Tetrabromobisphenol A (TBBPA) is a fire retardant with endocrine disruption properties, which unfortunately has been detected in trace amounts in various environmental samples. Studies have shown that TBBPA enters the aquatic environment, sediments, soils and biota predominantly through the wastewater treatment process. The dominant contributor to the fate and transport of TBBPA in wastewater is biosorption by the sludge matrix. Accordingly, the goal of this research was to determine the fate kinetics, biosorption correlation models and mechanisms of TBBPA from both conventional activated sludge (CAS) and membrane bioreactor (MBR) systems. Experiments were conducted to determine different types of partitioning and fate kinetic coefficients such as solid–liquid phase coefficient, log Kp (4.53), sludge matrix organic carbon coefficient, log Koc (6.05) and glass wall partition coefficient, Kw (0.053). The experimental values of these coefficients for TBBPA are not available in the literature for application to the models. The biosorption data were also verified and fitted to Freundlich, Langmuir, Temkin, Toth, Redlich–Peterson, Sips, Brunauer–Emmett–Teller (BET), Jovanovic and Dubinin–Radushkevich biosorption isotherm models to estimate the isotherm constants. These coefficients will allow engineering application for validation of various fate models for TBBPA.

MODEL NOTATION AND NOMENCLATURE

  • qe

    Equilibrium sorbate conc. in the solid (ng of sorbate TBBPA/mg of sludge)

  • Ce

    Equilibrium sorbate conc. in the liquid (ng of sorbate/L of liquid phase)

  • KF

    Freundlich isotherm constant (L liquid phase/mg sludge

  • n

    Freundlich biosorption intensity (-)

  • qm

    Maximum biosorption capacity (ng sorbate/mg of sludge)

  • KL

    Langmuir isotherm constant (L liquid/mg sludge)

  • RL

    Separation factor (-)

  • C0

    Initial sorbate concentration (ng sorbate/L liquid phase)

  • KT

    Temkin isotherm binding constant (L liquid phase/mg sludge)

  • b

    Sorbate–biosorbent interaction energy (kJ energy/mol sorbate)

  • R

    Universal gas constant (kJ energy/mol sorbate/K temperature)

  • T

    Temperature (K)

  • Kt

    Toth equilibrium constant (ng sorbate/mg sludge)

  • t

    Toth isotherm heterogeneity factor (fractional value)

  • KR

    Redlich–Peterson isotherm constant (L liquid/mg sludge)

  • αR

    Redlich–Peterson model coefficient (L liquid/mg sludge)

  • β

    Redlich–Peterson model exponent (-)

  • KS

    Sips isotherm constant (L liquid phase/mg sludge)

  • m

    Sips isotherm exponent (-)

  • B

    BET isotherm constant (L liquid phase/mg sludge)

  • qs

    Biosorption saturation capacity (ng sorbate/mg of sludge)

  • Cs

    Saturation conc. for monolayer formation (ng sorbate/L liquid)

  • KJ

    Jovanovic isotherm constant (-)

  • BD

    Dubinin–Radushkevich constant (kJ energy/mol of sorbate)

  • Ε

    Polyani potential energy (kJ energy/mol of sorbate)

  • KHJ

    Harkin-Jura isotherm constant (L liquid/mg sorbate)

  • AHJ

    Harkin–Jura model coefficient (-)

  • KH

    Halsey isotherm constant (-)

  • K

    Biosorption equilibrium constant (L liquid/mg sludge)

  • ΔG

    Gibbs free energy change (kJ energy/mol sorbate)

  • ΔH

    Enthalpy change of biosorption (kJ energy/mol sorbate)

  • ΔS

    Entropy change of biosorption (kJ energy/Kelvin temperature)

INTRODUCTION

The environmental fate and transport of tetrabromobisphenol A (TBBPA) has a strong inter-relation to the partitioning of TBBPA in aqueous phase and wastewater sludge matrix. TBBPA is introduced to the aquatic environment and ultimately to the other environmental compartments through wastewater (Lee & Peart 2002; Covaci et al. 2009; Labadie et al. 2010; McPhedran et al. 2013). Essential plastic commodities of daily life containing TBBPA as a fire retardant component are the major cause of its increasing abundance either in wastewater or in landfills. TBBPA has been found in biota, blood serum, human milk and adipose tissues (Thomsen et al. 2002; Cariou et al. 2005; Frederiksen et al. 2007). The adverse health and environmental consequences of TBBPA due to its ubiquitous nature is therefore an emerging concern. Accordingly, the investigation of the fate models and mechanisms of TBBPA was accomplished in this present study.

Organic carbon content may be the dominant constituent for the adsorption of TBBPA in activated sludge (Gao et al. 1997; Field 2006). As such, sludge-water partitioning and the fate of the analyte TBBPA was investigated in this research from experimentally determined pertinent physical parameters. Schwarzenbach & Westall (1981) reported the relationship between partitioning of chemical pollutants and the fraction of organic carbon of the sorbent, where sorption is greatly affected by the presence of organic matter. The partition coefficients were significantly high for the sorbents containing greater than 0.1% organic carbon. However, relatively small partition coefficients were found when the organic fraction was less than 0.1%.

A great deal of sorption work has been reported in the literature with soils, but the structure, texture, physical and biological composition of the activated sludge matrix is quite different from the soil matrix. In the activated sludge process, the sorbent is completely suspended in wastewater. There are many other factors that may influence the normal distribution of pollutants differently in the sludge matrix. Activated sludge containing microbial consortia provides enhanced sorption, referred to as biosorption (Aksu 2005; Vijayaraghavan & Yun 2008; Srinivasan & Viraraghavan 2010; Islam et al. 2014). Activated sludge is a composite of live and dead microorganisms with a large surface area between 0.82 and 1.66 m2/g (Wang & Govind 1993). Biosorption accompanied by diffusive penetration inside the cell membrane allows increased accumulation of organic contaminants into the sludge.

