Impact of irreversible sorption of phthalate acid esters on their sediment quality criteria

Abstract

The equilibrium partitioning (EqP) method has been applied to establish sediment quality criteria (SQC); however, it does not consider the nonlinear irreversible sorption of many organic contaminants. In this research, the sorption and desorption of two phthalate esters (PAEs), dimethyl phthalate (DMP) and di(2-ethylhexyl) phthalate (DEHP), in four natural sediments collected from the Yangtze River and the Yellow River were studied; the impact of irreversible sorption of DMP and DEHP on SQC has been evaluated. Based on the reversible and irreversible biphasic sorption model, the values of maximum irreversible sorption capacity (q^irr_max) were 125.19 μg g⁻¹–337.37 μg g⁻¹ for DMP and 515.87 μg g⁻¹–591.40 μg g⁻¹ for DEHP. The q^irr_max value was positively related to the organic carbon and black carbon contents, cation exchange capacity, and surface area of the sediments. The values of the irreversible sorption coefficient K^irr_oc for both DEHP and DMP in the four sediments approximated to a constant of 10^6.46±0.38, which was 1–2 orders of magnitude higher than their reversible sorption coefficient K^rev_oc. The values of SQC for PAEs based on the EqP method were modified by involving the irreversible sorption. The modified SQC of DEHP could be 2 to 20 times higher than the value predicted by the EqP method, and the assessment results for DEHP contamination in the sediments with the modified SQC were more reasonable than those with the non-modified SQC. It indicated that the current SQC based on the EqP method may be unnecessarily strict for specific organic compounds and the irreversible sorption should be taken into account.

---

Environmental impact

The equilibrium partitioning (EqP) method has been applied to establish sediment quality criteria (SQC); however, it does not consider the nonlinear irreversible sorption of many organic contaminants. This paper has studied the irreversible sorption of two phthalate esters (PAEs), dimethyl phthalate (DMP) and di(2-ethylhexyl) phthalate (DEHP), on natural sediments. The values of SQC for PAEs based on the EqP method were modified by incorporating irreversible sorption, and the assessment results for DEHP contamination in the sediments with the modified SQC were more reasonable than those with the non-modified SQC. This study has indicated that the current SQC based on the EqP method may be unnecessarily strict for specific organic compounds and the irreversible sorption should be taken into account.

1. Introduction

In aquatic environments, sediments serve as reservoirs for organic pollutants. Sorption and desorption are important processes for the interaction between hydrophobic organic contaminants and sediments in the aquatic environment.¹ The sorption behavior of organic substances has significant impact on their distribution and bioavailability and therefore affects the level of risk posed by them to benthic organisms. Based on the partition of organic compounds between sediment organic carbon, interstitial (pore) water and benthic organisms, the equilibrium partitioning (EqP) method² has been applied to establish sediment quality criteria (SQC) for the protection of benthic organisms, human health and the aqueous environment.^3–5

Recent studies, however, have shown that the release of many sorbed compounds is hysteretic, and the desorption deviates from the sorption.^6–10 A significant fraction of sorbed compounds would remain on sediments and be blocked in the irreversible sites or rearranged pores.^5–8,11 That is to say, the compounds, which resist moving into the aqueous phase, might do less harm to the aquatic ecosystem.^11,12 Therefore, the equilibrium partitioning method might not be able to precisely predict the desorption behavior of organic contaminants; it is necessary to re-evaluate the currently adopted EqP method for the establishment of SQC.

Some researchers have suggested that the phenomena of irreversible sorption are attributed to the chemical or structural rearrangement of the sorbents and the conversion of low-energy binding sites (reversible sites) to high-energy sites (irreversible sites) with increasing contact time.^9,11,13–16 Those sorbents provide functional groups and/or specific sites where chemical bonding is taking place, leading to bound residues and sorbent structure rearrangement.^9,17 In addition, several studies have inferred that the heterogeneity of organic matters (OM) in soil is one of the main factors for the irreversible sorption. There is a rigid or glassy state of OM present in soil, which contains excessive micropores serving as molecular-scale cavities for the entrance of organic molecules. When guest molecules occupy these cavities, the processes might cause irreversible matrix expansion for glassy OM.^18–20

For the irreversible sorption, there is a maximum sorption capacity for one specific compound in one soil or sediment.^11,12 Some researchers recently suggested that the observed biphasic desorption can be predicted by one complex model in which the irreversible model is combined with the linear partition model, and the complex model has been used to predict the amount of compounds irreversibly adhered on sediments and the maximum irreversible sorption capacity.^12,13

