Environmental Technology
ISSN: 0959-3330 (Print) 1479-487X (Online) Journal homepage: http://www.tandfonline.com/loi/tent20
Effective adsorption of phosphate from wastewaters by big composite pellets made of reduced steel slag and iron ore concentrate Hongjuan Wang, Shaobo Shen, Longhui Liu, Yilong Ji & Fuming Wang To cite this article: Hongjuan Wang, Shaobo Shen, Longhui Liu, Yilong Ji & Fuming Wang (2015) Effective adsorption of phosphate from wastewaters by big composite pellets made of reduced steel slag and iron ore concentrate, Environmental Technology, 36:22, 2835-2846, DOI: 10.1080/09593330.2015.1050069 To link to this article: http://dx.doi.org/10.1080/09593330.2015.1050069
Accepted online: 03 Jun 2015.Published online: 10 Jun 2015.
Submit your article to this journal
Article views: 27
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tent20 Download by: [University of Nebraska, Lincoln]
Date: 26 September 2015, At: 21:27
Environmental Technology, 2015 Vol. 36, No. 22, 2835–2846, http://dx.doi.org/10.1080/09593330.2015.1050069
Effective adsorption of phosphate from wastewaters by big composite pellets made of reduced steel slag and iron ore concentrate Hongjuan Wanga,b , Shaobo Shena,b∗ , Longhui Liua,b , Yilong Jia,b and Fuming Wanga,b a State
Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing, Beijing 100083, People’s Republic of China; b School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing, Beijing 100083, People’s Republic of China
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
(Received 15 November 2014; accepted 21 April 2015 ) In order to remove phosphate from wastewater, a large plastic adsorption column filled with big phosphate-adsorbing pellets with diameters of 10 mm, heated by electromagnetic induction coils, was conceived. It was found that the prepared big pellets, which were made of reduced steel slag and iron ore concentrate, contain magnetic Fe and Fe3 O4 . The thermodynamics and kinetics of adsorption of phosphate from synthetic wastewaters on the pellets were studied in this work. The phosphate adsorption on the pellets followed three models of Freundlich, Langmuir and Dubinin–Kaganer–Radushkevick. The maximum phosphate adsorption capacity Qmax of the pellets were 2.46, 2.74 and 2.77 mg/g for the three temperatures of 20°C, 30°C and 40°C, respectively, based on the Langmuir model. The apparent adsorption energies were − 12.9 kJ/mol for the three temperatures. It implied that ion exchange was the main mechanism involved in the adsorption processes. The adsorbed phosphate existed on the pellet surface mainly in the form of Fe3 (PO4 )2 . A reduction pre-treatment of the pellet precursor with H2 greatly enhanced pellet adsorption for phosphate. The adsorption kinetics is better represented by a pseudo-first-order model. The adsorbed phosphate amounts were similar for both real and synthetic wastewaters under similar adsorption conditions. The percentage of adsorbed phosphate for a real wastewater increased with increasing pellet concentration and reached 99.2% at a pellet concentration of 64 (g/L). Some specific phosphate adsorption mechanisms for the pellets were revealed and the pellets showed the potential to efficiently adsorb phosphate from a huge amount of real wastewaters in an industrial scale. Keywords: phosphate adsorption; big composite pellets; reduced steel slag and iron ore concentrate; thermodynamic and kinetics; adsorption mechanisms
1. Introduction Since the late 1990s, 77% of the lakes and 30.8% of the reservoirs in China have been reported to be polluted with eutrophication.[1] Phosphorus (P) is usually considered to be the limiting nutrient for the eutrophication of water bodies. For this reason, legislation on P rejects for wastewater treatment plants (WWTPs) is becoming stricter around the world.[2] The discharging limit of phosphorus in WWTPs is between 1.0 and 2.0 mg P/L.[3,4] When plants and animals excrete wastes or die, the various forms of organic phosphorus in the decaying matter are mineralized by microorganisms and thus transformed to inorganic phosphorus, mostly in forms of orthophosphate and polyphosphate.[5] Subsurface flow and other treatment wetlands have been used for the purification of industrial and domestic wastewater.[6] They were the main means to remove phosphate from a huge amount of industrial or domestic wastewaters.[6,7] However, phosphate removal with the wetlands is somewhat problematic and phosphate concentration left in the treated wastewater was still high.[7] Thus, some additional separated
*Corresponding author. Email: [email protected] © 2015 Taylor & Francis
filter units containing adsorbents for phosphate are suitable techniques to upgrade phosphate removal by wetlands.[8] Steel slag is molten oxides composite obtained at about 1600°C. It has very high hardness and can resist the corrosion of acidic and basic solutions. Its density is about 3.1–3.6 g/cm3 . It mainly contains about 40wt% of CaO, 10 wt% of SiO2 and 10 wt% of FeO. Many laboratoryscale studies have demonstrated that steel slag is a suitable adsorbent for efficient and economical phosphate removal from wastewater.[9–13] Zeng et al. studied P removal from wastewaters by adsorption using the iron oxide tailing,[9] which had an average particle size of 0.068 mm, derived from a mineral-processing industry and found that the maximum phosphate adsorption capacity Qmax of the iron oxide tailing was 8.21 mg/g at room temperature based on the Langmuir model and P adsorption increased with increasing temperature. The majority of crystalline iron oxides in the tailings was magnetite (Fe3 O4 ). The adsorption mechanism was not disclosed in that study. Jha et al. studied P adsorption using thermally activated steel-making slag under air,[10] which had an average particle size