Biosorption depends on the chemical nature of the contaminant, the surface properties of biosorbent, diverse interactions and environmental conditions (Aksu & Isoglu 2007). As such, biosorption is a complex phenomenon involving physical, biological, chemical, kinetic, thermodynamic and mass transfer processes. Depending on the interpretation of the system interactions, there are numerous equations and models available in the literature (Foo & Hameed 2010; Paska et al. 2014). The applicability of various models was tested to determine different partitioning coefficients for the fate and transport of TBBPA. Isotherm model fittings demonstrate the potential of biosorption capacity for the candidate compound with respect to the sorbate–sorbent interactions of the fitted model. If the experimental data show goodness of fit with multiple models, then the potential of biosorption for the candidate compound becomes more obvious, i.e., highly favourable in terms of the theoretical model assumptions. Multiple model fittings indicate high possibilities of diverse combinations of binding sites availability and layer formation for biosorption process (Samarghandi et al. 2009; Foo & Hameed 2010; Paska et al. 2014). Therefore, it can be concluded from the excellent curve fittings for TBBPA with a number of models that the dominant mechanism for the removal of TBBPA from wastewater process is biosorption.

Studies demonstrated that the partitioning parameters are directly related to the partition coefficients and constants (Birkett & Lester 2003). The distribution of the target compound depends on the extent of the increase or the decrease in the model constants at equilibrium. An excellent linear relationship from the regression analysis indicates high potential for the prediction of the parametric values by the correlation models (Schwarzenbach et al. 2003). The models developed from the experimental data produce more reliable results than empirical models and will be better predictors of what takes place in the field. Therefore, this research is valuable and promising for potential engineering application in the wastewater treatment process.

The biosorption isotherm models (Foo & Hameed 2010; Hadi et al. 2010; Paska et al. 2014) are generally justified from the curve fittings by plotting equilibrium sorbate concentrations in two phases (qe, Ce). The equilibrium concentrations are measured after establishing the equilibrium in the system with sufficient mechanical mixing and reaction time.

The Freundlich isotherm equation is the best suited model for partitioning of the most organic compounds in activated sludge (Dobbs et al. 1989; Clara et al. 2004). According to Freundlich adsorption, the heat of sorption decreases exponentially with the coverage of the sorbent surface. In this case, the sorbate TBBPA undergoes interactions with the heterogeneous surface of the biosorbent activated sludge. The Freundlich isotherm model for activated sludge biosorption can be expressed as follows: 
formula
1
where KF (L/mg) is the Freundlich constant, indicating the biosorption capacity of sludge and 1/n is the exponent, indicating biosorption intensity of TBBPA.
The Langmuir isotherm model (Langmuir 1916) was formulated for the monolayer formation with the fixed maximum capacity of the surface coverage. The Langmuir isotherm equation can be represented as: 
formula
2
where qm (ng/mg) is the maximum biosorption capacity of the sludge to form a monolayer with TBBPA and KL (L/ng) is the Langmuir constant, indicating the affinity of the binding sites of sludge for TBBPA.

There are multiple models available from the modifications of widely used Freundlich and Langmuir isotherm equations. As such, suitability of Temkin, Toth, Redlich–Peterson, Sips, Brunauer–Emmett–Teller (BET), Jovanovic, Dubinin–Radushkevich, Harkin–Jura and Halsey isotherm models were verified in addition to Freundlich and Langmuir models.

Biosorption is a consequence of surface free energy change due to the interaction of sorbate–biosorbent molecules. The free energy change (ΔG0), the enthalpy change (ΔH0), and the entropy change (ΔS0) were studied and calculated from the biosorption equation and thermodynamic relationships.

The fate and transport properties of organic pollutants in wastewater depend on the magnitude of partitioning coefficients. A number of transport and fate models are available for hydrophobic organic compounds in soil (Site 2001), whereas the fate model for the brominated flame retardants in activated sludge is very limited. The objective of this research was achieved by estimating the partition coefficients and developing the correlation models from the relevant experimental data. The ultimate engineering application of these outcomes will allow control and mitigation of environmental proliferation for TBBPA.

METHODOLOGY

Materials

All chemicals and reagents were purchased from Fischer Scientific Canada unless otherwise specified. Most chemicals were ACS grades and all solvents used were HPLC grades. Whatman GF/B filters, Whatman GF/F, Whatman 0.2 nylon membrane and nucleopore polycarbonate filters (Cat. No. 0930069) were purchased from Fisher Scientific. Helium (99.9999%) and methane (99.9999%) were supplied by BOC Canada (Guelph, Ontario). Milli-Q water was produced from a Milli-pore Ultrapure Water System.

Extraction and analysis

The extraction and analysis of aqueous samples for injecting into the gas chromatography–mass spectrometry (GC-MS) system require multiple steps of processing. The first step of wastewater processing was filtration for applying to solid phase extraction. The captured fractions were eluted and liquid–liquid extraction (LLE) clean-up was performed before adding internal standards (ISTD). The sample was then evaporated to dryness for obtaining quantitative transfer of the reconstituted volume to the GC-MS system. The entire analysis process flow diagram is shown in Figure 1.

Figure 1

Process flow diagram for wastewater sample.

Figure 1

Process flow diagram for wastewater sample.

The quantitative estimation was performed by an Agilent 6890N GC equipped with an autosampler injector. The signal detector for quantification was an Agilent 5975 MSD. Helium was used as the carrier gas, under a constant pressure of 80 kPa. Methane was used as the reagent gas for the negative ion source and was maintained at 40% of flow. The system was run in negative chemical ionization (NCI) mode. The select ion monitoring (SIM) acquisition was chosen for higher sensitivity of the method. The interface, inlet and quadrupole temperatures were maintained at 250 °C, 250 °C and 150 °C, respectively. The ions monitored were 79 and 81 m/z for TBBPA, 71 and 74 m/z for DDT.

