery Technologies
.
Wiley
,
Hoboken, NJ, USA
, pp.
117
154
.
Tchobanoglous
G.
,
Burton
F. L.
&
Stensel
H. D., Metcalf & Eddy
2003
Wastewater Engineering: Treatment and Reuse
.
McGraw-Hill Education
,
New York, NY, USA
.
van der Hoek
J. P.
,
de Fooij
H.
&
Struker
A.
2016
Wastewater as a resource: strategies to recover resources from Amsterdam's wastewater
.
Resour. Conserv. Recycl.
113
,
53
64
.
Wunsch
B.
,
Heine
W.
&
Neis
U.
2002
Combating Bulking Sludge with Ultrasound in Ultrasound in Environmental Engineering II
.
(U. Neis, ed.). TU Hamburg-Harburg, Hamburg, Germany
.
Zielińska
M.
,
Sobolewska
J.
,
Bułkowska
K.
,
Wojnowska-Baryła
I.
&
Szelążek
P.
2012
Removal of phenanthrene and 4-phenylphenanthrene from wastewater in an integrated technological system
.
Desalination Water Treat.
50
,
78
86
.

Supplementary data