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2836
H. Wang et al.
of 0.015 mm. It was found that the measured phosphate uptake was 261 mg/g with a pre-treatment temperature of 700°C. The majority of crystalline iron oxides in the calcined slag was kirschsteinite (CaFeSiO4 ). The majority of phosphate was adsorbed in the form of calcium phosphate. Drizo et al. studied P removal from wastewaters by column adsorption using the electric arc furnace steel slag,[11] which had a particle size of 2.5–10 mm, and found that the phosphate adsorption was very efficient. Adding limestone to the electric arc furnace steel slag did not significantly improve P removal efficiency, probably because steel slag is already rich in Ca. Kim et al. studied P removal from wastewaters by adsorption using the converter steel slag,[12] which had a particle size of 0.075–0.015 mm, and by precipitation using pH adjustment, under the assumption that phosphates can be removed by producing a hydroxyapatite precipitate, Ca5 (OH)(PO4 )3 . It was found that phosphate from wastewater was removed by adding Ca2+ to the mixture of steel slag and wastewater and P removal increased with increasing temperature. Xiong et al. studied P removal from wastewaters by adsorption using the steel slag,[3] which had particle size less than 2 mm and found that the maximum phosphate adsorption capacity Qmax was 5.3 mg/g at 25°C and P adsorption increased with increasing temperature. Barca et al. studied P removal from wastewaters by column adsorption using the steel slags as received,[2] which had a particle size of 5–12 mm, and found that the phosphate adsorption was very efficient. The main mechanism of P removal was related to CaO dissolution from slag followed by calcium phosphate precipitation and accumulation of the precipitates into the filters. However, large differences in experimental conditions lead to difficulties in comparing the results of the different studies.[13] The steel slag is very cheap and available in large quantity in China and other countries.[1,3,14,15] An attractive feature of the steel slag adsorbent is that the phosphate-loaded steel slag can be used in agriculture as phosphorus fertilizer and soil conditioner.[16] Thus, phosphate adsorption using an adsorption column made of plastic was conceived by us. Initially, a converter slag was used as the adsorbent in the column. It was found by us that phosphate adsorption increased with increasing solution temperature. This phenomenon was also reported by other researchers.[2,3,9,15,16] In cold weather, the flow rate of wastewater through a filter column will be reduced due to increasing water viscosity at low temperature. In order to increase phosphate adsorption and wastewater flowrate in the column during cold weather, the adsorbent and the wastewater passing through the column must be heated. An electromagnetic induction heating coil was preferred for this purpose. Compared to the traditional electrical resistance wire heating, the electromagnetic induction heating technology can save more than 70% of energy and its heating rate can be increased by more than 60%. Additionally, this heating equipment is free of maintenance. In order
to realize electromagnetic induction heating, the adsorbent must contain magnetic Fe or Fe3 O4 . The iron content from steel slags was about 10 wt%, which was too low. Thus, an iron ore concentrate containing 68 wt% of metallic iron was added to the steel slag to increase the iron content. In order to reduce the flowing resistance of wastewater through the filter column, a big cylinder pellet with a diameter of 10 mm was used in this work. Some deionized water was added to a powder mixture composed of steel slag, iron ore concentrate and bentonite. Then part of this mixture was compressed into a big cylinder pellet precursor. After that, this pellet precursor was sintered under air and then reduced with H2 . A reduced composite pellet with iron content of about 48 wt% was thus obtained. Previously, only steel slags as received without chemical H2 reduction pre-treatment were reported to adsorb phosphate from wastewaters. It was first reported in this work that chemical reduction pre-treatment with H2 greatly increased steel slag adsorption for phosphate and the mechanism involved was never reported before. Additionally, composite pellet made of reduced steel slag and iron ore concentrate has also never been studied for phosphate adsorption and the adsorption capacity of the big composite pellet was unknown. The objectives of this study were to investigate the adsorption capacities of the synthetic big composite pellets and study thermodynamic, kinetic and mechanisms of phosphate adsorption on the novel composite pellets.
2. Experimental 2.1. Materials 2.1.1. Preparation of composite pellets A steel slag (converter steel slag) and an iron ore concentrate from two metallurgical plants in China were used in this work. They were dried, ground and sieved. The samples of steel slag and iron ore concentrate with a particle size fraction of 74 − 106 μm were used in this work. The samples were dried at 105°C for 2 h and kept in a desiccator. The chemical compositions of the samples are listed in Table 1. The contents of iron and phosphate from the steel slag were 10.48 and 0.48 wt%, respectively. The contents of iron and phosphate from the iron ore concentrate were 68.78 and 0.00 wt%, respectively. In order to make pellets, about 10 g of deionized water was added into a powder mixture composed of 50 g of converter slag, 50 g of iron ore concentrate and 1 g of bentonite. The mixture thus obtained was mixed sufficiently to make a precursor sample. After that, about 1.5 g of precursor sample was taken from the mixture and pressed under a pressure of 40 MPa into a cylinder pellet with a diameter of 10 mm and a height of 7 mm. This pellet was placed in a quartz boat. The quartz boat was put at the centre of a horizontal tube furnace. The furnace temperature was first raised to 150°C and kept at this temperature for 30 min. Then, the furnace temperature was gradually raised to 600°C.
Environmental Technology
2837
Table 1. Chemical compositions of the samples analysed by XRF (wt%). Samples A B C
Fe
O
Ca
P
Mg
Al
Si
S
Ti
Mn
10.48 68.78 48.34
45.83 24.47 26.7
29.88 0.81 12.88
0.48 0 0.34
2.04 0.78 1.16
1.17 0.34 0.88
6.47 4.57 7.48
0 0.24 0.13
0.68 – 0.33
2.60 – 1.17
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Notes: A, steel slag; B, concentrated iron ore; C, pellet before adsorption.
