Adsorption of pyridine from aqueous solutions onto polyaluminium chloride and anionic polyacrylamide water treatment residuals

The adsorption performance of pyridine onto polyaluminium chloride (PAC) and anionic polyacrylamide (APAM) water treatment residuals (WTRs) was investigated by batch experiments. This study confirmed the assumption that PAC–APAM WTRs had the ability to remove pyridine. The non-linear Dubinin–Radushkevich model and non-linear Freundlich model better described the isotherms, indicating that the adsorption was a chemically controlled multilayer process. The pyridine adsorption rate was simultaneously controlled by external film diffusion and intraparticle diffusion. The adsorption of pyridine was an endothermic reaction with randomness increase. The pyridine adsorption decreased with pH increase. Pyridine removal was observed to be a linear increase from 6.16% to 96.18%, with the increase of dosage from 2.5 g/L to 15 g/L. The Langmuir maximum adsorption capacity was 3.605 mg/g while the theoretical isotherm saturation capacity was 9.823 mg/g. Therefore, PAC–APAM WTRs recycled into contaminated soils for remediation is expected to be an innovative alternative disposal method. More research is recommended in the future to identify detailed adsorption mechanisms and the most appropriate mixing ratio of PAC– APAM WTRs to contaminated soils under various climatic conditions.


INTRODUCTION
Pyridine has been widely regarded as a typical and concerning pollutant found in surface water, soil water and groundwater around the world. Although it occurs naturally in the environment at extremely low concentrations, the anthropogenic amounts of pyridine are more significant in various environmental systems due to the discharge found in industrial wastewater (Zhu et al. a). Those industries include coal and coking processing, coal-fired power generating, shale oil processing, refinery, dye and paint production, agricultural chemicals manufacturing including insecticides, herbicides and fungicides, pharmaceuticals production, textile water repellent production, rubber production, and food additive production (Zhu et al. a; Zhu et al. b; Singh et al. ). As a typical aromatic N-heterocyclic compound, pyridine has strong toxicologic effects on aquatic and terrestrial ecological systems and also on humans, even at low concentrations because of the strong carcinogenic potential (Zhu et al. b) and the harmful influence on the functioning of reproductive systems, immune systems, livers and kidneys (Singh et al. ). Because pyridine has a high solubility in water with a lasting persistence in the environment, it can easily enter nearby surface water bodies after the release by industrial wastewater discharge and infiltrate soils where it can leach downwards into underlying aquifers. Therefore, pyridine removal or immobilization is an important concern for minimizing the threat of organic pollutants to the health of human and ecological systems.
Water treatment residuals (WTRs) are wastes produced from municipal water treatment plants. The quantity of WTRs production has increased sharply in response to rapid urbanization, population increase, and life quality improvements. Currently, the main disposal method of WTRs is landfilling. However, the sustainability of landfilling WTRs is questioned from the current sustainable development and management context due to limited land resources, high cost for transporting WTRs from municipal water treatment plants to the remote landfilling sites, and the potential environmental and human health risks (Ahmad et al. ). Therefore, beneficial reuse of WTRs have become the primary interest for sustainable waste management in local and global water treatment industries (Duan & Fedler ). In addition, alternative disposal approaches for WTRs are being developed with an aim to replace landfilling. Due to the large particle surface area and strong absorption affinity to some pollutants, WTRs have been proposed as a low-cost sorbent for removing pollutants from wastewater (Ahmad et al. ) and remediating various contaminated soils (Tay et al. ; Zhao et al. ). High adsorption capacities of WTRs have been confirmed in previous studies for phosphorus, arsenate, heavy metals (Ahmad et al. ) and dye (Gadekar & Ahammed ).
Removal of pyridine from wastewater by adsorption processes has been considered because of its relatively lower operational cost with satisfying effects and it can be accomplished without the production of potentially dangerous byproducts often produced in oxidative or reductive processes (Zhu et al. a). Although previous studies have investigated the adsorption process for pyridine by using synthetic porous materials (Zhu et al. b), there are no reports available on pyridine adsorption by directly using wastes residuals without pre-processing by physico-chemical means. Our previous studies showed that WTRs had the ability to immobilize heavy metal ions in the contaminated soils (Duan & Fedler ). Moreover, more WTRs are produced by utilizing polyaluminium chloride (PAC) as the coagulant for clarifying raw water and anionic polyacrylamide (APAM) as the dewatering agent during sludge handing in water treatment plants worldwide (Duan & Fedler ). This type of WTR is typically called PAC-APAM WTR (Duan & Fedler ).
In this study, one hypothesis is that PAC-APAM WTRs have the ability to remove pyridine by adsorption, yet its adsorption performance for pyridine remains unknown. Therefore, the adsorption of pyridine onto PAC-APAM WTRs was investigated by batch experiments. The objectives of this study were: (1) to measure the adsorption capacity of PAC-APAM WTRs for pyridine by conducting an adsorption isotherm study, (2) to identify the critical process that controls the pyridine adsorption rate onto PAC-APAM WTRs by conducting adsorption kinetic study, and (3) to investigate the effects of temperature and pH on the pyridine adsorption by PAC-APAM WTRs. Overall, the study provides data about the remediation of pyridine contaminated soils for practicing engineers with an aim to recycling PAC-APAM WTRs in a sustainable manner.

