The CaO dissolution from slag, as well as the effects of influencing parameters (i.e. pH and Ca2+ concentration) on the ammonium adsorption onto zeolite, was systematically studied in this paper. Modeling results of Ca2+ and OH release from slag indicated that pseudo-second-order reaction had a better fitness than pseudo-first-order reaction. Changing pH value from 7 to 12 resulted in a drastic reduction of the ammonium adsorption capacity on zeolite, from the peak adsorption capacity at pH 7. High Ca2+ concentration in solution also inhibited the adsorption of ammonium onto zeolite. There are two proposed mechanisms for steel slag inhibiting the ammonium adsorption capacity of zeolite. On the one hand, OH released from steel slag can react with ammonium ions to produce the molecular form of ammonia (NH3·H2O), which would cause the dissociation of NH4+ from zeolite. On the other hand, Ca2+ could replace the NH4+ ions to adhere onto the surface of zeolite. An innovative substrate filling configuration with zeolite placed upstream of the steel slag was then proposed to eliminate the disadvantageous effects of steel slag. Experimental results showed that this novel filling configuration was superior to two other filling configurations in terms of ammonium removal.

Nitrogen and phosphorus are essential nutrients for biomass growth. But excessive nitrogen and phosphorus can contribute to accelerated eutrophication of lakes and rivers. The eutrophication causes a rapid growth of algae and other microorganisms, resulting in dissolved oxygen depletion and fish toxicity. The removal of nitrogen and phosphorus has not recently become an important and urgent wastewater treatment task but has, in fact, been studied for a long time now.

Constructed wetlands are an efficient and eco-friendly form of wastewater treatment technology with low operating cost and low-energy consumption. This technology has been widely used to relieve the eutrophication of waters and restore ecosystems for dozens of years (Tanner et al. 2002; Bezbaruah & Zhang 2003; Mateus et al. 2012). As one of the most important parts of constructed wetlands, the substrate can partly determine the removal or form-converting efficiency of nitrogen and phosphorous. Thus, to increase the nutrient removal efficiency of constructed wetlands, substrates should be selected and grouped carefully.

Because of high affinity with ammonium ion, zeolite, as a common substrate, is applied in constructed wetlands worldwide. Many studies have confirmed that zeolite has a good performance of nitrogen removal (Du et al. 2005; Alias et al. 2010; Huang et al. 2010; Beebe et al. 2013). Ion-exchange is considered to be a significant mechanism for adsorption of ammonium ion onto zeolite (Wang et al. 2007). In constructed wetland, ammonium ion could be adsorbed by zeolite via ion exchange, and then under proper conditions oxidized to nitrite first and then nitrate. Nitrogen can be removed from water in a biological denitrification process when nitrate and electron donor (i.e. chemical oxygen demand, COD) are present. The adsorption of ammonium ion by zeolite might be a step for nitrogen removal enhancement (Beebe et al. 2013).

Another substrate seen commonly for constructed wetlands is steel slag, which is the molten byproduct of many metallurgical operations. It has high contents of Ca, Al, or Fe, and possesses a strong affinity with phosphate (i.e. ) (Kostura et al. 2005; Johansson-Westholm 2006; Chris et al. 2007). Several authors have confirmed that steel slag is a better substrate for phosphorus removal compared to soil, zeolites, and so on (Mann 1997; Brix et al. 2001). The major removal mechanism is precipitation of Ca-PO4-P complexes with the release of Ca2+ from steel slag (Drizo et al. 2002, 2006; Barca et al. 2012, 2014; Claveau-Mallet et al. 2013). Iron-phosphate formation was also confirmed as another way for phosphorus removal (Pratt et al. 2007). Claveau-Mallet et al. (2014) have modeled the dissolution and precipitation of phosphorus removal by slag filter.

Due to the excellent ammonium removal ability of zeolite and phosphate removal ability of steel slag, they have been combined in the substrate bed of a constructed wetland in order to enhance the simultaneous removal efficiencies of nitrogen and phosphorus. Actually, in the study of Olfa et al. (2002), a porous carrier made from artificial zeolite and acidic treated blast furnace slag had better removal efficiency for phosphate and ammonium.