After that, H2 was passed through the quartz tube with a flowrate of 150 mL/min; meanwhile, the furnace temperature was kept at 600°C. After 2 h, the furnace power was shut down and the furnace temperature was decreased naturally to room temperature. Then, the pellet in the boat was taken out and put in a plastic bag, which was then placed in a desiccator. A magnetic pellet composed of reduced steel slag and iron ore concentrate was thus synthesized. The pellet weight was about 1.2 g. This pellet was used in the adsorption experiments. In another experiment regarding the effects of temperature and time of H2 reduction pre-treatment on phosphate adsorption, some bigger pellets with a diameter of 30 mm and a weight of about 20 g were prepared in a very similar way. 2.1.2.
Powder of adsorbed substances
After adsorption at 20–40°C, some white powder was found to be deposited on the pellet surface. The white powder on the pellet surface was scrapped off and collected. The collected powder was characterized with X-ray diffractometer (XRD) and scanning electron microscope (SEM). 2.1.3.
Synthetic phosphate-containing wastewaters
Some synthetic phosphate-containing wastewaters used in this work were prepared by dissolving NaH2 PO4 ·2H2 O in deionized water. The pH of wastewaters were adjusted to 7.0 by HCl or NaOH solution. 2.1.4. Real phosphate-containing wastewater A real phosphate-containing wastewater from a secondary activated sludge process of a WWTP in Beijing was used in this work. The wastewater pH was 7.01, the concentration of phosphate in this wastewater was 15.64 mg/L. 2.2. Adsorption thermodynamics experiment One hundred milliliter of synthetic phosphate-containing wastewaters of different phosphate concentrations (18.34, 38.19, 55.22, 72.25, 92.09, 112.87, 132.09, and 154.75 ppm) was placed in each of 8 plastic bottles of 250 mL with tight screw lids. One pellet of about 1.2 g with a diameter of 10 mm as described before was placed in each of the 8 bottles. Then, the 8 bottles were put sequentially in an incubator (BS-1E, China) with a time interval of 0.5 h
agitated at 150 rpm and controlled at a preset temperature. After 24 h, the 8 bottles were taken out of the incubator sequentially and filtrated immediately on a Whatman GFA membrane. The filtrate thus obtained was analysed for phosphate concentration. The thermodynamic experiments of phosphate adsorption on big pellets were carried out at 20°C, 30°C and 40°C, respectively. The adsorbed phosphate amount, Qt in mg/g, and adsorbed phosphate percentage, At in %, were calculated by Equations (1) and (2), respectively (C0 V0 − Ct Vt ) × 0.001, W (C0 V0 − Ct Vt ) At = × 100, C0 V0
Qt =
(1) (2)
where C0 and Ct are the phosphate concentrations in an initial phosphate-containing synthetic wastewater and the filtrate in mg/L, respectively; V0 and Vt are the volumes of the initial synthetic wastewater and the filtrate in mL, respectively; W is the weight of the pellet in g. The distribution coefficient of phosphate ions, K d (L solution/kg pellet), is defined as the ratio of the phosphate concentration in the solid pellet (C1 ) to that in the solution phase (C2 ) and is calculated by the following equation: Kd =
C1 × 103 , C2
(3)
where C1 is the phosphate ions mass (mg) adsorbed in 1 g of the solid pellet (mg/g) and C2 is the phosphate ions mass (mg) left in 1 l of solution after adsorption (mg/L). 2.3. Adsorption kinetics experiment One hundred milliliter of synthetic phosphate-containing wastewater with a phosphate concentration of 39.09 ppm was placed in each of 11 plastic bottles of 250 mL with tight screw lids. One pellet of about 1.2 g with a diameter of 10 mm as described before was placed in each of the 11 bottles. Then, the 11 bottles were put simultaneously in an incubator (BS-1E, China) agitated at 150 rpm and controlled at a preset temperature with a variation of ± 1°C. After a certain time (0.5, 1, 2, 3, 4, 6, 8, 9, 10, 12 and 24 h), one of the bottles was taken out of the incubator and filtrated immediately on a Whatman GF-A membrane. The filtrate thus obtained was analysed for phosphate concentration. The adsorption kinetics experiments were carried out at 20°C, 30°C and 40°C.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2838
H. Wang et al.
2.4. Phosphate adsorption in a real wastewater One hundred milliliter real wastewater from a WWTP in Beijing as described previously was put in each of 5 plastic bottles of 250 mL with tight screw lids. Then, one, two, three, four and five weighed pellets (about 1.2 g with a diameter of 10 mm for each pellet) were placed in each of the 5 bottles. After that, the 5 bottles were put sequentially in an incubator (BS-1E, China) with a time interval of 0.5 h agitated at 150 rpm and controlled at 30°C. After 24 h, the 5 bottles were taken out of the incubator sequentially and filtrated immediately on a Whatman GFA membrane. The filtrate thus obtained was analysed for phosphate concentration.
adsorption temperature (Figure 2(b)). Thus, a higher temperature facilitates the phosphate adsorption on the pellets.
2.5. Analysis and characterization The phosphate concentrations in the solutions were analysed with inductively coupled plasma-atomic emission spectroscopy (SPECTRO ARCOS EOP, SPECTRO Analytical Instruments GmbH). After adsorption, the white substances coated on the surface of several pellets (Figure 1(a)) were scrapped off. The white powder substances thus collected were used in the characterization of adsorbed substances with XRD and SEM. The SEM samples were prepared by dispersing the ground white powder on one side of double-sided carbon conductive tape. The other side of the tape was supported on a pure copper block. SEM and energy dispersive spectrometer (EDS) observation was performed with Zeiss Ultra 55. XRD patterns of the samples were recorded with a Rigaku D/max-TTRIII X-ray diffractometer equipped with a Cu Kα radiation source (λ = 0.15405 nm). The sample patterns of XRD were analysed using the software MDI Jade5.0 with the aid of the JCPDF database.