Collection and characterization of PAC-APAM WTRs
The air-dried PAC-APAM WTRs cakes were collected at the WTRs cake piling site of a municipal water treatment plant in the city of Taiyuan, Shanxi Province, China. The sampled cakes were meshed and dried overnight at 105 C, finely ground, and screened through a 0.25-mm sieve (0 < particle size < 0.25 mm). The obtained PAC-APAM WTR powder was stored in amber glass containers for use in the adsorption studies. Characteristics of PAC-APAM WTRs were determined by using the methods and instruments shown in Table 1. The metal elements were also measured and reported in Duan & Fedler ().

Study for effects of pH
Five standard pyridine solutions with pyridine concentrations between 10-50 mg/L were prepared by mixing pyridine (!99.5%) (Tianjin Kermel Chemical Reagent Co. Ltd) with distilled deionized water (DDI water) for developing a calibration curve. Pyridine concentrations in aqueous solutions used in the adsorption studies were analyzed using an ultraviolet-visible spectrophotometer (UV-2601, Beijing Beifen-Ruili Analytical Instrument Co. Ltd) at a wavelength of 256 nm.
The effects of pH on pyridine absorption were investigated over the pH range of 4-9, an initial pyridine concentration (C 0 ) of 30 mg/L, PAC-APAM WTRs dosage of 10 g/L, temperature (T) of 30 C, and time (t) of 10 days. Two grams of PAC-APAM WTRs were put into 200 mL of the pyridine water solution, sealed in a 250-mL flask, and shaken at a speed of 2.33 Hz in a rotary shaker in darkness. The initial pH of the mixture was adjusted with 0.1 M hydrochloric acid or 0.1 M sodium hydroxide. Additionally, NaN 3 (200 mg/L) was added to the flask to inhibit the potential biodegradation of pyridine. After a 10day adsorption period, the supernatant was withdrawn, separated by high-speed centrifugation at 50 Hz and filtered using a 0.45-μm syringe filter. The filtrate was then analyzed for the pyridine concentration.
The adsorbed pyridine by PAC-APAM WTRs (q t , mg/g) at a given time was determined using Equation (1), where C t (mg/L) is the pyridine concentration at t days, V (L) is the volume of the pyridine water solution and M (g) is the mass of PAC-APAM WTRs in the flask. The adsorption removal (AR, %) of pyridine by PAC-APAM WTRs was calculated using Equation (2).
For the requirements of quality assurance and quality control, blanks and standard solutions were used. All treatments in the batch adsorption study were replicated three times (n ¼ 3). The pH value was measured at the end of each adsorption test, and the difference of pH was in the range of À0.05 to 0.05 between the initial pH and the final pH in all tests.