Phthalate acid esters (PAEs) are widely used as plasticizers in polyvinyl chloride, polyvinyl acetates, cellulosics and polyurethanes, and as nonplasticizers in consumer products such as lubricating oils, automobile parts, paints, and food packaging all over the world.^21,22 The contamination of environmental media by PAEs has attracted attention throughout the world.^23–29 The sorption of PAEs to natural sediments was investigated in several studies. Schulten and Thomsen utilized molecular modeling to analyze the sorption of diethyl phthalate onto humic complexes and suggested that the humic acid structure offered favorable internal surface-association opportunities for sorption.³⁰ Furthermore, the understanding of sorption mechanisms of PAEs was improved by investigating the sorption of PAEs onto artificial sorbents, such as activated carbon, polystyrene resin, and carbonaceous nanotubes (CNTs).^21,31–33 In these studies, nonlinear models, the Freundlich model or Langmuir model were used to describe the sorption behavior of PAEs onto these sorbents. However, very little research has been conducted to study the irreversible sorption behavior of PAEs in natural sediments.

In this research, the irreversible sorption of di(2-ethyl hexyl) phthalate (DEHP) and dimethyl phthalate (DMP), two PAEs exhibiting different solubilities and physicochemical properties, was studied using four natural sediments collected from the Yangtze and Yellow Rivers. The irreversible maximum capacities of PAEs on sediments were calculated with the irreversible desorption model based on experimental data of multiple sorption–desorption cycles. Sediment characteristics as well as the physicochemical properties of the organic chemicals were considered as part of the interpretation of differences in the irreversible sorption capacity of the sediment samples. Furthermore, values of SQC for PAEs based on the EqP method were modified by involving the irreversible sorption. The contamination of DEHP in the four sediments was assessed with the modified SQC, and the evaluation results of DEHP in the sediments were compared with that in the porewater.

2. Materials and methods

2.1 Materials

Collection of water and sediment samples. Four sediments collected from four different locations, Hua yuankou (E113°39′25.60′′; N34°54′20.60′′) and Xiao langdi (E112°22′44′′; N34°55′12′′) of the Yellow River, and Dong fengzha (E114°12′44.52′′; N30°28′35.34′′) and Zhuan kou (E114°11′46.56′′; N30°27′25.08′′) of the Yangtze River in April of 2006, were used as sorbents in this study. The samples were air dried, ground and sieved (100-mesh) in turns before use. A total of 5 g individual sediment sample was used to determine the background concentrations of DMP and DEHP. Aqueous DMP and DEHP concentrations were determined for surface water samples collected from Xiao langdi of the Yellow River which was used as dilution water for the aqueous sorption and desorption experiments. In order to measure the concentrations of DMP and DEHP in porewater of the sediments for each sampling site, the porewater sample was separated from the fresh sediment samples by centrifugation at 4000 rpm for 20 min. About 100 ml porewater sample in total collected from each sediment sample was used to determine the concentrations of DMP and DEHP.

Chemicals. Both DMP and DEHP of >99% purity were obtained from Beijing Reagent Company of China. Selected physical and chemical properties^21,34–36 of DMP and DEHP are listed in Table 1. Standard solutions of DMP and DEHP at concentrations of 1 mg L⁻¹ were obtained from New Haven, CT, USA. NaN₃ used to inhibit microbial metabolism³⁷ was from Sigma Chemical Company; methanol (HPLC grade) and hexane (HPLC grade) from Fisher Scientific. Anhydrous sodium sulfate used as desiccant for the extract was from Beijing Reagent Company, China. A stock solution of DMP or DEHP was prepared in methanol at 100 mg L⁻¹ level by adding 100.0 mg DMP or DEHP into a 1 L flask containing methanol. Aqueous solutions of DMP and DEHP for the sorption and desorption experiments were prepared by adding the stock solution into the sterilized diluting water containing 0.1% NaN₃. The background concentrations of DMP and DEHP were 0.2 and 0.3 μg L⁻¹, respectively, in the diluting water and have been factored into the DMP and DEHP concentrations of the aqueous solutions.

Table 1 Selected physical and chemical properties of the studied chemicals

Chemicals	Molar mass	Molecular width (nm)	Molecular length (nm)	Water solubility (mg L^−1)	Log Kow	Log Koc
DMP	194.2	0.366^32	1.048^32	4200^19	1.46–1.9^19	1.9–2.3^19
DEHP	390.56	0.525^34	1.658^34	<0.4^19	7.0–7.8^19	4.9–6.13^19,33