3.3.1.
3.
Results and discussion
3.1. Effect of the contact time The distribution coefficients (K d ) of phosphate ions reached plateaus for 24 h at 20°C, 30°C and 40°C (Figure 2(a)). It implies that the adsorption equilibrium time for phosphate ions was 24 h at 20°C, 30°C and 40°C under the adsorption conditions. Based on these findings, an agitation time of 24 h was used in all thermodynamic experiments.
3.3. Thermodynamic studies of phosphate adsorption It is important to evaluate the most appropriate correlations for equilibrium data to optimize the design of an adsorption system. Langmuir, Freundlich, Dubinin– Kaganer–Radushkevich (DKR) isotherm models were used to describe the adsorption equilibrium. Experimental isotherm data were obtained at an equilibrium time of 24 h for different initial concentrations of phosphate ions in the solution. Langmuir adsorption model
The Langmuir adsorption isotherm is based on the assumption that all sites possess equal affinity for the adsorbate and monolayer of the adsorbate on the adsorbent.[17] The Langmuir isotherm is expressed by the following equation; Ceq Qeq
=
1 bQmax
Ceq , Qmax
(4)
where b (L/mg) and Qmax (mg/g) are the Langmuir constant and the theoretical monolayer saturation capacity of the pellet, respectively. High b values indicate high adsorption affinity. The equilibrium data were fitted to the Langmuir equation. The plot for this is shown in Figure 2(c). The linear plots of Ceq /Qt versus Ceq (Figure 2(c)) with regression correlation coefficients greater than 0.95 indicated the applicability of the Langmuir adsorption isotherm for the cases at three temperatures. The values of b and Qmax calculated from the intercept and slope of the linear plots are shown in Table 2. The values of Qmax were 2.46, 2.73 and 2.77 mg/g for the temperatures 20°C, 30°C and 40°C, respectively (Table 2). It implied that the saturation adsorption capacity of phosphate ions Qmax increased with an increase in the temperature. The values of b were 0.05, 0.06 and 0.11 for the temperatures 20°C, 30°C and 40°C, respectively (Table 2). It indicated that the bonding strength of phosphate ions to the pellet increased with an increase in the temperature. In order to predict the adsorption efficiency of the adsorption process, the dimensionless equilibrium parameter RL was determined by using the following equation [18]: RL =
3.2. Variation of Qeq with Ceq Qeq (mg/g) and Ceq (mg/L) are the amount of adsorbed phosphate ions per unit weight of adsorbent and unabsorbed phosphate ions concentration in solution at equilibrium, respectively. The variation of Qeq with Ceq is shown in Figure 2(b). Qeq increased with increasing Ceq . Ceq was proportional to the initial phosphate concentration C0 . Thus, Qeq increased with increasing initial phosphate concentration C0 . Additionally, Qeq increased with increasing
+
1 , 1+bC0
(5)
where C0 is the initial concentration of phosphate ions and b is the Langmuir constant. The values of RL < 1 represent favourable adsorption and the values greater than 1.0 represent unfavourable adsorption.[18] Also, RL values equal to 0 indicate irreversible adsorption, From this study, RL values for phosphate ions adsorption ranged from 0.056 to 0.52 for the three temperatures (Figure 4(b)). This is for initial phosphate concentrations ranging from 18.34 to 154.75 mg/L. Therefore, the adsorption process is favourable.
2839
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Environmental Technology
Figure 1. (a) Photos of the pellets before and after adsorption; (b) SEM image of the white powder (a) scrapped from the surface of the pellets after adsorption; (c) EDS of the red rectangle frame region indicated in (b); (d) elemental contents obtained according to the EDS analysis of (e) XRD of the white powder (a) scrapped from surface of the pellets after adsorption; and (f) variation of adsorbed phosphate amount and percentage with pellet concentration in a secondary wastewater from a WWTP in Beijing (Wastewater = 100 mL; initial phosphate concentration = 15.6 ppm; pH = 7.01; 30°C; 150 rpm; 24 h).
Moreover, RL values decreased with an increase in the initial phosphate concentration. A higher initial phosphate concentration was more favourable for the adsorption of phosphate on the pellet. So, the pellet can be used as a potential source for adsorption of phosphate ions from the phosphate-containing wastewater.
3.3.2. Freundlich adsorption model The empirical Freundlich equation is based on adsorption on a heterogeneous surface and is derived from the assumption that the adsorption sites are distributed exponentially with respect to the heat of adsorption.[17,18] The logarithmic linear form of the Freundlich equation is
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2840
H. Wang et al.
(a)
(b)
(c)
(d)
(e)
(f)
Figure 2. (a) Variation of distribution coefficient K d of phosphate with time (phosphate solution = 100 mL; pH = 7.0; pellet = 1.2 g; 150 rpm); (b) variation of equilibrium adsorption amount Qeq with equilibrium concentration of phosphate Ceq ; (c) Langmuir plot for phosphate adsorption on pellets; (d) variation of equilibrium parameter RL with initial concentration of phosphate; (e) Freundlich plot for phosphate adsorption on pellets and (f) DKR plot for phosphate adsorption on pellets (phosphate solution = 100 mL; pH = 7.0; pellet = 1.2 g; 150 rpm; 24 h).