Isotherms and the maximum adsorption capacity
Based on preliminary studies, 10 days was determined as the adsorption equilibrium time. The adsorption isotherm studies were performed under the conditions of C 0 of pyridine ¼ 10, 30, 50, 80 and 100 mg/L, pH ¼ 4, T ¼ 30 C, t ¼ 10 days, and dosage of PAC-APAM WTRs ¼ 2.0 g. The isotherm models in this study included non-linear Langmuir isotherm model, non-linear Freundlich isotherm model nonlinear Dubinin-Radushkevich (D-R) isotherm model, and non-linear Temkin isotherm model (Foo & Hameed ). The obtained data were directly fitted into the four models using the software R (R Core Team ) to investigate the adsorption mechanisms of pyridine onto PAC-APAM WTRs. The maximum adsorption capacity, q m (mg/g), of pyridine by PAC-APAM WTRs was acquired using the parameters of the Langmuir model.

Kinetics of pyridine adsorption onto and desorption from PAC-APAM WTRs
The kinetic study of pyridine adsorption was conducted at C 0 ¼ 30 mg/L, pH ¼ 4, T ¼ 30 C, dosage ¼ 2.0 g PAC-APAM WTRs and t ¼ 0-10 days. The measured C t (mg/L) and calculated q t (mg/g) were fitted into a non-linear pseudo-first-order model, pseudo-second-order model and an intraparticle diffusion (Weber-Morris) model using the software R. Prior to the desorption study, all PAC-APAM WTRs used in each test in the kinetics study were collected, rinsed three times with DDI water and dried at 105 C for 24 h. During the preparation of desorption study, a mass balance of PAC-APAM WTRs was executed. The mass loss of PAC-APAM WTRs in each test was less than 1% compared with the corresponding mass used in the kinetics study. The following desorption experiments were the same as those in the kinetics study, with DDI water replaced with the pyridine aqueous solutions.

Thermodynamic study
The effects of temperature on the adsorption of pyridine onto PAC-APAM WTRs were studied at T ¼ 20, 25 and 30 C. The other adsorption conditions were C 0 ¼ 30 mg/L, pH ¼ 4, dosage ¼ 2.0 g PAC-APAM WTRs and t ¼ 10 days. Three important valuesfree energy (ΔG , kJ/ mol), enthalpy (ΔH , kJ/mol) and entropy (ΔS , kJ/mol/K) were determined from the batch experiments that were used to indicate the type of pyridine adsorption process onto PAC-APAM WTRs.

Characterization of PAC-APAM WTRs
The PAC-APAM WTRs used in this study was the same as those in a previous study (Duan & Fedler ). The physical characteristics of PAC-APAM WTRs included bulk density of 1.144 g/cm 3 , particle size of less than 0.25 mm, Brunauer-Emmett-Teller surface area of 36.69 m 2 /g, t-plot micropore area of 1.89 m 2 /g, t-plot external surface area of 34.80 m 2 /g, t-plot micropore volume of 1.26 × 10 À3 cm 3 /g, single point adsorption total pore volume of pores of 62.34 × 10 À3 cm 3 /g and adsorption average pore width of 6.80 nm. The structure and micro-morphology of PAC-APAM WTRs are shown in the scanning electron microscope (SEM) image ( Figure 1). The irregular surface comprising multiple micro-pores and macro-pores was regarded as one of the important factors in favour of the adsorption of various pollutants. Chemical characteristics of the PAC-APAM WTRs included pH (m/v ¼ 1:2) of 7.64, electrical conductivity (EC) (m/v ¼ 1:2) of 5.13 mS/ cm, 11.38% carbon, 0.95% nitrogen, 2.90% hydrogen, 34.43% oxygen, and 0.35% sulfur.