In our early tests, zeolite and steel slag were combined with each other in order to simultaneously achieve a high removal efficiency of ammonium-nitrogen and phosphate-phosphorus. However, compared with the results when only zeolite was used, the removal efficiency of ammonia was sharply decreased when zeolite and steel slag were used together. No similar reports were made in studies of this field as a reason of differently treated substrates being used (e.g. Olfa et al. 2002; Renman et al. 2009; Lee et al. 2010). Until now, the mechanism of steel slag affecting the adsorption of ammonium onto zeolite still lacks a systematic study.

The objective of this paper was to explore the interaction mechanism behind steel slag and zeolite when targeting the removal of ammonium. Based on the interaction mechanism, a new substrate bed configuration was going to be proposed to eliminate the disadvantageous effects of steel slag, and to simultaneously achieve higher removal efficiencies for nitrogen and phosphorus.

Substrates collection and preparation

Zeolite and steel slag were selected as substrate subjects in this study according to their and removal ability obtained in pre-study tests. In the pre-study tests, zeolite showed best removal ability and steel slag showed best P removal ability among all eight types of substrates including shale, ceramisites made of various materials, steel slag and zeolite. Zeolite was obtained from KaiBiYuan Trading Co. Ltd (Beijing, China), and steel slag was supplied by a steel plant in Jiangxi Province. To prepare similarly sized substrate particles for tests, the samples were first screened on an automatic grinder according to three ranges of granular sizes (i.e. <2, 2–5 and >5 mm). The granular size of 2–5 mm was then selected to perform the batch experiments. This guarantees the results to be comparable. Before used in tests, substrates were rinsed with tap water to remove impurities and fine particles attached to their surface. The clean substrates were dried at 105 °C for 24 h. Energy dispersive spectrometer X-ray (EDS, 7021-H, Horiba, Japan) was performed to quantify the elemental composition on substrate surfaces.

Kinetic experiments on Ca2+ and OH release from steel slag

Two series of batch experiments were carried out to verify the kinetics of Ca2+ and OH release from slag. Batch experiments were to verify if the increase of Ca2+ and pH of the solutions depended primarily on CaO-slag dissolution. Forty grams of slag were added into triangular flasks, which contained 1 L of deionized water (conductivity <30 μS/cm) and 25 mg/L ammonium nitrogen solution, separately. Two drops of chloroform were dosed into the prepared solutions for the inhibition of microorganisms' growth. The flasks were then placed on an agitation table and shaken at 150 rpm under the controlled temperature condition of 25 °C for 4 days. During 4 days agitation, samples were taken at the end of the 2nd, 4th, 8th, 12th, 24th, 36th, 48th, 72th, and 96th hour for pH and Ca2+ concentrations measurements. The pH of the solutions could be expressed in OH concentration. Equations (1) and (2) were used in this study to calculate the experimental capacities of Ca2+ and OH released from steel slag.
formula
1
formula
2
where Qt (mg/g) are the capacities of Ca2+ and OH release at time t (h), V (L) is the volume of the solution, M (g) is the mass of steel slag, Cat and OHt (mg/L) are the Ca2+ and OH concentration of the solutions at time t (h).

Influence of Ca2+ and OH on ammonium adsorption onto zeolite

In order to examine the various effects of Ca2+ and OH on ammonium adsorption performance, two series of tests were conducted. For each series of tests, 20 g of zeolite was immersed in 500 mL 25 mg N/L ammonium solution contained in triangular flasks, thus leading to a zeolite to solution ratio of 0.04 g per mL. For OH tests, the initial pH values in flasks were adjusted to 4, 5, 6, 7, 8, 9, 10, 11, and 12 with 0.5 mol/L HCl and 0.1 mol/L NaOH. For Ca2+ tests, the concentration of Ca2+ was prepared at 0, 5, 10, 20, 40, 60, 80, 120, 140, 180 mg/L, respectively. Two drops of chloroform were also dosed for the inhibition of microorganism in each flask. The flasks were then placed on agitation tables and shaken at 150 rpm under controlled temperature conditions (25 °C) for 4 days.

Influence of steel slag on zeolite saturated with ammonium

To prepare zeolite saturated with ammonium, 40 g of zeolites was immersed in 1 L of 25 mg/L ammonium solution, and then was shaken at 150 rpm under 25 °C for 4 days under sterilized circumstances. Forty grams of steel slag was added into the flask with ammonium-saturated zeolite for another 4 days' agitation. During the second 4 days' agitation, solution samples were taken at a serial hours for pH, ammonium nitrogen and Ca2+ measurements.