given below: ln Qeq = ln KF +
1 ln Ceq , n
(6)
where Qeq (mg/g) and Ceq (mg/L) are the amount of adsorbed phosphate ions per unit weight of adsorbent
and unabsorbed phosphate ions concentration in solution at equilibrium, respectively, and K F and 1/n the Freundlich constants representing the adsorption capacity (mg1−n g−1 Ln ) and intensity (dimensionless) of the adsorbent, respectively. The values of 1/n less than 1 represent a favourable adsorption.[19]
Environmental Technology Table 2. Constants of adsorption isotherms of various models for phosphate on pellet (phosphate solution = 100 mL; pellet = 1.2 g; 150 rpm; 24 h). Temperature (°C)
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Model
Parameter
20
30
40
Freundlich K F (mg1−n g−1 Ln ) 0.4821 0.7046 0.9655 1/n 0.2010 0.2616 0.3125 R2 0.9561 0.9720 0.9786 Langmuir Qmax (mg/g) 2.46 2.73 2.77 b (L/mg) 0.050 0.061 0.107 R2 0.9548 0.9792 0.9844 DKR X m (mg/g) 4.84 5.21 5.23 B (mol2 /kJ2 ) 3 × 10−9 3 × 10−9 3 × 10−9 E ad (kJ/mol) 12.9 12.9 12.9 R2 0.9620 0.9597 0.9794
The equilibrium data were fitted to the Freundlich equation. The plot for this is shown in Figure 2(e). The linear plots of ln Qcq versus ln Ceq with regression correlation coefficients greater than 0.95 indicated the applicability of the Freundlich adsorption isotherm for the cases at the three temperatures. The values of K F and 1/n were calculated from the intercept and slope of the linear plots between ln Qeq and ln Ceq and are shown in Table 2. The values of K F were 0.48, 0.70 and 0.96 for the temperatures 20°C, 30°C and 40°C, respectively (Table 2). The values of 1/n were 0.20, 0.26 and 0.31 for the temperatures 20°C, 30°C and 40°C, respectively (Table 2), which indicated that the phosphate ions adsorption on the pellet was favourable for the three temperatures. 3.3.3. DKR model Langmuir and Freundlich isotherms do not give any idea about the adsorption mechanism. In order to distinguish between physical and chemical adsorption, the data were simulated with the DKR isotherm model.[20–22] The DKR equation is expressed as ln Qeq = ln Xm − Bε2 ,
(7)
where ε (Polanyi Potential) is [RT ln(1 + (1/Ceq )] (kJ/mol), Ceq are the unabsorbed phosphate ions concentration in solution at equilibrium (mol/L), X m is the DKR constant related to the maximum adsorption capacity of the pellet with respect to phosphate ions (mg/g) and B is the DKR constant related to the free energy of adsorption per mole of the adsorbate as it migrates to the surface of the adsorbent from an infinite distance in the solution (mol2 /kJ2 ). The plots of ln Qeq versus ε2 (Figure 2(f)) yielded straight lines with regression correlation coefficients greater than 0.98, which indicated the applicability of the DKR adsorption isotherm. The values of X m and B were calculated from the intercept and slope of the linear
2841
plots are shown in Table 2. The values of X m calculated from DKR adsorption isotherms were 4.84, 5.21 and 5.23 mg/g for the temperatures 20°C, 30°C and 40°C, respectively, which implied that the adsorption capacity (mg/g) of pellet for phosphate ions increased with an increase in the temperature. This finding was consistent with that obtained from the Langmuir adsorption isotherm (Table 2). The apparent adsorption energy (E ad ) from the DKR isotherm,[20–22] which was calculated using the following equation: −1 Ead = √ . 2B
(8)
The calculated values of E ad were − 12.9 kJ/mol for the temperatures of 20°C, 30°C and 40°C (Table 2). It is implied that the adsorption was an exothermic process. It is known that the magnitude of apparent adsorption energy E ad is useful for estimating the type of adsorption and if this value is below 8 kJ/mol the adsorption type can be explained by physical adsorption, between 8 and 16 kJ/mol the adsorption type can be explained by ion exchange, and over 16 kJ/mol the adsorption type can be explained by chemical adsorption.[20–22] The absolute values of E ad found in this study were 12.9 kJ/mol (Table 2), which implied that ion exchange was the main mechanism involved in the adsorption processes for the three temperatures. The effects of H2 reduction temperature and reduction time of the pellets on phosphate adsorption are presented in Figure 3. Both adsorbed phosphate amount and adsorbed phosphate percentage increased significantly with increasing reduction temperature and time (Figure 3). Thus, H2 reduction pre-treatment greatly enhanced the pellet adsorption capacity for phosphate. After adsorption, some white powder was left on the pellet surface (Figure 1(a)). The SEM and EDS of the white powder scrapped from the pellet surface are shown in Figure 1(b) and 1(c), respectively. The phosphate content in the pellet before adsorption was 0.34 wt% (Table 1). After adsorption, this value changed to 0.86 wt% (Figure 1(d)). Thus, at least part of phosphate ions were adsorbed on the pellet surface and existed in the white powder. The XRD of the white powder (Figure 1(a)) is shown in Figure 1(e). The main phases existing in the white powder were A-CaO (PDF no. 17-912), B-CaSiO3 (PDF no. 2-689), C,D,F,I,K-Fe3 O4 (PDF no. 75-1609), N-Al2 O3 (PDF no. 4-877), E,J,M-FeO (PDF no. 74-1886), G-Fe (PDF no. 87-722), H-Fe3 (PO4 )2 (PDF no. 48-1880) and L-Ca2 Fe9 O13 (PDF no. 83-1902). Therefore, most of the phosphates were adsorbed on the pellet surface probably in the form of Fe3 (PO4 )2 . Obviously, the scraped white powder contained both the adsorbed substances and the pellet surface layer. Thus, the ion exchange occurred on the pellet surface during the phosphate adsorption can be represented
2842
H. Wang et al.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
(a)
(b)
Figure 3. Effect of (a) H2 reduction temperature and (b) H2 reduction time of the pellets on phosphate adsorption (phosphate solution = 100 mL; pellet = 20 g; pellet diameter = 30 mm; initial phosphate concentration = 15.55 ppm; pH = 7.00; 30°C; 150 rpm; 4 h).