Effects of pH
The uptake of pyridine at equilibrium by PAC-APAM WTRs (q e , mg/g) and the adsorption removal (%) as a function of the initial solution pH are illustrated in Figure 2. Both values were calculated using Equations (1) and (2). It can be seen from Figure 2 that the equilibrium adsorption uptake of pyridine by PAC-APAM WTRs decreased with the increase in pH from 4 to 9. The significant drop (P-value <0.05) occurred immediately after pH 7. This finding about pH effects is different from that reported on polymeric  This means that the surface of PAC-APAM WTRs is positively charged when the solution initial pH is <9.1. It indicated that the surface of PAC-APAM WTRs was positively charged in the pH range used in this study. In acidic aqueous solutions, pyridine has the remarkable protonating trend (the acid dissociation constant, pK a ¼ 5.2) (Zhu et al. a). More positively charged pyridinium ions form in lower pH solutions and, conversely, less positively charged pyridinium ions exist in higher pH conditions.
The possible interactions occurring in the pyridine solutions with the addition of PAC-APAM WTRs are shown in Equations (3a)-(3f) based on the explanation provided by Weber (Weber ) and the following researchers (Singh et al. ). In Equations (3a)-(3f), Py means pyridine, H þ means the hydronium ions, WTRs are the PAC-APAM WTRs and M þ refers to the exchangeable cations on the PAC-APAM WTRs.
At pH ¼ 4 and 5 that are lower than pK a of pyridine, more pyridinium ions exist than neutral pyridine molecules in the solutions due to the reactions given by Equation (3a). The PAC-APAM WTRs are more positively charged at a lower pH because of the reactions given by Equation (3e) as well as possibly the reactions given by Equation (3d).
As a result, the pyridine adsorption should be higher at pH ¼ 5 due to the weaker electrostatic repulsion which causes more reactions given by Equation (3c) if the exchangeable cations on the PAC-APAM WTRs are sufficient. However, it was observed that the pyridine adsorption was higher at pH ¼ 4. The reactions described by Equation (3b) were the predominant adsorption mechanism, and they were negatively impacted by the pH increase. Singh et al. () suggested that the adsorption decrease with the pH increase was because some neutral pyridine molecules might be negatively charged and, at lower pH, the surface of PAC-APAM WTRs was more positively charged, which was more suitable for the adsorption of the negatively charged pyridine molecules.
At pH ¼ 6 and 7 that are higher than pK a of pyridine, more neutral pyridine molecules than pyridinium ions are in the solutions. The main adsorption reactions should be those given by Equation (3b). At higher pH, the surface of PAC-APAM WTRs was less positively charged and there were possibly more OH À competing with pyridine molecules for adsorption sites (Singh et al. ), resulting in the decrease of the pyridine adsorption. The competition of more OH À with pyridine molecules could possibly explain the significant adsorption difference between pH ¼ 7 and pH ¼ 8.
In this study, with the increase of pH, pyridine adsorption became more difficult, but adsorption still took place. It might be due to π-π interactions ( should be performed to identify those compounds in PAC-APAM WTRs supplying π-π interactions with pyridine and identify those additional adsorption mechanisms to fully explain the findings in this study.

Dosage effects
The dosage effects on the pyridine adsorption by PAC-APAM WTRs for C 0 ¼ 30 mg/L at pH 4.0 is shown in Figure 3. It can be seen that the pyridine removal (%) was observed to be a linear increase from 6.16% to 96.18% with the increase of WTRs dosage from 2.5 g/L to 15 g/L. The q e of pyridine on PAC-APAM WTRs also increased approximately linearly from 0.185 mg/g to 2.885 mg/g. The increase of adsorption with the dosage increase could be attributed to the increase in the available adsorption sites resulting from the increase of adsorption surface area.