Optimization of the substrate bed configuration

Three identical cylinders with the height and diameter 40 × 6 cm were used as laboratory-scale reactors (marked as A, B, and C, respectively). Each cylinder was filled with exactly the same quantity of zeolite and steel slag by using different configurations (Table 1). The placing order of substrates in reactor A was that, 10 cm of steel slags first and then 20 cm of zeolites from the bottom up. The B reactor was filled with 30 cm mixture of zeolites and steel slags. For reactor C, the sequence was opposite to reactor A. Figure 1 shows the schematic diagram of all three reactors. In all three tests, synthetic wastewater was used as influent, which was pumped into reactors from the bottom inlet. Main parameters of the synthetic wastewater was listed as pH 6.8–7.4, 25 mg·L−1, 5 mg·L−1. The inflow rate was set at 0.15 mL·min−1. Both effluent and influent samples were taken from the three reactors every other day for pH, and measurements. The cylinders were operated for 2 months running.
Table 1

Substrate configuration in reactor A through C

ReactorDescription
From bottom up: 10 cm steel slag; 20 cm zeolite 
30 cm mixture of zeolites and steel slags, volumetric ratio 2:1 
From bottom up: 20 cm zeolite; 10 cm steel slag 
ReactorDescription
From bottom up: 10 cm steel slag; 20 cm zeolite 
30 cm mixture of zeolites and steel slags, volumetric ratio 2:1 
From bottom up: 20 cm zeolite; 10 cm steel slag 
Figure 1

The schematic diagram of all three reactors.

Figure 1

The schematic diagram of all three reactors.

Analytical methods

Phosphate, and Ca2+ analyses were performed according to the Ammonium Molybdate Spectrometric Method, Nessler's reagent Spectrometric Method and the Atomic Absorption Spectrometric Method, respectively. Phosphate and concentrations were measured using a MI-parameter Meter (5B-3B, LianHua, China), whereas Ca2+ concentrations were measured using an Atomic Absorption Spectrophotometer (AA6300, Shimadzu, Japan). The pH value was measured with a pH meter (WTW, Germany).

Chemical composition of substrates

The chemical composition of substrates plays an important role in their pollutant removal ability. The results of EDS analyses (shown in Table 2) show the main chemical composition of the surfaces of zeolite and steel slag. Based on Table 2, aluminosilicate minerals dominate the surface of zeolite, which are mainly composed of silicon, oxygen and aluminum. According to other related reports (Alias et al. 2010; Huang et al. 2010), the structure of zeolite is networks of tetrahedral or polyhedral arrangements of atoms, as well as microporous, which works as an mechanical sieve. These structures decide that zeolite has a better adsorption capacity for ammonium. Steel slag is rich in CaO. These results are in agreement with findings of other reported studies on slag composition (Drizo et al. 2002; Xue et al. 2009; Wang et al. 2010; Barca et al. 2012).

Table 2

EDS analyses: chemical composition of zeolite and steel slag (weight %)

 ZeoliteSteel slag
SiO2(%) 51.54–60.13 0.60–1.52 
Fe2O3(%) 2.03–4.73 – 
Al2O3(%) 9.39–10.50 – 
CaO(%) 2.42–3.44 35.98–37.59 
MgO(%) 0.47–0.93 0.32–0.52 
C(%) 2.78–3.01 9.12–9.77 
N(%) 5.01–5.56 8.88–9.02 
 ZeoliteSteel slag
SiO2(%) 51.54–60.13 0.60–1.52 
Fe2O3(%) 2.03–4.73 – 
Al2O3(%) 9.39–10.50 – 
CaO(%) 2.42–3.44 35.98–37.59 
MgO(%) 0.47–0.93 0.32–0.52 
C(%) 2.78–3.01 9.12–9.77 
N(%) 5.01–5.56 8.88–9.02 