Total organic carbon (TOC) was measured by a Shimadzu TOC-V unit equipped with a solid sample module 5000A. Calibration curves were developed and updated frequently as specified in Standard Methods (APHA 2012). Basic water quality parameters such as dissolved oxygen (DO), pH and temperature were recorded immediately after sample collection using a pH/conductivity/DO meter (probe model sensION6, Hach, USA).

Pertinent equations for experimental measurements

Adsorption losses on to the container wall for organic compounds have been reported (Kerr et al. 2000; Kokkola et al. 2014). Similar to other organic compounds, TBBPA has a tendency to be adsorbed onto the surface of the glass vessel, indicating adsorptive losses (Covaci et al. 2009). Thus, care must be taken to compensate for these adsorption losses. As such, the vessel-wall partition coefficient KW (mL/cm2) was determined, which represents the partition coefficient between glass wall and the aqueous phase as represented by Equation (3) (Backhus & Gschwend 1990): 
formula
3
The organic carbon partition coefficient (KOC) of activated sludge was estimated from Equation (4): 
formula
4
where KOC = organic carbon partition coefficient; KW = vessel-wall partition coefficient; [OC] = organic carbon concentration; [SA/V] = surface area to volume ratio of the container.
Sorptive partitioning to activated sludge was estimated from the sludge-water partition coefficient, KP (L/mg). The solid-liquid partition coefficient is given by the following Equation (5): 
formula
5
It should be noted that the solid phase concentration, Cs (ng/mg) was not measured in this study for the determination of sludge-water partition coefficient. However, it was calculated from the difference of the total concentration and the dissolved phase aqueous concentration.

Batch experiments

Batch experiment for KW

The wall loss coefficient was determined by spiking 100 μL of TBBPA solution in acetone to 500 mL Milli-Q water in borosilicate glass Erlenmeyer flasks. The equilibration period was over 24 h to achieve complete equilibrium before extraction and analysis. The time recommended for establishing equilibrium was the minimum mixing of half an hour (US EPA 1996). The flasks were wrapped with aluminum foil and placed in the dark chamber for 36 h in order to avoid photo-degradation losses. The concentration of the content by calculation was 100 ng/L after spiking. The initial aqueous concentration was estimated experimentally from the instantaneous extraction and quantification of the spiked solutions. Wall loss coefficient (KW) was determined from the geometry and equilibrium concentrations between aqueous phase and borosilicate glass surface by using Equation (3).

Batch experiment for KOC and KP

The fraction of organic carbon in the sludge was obtained from the measurement of the organic carbon concentration. Organic carbon content in activated sludge was measured by a TOC analyser while the spiked concentration of TBBPA was extracted and analysed in parallel by GC-MS following the method protocol (Figure 1) for the determination of KOC. For each sample analysis, diluted sludge (not filtered sludge) was used for organic carbon measurement by TOC analyser. One mL of fresh activated sludge was diluted to 1,000 mL in a volumetric flask. The diluted sludge was transferred to the bioreactor (300 mL biochemical oxygen demand (BOD) bottles) in varying amounts. TBBPA standard solution (90 μL) was spiked to each bioreactor while stirring. The TOC analyser suction tube was inserted into the continuously stirred bioreactors approximately at the midpoint for actual representation of the organic carbon. The organic carbon concentration was recorded from the analyser reading and the remaining solution was used for extraction, analysis and quantification by GC-MS. Equation (4) was used for the computation of the organic carbon coefficient, KOC.

Fate partitioning and distribution of TBBPA to activated sludge was estimated from the sludge-water partition coefficient (KP). The concentrations of TBBPA in the aqueous phase and sludge were determined as detailed in the process flow diagram of Figure 1. Fresh activated sludge and membrane bioreactor (MBR) sludge was added to BOD bottles in varying amounts. A standard solution of TBBPA (50 μL) was added to each bioreactor bottles in stirred condition. The mixtures were extracted and analysed for TBBPA quantification in sludge phase and aqueous phase. The sludge phase TBBPA was estimated from the concentration difference of the total TBBPA and the dissolved phase TBBPA. In this study, only aqueous phase determination was accomplished by the developed method. Equation (5) was used for the computation of the required parameters.

Batch experiment for biosorption coefficients

Biosorption of TBBPA was studied by using fresh sludge collected from the aeration basin of the conventional activated sludge (CAS) process and from the aeration tank of the MBR pilot plants at the Guelph municipal wastewater treatment plant (GWWTP). Biosorption isotherms were developed to explain the partitioning behaviour of TBBPA in both MBR and CAS sludge. Different mathematical models have been applied to evaluate the goodness of fit for the experimental data. A set of biosorption partition coefficients for TBBPA were generated from the analysis of experimental data using isotherm equations.

Sludge samples from both biological reactors were collected in amber glass containers, and placed in an ice chamber. Samples were then brought immediately to the environmental engineering laboratory, University of Guelph for analysis.

US EPA (1996) guidelines recommend a minimum half an hour of stirring of the mixture to establish equilibrium for an adsorption study. Accordingly, the extraction was carried out after 2 h of spiking to allow enough time for achieving equilibrium. The reacting vessels were continuously stirred during the equilibration period. Fresh sludge in varying amounts was added to each reaction vessel for developing the biosorption isotherms. The volumes of sludge directly added to each reactor were 0 mL, 1 mL, 2 mL, 4 mL, 6 mL, 8 mL, 10 mL, 12 mL and 15 mL, respectively. The spiked concentration of TBBPA in the reacting vessels was 30 ng/L. The pH and temperature of the test solutions were recorded at 7.1 °C and 21.5 °C for CAS sludge and 6.82 °C and 21.3 °C for MBR sludge, respectively.

The extraction, analysis and estimation of experimental samples were accomplished as shown in the process flow diagram of Figure 1. Biosorption coefficients were estimated from the experimental data using various isotherm model equations.