Isotherms and the maximum adsorption capacity
Non-linear Langmuir isotherm model (Equation (4)), nonlinear Freundlich isotherm model (Equation (5)), nonlinear D-R isotherm model (Equations (6) and (7)) and non-linear Temkin isotherm model (Equation (8)) can describe the adsorption equilibrium between the concentration of pyridine in the bulk solution and the quantity of the adsorbed pyridine at the PAC-APAM WTRs interface. The Langmuir model is a classic isotherm model that assumes uniform adsorption energies are distributed on the adsorbent surface. One of advantages of the Langmuir model is that it provides a value of maximum adsorption capacity (q m ). As another of the earliest known adsorption isotherm model, the Freundlich isotherm model assumes multilayer adsorption with non-uniform affinities and adsorption energies on a heterogeneous adsorbent surface (Foo & Hameed ). The D-R isotherm model is an empirical model that is often used to identify the adsorption mechanism (Foo & Hameed ). The Temkin isotherm model assumes that the heat of adsorption is a function of temperature and will linearly decrease with the coverage (Foo & Hameed ).
q e ¼ q m bC e 1 þ bC e (4) In Equations (4)- (8), q e is the adsorption capacity of pyridine onto PAC-APAM WTRs at equilibrium (mg pyridine/g PAC-APAM WTRs); q m is the Langmuir maximum monolayer adsorption capacity of pyridine by PAC-APAM WTRs (mg pyridine/g PAC-APAM WTRs); b is the Langmuir isotherm constant (L/mg); C e is the equilibrium concentration of pyridine in bulk solutions (mg/L); K F is the Freundlich isotherm constant ((mg/g)(L/mg) 1/n ); n is the Freundlich dimensionless empirical parameter; q s is the theoretical isotherm saturation capacity (mg pyridine/g PAC-APAM WTRs); k ad is the D-R isotherm adsorption energy constant (mol 2 /J 2 ); ε is Polanyi potential (J/mol), R is the gas constant (8.314 J/mol/K), T is the adsorption absolute temperature (K), b T is the Temkin isotherm constant (J/mol) and A T is the Temkin isotherm equilibrium binding constant (L/mg). To minimize the possible errors caused by linearization (Foo & Hameed ; Karri et al. ), these isotherm models were fitted with the obtained data from the batch experiments in the non-linearized forms rather than linearized forms.
The data fitting results are listed in Table 2. It can be clearly seen that the D-R isotherm model better describes  the data of the isotherm adsorption batch experiments because the R 2 (0.996, the coefficient of determination) was higher than for the other isotherm models. However, no isotherm model tested was identified to be significantly superior in explaining the isotherm adsorption mechanisms. Although the coefficient of determination (R 2 ) of the Langmuir model was lower than that of the others, the Langmuir model still provided an excellent fit to the data and the obtained parameters still provided useful information for the potential application of PAC-APAM WTRs for pyridine removal. The q m was 3.605 mg pyridine/g PAC-APAM WTRs, which was comparable to the laboratory-measured maximum adsorption capacity, 3.253 mg pyridine/g PAC-APAM WTRs. In this study, the separation factor (R L ), which is defined as 1/(1 þ b C 0 ), was always a positive number less than 1. These results indicate that the adsorption of pyridine onto PAC-APAM WTRs was inherently favourable. Although the q m of PAC-APAM WTRs for pyridine adsorption was smaller than that of artificially synthesized materials such as MN 200 and MN 500 (Zhu et al. a) or materials made by high energy inputs such as activated carbon from various raw materials (Lataye et al. ; Alonso-Davila et al. ; Zhu et al. a), it was higher than other adsorbents such as sodium-montmorillonite and sodium-kaolinite (2.52 mg/g), rundle oil shale (0.26 mg/g) and granular activated carbon (1.20 mg/g) (Lataye et al. ). Compared to the processed or artificial synthesized adsorbents, one of the advantages of using PAC-APAM WTRs for pyridine adsorption over other adsorbents is that it requires minimal additional energy for pretreatment and does not require new raw materials because PAC-APAM WTRs are waste products.
Compared with the Langmuir model, the Freundlich model had a slightly higher coefficient of determination (P-value < 0.05). It shows that the Freundlich model described the isotherm adsorption experiments of pyridine onto PAC-APAM WTRs better than the Langmuir model. The value of 1/n was 0.38 and within the expected range of 0-1. This indicates that the adsorption of pyridine onto PAC-APAM WTRs is a chemisorption process and is favourable (Foo & Hameed ). The value of 1/n is indicative of the adsorption intensity or surface heterogeneity if it is in the range of 0-1 (Foo & Hameed ). If the value of 1/n is closer to 0, it means that the adsorption surface is more heterogeneous (Foo & Hameed ). This confirmed that the surface of PAC-APAM WTRs was more heterogeneous with respect to pyridine adsorption because this type of waste contains multiple complicated compounds for pyridine adsorption and is not uniformly distributed on the adsorption surface. As a result, the adsorption sites for pyridine adsorption might be on the multiple layers with different adsorption energy and affinities.
The D-R model was also slightly better than the Langmuir model and the Temkin model because it had the highest R 2 fit ( Table 2). The D-R model assumes that the energy distribution on the heterogenous adsorption surface is a Gaussian energy distribution (Foo & Hameed ). The parameter, E, is the mean free energy of per molecule of adsorbate (Foo & Hameed ). A value of E in the range of 8-16 kJ/mol shows a chemical adsorption process, while if it is in the range of 1-8 kJ/mol, it would mean a physical adsorption process (Ozdes et al. ; Alberti et al. ). The E value (11.855 kJ/mol) in this study indicated that the adsorption of pyridine onto the PAC-APAM WTRs was a chemically controlled adsorption process. This was consistent with the Freundlich model with respect to chemisorption. The theoretical isotherm saturation capacity, q s , was 9.823 mg pyridine/g PAC-APAM WTRs, which was higher than the laboratory-measured maximum adsorption capacity and the Langmuir model determined maximum adsorption capacity. This indicates that, given more time under favourable conditions, the PAC-APAM WTRs might have the ability to adsorb more pyridine than was found in this study on a per unit mass basis.

