Evaluation and mechanism of ammonia nitrogen removal using sediments from a malodorous river

Malodorous rivers are among the major environmental problems of cities in developing countries. In addition to the unpleasant smell, the sediments of such rivers can act as a sink for pollutants. The excessive amount of ammonia nitrogen (NH3−N) in rivers is the main factor that causes the malodour. Therefore, a suitable method is necessary for sediment disposition and NH3−N removal in malodorous rivers. The sediment in a malodorous river (PS) in Beijing, China was selected and modified via calcination (PS-D), Na+ doping (PS-Na) and calcination–Na+ doping (PS-DNa). The NH3−N removal efficiency using the four sediment materials was evaluated, and results indicated that the NH3−N removal efficiency using the modified sediment materials could reach over 60%. PS-DNa achieved the highest NH3−N removal efficiency (90.04%). The kinetics study showed that the pseudo-second-order model could effectively describe the sorption kinetics and that the exterior activated site had the main function of P sorption. The results of the sorption isotherms indicated that the maximum sorption capacities of PS-Na, PS-D and PS-DNa were 0.343, 0.831 and 1.113 mg g−1, respectively, and a high temperature was favourable to sorption. The calculated thermodynamic parameters suggested that sorption was a feasible or spontaneous (ΔG < 0), entropy-driven (ΔS > 0), and endothermic (ΔH > 0) reaction.


Introduction
The malodorous phenomenon in urban rivers is becoming increasingly serious in developing countries [1,2]. Water bodies are highly polluted by organic pollutants, nutrients (e.g. nitrogen and phosphorus) and heavy metals. As one of the main indexes of malodorous rivers, ammonia nitrogen (NH 3 Table 1. Physico-chemical properties (mean value ± standard deviation) of sediments and water (N = 7) in Liangshui River. TP, total phosphorus; TN, total nitrogen; OM, organic matter; DO, dissolved oxygen. and analysed to determine NH 3 −N concentration using the Nessler reagent spectrophotometric method [19]. Three groups of parallel experiments were set up.

Sorption isotherm tests
The sorption isotherm of NH 3 −N was obtained in batch experiments. Sediment materials (1 g in triplicate) of PS, PS-Na, PS-D and PS-DNa were added to 50 ml NH 4 Cl solution with different concentrations (0, 1, 2, 5, 10, 20 and 50 mg l −1 ). The samples were continuously agitated in a shaker at a speed of 220 rpm at constant temperatures of 15°C, 25°C and 35°C, for 480 min. The suspensions were centrifuged, filtered (0.45 µm), and analysed to determine NH 3 −N concentration.

Thermodynamic parameters
Sediment materials (1 g) of PS, PS-Na, PS-N and PS-DNa were mixed into 50 ml NH 3 −N solutions with different initial concentrations (0, 1, 2, 5, 10, 20 and 50 mg l −1 ) at 15°C, 25°C and 35°C. Batch samples were shaken in a temperature-controlled shaker for 480 min. The thermodynamic parameters of NH 3 −N sorption, such as enthalpy ( H), Gibbs energy ( G) and entropy ( S), were determined by fitting linear equations into the thermodynamic data obtained under different concentrations. Three groups of parallel experiments were set up.