RESULTS AND DISCUSSION

Partitioning coefficients KW, KOC, KP

The vessel-wall partition coefficient was determined in the laboratory. The measurement was done in triplicate experiments and the average experimental value was 0.053 mL/cm2. This value serves as the correction coefficient for a unit volume of TBBPA solution per unit surface area of borosilicate glass. The correction coefficient compensates for the adsorptive losses onto the vessel wall for the determination of TBBPA. Table 1 shows the experimental result for the determination of the glass wall correction coefficient.

Table 1

Experimental data for wall loss coefficient

Aqueous phase TBBPA conc. [TBBPA]Aq.ng/L Vessel wall TBBPA conc. [TBBPA]Wall ng/L Surface area to volume ratio [SA/V] cm2/L Wall loss coefficient KW = [TBBPA]Wall [TBBPA]Aq.[SA/V] mL/cm2 
83.58 16.42 4.26 0.047 
81.53 18.48 4.26 0.053 
79.98 20.02 4.26 0.059 
Average 0.053 
Aqueous phase TBBPA conc. [TBBPA]Aq.ng/L Vessel wall TBBPA conc. [TBBPA]Wall ng/L Surface area to volume ratio [SA/V] cm2/L Wall loss coefficient KW = [TBBPA]Wall [TBBPA]Aq.[SA/V] mL/cm2 
83.58 16.42 4.26 0.047 
81.53 18.48 4.26 0.053 
79.98 20.02 4.26 0.059 
Average 0.053 

Backhus & Gschwend (1990) were the first to indicate the partitioning of polycyclic aromatic compounds to glass vessels with a wall loss coefficient, KW = 0.33 mL/cm2. TBBPA has a low KW value as it is less hydrophobic than polycyclic aromatic compounds since it has two polar hydroxyl groups attached to each benzene ring. It was expected that the adsorption losses to glass vessels would be much less than polycyclic aromatic hydrocarbons. Accordingly, the estimated value KW = 0.053 mL/cm2 is acceptable for the compound TBBPA.

The organic carbon partition coefficients (KOC) and sludge-water partition coefficients (KP) for activated sludge determined in the laboratory are given in Table 2. The experimental data for the organic carbon partition coefficients (KOC) and sludge-water partition coefficients (KP) for MBR sludge are given in Table 3.

Table 2

Experimental data of activated sludge log KOC and log KP

No. Vol. of sludge mL [OC] mg/L log KOC L/mg log KP L/mg 
0.523 5.80 4.32 
0.615 5.87 4.50 
1.427 5.98 4.51 
1.743 6.00 4.54 
2.161 6.01 4.57 
10 2.299 6.10 4.61 
12 3.220 6.20 4.62 
15 3.717 6.21 4.63 
Average 6.02 4.54 
No. Vol. of sludge mL [OC] mg/L log KOC L/mg log KP L/mg 
0.523 5.80 4.32 
0.615 5.87 4.50 
1.427 5.98 4.51 
1.743 6.00 4.54 
2.161 6.01 4.57 
10 2.299 6.10 4.61 
12 3.220 6.20 4.62 
15 3.717 6.21 4.63 
Average 6.02 4.54 
Table 3

Experimental data of MBR sludge log KOC and log KP

No Vol. of sludge mL [OC] mg/L log KOC L/mg log KP L/mg 
0.96 5.64 4.21 
1.414 5.65 4.24 
1.556 5.71 4.30 
2.427 5.73 4.31 
3.050 5.82 4.32 
10 3.364 5.84 4.33 
12 3.914 5.85 4.34 
15 4.106 5.86 4.39 
Average 5.76 4.31 
No Vol. of sludge mL [OC] mg/L log KOC L/mg log KP L/mg 
0.96 5.64 4.21 
1.414 5.65 4.24 
1.556 5.71 4.30 
2.427 5.73 4.31 
3.050 5.82 4.32 
10 3.364 5.84 4.33 
12 3.914 5.85 4.34 
15 4.106 5.86 4.39 
Average 5.76 4.31 

The resulting log KOC value is 6.02, with log KP equal to 4.54 for activated sludge. The corresponding values for MBR sludge are 5.76 and 4.31, respectively. The organic carbon partition coefficient, KOC is more of an intrinsic parameter than KP. KOC is specific to the type of organic matter present in the biological process. The value of log KOC > 3 provides a high potential for sorption to the sludge (Birkett & Lester 2003). Accordingly, the higher average value of KOC for TBBPA indicates the higher biosorption potential in activated sludge (Islam et al. 2014).

The solid-water partition coefficient, KP is a phase equilibrium constant which approaches to Freundlich constant (KF) if the exponent of biosorption intensity (1/n) becomes unity. The linear partition coefficient KP can be combined with Langmuir isotherm or Freundlich isotherm to model complex biosorption mechanisms covering a wide range of concentrations (Wang & Govind 1993; Schwarzenbach et al. 2003).

Partitioning parameters are empirically related to each other for developing the fate model. If the proportion of organic carbon content increases, the propensity of sorptive partitioning to sludge increases (Birkett & Lester 2003). The hydrophobicity of organic compounds is dependent on the organic carbon partition coefficient, KOC (Field 2006). The sludge-water partition coefficient (KP) is directly related to the organic fraction of the sludge. The relationship was expressed empirically by Karickhoff et al. (1979). 
formula
6
where fOC is the organic fraction of the sludge.
KP is a useful parameter for prediction partitioning of organic compounds in biological wastewater treatment systems. However, experimentally measuring the octanol-water partition coefficient (KOW) is much easier than measuring KOC and KP from the sludge matrix. Accordingly, the empirical model given in Equation (7) was proposed for predicting KP from KOW, which is as follows (Karickhoff 1981): 
formula
7
To overcome the limitation of the empirical modelling and to establish reliable experimental values for the partitioning fate model of TBBPA, the partition coefficients have been estimated in the laboratory.

Table 4 provides a comparison of the experimental results of this study with that of the model estimated literature values. The values obtained in this research were within the expected range of the estimated literature values.