Kinetics and desorption
The adsorption kinetic models describe the relationship between adsorption time and the uptake of adsorbate under various specific adsorption conditions. They can provide information about the adsorption rate. Four adsorption kinetic models were used in this study to investigate the adsorption kinetic for pyridine adsorption onto PAC-APAM WTRs. Two of them were reaction orderbased sorption kinetic models, including pseudo-firstorder model (Equation (9)) and pseudo-second-order model (Equation (10)). The other two models were intraparticle diffusion modelsthe Weber-Morris model (Equation (11)) and the Elovich model (Equation (12)) (Alberti et al. ), which are based on reaction mechanisms. One of advantages of the reaction order-based sorption kinetic models is that they can describe the complete kinetic experimental data as well as the occurred discerned deviations (Alberti et al. ). The adsorption reaction mechanism-based kinetic models provide information about understanding the adsorption mechanisms involved. q t ¼ q e (1 À e Àk 1 t ) (9) q t ¼ q 2 e k 2 t 1 þ q e k 2 t (10) In Equations (9)-(12), q t (mg pyridine/g PAC-APAM WTRs) was the mass of pyridine uptake at time t per unit mass of PAC-APAM WTRs; k 1 (1/day), k 2 (g/(mg·day)), K P (mg/(g·day 1/2 )), C (mg/g), α (mg/(g·day)) and β (g/mg) were the rate constants of the pseudo-first-order adsorption, the rate constant of pseudo-second-order adsorption, the Weber-Morris intraparticle diffusion rate, the Weber-Morris constant, the initial adsorption rate related to activation energy, and the desorption constant related to surface coverage (Alberti et al. ), respectively. Nonlinear forms of all four kinetic models were utilized during data fitting. Similar to the data fitting for isotherms, it is recognized and reported that linearization of sorption kinetic models in some cases possibly causes a variation in the error structure and the error distribution (Alberti et al. ). This frequently results in different results and conclusions from non-linearized forms, even if different linearized forms are used for the same sorption kinetic model (Alberti et al. ). However, the data fitting for non-linearized forms requires a trial and error procedure (Alberti et al. ).
The data fitting results are shown in Table 3. It can be clearly seen that the pseudo-first-order, pseudo-secondorder and Elovich kinetic models were consistent in describing the kinetic batch study data with similar values of the coefficient of determination (R 2 ). The Elovich model is mainly used to describe chemisorption. It assumes that the adsorption occurs on fairly heterogeneous adsorbents, and the adsorption is a multilayer process with each layer having different activation energies and adsorption sites exponentially increasing (Alberti et al. ). Since the Elovich model had a high correlation efficient in this study (Table 3), it can be inferred that the adsorption of pyridine onto PAC-APAM WTRs might be a multilayer chemisorption process. This inference from the kinetic study was in agreement with that from the isotherm study.
Non-linear Weber-Morris model was fitted in two cases: one for C ! 0, and the other was C < 0 (Table 3).
Previous studies on adsorption have reported that C was not a negative number (Alberti et al. ). For C ¼ 0, the adsorption rate limiting step is the intraparticle diffusion, while C > 0 means that both external mass transfer and intraparticle diffusion are the rate limiting steps (Alberti et al. ). In this study, when C was set as a non-negative number, the results showed that C was 5 × 10 À15 mg/g and thus reported as zero with the coefficient of determination of (R 2 ) 0.798. In this case, the Weber-Morris model had a lower R 2 than the other three models. This means that the intraparticle diffusion was the rate limiting step because C was zero. However, in the other case, the Weber-Morris model described the batch study data quite well, with an R 2 ¼ 0.9914 when C was set as a negative real number. The negative value of C was reported by previous adsorption studies and showed that external film diffusion limited the adsorption rate, in addition to intraparticle diffusion (Zhu et al. b). The measured kinetic data and the Weber-Morris plots for these two cases of pyridine adsorption onto PAC-APAM WTRs are shown in Figure 4. It can be clearly seen that the prediction of the Weber-Morris model with C < 0 was much better than that with C ¼ 0. The negative value of C (negative y-axis intercept in Figure 4) can be interpreted as an external film diffusion resistance that led to the time lag for the pyridine adsorption.
In the desorption study, no pyridine was detected in the liquid phase after a 10-day adsorption. It showed that the adsorption of pyridine onto the PAC-APAM WTRs was relatively stable under the same adsorption conditions, and the leaching potential of pyridine is extremely low from PAC-APAM WTRs.