Data analysis
The NH 3 −N uptake amount Q t (mg g −1 ) in different samples at equilibrium was calculated as follows: where C 0 (mg l −1 ) is the initial liquid phase NH 3 −N concentration, C t (mg l −1 ) is the blank corrected concentration of NH 3 −N at time t, W (g) is the amount of dried sediment materials and V (l) is the volume of NH 3 −N solution. The sorption kinetics was described by the pseudo-first-order, pseudo-second-order and power function models as follows [19,21,22]: and where Q t and Q e (mg g −1 ) are the uptake amounts of NH 3 −N adsorbed at time point t and at equilibrium, respectively. K 1 (h −1 ) is the first-order kinetic rate constant, K 2 (g mg −1 h −1 ) is the sorption rate constant of the pseudo-second-order kinetic model, and a and b are the power function models of the kinetic rate constant. The sorption isotherms were fitted by the Langmuir and Freundlich models, and the equations of the isotherm parameters are shown as follows [23,24]: and where Q e and Q m (mg g −1 ) are the adsorbed amounts of NH 3 −N at equilibrium and the maximum NH 3 −N uptake amount, respectively; C e (mg l −1 ) is the NH 3 −N concentration in aqueous phase at equilibrium; K (l mg −1 ) is the affinity parameter; K f (l g −1 ) is the sorption coefficient; and n is a constant used to measure sorption intensity or surface heterogeneity. The thermodynamic parameters were analysed using the following equations [25,26]: where G°(kJ mol −1 ) is the change in Gibbs free energy, S°(kJ mol −1 K −1 ) is the change in entropy, H°(kJ mol −1 K −1 ) is the change in enthalpy, K D is the equilibrium constant (dimensionless), R (8.314 J mol −1 K −1 ) is the gas constant, T (K) is the absolute temperature, C 0 (mg l −1 ) is the initial solution concentration, C e (mg l −1 ) is the solution equilibrium concentration, V (ml) is the solution volume and m (g) is the gravel sand mass.

Result and discussion
3.1. Effects of temperature on the removal of NH 3 −N using different modified sediment materials Figure 1 shows the effect of temperature on the removal efficiency of NH 3 −N. The results illustrated that PS-DNa achieved the highest removal efficiency (average value: >90%), followed by PS-D (71%-76%) and PS-Na (64%-70%). The order of removal efficiency of NH 3 −N was PS-DNa > PS-D > PS-Na > PS. The NH 3 −N removal efficiency using the modified sediment materials increased rapidly when temperature increased from 15°C to 35°C. The modified materials demonstrated considerably higher NH 3 −N removal efficiency than the raw sediment samples. This observation could be explained by the fact that PS-DNa and PS-D had higher surface areas and pore volumes on the surface of the sediment owing to calcination.  was loaded with a greater layer of active substances that could change its physico-chemical properties. Therefore, the NH 3 −N removal efficiency that used PS-Na and PS-DNa dramatically improved. The removal efficiency of the four sediment materials exhibited a rising trend when temperature increased from 15°C to 35°C, thereby indicating that increasing the temperature within limits could intensify molecular motion, accelerate diffusion rate and enhance the opportunity of NH 4 + to impact on the sediment surface. In addition, NH 4 + with high energy adsorbed on the surface of particles improved removal efficiency under high temperatures [27,28]. Previous studies indicated that natural zeolite had high NH 3 -N sorption capacity, and researchers focused on sorbents based on natural or modified zeolites [19,29]. The results indicated that PS-DNa had a similar removal efficiency of NH 3 -N (>90%) to the natural or modified zeolite (the highest value of 99.8%) [19,30]. In addition, PS-DNa had greater economic benefits than the natural or modified zeolite, and therefore PS-DNa might be a potential material for NH 3 -N removal. . NH 3 −N sorption on the four modified materials was initially fast and then rose rapidly during the first few minutes (0-10 min), but slowed down immediately after the initial uptake. The amount of NH 3 −N sorption tended toward equilibrium at approximately 40 min. The rapid original sorption could be attributed to physical sorption mechanisms, such as electrostatic interactions, which led to the adhesion of NH 4 + on the material surface during the initial stage. Thereafter, the decreasing sorption rate could be interpreted as a ligand exchange [31]. Table 3 shows that the kinetic parameters of NH 3 −N sorption derived from the four modified sediment materials are acquired through nonlinear fitting with the three models. The correlation coefficients (R 2 ) demonstrated that the pseudo-second-order model achieved better fitting data (R 2 > 0.96) than the pseudo-first-order and power function kinetic models. The fitting data showed that PS-DNa exhibited a relatively high equilibrium sorption capacity at different temperatures and reached its maximum value of 0.485 mg g −1 at 35°C. In addition, table 3 illustrates the change in the sorption rate constant K 2 . When ambient temperature is high, the K 2 value is large. The result indicated that the crystal structure, surface area and pore volume of the sediment materials changed due to calcination or sodium doping, which improved NH 3 −N sorption on the surface of materials. Furthermore, sodium demonstrated greater capability to complex with the surface activity site of material particles and reacted more easily with the oxygen-containing functional groups when temperature was rising. Therefore, the modified sediment materials of PS-DNa have higher K 2 values than those of PS-D and PS-Na.