Based on the literature,[26] K sp = 9.91 × 10−16 , the −8 M. calculated soluble [PO3− 4 ] = 3.15 × 10 Thus, the formation of phosphate precipitate has the following decreasing order based on the thermodynamic calculations:
by the following equation:
(9) It means that Fe2+ on the pellet surface adsorbed phos3+ in the pellet precursor was phate (PO3− 4 ). When Fe reduced by H2 to a larger extent, more Fe2+ can be produced. This explained why a reduction pre-treatment of the pellet precursor by H2 greatly enhanced phosphate adsorption on the pellet as shown in Figure 3. Fe2+ on the pellet surface originated from Fe3 O4 (PDF no. 75-1609) and FeO (PDF no. 74-1886) on the pellet surface. Some studies indicated that most of the phosphates from the wastewaters was removed using steel slags in the form of calcium phosphates.[10,12] However, Drizo et al. found that adding Ca2+ to the wastewaters containing steel slags did not enhanced phosphate removal.[11] It implied that the formation of calcium phosphates probably did not completely govern the removal of phosphate. The solubility product constants (K sp ) of some common phosphate mineral precipitates are listed as follows: Ca3 (PO4 )2 → 3Ca2+ + 2PO3− 4 .
(10) −29
Based on the literature,[23] K sp = 2.07 × 10 −6 M calculated soluble [PO3− 4 ] = 1.44 × 10 Ca5 (OH)(PO4 )3 → 5Ca
2+
−
+ OH +
3PO3− 4 .
For the composite pellet used in this work, the atomic percentages of Fe and Ca were 15.06% and 10.47%. It implied that the Fe atom number was larger than that of Ca in the composite pellet. Thus, FePO4 · 2H2 O and Fe3 (PO4 )2 · 8H2 O were more readily to be formed than Ca3 (PO4 )2 or Ca5 (OH)(PO4 )3 for this kind of composite pellet. In addition, the number of Fe2+ from the composite pellet was likely much larger than that of Fe3+ due to H2 reduction pre-treatment based on XRD characterization (Figure 2(e)). This explained why only Fe3 (PO4 )2 phase from the pellet surface was detected by XRD (Figure 2 (e)). It is noteworthy that the Fe3 (PO4 )2 precipitate can be formed in the neutral solution,[25] although most of the − phosphates probably exist in the form of HPO2− 4 or H2 PO4 at this pH.
3.4. (11)
(12)
Based on the literature,[25] K sp = 1 × 10−35.77 , the −8 M calculated soluble [PO3− 4 ] = 5.50 × 10 FePO4 · 2H2 O → Fe3+ + PO3− 4 + 2H2 O.
Ca5 (OH)(PO4 )3 .
, the
Based on the literature,[24] K sp = 1 × 10−53.28 , the −5 M calculated soluble [PO3− 4 ] = 1.32 × 10 Fe3 (PO4 )2 · 8H2 O → 3Fe2+ + 2PO3− 4 + 8H2 O.
FePO4 · 2H2 O ≈ Fe3 (PO4 )2 · 8H2 O > Ca3 (PO4 )2 >
(13)
Kinetic studies of phosphate ions adsorption
Kinetic adsorption experiments were carried out in this work to investigate the adsorption rate and adsorption mechanism of phosphate ions. 3.4.1. Variation of Qt with t The variation of phosphate adsorption amount Qt with adsorption time t is shown in Figure 4(a). Qt increased significantly with t before 9 h. After that, Qt increased slowly with t between 9 and 24 h.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Environmental Technology (a)
(b)
(c)
(d)
(e)
(f)
2843
Figure 4. (a) Variation of Qt with t; (b) pseudo-second-order kinetics plot for phosphate adsorption on the pellets; (c) pseudo-first-order kinetics plot for phosphate adsorption on the pellets; (d) Arrhenius equation plot for pseudo-first-order adsorption of phosphate on the pellet; (e) external diffusion model for phosphate adsorption on the pellets and (f) interparticle diffusion model for phosphate adsorption on the pellets (phosphate solution = 100 mL; initial phosphate concentration = 39.1 ppm; pH = 7.0; pellet = 1.2 g; 150 rpm).