Table 4

Comparison of partitioning coefficients with literature values

Partitioning coefficients Experimentally measured values Literature values (estimated) 
log KOC (L/mg) 6.02 5.43a (KOCWIN, V.2.0, MCI) 6.80a (QSAR using log KOW) 4.52a (derived from log KOW
log KP (L/mg) 4.54 6.49b (EPIWIN, V.3.04; BFRIP report) 
KW (mL/cm20.053 Not available 
Partitioning coefficients Experimentally measured values Literature values (estimated) 
log KOC (L/mg) 6.02 5.43a (KOCWIN, V.2.0, MCI) 6.80a (QSAR using log KOW) 4.52a (derived from log KOW
log KP (L/mg) 4.54 6.49b (EPIWIN, V.3.04; BFRIP report) 
KW (mL/cm20.053 Not available 

aEnvironment Canada: Physical and Chemical Properties of TBBPA.

bBFRIP (2001): Data Summary and Test Plan for TBBPA.

Correlation models

It was observed that the organic carbon partition coefficient for TBBPA was linearly influenced by the increase of sludge organic carbon concentration. The fate and partitioning correlation models for TBBPA in the MBR and activated sludge processes are shown in Figures 2 and 3. The log KOC for TBBPA increases linearly (R2 = 0.94 for MBR, R2 = 0.91 for CAS) with the organic carbon content exhibiting similar correlation with the solid-water partition coefficient KP (R2 = 0.83 for MBR, R2 = 0.82 for CAS) in both biological processes. Figure 3 shows the linear relationship of log KOC with organic carbon for MBR sludge and CAS sludge.

Figure 2

Correlation of log KOC with KP for MBR sludge and activated sludge.

Figure 2

Correlation of log KOC with KP for MBR sludge and activated sludge.

Figure 3

Correlation of log KOC with organic carbon for MBR sludge and CAS sludge.

Figure 3

Correlation of log KOC with organic carbon for MBR sludge and CAS sludge.

Xia et al. (2005) reported a linear relationship of log KP with log KOW for selected compounds similar to the TBBPA experimental results observed in this study. The parameter log KOW is indeed indicative of the partitioning of the compound between organic and aqueous phases (Karickhoff 1981; Dobbs et al. 1989; Schnoor 1996). The relationship between organic carbon and octanol-water partition coefficient was reported for the non-polar organic compounds (Schwarzenbach & Westall 1981; Birkett & Lester 2003). The experimental data for a large set of organic compounds (n = 186) were summarized by Schwarzenbach et al. (2003) that follows the linear model equation: 
formula
8
The correlation models developed in this study from the plotting of the experimental data for TBBPA in activated sludge can be represented as: 
formula
9
 
formula
10
The correlation models developed for TBBPA in MBR sludge can be given as: 
formula
11
 
formula
12
Experimentally determined model coefficients will be helpful for environmental engineers to predict the fate of similar hydrophobic compounds in biological wastewater treatment processes (Dobbs et al. 1989; Struijs et al. 1991). Schwarzenbach et al. (2003) have already shown the usefulness of the identical models developed from the experimental results. Table 5 shows the model constants for TBBPA of this study and other studies of linear regression fittings.
Table 5

Comparison table of the model constants with other studies

Compounds Slopes Intercepts Correlations R2 
aBrominated compounds 0.50 0.81 log KOC with log KOW 0.49 
aPCBs, chlorinated benzenes, 0.74 0.15 log KOC with log KOW 0.96 
aChlorinated phenols 0.89 −0.15 log KOC with log KOW 0.97 
aPAHs 0.98 −0.32 log KOC with log KOW 0.98 
bTBBPA 1.37 −0.20 log KOC with log KP (CAS) 0.82 
bTBBPA 0.15 5.76 log KOC with [OC] (CAS) 0.91 
bTBBPA 1.36 −0.10 log KOC with log KP (MBR) 0.83 
bTBBPA 0.07 5.58 log KOC with [OC] (MBR) 0.94 
Compounds Slopes Intercepts Correlations R2 
aBrominated compounds 0.50 0.81 log KOC with log KOW 0.49 
aPCBs, chlorinated benzenes, 0.74 0.15 log KOC with log KOW 0.96 
aChlorinated phenols 0.89 −0.15 log KOC with log KOW 0.97 
aPAHs 0.98 −0.32 log KOC with log KOW 0.98 
bTBBPA 1.37 −0.20 log KOC with log KP (CAS) 0.82 
bTBBPA 0.15 5.76 log KOC with [OC] (CAS) 0.91 
bTBBPA 1.36 −0.10 log KOC with log KP (MBR) 0.83 
bTBBPA 0.07 5.58 log KOC with [OC] (MBR) 0.94 

PCBs: polychlorinated biphenyls; PAHs: polycyclic aromatic hydrocarbons.

bPresent study.

The developed linear correlation models with high values of R2 indicate strong correlation among the measured partitioning parameters, supporting the evidence for the use of model constants for predicting the fate and partitioning of TBBPA in wastewater sludge. The environmental impacts of related compounds can be estimated from the experimental model coefficients. The equations presented in the present study will allow prediction of the fate, persistence and partitioning of nonvolatile, nonbiodegradable organic pollutants in wastewater process, sludge treatment process and landfill operation process (Dobbs et al. 1989; Wang & Govind 1993).

Isotherm coefficients

Isotherm models were verified for the curve fittings to extract the estimate of the isotherm coefficients. The experimental data for TBBPA biosorption have been fitted to linear regression analysis for determining isotherm model constants. These values are reported in Table 6 and compared with other studies for TBBPA and similar compounds. BPA is structurally similar to TBBPA with endocrine disrupting properties. Compounds having identical structural features generally display similar physicochemical and biological characteristics.