---

2.2 Sorption and desorption experiments

In order to study the irreversible sorption of PAEs on sediments, multiple cycle sorption/desorption experiments were conducted according to Chen et al.¹² The experimental procedures are summarized in Table 2 and briefly described in the remainder of this section. A total of 0.5 g sediment and 50 mL PAEs solution with known DEHP and DMP concentrations, were added in a series of 80 mL glass vials. These vials were sealed by hexane-rinsed tin-foil, and placed in an orbital shaking incubator (125 rpm) at 25 °C in dark for 3–7 days. After equilibrium, these vials were centrifuged at 4000 rpm for 15 min, and the supernatant was taken out for DEHP and DMP analysis. Then, another 50 mL PAEs solution was added into the vials for the second step of sorption. Six steps of such successive sorption experiments were conducted in the first sorption cycle. After the sorption cycle, the desorption experiment was conducted by replacing supernatant with distilled water containing 0.1% NaN₃. After 3–7 days incubation period, the vials were centrifuged at 4000 rpm for 15 min, and the supernatant was taken out for DEHP and DMP analysis. Then another 50 mL distilled water was added again for the second step of desorption. Similar to the sorption cycle, there were also six successive desorption step experiments which constituted the desorption cycle. After the first sorption–desorption cycle, aqueous PAEs solution was added to vials containing the sediments to start the second cycle. Unlike the first cycle, only one sorption step was conducted and was followed by six desorption steps. Supernatant was collected from each sorption and desorption step in the second cycle to analyze the concentrations of DMP and DEHP. At the end of the experiment, the concentrations of DMP and DEHP in the four sediments were determined. Each treatment of the experiments was conducted in triplicate, and the results were expressed as the mean of the triplicate.

Table 2 Multiple sorption–desorption experimental protocols of DMP and DEHP

	Steps	Equilibrium time	Initial solution concentration C0 (μg mL^−1)
First cycle	1st–6th steps for adsorption	3–7 d per step	DMP: 1.0–2.4; DEHP: 0.8–2.0
	1st–6th steps for desorption	3–7 d per step	0
Second cycle	7th adsorption	5 d	DMP: 1.8; DEHP: 1.5
	7th–12th desorptions	4–7 d per step	0

---

In order to get the equilibrium partition coefficients of DMP and DEHP in the sediments for the calculation of SQC based on the EqP method, equilibrium sorption experiments were conducted. A total of 0.5 g sediment and 50 mL PAEs solution, with known concentrations of DMP and DEHP (0.02–0.2 μg mL⁻¹), were added in a series of 80 mL glass vials. These vials were sealed by solvent-rinsed tin-foil, and placed in an orbital shaking incubator (125 rpm) at 25 °C in the dark for 48 h. The preliminary experiment indicated that sorption equilibrium could be achieved within 12 h (see ESI, Fig. S1†). After equilibrium, these vials were centrifuged at 4000 rpm for 15 min, and the supernatant was taken out for DEHP and DMP analysis. Each treatment was conducted in triplicate, and the results were expressed as the mean of the triplicate.

2.3 Chemical analysis

Selected properties of sediments. The contents of total organic carbon (TOC) and carbon black (BC) of the sediments were determined with Vario EL, a CHN elemental analyzer (Elementar, Germany) using methods in the literature,³⁸ and the specific surface area (SSA) of the sediments was determined using the Brunauer, Emmett, and Teller (BET) nitrogen adsorption technique with a Micromeritics ASAP-2020 (Micromeritics, USA). Sediment pH was measured by the potential method, with 2.5 : 1 ratio of water to soil. The cation exchange capacities of the sediment samples were measured with the ammonium acetate method (GB7864-87). The main physicochemical properties of the sediment samples are listed in Table 3. The particle size distribution of the sediments was analyzed by the hydrometer method (LY/T 1225-1999), and the results are shown in Table S1 (ESI†).

Table 3 Physicochemical properties of sediment samples and partition coefficients of PAEs on sediments

Samples	Cation exchange capacity (cmol kg^−1)	Surface area (m^2 g^−1)	TOC (%)	Black carbon (%)	Pore diameter (nm)	pH	DMP Kp (L kg^−1)	DEHP Kp (L kg^−1)
Xiao langdi	4.91	6.38	0.2	0.04	6.54	8.80	664	4655
Hua yuankou	10.65	7.15	0.19	0.06	7.01	8.81	556	4811
Zhuan kou	24.09	9.4	0.81	0.24	10.10	8.24	461	8658
Dong fengzha	34.47	22.22	4.62	1.23	13.48	7.23	332	5800

---

Extraction and analysis of PAEs in water and sediment samples. Each water or sediment sample was extracted three times in succession with 15 ml hexane. All extracts were collected and dried with anhydrous sodium sulfate. Then the volume of each extract was concentrated to 2 mL by the RE-52 rotating evaporator (Shanghai Yarong Bio-chemical Instrument Plant, China). Concentrations of DEHP and DMP in extracts were determined by Varian GC 3800 equipped with a DB-5 elastic quartz capillary chromatographic column (30 m × 0.32 mm and 0.11 μm film thickness, Varian, USA). Nitrogen was used as the carrier gas. The analysis was performed in the splitless mode, and the injection volume was 1 μL. The injector temperature was 250 °C, and the ECD detector temperature was 300 °C. The temperature program of the column oven was set at 100 °C for 1 min, then at 10 °C min⁻¹ to 250 °C and held for 10 min. The PAEs concentration was calculated by the external standard method.