Fitting of sorption isotherms
In this study, the isotherm data were fitted by the Langmuir and Freundlich models (equations (2.6) and (2.7)). Figure 3 and Table 4 show the results of the NH 3 −N sorption isotherm experiments. Both the Langmuir and Freundlich models could fit the data and estimate model parameters according to the correlation coefficients (R 2 ). However, the Langmuir model can describe the NH 3 −N sorption isotherm better than the Freundlich model. The Langmuir sorption affinity parameter (K) and the maximum adsorption capacity (Q m ) of the four materials appeared as continuous rising trends with increasing temperature, and the values for PS-DNa reached 4.986 l mg −1 and 1.113 mg g −1 at 35°C. At the same temperature, the K and Q m of PS-DNa reached higher values than those of the other samples. The sorption capacity of the four materials increased with rising equilibrium concentration, which could be interpreted as the equilibrium concentration of the solution being closely related to the surface coverage density of the adsorbent. Under the condition of high equilibrium concentration, a large number of active sites on the surface of the material particles were occupied by the adsorbate, and the driving force of adsorption decreased and finally reached saturation state [32]. An increasing temperature could accelerate the diffusion rate, and a higher amount of high-energy NH 4 + was adsorbed onto the surface of particles. In addition, the value of n obtained by the Freundlich model can reflect the strength of sorption capacity; when n is small, the adsorbate can be easily adsorbed by the materials [33]. PS-DNa achieved a higher K f value than the other sediment materials, which indicated that PS-DNa had a higher sorption distribution coefficient and greater capacity to combine with NH 3 −N (table 4).

Thermodynamic parameters
We acquired the values of H 0 and S 0 /R (figure 4) after fitting lnK D and 1000/T according to equation (2.10). The thermodynamic parameters were calculated, and the results are presented in  Table 3. Fitting kinetics and mechanism parameters of NH 3 −N sorption on the four sediment materials according to the pseudo-first-order, pseudo-second-order and power function models at three different temperatures.
pseudo-first-order model pseudo-second-order model     Table 5. Thermodynamic parameters of NH 3 −N sorption on the four sediment materials.  indicated that the NH 3 −N sorption of the modified sediment materials was better at high temperatures [34]. The positive values of H 0 indicated that NH 3 −N sorption was an endothermic process. In addition, S 0 values were higher than 0, which indicated that NH 3 −N tended to be adsorbed onto the surface of sediment materials [35]. In general, entropy decreases when molecules are adsorbed onto the surface of solid materials. In the solute sorption process, the solute molecular degree of freedom dropped and the entropy of the modified sediment materials was lower than that of the raw sediment. Consequently, the sorption of NH 3 −N is a complex process, and the entropy effect is the driving force of the sorption process [19].

Conclusion
Solutions must be obtained to address problems of excessive NH 3 −N and the utilization of sediment resources in malodorous rivers. The sediment from a malodorous river was modified and used to remove NH 3 −N. The NH 3 −N removal efficiency was evaluated, and the NH 3 −N sorption mechanism was studied through kinetics, equilibrium and thermodynamic experiments. The results indicated that the calcination-sodium-doped materials achieved the highest NH 3 −N removal efficiency. The sediment sorption rate followed the pseudo-second-order model, and high temperatures favoured NH 3 −N uptake. Overall, the data that described the sorption isotherms were better fitted by the Langmuir model. The maximum P sorption capacities exhibited the following order: PS-DNa > PS-D > PS-Na > PS. PS-DNa yielded the highest value of 1.113 mg g −1 . Sorption was feasible or spontaneous ( G < 0), randomly entropy-driven ( S > 0), and endothermic ( H > 0) according to the calculation of the thermodynamic parameters of the sediment materials. This study provides a new method for the utilization of sediment resources and a theoretical foundation for NH 3 −N removal.