Thermodynamic parameters
The uptake of pyridine by PAC-APAM WTRs increased with an increase in temperature. Calculation of the Gibbs free energy change (ΔG , kJ/mol) of pyridine adsorption by PAC-APAM WTRs is required to determine the adsorption equilibrium distribution coefficient (K d , L/g). K d was calculated by using Equation (13). ΔG can be determined by the classical Van't Hoff equation (Equation (14)) (Lataye et al. ). Enthalpy change (ΔH , kJ/mol) and entropy change (ΔS , kJ/mol/K) were determined by fitting thermodynamic batch study data into Equation (15) (Lataye et al. ). ln In this study, the Gibbs free energy change for pyridine adsorption by the PAC-APAM WTRs was determined as 5.308 kJ/mol (20 C), 4.966 kJ/mol (25 C) and 4.759 kJ/ mol (30 C). The positive values of ΔG showed that the pyridine adsorption by PAC-APAM WTRs was not a spontaneously process. In reality, although the pyridine adsorption rate onto PAC-APAM WTRs was lower than that onto reported artificial synthesized materials which have negative values of Gibbs free energy change (Zhu et al. a), the batch experiments showed that the unprocessed PAC-APAM WTRs used in this study had the ability to adsorb pyridine.
The enthalpy change (ΔH ) and the entropy change (ΔS ) were 21.400 kJ/mol and 0.055 kJ/mol/K, respectively. The coefficient of determination (R 2 ) was 0.988. The positive value of ΔH indicated that the pyridine adsorption by PAC-APAM WTRs was an endothermic adsorption process and was an activated process probably caused by the activation of some functional groups on the PAC-APAM WTRs which were available to interact with pyridine molecules.
The positive value of ΔS implied that the pyridine adsorption by the PAC-APAM WTRs increased the randomness at the reaction interface. With the increase of temperature, some functional groups on the PAC-APAM WTRs favouring interaction with pyridine molecules were gradually activated and that caused the increase in the randomness at the liquid-solid interface (Zhu et al. a). Also, it was possible that the temperature increase caused some structural changes in the PAC-APAM WTRs resulting in affinity changes between pyridine and the PAC-APAM WTRs which contributed to the randomness increase on the interface (Lataye et al. ).