Quality control. A strict regime of quality control was operated in the experiments. The correlation coefficients of the calibration curves were all higher than 0.99. The detection limits of DMP and DEHP were 0.5 and 1.0 μg L⁻¹ for the GC-ECD, respectively. Recoveries of DMP and DEHP were 91.6–99.1% in the water phase and 94.9–99.4% in the sediment phase. The blank sample only containing the distilled water was set to identify the background of DEHP and DMP during the experiment process, and the results indicated that the concentrations of DEHP and DMP in blanks were about 20 and 40 ng L⁻¹, respectively, which were less than 5% of its lowest equilibrium concentrations in the isotherms. The background concentrations of DMP and DEHP in the four sediments ranged from 0.32 to 0.56 μg g⁻¹, and from 1.05 to 15.02 μg g⁻¹, respectively, which were about two orders of magnitude less than the maximum capacities of their irreversible sorption. In addition, the background concentrations of DMP and DEHP in the sediments were counted in the equilibrium sorption experiments. In the control experiments without sorbents, the variations of DEHP and DMP were less than 5%.

3. Results and discussion

3.1 Multiple sorption/desorption of PAEs

For the irreversible compartments in sediments, there were maximum sorption capacities and it was reported that the irreversible sorption compartments could be completely filled when the aqueous equilibrium concentration approached to one third or half of the solubility of compounds.^11,12 For the first cycle of sorption, the aqueous equilibrium concentrations of DEHP in the end were 0.084 μg mL⁻¹, 0.05 μg mL⁻¹, 0.023 μg mL⁻¹ and 0.185 μg mL⁻¹ for the four sediments, respectively, which might be lower than half of its solubility (Table 1); the aqueous equilibrium concentrations of DMP at the end of the first sorption cycle were less than 2 mg L⁻¹ for the four sediment samples, which were lower than half of its solubility (4000 mg L⁻¹).²¹ Therefore, the second sorption cycle was conducted to make sure the maximum irreversible sorption capacity can be reached.

As shown in Fig. 1 and Fig. 2, desorption isotherms of the two cycles obviously deviated from the sorption isotherms, which inferred that the desorption process was irreversible to the sorption process for both DMP and DEHP on natural sediments. For example, the sediment–water partition ratios of DMP and DEHP for the first desorption step were approximately 40 and 3 times that for the six sorption steps, respectively. During the desorption steps of the two cycles in Hua yuankou sediment (Fig. 1), the aqueous concentration of DMP was only about 0.013–0.047 μg mL⁻¹, but more than 240 μg g⁻¹ DMP remained in the sediment at the end of multiple cycles of sorption–desorption. Furthermore, as the successive desorption steps increased, the desorbed amount of DMP decreased. As an example, the amount of desorbed DMP was approximately 40 μg g⁻¹ for the first desorption step of the second cycle, dropping to 3 μg g⁻¹ for the last three steps. For DEHP, the equilibrium concentration was only 0.0022–0.046 μg mL⁻¹ for the two cycles of desorption in Hua yuankou sediment (Fig. 2). At the last three desorption steps (10th–12th steps) in the second cycle, the amount of desorbed DEHP was only about 2 μg g⁻¹, and the concentration of DEHP remaining in the sediment was about 567 μg g⁻¹. The above results imply that the reversible fraction of DMP and DEHP sorbed on sediment could quickly release into the aqueous phase while the irreversible fraction remained on the sediment even after several desorption steps.

Fig. 1 Plots of the multiple sorption/desorption experiments for DMP with irreversible sorption fraction. ◆ represents the 1st–6th sorption; ■ represents the 1st–6th desorption; ▲ represents the 7th sorption; × represents the 7th–12th desorption.

Fig. 2 Plots of the multiple sorption/desorption experiments for DEHP with irreversible sorption fraction. ◆ represents the 1st–6th sorption; ■ represents the 1st–6th desorption; ▲ represents the 7th sorption; × represents the 7th–12th desorption.