Kinetics of Ca2+ and OH release from steel slag

To elucidate the kinetics of Ca2+ and OH release from steel slag, and the interrelationship between steel slag and ammonium solution, two series of batch experiments were conducted, the methods of which were described in the Methods section above. The experimental results showed that Ca2+ concentrations and pH values of the solutions changed with soaking time (shown in Figure 2). It can be seen from Figure 2 that, no matter whether in deionized water or in ammonium solution, Ca2+ concentrations and pH values of the solutions increased quickly until a (pseudo-) equilibrium was reached at some time between the 36th and 48th hour of agitation and had small variations after then. These curves can be explained by following CaO-slag dissolution Equation (3). Several authors have also confirmed that steel slag had a particularly strong tendency to produce basic leachates as a result of a mixture of dissolution reactions of various metal oxides, including CaO, Al2O3, and MgO (Bowden et al. 2009; Xue et al. 2009).
formula
3
Obviously, the existence of ammonium ions influenced the CaO dissolution from steel slag. From Figure 2(a), it can be seen that pH value in deionized water was higher than that in ammonium solution. However, the situation was opposite for Ca2+. From Figure 2(b), the Ca2+ release capacity of steel slag in ammonia solution was higher than in deionized water. Equation (4), which describes ammonia dissociation in water, may explain the phenomenon that the variation of Ca2+ concentration differed with that of OH concentration. In ammonia solution, the OH- produced by CaO dissolution was consumed by ammonium ion via transformation to molecular form ammonia (NH3·H2O) (see Equation (4)). Therefore, the pH curve of ammonia solution lies below its counterpart of deionized water. At the same time, the consumption of OH leads to further dissociation of CaO-slag (Equation (3)). As a result, it can be explained why Ca2+ curve of steel slag in ammonia solution was above that in deionized water. As shown in Table 2, no matter in deionized water or in ammonium solution, the equilibrium capacities Qe and the rate constants k2 of Ca2+ and OH release do not agree with each other. This could be attributed to the influence of ammonia on Equation (3) as stated in paragraph above.
formula
4
The kinetics of CaO-slag dissolution has been modeled by Barca et al. (2012), and in their study, high correlation coefficients were obtained when a pseudo-first-order kinetic model was used for European steel slag. Thus, in this study, the kinetics of Ca2+ and OH release was first modeled based upon experimental results in Figure 2 with the pseudo-first-order kinetic reaction (see Equation (5)). However, the fitting results (R2 < 0.9) were not as good as expected. Then the pseudo-second-order kinetic reaction (Equation (6)) was employed for further modeling.
Figure 2

Variation of Ca2+ and OH concentration with time in different original solutions: pH (a) and Ca2+ concentration (b).

Figure 2

Variation of Ca2+ and OH concentration with time in different original solutions: pH (a) and Ca2+ concentration (b).

The pseudo-first-order kinetic equation:
formula
5
The pseudo-second-order equation:
formula
6
where k1 (h−1) and k2 (g·mg−1 h−1) are constants of release rate, Qt (mg·g−1) is the release capacity at time t (h), Qe (mg·g−1) is the release capacity at equilibrium (Ho & McKay 1999).

The modeling results of Ca2+ and OH release from slag indicated that pseudo-second-order reaction had a better fitness (much higher R2 shown in Table 3). Barca et al. (2012) reported that CaO was present in great excess over the other reactants in the reaction mixture, and the dissolution of CaO-slag was the primary reaction explaining the increase of Ca2+ and pH in solution. However, in this study, the lower applicability of the pseudo-first-order model to describe Ca2+ release (R2 < 0.9, data not shown) gave another different picture for the dissolution of CaO-slag.

Table 3

Correlation coefficients and rate constants for the pseudo-second-order kinetic model describing Ca2+ and OH release

 Ca2+ release
OH release
Steel slagQe (mg Ca/g)k2 (g/mg h)R2Qe (mg OH/g)k2 (g/mg h)R2
In deionized water 2.41 92.01 0.998 8.18 3.90 0.997 
In 25 mg/L ammonia 4.28 6.64 0.925 2.43 22.88 0.987 
 Ca2+ release
OH release
Steel slagQe (mg Ca/g)k2 (g/mg h)R2Qe (mg OH/g)k2 (g/mg h)R2
In deionized water 2.41 92.01 0.998 8.18 3.90 0.997 
In 25 mg/L ammonia 4.28 6.64 0.925 2.43 22.88 0.987 

Weathering has an influence on steel slag changing the Ca2+ and OH release capacity (Wang et al. 2010; Barca et al. 2012). This might have affected the steel slag used within this study as it was exposed to weathering over an uncertain period of time before being used in the experiments. During the weathering of steel slag, a certain amount of Ca(OH)2 and CaCO3 might be converted from CaO. The dissolution of CaO-slag in this study might contain three parts: (1) the dissolution of CaO-slag, (2) the dissolution of Ca(OH)2-slag, and (3) the constraint from CaCO3 covered on the surface of steel slag. The dissolution of Ca(OH)2 is probably different from CaO. CaCO3 is not dissolved in water, therefore, it would act like a cover to inhibit the dissolution of CaO or Ca(OH)2.