Table 6

Comparison of biosorption coefficients with other studies

Compound (isotherm) Constant Unit/Unitless Estimate R2 Sorbent Reference 
Bisphenol A (Freundlich) log KF L/mg −0.61 0.97 Activated sludge Clara et al. (2004)  
n – 1.49 
Bisphenol A (Freundlich) log KF L/mg 1.83 0.94 Zeolite Tsai et al. (2006)  
n – 4.55 
(Langmuir) qmax ng/mg 111.10 0.96 
KL L/mg 3.10 
(Redlich–Peterson) KR L/mg 140.85 0.99 
α – 1.092 
β – 0.59 
TBBPA (Freundlich) log KF L/mg 0.58 0.99 Carbon nanotube/CoFe2O4-NH2 Zhou et al. (2014)  
n – 1.57 
(Langmuir) qmax ng/mg 30.65 0.98 
KL L/mg 0.12 
TBBPA (Freundlich) log KF L/mg 1.74 0.94 Carbon nanotubes Fasfous et al. (2010)  
n – 3.60 
(Langmuir) qmax ng/mg 64.40 0.97 
KL L/mg 8.00 
TBBPA (Freundlich) log KF L/mg 1.89 0.99 Soil Sun et al. (2008)  
n – 0.75 
TBBPA (Freundlich) log KF L/mg −1.49 0.74 CAS sludge This study 
n – 1.09 
(Langmuir) qmax ng/mg 4.90 × 10−1 0.65 
KL L/mg 2.00 × 10−1 
RL – 7.00 × 10−2 
(Temkin) KT L/mg 1.10 × 10−1 0.75 
b kJ/mol 2.00 × 10−3 
TBBPA (Freundlich) log KF L/mg −2.92 0.86 MBR sludge This study 
n – 1.75 
(Langmuir) qmax ng/mg 4.00 × 10−1 0.90 
KL L/mg 2.70 × 10−3 
RL – 8.80 × 10−1 
(Temkin) KT L/mg 8.60 × 10−1 0.85 
b kJ/mol 5.10 × 10−3 
Compound (isotherm) Constant Unit/Unitless Estimate R2 Sorbent Reference 
Bisphenol A (Freundlich) log KF L/mg −0.61 0.97 Activated sludge Clara et al. (2004)  
n – 1.49 
Bisphenol A (Freundlich) log KF L/mg 1.83 0.94 Zeolite Tsai et al. (2006)  
n – 4.55 
(Langmuir) qmax ng/mg 111.10 0.96 
KL L/mg 3.10 
(Redlich–Peterson) KR L/mg 140.85 0.99 
α – 1.092 
β – 0.59 
TBBPA (Freundlich) log KF L/mg 0.58 0.99 Carbon nanotube/CoFe2O4-NH2 Zhou et al. (2014)  
n – 1.57 
(Langmuir) qmax ng/mg 30.65 0.98 
KL L/mg 0.12 
TBBPA (Freundlich) log KF L/mg 1.74 0.94 Carbon nanotubes Fasfous et al. (2010)  
n – 3.60 
(Langmuir) qmax ng/mg 64.40 0.97 
KL L/mg 8.00 
TBBPA (Freundlich) log KF L/mg 1.89 0.99 Soil Sun et al. (2008)  
n – 0.75 
TBBPA (Freundlich) log KF L/mg −1.49 0.74 CAS sludge This study 
n – 1.09 
(Langmuir) qmax ng/mg 4.90 × 10−1 0.65 
KL L/mg 2.00 × 10−1 
RL – 7.00 × 10−2 
(Temkin) KT L/mg 1.10 × 10−1 0.75 
b kJ/mol 2.00 × 10−3 
TBBPA (Freundlich) log KF L/mg −2.92 0.86 MBR sludge This study 
n – 1.75 
(Langmuir) qmax ng/mg 4.00 × 10−1 0.90 
KL L/mg 2.70 × 10−3 
RL – 8.80 × 10−1 
(Temkin) KT L/mg 8.60 × 10−1 0.85 
b kJ/mol 5.10 × 10−3 

Biosorption is a complex process of physical, chemical and biological phenomena dependent on multiple factors of the system under study. Therefore, the results reported in the literature fall under a wide range in orders of magnitude. BPA is structurally similar to TBBPA, containing additional bromine atoms. The coefficients measured by Clara et al. (2004) were studied in sewage sludge which resembled the present study. The values of isotherm coefficients for TBBPA are 2.5 to 4.5 times higher in order of magnitude than BPA. This is reasonable with respect to the higher hydrophobicity of TBBPA compared to BPA. Other literature values were studied for the different sorbents but were very similar to the current study. Multiple model fittings were not available in the literature for TBBPA. It may be that the experimental data did not fit with models or the authors did not apply models other than the Freundlich and the Langmuir models.

The RL value of 0.07 for activated sludge indicates favourable biosorption for TBBPA and 0.88 for MBR sludge indicates favourable biosorption but less favourable than CAS sludge. The RL values for TBBPA on a carbon nanotube sorbent surface were found to be 0.01 and 0.04 at temperature 25 °C and 45 °C, respectively (Fasfous et al. 2010).

The negative value of ΔG0 refers to the thermodynamic spontaneity of the biosorption process. Generally for physical biosorption, the values of ΔG0 are in the range of 0 to −20 kJ/mol (Yu et al. 2004). The negative values of ΔG0 = −7.79 KJmol−1 (CAS) and ΔG0 = −10.84 kJmol−1 (MBR) were calculated from the biosorption equilibrium. Therefore, the biosorption removal of TBBPA is a favourable thermodynamic process. Fasfous et al. (2010) reported ΔG0 values for TBBPA biosorption in the range of 10.2 to −17.9 kJ/mol.

The values of Freundlich intensity constant, n, equal 0.92 and 0.57 for the two types of sludge, indicate favourable biosorption (potentially chemisorption) and cooperative adsorption (Samarghandi et al. 2009; Foo & Hameed 2010). The positive values of adsorption energies: 2 kJ/mol and 5.1 kJ/mol, for the Temkin model constant b predict an exothermic biosorption process (Samarghandi et al. 2009). Therefore, the biosorption removal of TBBPA is a favourable process both kinetically and thermodynamically.