3.2 Irreversible sorption parameters of PAEs

The irreversible sorption models^12–14 were applied to study the irreversible sorption of DEHP and DMP in the four sediments. As shown in eqn (1), it was assumed that there were two compartments for the sorption of PAEs onto sediments; one compartment was reversible linear partition (eqn (2)) and the other was irreversible sorption. The irreversible compartment could be described by the Langmuir-type sorption model, and the biphasic sorption could be expressed by eqn (3).

q^total = q^rev + q^irr (1)

q^rev = K^rev_p × C (2)

where q^total is the total sorption amount (μg g⁻¹); q^rev is the reversible linear partition amount (μg g⁻¹); q^irr is the irreversible sorption amount (μg g⁻¹); K^rev_p is reversible sorption coefficient (L kg⁻¹); C is the aqueous concentration of PAEs (μg L⁻¹); K^rev_oc is reversible sorption coefficient normalized by organic carbon percentage (L kg⁻¹); K^irr_oc is analogous to K^rev_oc, but for the irreversible compartment (L kg⁻¹); q^irr_max is the maximum irreversible sorption (μg g⁻¹); f (0 ≤ f ≤ 1) is the fraction of q^irr_max filled during sorption and was assumed to be 1 for all the experiments in this study; f_oc is organic carbon percentage. K^irr_oc and q^irr_max are two important parameters for irreversible sorption of a specific organic compound on sediment. The irreversible partition coefficient K^irr_oc is defined as the ratio of the organic carbon-normalized solid-phase concentration to the solution-phase concentration after extensive desorption when the aqueous-phase concentration remains at significantly low or constant values.

After multiple sorption steps, the irreversible sorption of DEHP and DMP would be saturated. The fraction of irreversibly sorbed PAE would not change significantly during subsequent desorption steps, and the irreversible fraction would remain essentially constant as q^irr_max. This has been manifested by the results shown in Fig. 1 and Fig. 2, the average variations of sorption quantities of DEHP and DMP were approximately only 8 μg g⁻¹ and 10 μg g⁻¹ for the 7th–12th desorption steps of the second cycle, respectively. Therefore, for the desorption process in the second cycle, eqn (3) could be simplified as eqn (4):

q^total = K^rev_oc × f_oc × C + q^irr_max (4)

According to eqn (4), the value of q^irr_max could be determined by fitting the desorption data from the second cycle of the desorption experiments with a linear model. And each desorption isotherm was extrapolated to the y axis and the intercept was the value of q^irr_max for DMP or DEHP.¹⁴ Based on the total sorption and irreversible sorption quantities, the reversible sorption quantity (q^rev) of each desorption step in the second cycle could be determined with eqn (1) and eqn (4). Then K^rev_p could be obtained by fitting q^rev with eqn (2), and K^rev_oc was K^rev_p divided by OC content. The K^irr_oc values shown in Table 4 were the average of K^irr_oc values in the last six desorption steps, and K^irr_p values were obtained through multiplying K^irr_oc by OC content.

Table 4 Sorption and desorption parameters of PAEs on sediment samples obtained from the multiple sorption/desorption experiments

	K ^revp (L kg^−1)	K ^irrp (L kg^−1)	Log K^revoc	Log K^irroc	q ^max (μg g^−1)
a The measured sediment concentration of PAEs at the end of the experiments.
DMP
Xiao langdi	192	8758	4.98	6.64	125.19 (131.72^a)
Hua yuankou	177	11 805	4.97	6.79	240.10 (243.58^a)
Zhuan kou	130	13 770	4.20	6.23	243.24 (248.25^a)
Dong fengzha	230	18 995	3.77	5.68	337.37 (344.17^a)
DEHP
Xiao langdi	408	9919	5.31	6.69	515.87 (528.54^a)
Hua yuankou	197	12 188	5.02	6.81	565.95 (569.21^a)
Zhuan kou	568	25 760	4.85	6.50	564.87 (577.13^a)
Dong fengzha	608	150 316	4.19	6.58	591.40 (606.64^a)

---

At the end of the multiple sorption–desorption experiments, all sediment samples were analyzed for DEHP and DMP concentrations, which were equal to the maximum irreversible sorption capacities (q^irr_max). The mass balance analysis showed that the average loss of DMP and DEHP were 8.5% and 7.3%, respectively for the whole multiple sorption–desorption experiments. As shown in Table 4, the measured values of q^irr_max were very close to those calculated from eqn (4). According to the results shown in Fig. S2 (see ESI†), the calculated log K^rev_oc of DMP ranged from 3.69 to 4.98, which were higher than the reported values of K_oc (1.9–2.3),²¹ suggesting that other sorption mechanisms exist for DMP in addition to its partition in organic carbon. The calculated log K^rev_oc of DEHP ranged from 4.12 to 5.31, which were comparable to the reported values of K_oc (4.94–6.13),^21,35 suggesting that the main reversible sorption mechanism for DEHP is its partition in organic carbon. As shown in Table 4, the K^rev_oc values of both DMP and DEHP were about 1–2 orders of magnitude lower than the K^irr_oc values.