3.4.2. Pseudo-first and pseudo-second-order models The pseudo-first-order rate expression based on the solid capacity is generally expressed as follows [27]: dQ = k1 (Q − Qeq ), dt Qt = −k1 t. ln 1 − Qeq
The pseudo-second-order equation is also based on the adsorption capacity of the solid phase. It is expressed as follows [27]: dQ = k2 (Q − Qeq )2 , dt
(14)
1 = 1 + Qeq k2 t, 1 − (Qt /Qeq )
(15)
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2844
H. Wang et al.
where Qeq and Qt (both in mg/g) are the amount of phosphate ions adsorbed per unit mass of the resin at equilibrium and time t (h), respectively, and k 1 (1/h) and k2 (g/mg h) are the rate constants of the pseudo-first-order adsorption and pseudo-second-order adsorption, respectively. The kinetic experimental data were simulated using pseudo-second-order and pseudo-first-order models and the simulating curves are shown in Figure 4 (b) and 4(c), respectively. It is observed that for the pseudo-secondorder model there were not linear relationships between 1/(1 − Qt /Qeq ) and adsorption time t at the three temperatures (Figure 4(b)). However, such linear relationships occurred for the pseudo-first-order model (Figure 4(c)). The linear plots of ln(1 − (Qt /Qeq )) versus time t with regression correlation coefficients greater than 0.98 indicated the applicability of pseudo-first-order model at the three temperatures. Thus, it indicated that the adsorption kinetics data are better represented by the pseudo-firstorder model. The values for k1 for the three temperatures are calculated from the curve slopes shown in Figure 4(c). The adsorption rate constant k1 were 0.1947, 0.2166 and 0.2351 h−1 for temperatures 20°C, 30°C and 40°C, respectively (Figure 4(c)). The linear plot of lnk1 versus (103 /T) (K−1 ) (Figure 4(d)) with regression correlation coefficient 0.9968 indicated the applicability of the following Arrhenius equation: ln k1 = ln A −
Ea,1 , RT
(16)
where k 1 is the rate constant of pseudo-first-order reaction (1/h), A is the pre-exponential factor, E a,1 is the apparent adsorption activation energy (J/mol), R is the gas constant (8.314 J K−1 mol−1 ) and T is the temperature (K). The apparent adsorption activation energy E a,1 for the pseudo-first-order model was calculated from the slope of the linear plot of ln k 1 versus (103 /T) and found to be 7.2 kJ/mol (Figure 4(d)). 3.4.3. External diffusion model At early times of contact (e.g. between 0 and 10 min of contact) the system could be simplified by assuming that the concentration of phosphate ions on the pellet surface tended towards zero (pure external diffusion) and the internal diffusion was negligible. Thus, Ficks’ laws may be applied to describe the mass-transfer rate of external diffusion. The diffusion flux J of Ficks’ law could be expressed as follows [28]: J = kf Cb = −
dnb VdCb =− , Adt Adt
(17)
where Cb is the bulk concentration of phosphate ions (mol/m3 ), kf is the external mass-transfer coefficient of phosphate ions (m/h), A is the pellet surface (m2 ), V is the
volume of bulk solution (m3 ) and t is the diffusion time (h)
t
− 0
A kf dt = V ln
Ct
d ln Cb , C0
A C0 = kf t, Ct V
(18)
where C0 and Ct are the bulk concentrations of phosphate ions (mol/m3 ) at times of 0 and t, respectively. If external diffusion was a rate-controlling step at the beginning of phosphate adsorption process, there was a linear relationship between ln C0 /Ct and t as described by Equation (18) and the line pass through the original at a certain temperature. Such linear relationships at the time range between 0 and 8 h existed for the 3 temperatures (Figure 4(e)). However, after 8 h such linear relationships did not exist and the lines also did not pass through the original (Figure 4(e)). Thus external diffusion was the rate-controlling step only before 8 h. 3.4.4. Interparticle diffusion model Besides for phosphate adsorption on the external surface of adsorbent pellet, there is also a possibility of transporting adsorbate phosphate from the solution to the internal pores of the adsorbent pellet (intraparticle diffusion). The most commonly used technique for identifying the mechanism involved in the internal pore diffusion process is by using the intraparticle diffusion model [29–31] given by: Qt = kid t(1/2) + c,
(19)
where Qt (mg/g) is the amount of Pt ions adsorbed per unit mass of the resin at time t (h), kid is the intraparticle diffusion rate constant (mg g−1 h−1/2 ), and c is the intercept. For the intercept c, McKay et al. [30] have indicated that ‘extrapolation of the linear portion of the plot back to the y-axis provides intercepts which are proportional to the extent of the boundary layer thickness, that is, the larger the intercept the greater the boundary layer effect’. Plots of Qt versus t1/2 for the three temperatures are shown in Figure 4(f). From the figure, it is observed that there are two linear regions. One was between 0 and 8 h and the other was between 8 and 12 h. The intercepts of first linear portion were negative. Similar phenomena were also observed by McKay et al. [30,31]. They believed that the boundary layer thickness retarded intraparticle diffusion and led to negative intercepts.[30,31] The intercepts of second linear portion were positive, which indicated that intraparticle diffusion was a rate-controlling step between 8 and 12 h. In summary, the external diffusion was the main rate-controlling step of phosphate adsorption before 8 h and internal diffusion became the rate-controlling step between 8 and 12 h.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Environmental Technology 3.5. Adsorption of phosphate from real wastewater In order to investigate the adsorption performance of phosphate from a real wastewater on the pellets, the effect of pellet concentration in the wastewater on phosphate adsorption was studied and the results are presented in Figure 1(f). The adsorbed phosphate amount increased with increasing pellet concentration and reached a plateau value of 1.55 mg/g at a pellet concentration of 50 (g/L) (Figure 1(f)). The adsorbed phosphate amount was 0.95 (mg/g) for the real wastewater with an initial phosphate concentration of 15.6 ppm and a pellet concentration of 12 (g/L) at 30°C (Figure 1(f)). The adsorbed phosphate amount Qeq at equilibrium was 1.44 (mg/g) for the synthetic phosphate-containing solution with an initial phosphate concentration 18.3 ppm and a pellet concentration of 12 (g/L) at 30°C (Figure 2(b)). Thus, adsorbed phosphate amounts were similar for both synthetic and real phosphate-containing wastewaters. The adsorbed phosphate percentage also increased with increasing pellet concentration and reached 99.2% at pellet concentration of 64 (g/L) (Figure 1(f)). Thus, the pellets prepared in this work have the potential to be used as adsorption media filled in an adsorption column to remove phosphate from real wastewaters. 4.