Implications of engineering
This study proves that the low-cost PAC-APAM WTRs have the ability to remove one of the priority organic pollutants, pyridine. This finding provides information for engineers to comprehensively consider the technique by using PAC-APAM WTRs to remediate various pollutants in contaminated soils. Our studies have found that PAC-APAM WTRs can effectively remove phosphorus (Duan & Fedler ), copper(II), lead(II) (Ma et al. ), cadmium(II), zinc(II) (Duan & Fedler ) and phenol ( Jiao et al. ). One concern for the application of PAC-APAM WTRs is the aluminium ion (Al 3þ ) leaching or being released from PAC-APAM WTRs, which may lead to possible Al 3þ phytotoxicity (Zhao et al. ). However, our previous study did not find significant leaching. This study, plus our previous studies, demonstrated that PAC-APAM WTRs have the capability to be used as an adsorbent for the removal of PAC-APAM WTRs. Dosage ¼ 10 g/L, T ¼ 30 C, pH ¼ 4 and initial pyridine concentration ¼ 30 mg/L. All standard deviations (n ¼ 3) were less than 2.52 ×

10
À3 mg/g for measured qt, so the standard deviation bars approximately overlapped with the points of the average values. All coefficients of variation for measured value were less than 1.79%.
multiple organic and inorganic pollutants in soils. This study, as well as our other studies, also confirms that PAC-APAM WTRs recycling into contaminated soils is a sustainable disposal method for the municipal drinking water treatment industry by increasing the revenue and decreasing the cost on occupying land resources. More research is needed to understand the removal mechanisms and effects when PAC-APAM WTRs are used in the scenarios in which various pollutants simultaneously exist in the soils where there will be competition for adsorption sites. Also, when PAC-APAM WTRs is land-applied into contaminated soils, they are required to be mixed with the original contaminated soils rather than be entirely substituted for those soils (Hovsepyan & Bonzongo ). The appropriate ratio of PAC-APAM WTRs to the original contaminated soil should be investigated based on typical scenarios of contaminated soils under specific climate conditions.

CONCLUSIONS
This study investigated the adsorption of pyridine onto PAC-APAM WTRs and included an analysis of the adsorption isotherms, kinetics and thermodynamics of the adsorption in aqueous solutions. This study confirmed the assumption that PAC-APAM WTRs had the ability to remove pyridine. The D-R model and Freundlich model better described the pyridine adsorption isotherms. It indicated that pyridine adsorption by PAC-APAM WTRs was a chemically controlled multilayer adsorption process. An adsorption kinetic study showed that the pyridine adsorption rate was simultaneously controlled by external film diffusion and intraparticle diffusion. The pyridine adsorption was an endothermic process with randomness increase on the adsorption interface. The pyridine adsorption decreased with an increase in pH and increased with the increase in temperature and dosage of the PAC-APAM WTRs. Therefore, PAC-APAM WTRs recycled into contaminated soils for remediation is expected to be an innovative alternative disposal method for the WTRs produced from municipal drinking water treatment with an aim of achieving sustainability of the water industry. More research is recommended to identify the various compounds in PAC-APAM WTRs and their detailed adsorption mechanisms interacting with pyridine or other organic pollutants, and the proper mixing ratio of PAC-APAM WTRs to contaminated soils under various physical and climatic scenarios.