3.3 Factors affecting the irreversible sorption of PAEs

As shown in Table 4, although the value of log K_ow for DEHP was six orders of magnitude higher than that for DMP, the q^irr_max values of DEHP were only 2–4 times higher than that of DMP for the four sediments. In addition, the value of log K^irr_oc for DEHP was similar to that for DMP, with the average of 6.647 ± 0.048 for the former and 6.32 ± 0.72 for the latter, and the values of log K^irr_oc for both DEHP and DMP in the four sediments approximated to a constant of 6.46 ± 0.38. This was in agreement with the results reported by Chen et al.,¹² who also found the value of log K^irr_oc of five kinds of chlorobenzene on sediments was a constant. This suggested that the mechanisms and predominant influencing factor of irreversible sorption for DMP and DEHP were similar, and the value of K_ow was not the major factor affecting the irreversible sorption of PAEs.

Phthalate ester molecules are structurally composed of aromatic rings with carboxyl groups which provide the opportunity to form hydrogen bonds and have π–π electron-donor–acceptor (EDA) interactions with the sorbent surface. It has been reported that the PAE molecules could chemically bind with functional groups in heterogeneous sorbents.^31,32 Furthermore, n–π (nonbinding electron–π) EDA interaction³⁹ might occur during sorption of PAEs to sediment surface. Functional groups on the sediment surface, such as hydroxyl groups and carboxyl groups might serve as strong n-donors interacting with π-acceptor PAE molecules, leading to the irreversible sorption of PAEs. Another possible mechanism contributing to extensive irreversible sorption of PAEs is micropore filling in a glassy state of OM such as BC. It is suggested that the entrance of guest molecules may lead to the rearrangement of micropores in sorbents and bring about irreversible sorption phenomena.¹⁰ According to our previous research,⁴⁰ strong and nonlinear sorption was observed for the sorption of DEHP on both pure BC and environmental BC with the Freundlich exponent ranging from 0.55 to 0.75, and the measured K_BC (BC–water partition coefficient) of DEHP was about one order of magnitude higher than its organic carbon–water partition coefficient.

As shown in Fig. S3 (see ESI†), the values of q^irr_max for DMP and DEHP were positively related to the surface area, organic carbon content, and cation exchange capacity of the sediments, as there could be high densities of functional groups in sediments with higher organic carbon content, surface area, and cation exchange capacity. In addition, the increase of surface area might indicate the availability of more micropores for guest molecules or high-energy sites for irreversible sorption. Black carbon is known as a strong sorbent for many organic chemicals,^1,41 and the sorption of chemicals on black carbon possesses the characteristics of non-linear, competition and irreversible sorption. As shown in Fig. S3 (see ESI†), the values of q^irr_max for both DMP and DEHP rose up with the increase in black carbon content in sediment samples. In addition, the values of q^irr_max for DEHP and DMP in Zhuan kou sediment was close to that in Hua yuankou sediment, while the black carbon content of Zhuan kou was four times greater than that in Hua yuankou. It might result from the different composition of organic matter in these two sediments as they were collected from different rivers. Our previous research indicated that the log K_BC value of DEHP on the Hua yuankou sediment was highest among the sediments collected from the Yangtze River and the Yellow River.⁴⁰

3.4 Impact of irreversible sorption on sediment quality criteria

Sediment quality criteria (SQC) introduced by the USEPA is based on the linear equilibrium partitioning (EqP) model (eqn (5)), in which it is assumed that a nonionic chemical partitions between sediment and porewater.⁴²

SQC = K_p × WQC (5)

where K_p is the equilibrium partition coefficient of a specific chemical in sediment (L kg⁻¹); WQC is the value of the water quality criteria for this chemical. However, according to our investigations of PAE sorption onto natural sediments, there are other mechanisms driving the sorption and desorption process besides linear partitioning equilibria, and significant amounts of adsorbed PAEs remain associated with the sediment phase as opposed to desorbing into the aqueous phase. Therefore the irreversible sorption should be taken into account to calculate the SQC. Making reference to previous research,¹² a new SQC model composed of two sorption parts was proposed to calculate the modified SQC* of DEHP and DMP for the four sediments as eqn (6),where the first part represents the reversible partitioning fraction and the second part stands for the irreversible sorption fraction.

The values of K_p in eqn (5) were obtained from the equilibrium sorption experiments of DMP and DEHP (Table 3, Fig. S4 in the ESI†), and the values of K^rev_p, K^irr_oc, and q^irr_max in eqn (6) were obtained from the results of multiple cycles of sorption/desorption experiments for DMP and DEHP (Table 4). Based on eqn (5) and eqn (6), the SQC and modified SQC were calculated for DMP and DEHP in the four sediments. As the porewater concentrations of DMP in the four sites were much lower than its WQC (270 000 μg L⁻¹),⁴³ the SQC for DMP is not discussed here. As shown in Table 5, when considering the impact of irreversible sorption of DEHP on SQC, the value of modified SQC based on eqn (6) for each site was greater than that based on eqn (5). For Xiao langdi, Hua yuankou and Zhuan kou, the DEHP concentrations in the sediment and porewater were lower than the values of SQC and WQC, respectively. For Dong fengzha, although the aqueous concentration of DEHP (2.01 μg L⁻¹) was below WQC⁴³ (2.2 μg L⁻¹) for this site, the concentration of DEHP (15.02 μg g⁻¹) exceeded its EqP-SQC (12.77 μg g⁻¹). However, the sediment concentration of DEHP at Dong fengzha was much lower than the modified SQC (213.43 μg g⁻¹); this result coincided with the aqueous concentration of DEHP. Therefore, the modified SQC model seemed more reasonable to estimate the risk of DEHP contamination.