The higher order three parameter model equations did not allow linear plotting of the experimental data (Wang & Govind 1993). Thus, the complex model equations could not be solved for the linear regression curve fittings. Nonlinear regression analysis was applied to experimental data for extracting model constants using the MS Excel data solver (Wong et al. 2004). Most models were suitable for the curve fittings by nonlinear regression analysis except the Harkin–Jura and the Halsey model isotherms. These two models account for the heteroporous nature of the adsorbent which is not the case for CAS sludge (Samarghandi et al. 2009). Thus, it is obvious that these two adsorption models, employing pore-filling mechanisms for heterogeneous porous media, will not be suitable for explaining biosorption processes of TBBPA by activated sludge and aerobic MBR sludge.

The estimated isotherm coefficients obtained from the experimental data by nonlinear analysis are reported in Table 7. The nonlinear regression analysis adopts a minimization approach for the sum of the squared errors (SSE) to attain the closest matching of the model equation to the experimental data. KR (L/mg) is the isotherm equilibrium constant for the Redlich–Peterson equation similar to Freundlich constant, Langmuir constant or Henry's constant. The values of α (2.58, 2.57) are in the reasonable range as compared to the literature results (Wong et al. 2004). β is the exponent with a value between 1 and 0. When β = 0, then the isotherm equation becomes identical to the form of Freundlich equation. That was the case for the regression analysis of the experimental data for this study.

Table 7

Isotherm constants for TBBPA by nonlinear regression analysis

Isotherm model Isotherm constant Unit/Unitless Value for activated sludge Value for MBR sludge 
Redlich–Peterson KR L/mg 1.60 × 10−1 6.00 × 10−2 
αR – 2.58 3.57 
β – 
SSE – 6.50 × 10−1 5.00 × 10−2 
Toth qmax ng/mg 4.00 × 10−2 3.90 × 10−4 
Kt ng/mg 1.40 1.80 
t – 2.40 × 10−1 1.00 × 10−2 
SSE – 8.00 × 10−1 5.00 × 10−2 
Sips qmax ng/mg 3.80 × 103 0.32 × 10−2 
KS L/mg 6.50 × 10−3 0.43 
m – 5.40 × 102 1.00 × 103 
SSE – 6.50 × 10−1 5.00 × 10−2 
BET qmax ng/mg 4.00 × 103 5.80 × 102 
Cs ng/L 6.60 × 102 0.78 × 102 
B – 6.70 × 10−3 7.90 × 10−4 
SSE – 6.20 × 10−1 8.00 × 10−3 
Jovanovic qmax ng/mg −2.40 × 10−2 −5.80 × 10−2 
KJ – 1.00 × 10−1 1.10 × 10−1 
SSE – 3.00 × 10−2 7.70 × 10−1 
Dubinin–Radushkevich qs ng/mg 9.40 × 10−1 2.70 
BD kJ/mol 0.980 × 102 0.88 × 102 
SSE – 1.50 × 10−1 7.60 × 10−1 
Isotherm model Isotherm constant Unit/Unitless Value for activated sludge Value for MBR sludge 
Redlich–Peterson KR L/mg 1.60 × 10−1 6.00 × 10−2 
αR – 2.58 3.57 
β – 
SSE – 6.50 × 10−1 5.00 × 10−2 
Toth qmax ng/mg 4.00 × 10−2 3.90 × 10−4 
Kt ng/mg 1.40 1.80 
t – 2.40 × 10−1 1.00 × 10−2 
SSE – 8.00 × 10−1 5.00 × 10−2 
Sips qmax ng/mg 3.80 × 103 0.32 × 10−2 
KS L/mg 6.50 × 10−3 0.43 
m – 5.40 × 102 1.00 × 103 
SSE – 6.50 × 10−1 5.00 × 10−2 
BET qmax ng/mg 4.00 × 103 5.80 × 102 
Cs ng/L 6.60 × 102 0.78 × 102 
B – 6.70 × 10−3 7.90 × 10−4 
SSE – 6.20 × 10−1 8.00 × 10−3 
Jovanovic qmax ng/mg −2.40 × 10−2 −5.80 × 10−2 
KJ – 1.00 × 10−1 1.10 × 10−1 
SSE – 3.00 × 10−2 7.70 × 10−1 
Dubinin–Radushkevich qs ng/mg 9.40 × 10−1 2.70 
BD kJ/mol 0.980 × 102 0.88 × 102 
SSE – 1.50 × 10−1 7.60 × 10−1 

Sum of the squared errors, where = experimental TBBPA biosorption and = TBBPA biosorption predicted by the model isotherm.

The Toth isotherm is a modified form of the Langmuir equation with the heterogeneity factor, t. The magnitude of the factor can be between the values of a small fraction to unity. The other isotherm constants may acquire different values depending on the adsorption type. qm (ng/mg) is the maximum adsorption capacity parameter applicable to all types of isotherm equations such as Sips, BET and Jovanovic models. The characteristic constant values obtained in this study for different isotherm models were in good agreement with the other adsorption studies (Wong et al. 2004; Samarghandi et al. 2009). However, those studies have been designed to interpret dye adsorption phenomena. These models are widely used in the treatment of wastewater from textiles, dye industries, agricultural wastes and pulp and paper industries (Foo & Hameed 2010; Dada et al. 2012; Paska et al. 2014). The present study will extend these applications to municipal wastewater for the removal of TBBPA and similar nonvolatile, nonbiodegradable hydrophobic compounds.

Overall the models have similarities and differences in comparison to each other. The correlations and the magnitude of the coefficients from different models clearly demonstrated the usefulness and applications of the research in the engineering field. Experimental values will help to validate and verify the empirical models. However, there are drawbacks and constraints to these constants for transferring laboratory data to the field. Aspects of variations and limitations to the application should be considered. Some modifications, adjustments or optimization may be required in the real engineering applications.