The influence of Ca2+ and OH on the adsorptive performance of ammonia nitrogen onto zeolite

Two series of tests were conducted to examine the various effects of Ca2+ and OH concentrations on ammonia nitrogen adsorption performance. The experimental results are shown in Figure 3. In one of the series (shown in Figure 3(a)), initial pH ranges were set from 4 to 12. It can be seen from that the adsorption capacity of ammonium ion increased slightly with the pH changed from 4 to 7. When pH value of the solution was 7, the adsorption capacity of ammonium ion reached a peak. Moreover, pH changing from 7 to 12 resulted in a drastic reduction of the ammonium ion adsorption capacity. The experimental results indicated that the adsorption capacity for ammonium onto zeolite was obviously impacted by pH of ammonium solution. This phenomenon was also widely reported by other authors (e.g. Du et al. 2005; Beebe et al. 2013). The reason might be that pH can influence both the character of the exchanging ions and the character of zeolite itself.
Figure 3

Effects of initial solution conditions on the adsorption performance of zeolite: (a) pH; and (b) Ca2+ concentration.

Figure 3

Effects of initial solution conditions on the adsorption performance of zeolite: (a) pH; and (b) Ca2+ concentration.

In solutions at pH below 7, ammonium ion is the dominant specie of ammonium chloride solution. As a competitor for exchanging sites on zeolite with ammonium ions, the number of H+ ion also rises at low pH, which explains the slight increase in adsorption capacity of ammonium ion from pH 4 to 12. At pH above 7, the reduction of adsorption capacity can be understandable by considering the transformation among and NH3·H2O with pH changing (see Equation (4)). Certainly, when the molecular form ammonia (NH3) is dominant in solution at high pH, the ion exchanging adsorption onto zeolite would be damaged badly. On the other hand, Lee et al. (2010) and Renman et al. (2009) have stated that elevated pH (>10) caused by steel slag filter could shift the activity of ammonium species towards unionized ammonia (pKa for ammonium is 9.24, and above this value, NH3 is the predominant form), which might be more readily volatilized. This might be another reason for the reduction of adsorption capacity at pH > 7.

Compared with the initial solutions at pH below 7, the pH value rose after adsorption equilibrium (see Figure 3(a)). This is probably because parts of H+ ions were adsorbed. For solutions at pH above 9, the reduction of pH after adsorption processes is due to part of molecular-formed ammonia transforming into ammonia gas.

The concentration of Ca2+ also has a great impact on the adsorption capacity of ammonium ions onto zeolite. As observed in Figure 3(b), the adsorption capacity of ammonium onto zeolite decreased gradually with the increase of Ca2+ concentration. It indicates that high Ca2+ concentration in solution can inhibit the adsorption of ammonium onto zeolite. In general, the ion exchange preferential order goes as following: (Tsitsis & Ronikashvili. 1992). At lower Ca2+ concentration, ammonium ion is adsorbed by zeolite in preference to calcium ion. Therefore, calcium ions, which are already adsorbed onto zeolite, might be replaced by ammonium ions. However, at higher Ca2+ concentration, the ammonium ions have to compete with calcium ions for being adsorbed onto exchange sites as the Ca2+ concentration increases. Equilibrium between the two species would be reached at last. As a result, adsorption of calcium ions onto zeolite caused the reduction of adsorption capacity of ammonium ion. This theory could also be confirmed by comparing Ca2+ concentrations before and after adsorption experiments, which shows that (1) the Ca2+ concentration after adsorption was greater than the corresponding initial values when that initial values were below 80 mg Ca/L and (2) the Ca2+ concentration after adsorption was lower than the corresponding initial values when that initial values were above 80 mg Ca2+ per liter. Under the condition of 20 g zeolite in 500 mL aqueous solution, the equilibrium concentrations for Ca2+ and are 80 mg·L−1 and 0.8 mg·L−1, respectively.