Table 5 SQC and modified SQC of DEHP for natural sediments

Sediment	C s (μg g^−1)	C W (μg L^−1)	WQC (μg L^−1)	SQC – eqn (5) (μg g^−1)	Modified SQC – eqn (6) (μg g^−1)
Xiao langdi	1.05	1.96	2.2	10.23	21.83
Hua yuankou	2.11	1.07	2.2	10.58	26.03
Zhuan kou	6.27	1.57	2.2	19.05	52.75
Dong fengzha	15.01	2.01	2.2	12.77	213.43

---

As shown in Table 5, when the irreversible sorption was taken into account, the value of modified SQC for DEHP could be 2 to 20 times higher than that derived from the EqP model for the four sediments. This implied that the SQC based on the EqP method may be unnecessarily strict for some specific organic substances; the irreversible fraction should not be ignored for further sediment quality studies. In addition, whether the modified SQC should be used depends on the pollution status of organic substances. When sediment acts as a source for organic pollutants, transferring into the aqueous phase, desorption controls the effective pollutant concentration in the porewater. Therefore, the modified SQC should be used to evaluate the pollution risk in this case. On the contrary, when sediment acts as a sink for organic pollutants in the water phase and sorption processes control the pollutant concentration in sediment, the SQC based on the EqP method should be used to evaluate the pollution risk. For the Dong fengzha site in this research, the ratio of DEHP concentration in the sediment to the water phase was approximately 7500 L kg⁻¹, which was higher than its equilibrium distribution coefficient (5800 L kg⁻¹), therefore, the desorption process would control the DEHP concentration in the water phase, and the modified SQC should be applied to evaluate the DEHP pollution at this site.

4. Conclusions

This study showed that the sorption of PAEs included both linear and non-linear (signifying irreversible) regimes. For the four sediment samples, the irreversible sorption maxima were 126.19–337.37 μg g⁻¹ and 515.87–591.40 μg g⁻¹, respectively for DMP and DEHP which positively correlated to the surface areas, cation exchange capacities, and carbon black contents of the sediments. The OC-normalized partition coefficients for the reversible compartment (K^rev_oc) were 10^3.69–10^4.98 for DMP which were higher than K_oc values reported in the literature. This suggested that other reversible surface-association mechanisms exist besides the linear partitioning onto the organic carbon fraction of the sediment. The values of K^rev_oc were 10^4.12–10^5.31 for DEHP; they were close to those (K_oc) reported in the literature, suggesting the linear partition on organic carbon is the main reversible sorption mechanism. Although the physicochemical properties of DMP and DEHP are different, the values of the irreversible sorption coefficient K^irr_oc for both DEHP and DMP in the four sediments approximated to a constant of 10^6.46±0.38.

The values of SQC for PAEs based on the EqP method were modified by incorporating irreversible sorption. The modified SQC of DEHP could be 2 to 20 times higher than the values predicted by the EqP method, and the assessment results for DEHP contamination in the sediments with the modified SQC were more reasonable than those with the non-modified SQC. This indicates that the current SQC which is based on the EqP method may be unnecessarily strict for highly hydrophobic contaminants like DEHP and that irreversible sorption should be factored into assessments of risk for these compounds. This study provides a direction for incorporation of irreversible sorption of hydrophobic organic contaminants into partitioning modeling so that better sediment quality criteria can be determined.

Acknowledgements

This work was supported by the Major State Basic Research Development Program (No. 2010CB951104), National Natural Science Foundation of China (No. 51079003), and Program for New Century Excellent Talents in University (NCET-09-0233).