CONCLUSIONS

The partitioning coefficients presented in this paper were successfully measured experimentally. The experimental values showed excellent goodness of fit with the theoretical models for applying to biosorption process modelling, measuring and predicting of the persistent pollutants. In addition, the isotherm constant values were in the acceptable range according to the theories of those models. The developed correlation models for partition coefficients will be useful to determine the fate, persistence and partitioning of TBBPA in wastewater treatment processes, sludge treatment processes and landfill applications.

The correlation models provide an easy means of assessing fate, partitioning and transport of the contaminant without time-consuming laboratory effort. The important purpose of the correlation models is to develop engineering tools for environmental management planning, monitoring, problem identification, impact assessment, risk assessment, pollution load evaluation and mitigation of transboundary migration. The kinetic coefficients determined in this research are valuable for potential engineering applications in tackling TBBPA proliferation in the environmental compartments. These models will also be applicable to the biosorption removal of contaminants in wastewater with properties similar to TBBPA.

REFERENCES

REFERENCES
APHA
2012
Standard Methods for the Examination of Water and Wastewater
,
22nd edn
.
American Public Health Association
,
Washington, DC
,
USA
.
BFRIP
2001
Summary and Test Plan for Tetrabromobisphenol A, TBBPA.
Brominated Flame Retardant Industry Panel
,
Arlington, VA
,
USA
.
Birkett
J. W.
Lester
J. N.
2003
Endocrine Disrupters in Wastewater and Sludge Treatment Processes
.
CRC Press
,
Boca Raton, FL
,
USA
.
Cariou
R.
Antignac
J. P.
Marchand
P.
Berrebi
A.
Zalko
D.
Andre
F.
Bizec
B. L.
2005
New multiresidue analytical method dedicated to trace level measurement of brominated flame retardants in human biological matrices
.
J. Chromatogr. A
1100
,
144
152
.
Clara
M.
Strenn
B.
Saracevic
E.
Kreuzinger
N.
2004
Adsorption of bisphenol-A, 17 β-estradiol and 17α-estradiol to sewage sludge
.
Chemosphere
56
,
843
851
.
Covaci
M. A.
Voorspoels
S.
Abdallah
M. A.-E.
Greens
T.
Harrad
S.
Law
R. J.
2009
Analytical and environmental aspects of the flame retardant tetrabromobisphenol-A and its derivatives
.
J. Chromatogr. A
1216
,
346
363
.
Environment Canada
2013
Physical and chemical properties of TBBPA
.
Screening Assessment Report
,
Environment Canada, Health Canada
,
Toronto
.
Frederiksen
M.
Vorkamp
K.
Bossi
R.
Riget
F.
Dam
M.
Svensmark
B.
2007
Method development for simultaneous analysis of HBCD, TBBPA, and dimethyl-TBBPA in marine biota from Greenland and the Faroe Islands
.
Int. J. Environ. Anal. Chem.
87
,
1095
1109
.
Karickhoff
S. W.
Brown
D. S.
Scott
T. A.
1979
Sorption of hydrophobic pollutants on natural sediments
.
Water Res.
13
,
241
248
.
Kokkola
H.
Yli-Pirila
P.
Vesterinen
M.
Korhonen
H.
Keskinen
H.
Romakkaniemi
S.
Hao
L.
Kortelainen
A.
Joutsensaari
J.
Warsnop
D. R.
Virtanen
A.
Lehtinen
K. E. J.
2014
The role of low volatile organics on secondary organic aerosol formation
.
Atmos. Chem. Phys.
14
(
3
),
1689
1700
.
Lee
H. B.
Peart
T. E.
2002
Organic contaminants in Canadian municipal sewage sludge. Part I. Toxic or endocrine-disrupting phenolic compounds
.
Water Qual. Res. J. Can.
37
,
681
696
.
Paska
O. M.
Pacurariu
C.
Muntean
S. G.
2014
Kinetic and thermodynamic studies on Methylene Blue biosorption using corn-husk
.
RSC Adv.
4
,
62621
62630
.
Samarghandi
M. R.
Hadi
M.
Moayedi
S.
Askari
F. B.
2009
Two-parameter isotherms of methyl orange sorption by pinecone derived activated carbon
.
Iran J. Environ. Health Sci. Eng.
6
(
5
),
285
294
.
Schnoor
J. L.
1996
Environmental Modeling: Fate and Transport of Pollutants in Water, Air, and Soil
.
John Wiley & Sons
,
New York
,
USA
.
Schwarzenbach
P. R.
Gschwend
P. M.
Imboden
D. M.
2003
Environmental Organic Chemistry
,
2nd edn
.
Wiley
,
New York
,
USA
.
Site
A. D.
2001
Factors affecting sorption of organic compounds in natural sorbent/water systems and sorption coefficients for selected pollutants
.
J. Phys. Chem. Ref. Data
30
(
1
),
188
439
.
Srinivasan
A.
Viraraghavan
T.
2010
Decolorization of dye wastewaters by biosorbents: A review
.
J. Environ. Manage.
91
,
1915
1929
.
Thomsen
M.
Arias
P.
Sjodin
A.
Bergman
A.
2002
An overview of commercially used brominated flame retardants, their application, their use patterns in different countries/regions and possible modes of release
.
Environ. Int.
29
,
683
689
.
US EPA
1996
Fate, Transport and Transformation Guidelines – Activated Sludge Sorption Isotherm. Prevention, Pesticide and Toxic Substances
.
EPA-712-C-98-298
.
US Environmental Protection Agency
,
Washington, DC
,
USA
.
Vijayaraghavan
K.
Yun
Y.-S.
2008
Bacterial biosorbents and biosorption
.
Biotechnol. Adv.
26
,
266
291
.
Wong
Y. C.
Szeto
Y. S.
Cheung
W. H.
McKay
G.
2004
Adsorption of acid dyes on chitosan-equilibrium isotherm analyses
.
Process Biochem.
39
,
693
702
.
Yu
Y.
Zhuang
Y. Y.
Wang
Z. H.
Qiu
M. Q.
2004
Adsorption of water-soluble dyes onto modified resin
.
Chemosphere
54
,
425
430
.