The influence of steel slag on zeolite saturated with

In order to simulate the situation of steel slag being placed ahead of zeolite, two tests were carried out to further explore the inhibition mechanism of steel slag on adsorption onto zeolite. In one test, steel slag was added into aqueous solution with saturated zeolite. In another contrast test, there was only zeolite saturated with in solution. Pseudo-equilibrium of zeolite adsorbing was assumed being achieved after 4 days of agitation. After equilibrium, steel slag was added in. The variation of pH, Ca2+ and concentrations in both tests are illustrated in Figure 4. For the control sample (without steel slag), pH, Ca2+ and concentrations of the solution had no large fluctuation from the 96th hour on. But in another test (where steel slag was added), the pH and Ca2+ concentration skyrocketed to a high level in 4 to 8 h after the steel slag addition (steel slag was added at the 96th hour). Whereafter, both pH and Ca2+ concentration decreased gradually and slightly from the 104th hour to the end of the tests. The concentration increased after the addition of steel slag, and the final value was close to the initial concentration. The release was probably because of the change of pH and Ca2+ concentrations in solution. Equations (7) and (8) can be used to calculate the correlation parameter.
formula
7
formula
8
formula
9
Figure 4

The influence of steel slag on zeolite saturated with : (a) pH; (b) Ca2+ and .

Figure 4

The influence of steel slag on zeolite saturated with : (a) pH; (b) Ca2+ and .

The subscripts Z, S and G refer to on zeolite, in aqueous solution and in gas phase, respectively. Where N (mol) is molar mass, Cmax (mg/L) are the maximum values of the Ca2+, OH and concentrations, Cmin are minimum values of the Ca2+, OH and concentrations, V (L) is the volume of the solution, M is relative molecular mass.

According to Equation (7), the released capacity of should be equal to the reduced capacity of OH or to two folds of reduced capacity of Ca2+, respectively. The maximum and minimum values of pH, Ca2+ and concentrations were then selected to calculate their amount of substance according to Equation (9). The values of Cmax, Cmin and N (i.e. the amount of substance) of Ca2+, OH and are shown in Table 4. Due to molecular form ammonia being a volatile liquid, a portion of ammonia (NHH2O) was expected to transform into ammonia gas (NH3), which would volatilize into atmosphere and do not determine concentration in solution. Thus, it can be confirmed that the reduced capacity of OH is approximately equal to the released capacity of (1.43 mmol + volatile ammonia gas (NH3) ≈ 1.73 mmol OH). But, the reduced capacity of Ca2+ was much more than expected, i.e. a half of the released capacity of (1.35 mmol Ca2+ > 0.5*1.43 mmol ). The unaccounted-for reduced capacity of Ca2+, i.e. approximately 0.63 mmol, probably adhered onto the surface of zeolite in forms of solid crystal or cation. And the proof based upon scanning electron microscopy (SEM) morphological analysis of zeolite surface would be stated in the next sector. It can be told that the surface of zeolite in zeolite-steel group was covered by crystal substance. Based upon this evidence, a deduction can be made that steel slag inhibits the ammonium adsorption capacity of zeolite, mainly because of the release of Ca2+ and OH- from steel slag. Ammonium ions can react with hydroxyl ion to produce molecular form ammonia (NH3·H2O), which would cause the dissociation of from zeolite. Whereafter, a quantity of Ca2+ will replace the ions to adhere onto the surface of zeolite. Thus, the existence of steel slag would result in a poor adsorption capacity of ammonium ion onto zeolite.

Table 4

Values of Cmax, Cmin and N (i.e. the amount of substance) of Ca2+, OH and NH4+-N

 OH (as N)Ca2+
Cmax (mg/L) 77.69 20.81 165.26 
Cmin (mg/L) 48.25 0.83 117.65 
N (mmol) 1.73 1.43 1.19 
 OH (as N)Ca2+
Cmax (mg/L) 77.69 20.81 165.26 
Cmin (mg/L) 48.25 0.83 117.65 
N (mmol) 1.73 1.43 1.19 

Optimization of the substrates bed configuration

Based on the inhibition effects of steel slag on zeolite adsorbing , a new substrate bed configuration that zeolites are placed upstream, followed by steel slag downstream, was proposed to eliminate the disadvantageous effect of steel slag. Zeolite and steel slag were filled into reactor C according to this bed configuration. At the same time, as a contrast, two other bed configurations are also prepared (sequentially there was steel slag and zeolite in reactor A from the bottom up, and in reactor B zeolite and steel slag were all mixed up). During the period of operation, three reactors all showed relatively stable and high efficiency for phosphate removal, exhibiting the configuration of zeolite and steel slag having no effect on the phosphate removal. The P removal rate ranges for A, B, and C were all in the range of 80–90%.