References

1. G. Cornelissen, Ö. Gustafsson, T. D. Bucheli, M. T. O. Jonker, A. A. Koelmans and P. C. M. Van Noort, Environ. Sci. Technol., 2005, 39, 6881–6895.
2. US Environmental Protection Agency, 2004, EPA 823-R-97-006.
3. US Environmental Protection Agency, 2003, EPA 600-R-02-013.
4. R. A. Hoke, G. T. Ankley, A. M. Cotter, T. Goldenstein, P. A. Kosian, G. L. Phipps and F. M. Vander Meiden, Environ. Toxicol. Chem., 1990, 13, 157–166.
5. D. Shea, Environ. Sci. Technol., 1988, 22, 1256–1261.
6. A. T. Kan, G. Fu and M. B. Tomson, Environ. Sci. Technol., 1994, 28, 859–867.
7. A. S. Gunasekara and B. S. Xing, J. Environ. Qual., 2003, 32, 240–246.
8. L.Y and J. J. Pignatello, Environ. Sci. Technol., 2004, 38, 5853–5862.
9. W. L. Wu and H. W. Sun, Chemosphere, 2010, 81, 961–967.
10. X. Y. Yu, L. G. Pan, G. G. Ying and R. S. Kookana, J. Environ. Sci., 2010, 22, 615–620.
11. A. T. Kan, G. Fu, M. Hunter, W. Chen, C. H. Ward and M. B. Tomson, Environ. Sci. Technol., 1998, 32, 892–902.
12. W. Chen, A. T. Kan and M. B. Tomson., Environ. Sci. Technol., 2000, 34, 385–392.
13. A. T. Kan, G. Fu, M. Hunter and M. B. Tomson, Environ. Sci. Technol., 1997, 31, 2176–2185.
14. W. J. Braida, J. J. Pignatello, Y. F. Lu and P. I. Ravikovitch, Environ. Sci. Technol., 2003, 37, 409–417.
15. G. A. C. Ehlers and A. P. Loibner, Environ. Pollut., 2006, 141, 494–512.
16. G. Cornelissen, C. M. N. Paul, J. R. Parsons and H. A. J. Govers, Environ. Sci. Technol., 1997, 31, 454–460.
17. B. Gevao, K. T. Semple and K. C. Jones, Environ. Pollut., 2000, 108, 3–14.
18. M. Sander and J. J. Pignatello, Environ. Toxicol. Chem., 2009, 28, 447–457.
19. Y. Lu and J. J. Pignatello, Environ. Sci. Technol., 2002, 36, 4553–4561.
20. M. Sander, Y. Lu and J. J. Pignatello, Environ. Sci. Technol., 2006, 40, 170–178.
21. C. Stales, D. Peterson, T. Parkerton and W. Adams, Chemosphere, 1997, 35, 667–749.
22. S. Wang, P. Wang and C. Fan, J. Environ. Monit., 2009, 11, 882–886.
23. K. Asaoka and K. Hagihara, Environ. Toxicol. Chem., 2000, 64, 679–685.
24. F. Wang, X. H. Xia and Y. J. Sha, J. Hazard. Mater., 2008, 154, 317–324.
25. X. H. Xia, L. Y. Yang, Q. W. Bu and R. M. Liu, J. Environ. Qual., 2011, 40, 1–9.
26. Y. J. Sha, X. H. Xia and Z. F. Yang, Environ. Monit. Assess., 2007, 124, 277–287.
27. US Environmental Protection Agency 2009, EPA 816-F-09-0004.
28. S. Sirivithayapakorn and K. Thuyviang, Bull. Environ. Contam. Toxicol., 2010, 84, 503–506.
29. L. Ma, S. Chu and X. Xu, J. Environ. Monit., 2003, 5, 786–790.
30. H. R. Schulten, M. Thomsen and L. Carlsen, Chemosphere, 2001, 45, 357–369.
31. J. Wu, Z. Xu, W. Zhang, L. Lv, B. Pan, G. Nie and M. Li, AIChE J., 2010, 56, 2699–2705.
32. W. M. Zhang, Z. W. Xu and B. C. Pan, J. Colloid Interface Sci., 2008, 325, 41–47.
33. F. Wang, J. Yao, K. Sun and B. S. Xing, Environ. Sci. Technol., 2010, 44, 6985–6991.
34. Y. Kiso, Y. Nishimura, T. Kitao and K. Nishimura, J. Membr. Sci., 2000, 171, 229–237.
35. C. E. Mackintosh, J. A. Maldonado, M. G. Ikonomou and F. A. P. C. Gobas, Environ. Sci. Technol., 2006, 40, 3481–3488.
36. M. Bodzek, M. Dudziak and K. Luks-Betlej, Desalination, 2004, 162, 121–128.
37. D. Patureau and E. Trably, Biodegradation, 2006, 17, 9–17.
38. X. H. Xia, Y. R. Li, Z. Zhou and C. L. Feng, Chemosphere, 2010, 78, 1329–1336.
39. M. Keiluweit and M. Kleber, Environ. Sci. Technol., 2009, 43, 3421–3429.
40. X. H. Xia, Z. N. Dai and J. Zhang, J. Environ. Monit., 2011, 13, 2858–2864.
41. G. Cornelissen, J. Haftka, J. Parsons and O. Gustafsson, Environ. Sci. Technol., 2005, 39, 3688–3694.
42. US Environmental Protection Agency, 2003, EPA 600-R-02-014.
43. US Environmental Protection Agency, 2002, EPA 822-R-02-047.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c1em10649j

---