However, it can be seen from Figure 5 that different bed configurations of zeolite and steel slag have significant effects on removal. The average removal efficiency of reactor C is obviously higher than reactors A and B. This result has a good agreement with our expectation. In reactor C, the wastewater flowed through zeolite layer and steel slag layer sequentially. Zeolite layer can provide substantial removal of nitrogen from wastewater. In this progress, the Ca2+ and OH of steel slag had not yet released into wastewater. When wastewater flowed through the steel slag layer, the released Ca2+ and OH outflowed from the top of reactor with wastewater, and could not affect the adsorption of ammonium ion on zeolite.
Figure 5

Optimization of three substrates bed configurations.

Figure 5

Optimization of three substrates bed configurations.

The concentration in the effluent of reactor A kept rising from the 23th day (as shown in Figure 5). Figure 6(c) shows the SEM pictures of zeolite particles sampled from the bottom of zeolite layer before the 23th day. Crystal substance can be seen on the surface of zeolite particles. While, at the same time, this crystal substance could not be found on zeolite particles in reactor C, as well as on zeolite particles sampled from the top of zeolite layer in reactor A. The chemical composition of the crystals is calcium and oxygen based upon the results of EDS. Obviously, calcium crystal is prone to adhere onto the surface of zeolite particles immediately when they encounter each other. Therefore, as the operation goes on, more and more zeolite particles from the bottom up were covered by the crystal, and lost their ammonia nitrogen adsorption ability. Zeolite could be viewed as a ‘buffer’ to decrease the bad impact of high Ca2+ and OH concentrations. In reactor A, this buffer capacity of zeolite decreased gradually and disappeared at the 23rd day from reactor initiating. That explains the phenomenon that, removal efficiency declined rapidly from the 23th day in Figure 5. The removal efficiency of reactor B was always lower than that in reactor C, and also lower than that in reactor A before the 23th day. The reason might be due to the mixture of two substrates enabling the inhibition effect of steel slag from the very beginning.
Figure 6

SEM pictures of zeolite in: (a) column without steel slag; (b) column with steel slag; (c) bottom of zeolite of Reactor A.

Figure 6

SEM pictures of zeolite in: (a) column without steel slag; (b) column with steel slag; (c) bottom of zeolite of Reactor A.

In this study, the relationship and interaction when using steel slag and zeolite in combination on ammonium ions removal has been studied. As steel slag is rich in CaO, the dissolution of CaO off slag happens no matter in water or in solution, and it results in quick increase of pH and Ca2+ concentration in any solutions. An equilibrium was reached at some time between the 36th and 48th hour of agitation and modeling results of Ca2+ and OH release from slag indicated that pseudo-second-order reaction had a better fitness than pseudo-first-order reaction. The existence of ions influenced the CaO dissolution of steel slag via reacting with OH.

Both of pH and Ca2+ can change the adsorption capacity of ammonium ion onto zeolite. The pH value changing from 7 to 12 resulted in a drastic reduction of the ammonium ion adsorption capacity from the peak at pH7. The adsorption capacity of ammonium onto zeolite decreased gradually with the increase of Ca2+ concentration, indicating that high Ca2+ concentration in solution can inhibit the adsorption of ammonium onto zeolite.

The mechanism for steel slag inhibiting the ammonium adsorption capacity of zeolite is that, OH released from steel slag can react with ammonium ions to produce molecular form ammonia (NH3·H2O), which would cause the dissociation of from zeolite. And also, a quantity of Ca2+ could replace the ions to adhere onto the surface of zeolite.

A new substrate bed configuration of zeolites placed upstream, followed by steel slag downstream was proposed to eliminate the disadvantageous effect of steel slag. And experimental results showed that the new bed configuration was superior to the other two contrast patterns in terms of ammonium removal.

This research was supported by the Fundamental Research Funds for the Central Universities (No. 2016ZCQ03).

Barca
C.
Gérente
C.
Meyer
D.
Chazarenc
F.
Andrès
Y.
2012
Phosphate removal from synthetic and real wastewater using steel slags produced in Europe
.
Water Research
46
(
7
),
2376
2384
.
Barca
C.
Meyer
D.
Liir
M.
Drissen
P.
Comeau
Y.
Andrès
Y.
Chazarenc
F.
2014
Steel slag filters to upgrade phosphorus removal in small wastewater treatment plants: removal mechanisms and performance
.
Ecology Engineering
68
,
214
222
.
Beebe
D. A.
Castle James
W.
Rodgers
J. H.
Jr.
2013
Treatment of ammonia in pilot-scale constructed wetland systems with clinoptilolite
.
Journal of Environmental Chemical Engineering
1
(
4
),
1159
1165
.
Bowden
L. I.
Jarvis
A. P.
Younger
P. L.
Johnson
K. L.
2009
Phosphorus removal from waste waters using basic oxygen steel slag
.
Environmental Science & Technology
43
(
7
),
2476
2481
.
Brix
H.
Arias
C. A.
Bubba
M. D.
2001
Media selection for sustainable phosphorus removal in subsurface flow constructed wetlands
.
Water Science and Technology
44
(
11–12
),
47
54
.
Chris
P.
Andy
S.
Steven
P.
Richard
G.
Nanthi
S. B.
2007
Phosphorus removal mechanisms in active slag filters treating waste stabilization pond effluent
.
Environmental Science & Technology
41
(
9
),
3296
3301
.
Claveau-Mallet
D.
Courcelles
B.
Comeau
Y.
2014
Phosphorus removal by steel slag filters: modeling dissolution and precipitation kinetics to predict longevity
.
Environmental Science & Technology
48
(
13
),
7486
7493
.
Drizo
A.
Comeau
Y.
Forget
C.
Chapuis
R. P.
2002
Phosphorus saturation potential: a parameter for estimating the longevity of constructed wetland systems
.
Environmental Science & Technology
36
(
21
),
4642
4648
.
Drizo
A.
Forget
C.
Chapuis
R. P.
Comeau
Y.
2006
Phosphorus removal by electric arc furnace steel slag and serpentinite
.
Water Research
40
(
8
),
1547
1554
.
Du
Q.
Liu
S. J.
Cao
Z. H.
Wang
Y. Q.
2005
Ammonia removal from aqueous solution using natural Chinese clinoptilolite
.
Separation and Purification Technology
44
(
3
),
229
234
.
Ho
Y. S.
McKay
G.
1999
Pseudo-second-order model for sorption processes
.
Process Biochemistry
34
(
5
),
451
465
.
Huang
H. M.
Xiao
X. M.
Yan
B.
Yang
L. P.
2010
Ammonium removal from aqueous solutions by using natural Chinese (Chende) zeolite as adsorbent
.
Journal of Hazardous Materials
175
(
1–3
),
247
252
.
Kostura
B.
Kulveitová
H.
Leško
J.
2005
Blast furnace slags as sorbents of phosphate from water solutions
.
Water Research
39
(
9
),
1795
1802
.
Mateus
D. M. R.
Vaz
M. M. N.
Pinho
H. J. O.
2012
Fragmented limestone wastes as a constructed wetland substrate for phosphorus removal
.
Ecology Engineering
41
,
65
69
.
Olfa
K.
Yasunori
K.
Hitoshi
M.
Kengo
K.
Mamuro
N.
2002
Nutrients adsorption from seawater by new porous carrier made from zeolitized fly ash and slag
.
Marine Pollution Bulletin
45
(
1–12
),
311
315
.
Pratt
C.
Shilton
A.
Pratt
S.
Haverkamp
R. G.
Bolan
N. S.
2007
Phosphorus removal mechanisms in active slag filters treating waste stabilization pond effluent
.
Environmental Science & Technology
41
(
9
),
3296
3301
.
Tanner
C. C.
Kadlec
R. H.
Gibbs
M. M.
Sukias
J. P. S.
Nguyen
M. L.
2002
Nitrogen processing gradients in subsurface flow treatment wetlands – Influence of wastewater characteristics
.
Ecology Engineering
18
(
4
),
499
520
.
Tsitsis
H. G. V.
Ronikashvili
T. G.
1992
Natural Zeolites Chichester
.
Ellis Horwood Limited
,
Hertfordshire, UK
.
Wang
G.
Wang
Y. H.
Gao
Z. L.
2010
Use of steel slag as a granular material: volume expansion prediction and usability criteria
.
Journal of Hazardous Materials
184
(
1–3
),
555
560
.
Xue
Y.
Hou
H.
Zhu
S.
2009
Characteristics and mechanisms of phosphate adsorption onto basic oxygen furnace slag
.
Journal of Hazardous Materials
162
(
2–3
),
973
980
.