Conclusions
The phosphate adsorption on the pellets made of the composite composed of reduced steel slag and concentrated iron ore by H2 was studied. It was found that the adsorption equilibrium time was about 24 h. Phosphate adsorption increased with increasing temperature. The phosphate adsorption on the pellets followed the three models of Freundlich, Langmuir and DKR. The maximum phosphate adsorption capacity Qmax of the pellets were 2.46, 2.74 and 2.77 mg/g for the temperatures 20°C, 30°C and 40°C, respectively, based on the Langmuir model. The apparent adsorption energy (E ad ) was − 12.9 kJ/mol for the 3 temperatures based on the DKR model. It implied that ion exchange was the mechanism involved in the adsorption processes. The adsorbed phosphate existed on the pellet surface mainly in the form of Fe3 (PO4 )2 . The adsorption kinetics data are better represented by the pseudo-firstorder model. The external diffusion was the main ratecontrolling step of phosphate adsorption before 8 h and internal diffusion became the rate-controlling step between 8 and 12 h. The adsorbed phosphate amounts were similar for both real and synthetic wastewaters under similar adsorption conditions. The percentage of adsorbed phosphate for a real wastewater increased with increasing pellet concentration and reached 99.2% at a pellet concentration of 64 (g/L). Some specific phosphate adsorption mechanisms for the composite pellets were revealed and the pellets showed the potential to efficiently adsorb phosphate from a huge amount of real wastewaters in an industrial scale.
2845
Disclosure statement No potential conflict of interest was reported by the authors.
Funding This work was supported by the grants from the National Natural Science Foundation of China [grant number 50874011] and the Programs for Innovative Research Team of undergraduate students at University of Science and Technology Beijing [grant numbers 14020118 and 14020367].
References [1] Ma J, Li H. Preliminary discussion on eutrophication status of lakes, reservoirs and rivers in China and overseas. Resour Environ Yangtze Basin. 2002;11:575–577 (in Chinese). [2] Barca C, Meyer D, Liira M, et al. Steel slag filters to upgrade phosphorus removal in small wastewater treatment plants: Removal mechanisms and performance. Ecol Eng. 2014;68:214–222. [3] Xiong J, He Z, Mahmood Q, Liu D, Yang X, Islam E. Phosphate removal from solution using steel slag through magnetic separation. J Hazard Mater. 2008;152:211–215 (in Chinese). [4] Tanada S, Kabayama M, Kawasaki N, et al. Removal of phosphate by aluminum oxide hydroxide. J Colloid Interf Sci. 2003;257:135–140. [5] Lee SH, Vigneswaran S, Moon H. Adsorption of phosphorus in saturated slag media columns. Sep Purif Technol. 1997;12:109–118. [6] Kadlec RH, Knight RL. Treatment wetlands. Boca Raton (FL): Lewis, CRC Press; 1996; p. 893. [7] Brix H, Arias CA, Del Bubba M. Media selection for sustainable phosphorus removal in subsurface flow constructed wetlands. Water Sci Technol. 2001;44:47–54. [8] Vohla C, Kõiv M, Bavor HJ, Chazarenc F, Mander Ü. Filter materials for phosphorus removal from wastewater in treatment wetlands– a review. Ecol Eng. 2011;37:70–89. [9] Zeng L, Li X, Liu J. Adsorptive removal of phosphate from aqueous solutions using iron oxide tailings. Water Res. 2004;38:1318–1326. [10] Jha VK, Kameshima Y, Nakajima A, Okada K. Hazardous ions uptake behavior of thermally activated steel-making slag. J Hazard Mater. 2004;B114:139–144. [11] Drizo A, Forget C, Chapuis RP, Comeau Y. Phosphorus removal by electricarc furnace steel slag and serpentinite. Water Res. 2006;40:1547–1554. [12] Kim EH, Lee DW, Hwang HK, Yim S. Recovery of phosphates from wastewater using converter slag: kinetics analysis of a completely mixed phosphorus crystallization process. Chemosphere. 2006;63:192–201. [13] Chazarenc F, Brisson J, Comeau Y. Slag columns for upgrading phosphorus removal from constructed wetland effluents. Water Sci Technol. 2007;56:109–115. [14] Bowden LI, Jarvis AP, Younger PL, Johnson KL. Phosphorus removal from wastewaters using basic oxygen steel slag. Environ Sci Technol. 2009;43:2476–2481. [15] Claveau-Mallet D, Wallace S, Comeau Y. Model of phosphorus precipitation and crystal formation in electric arc furnace steel slag filters. Environ Sci Technol. 2012;46: 1465–1470. [16] Johansson-Westholm L. Substrates for phosphorus removal –potential benefits for on-site wastewater treatment? Water Res. 2006;40:23–36.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2846
H. Wang et al.
[17] Saeed MM, Bajwa SZ, Ansari MS. Investigation of the removal of lead by adsorption onto 1-(2-thiazolylazo)-2naphthol (TAN) imbedded polyurethane foam from aqueous solution. J Chin Chem Soc. 2007;54:173–183. [18] Basha S, Murthy ZVP. Kinetic and equilibrium models for biosorption of Cr(VI) on chemically modified seaweed, Cystoseira indica. Process Biochem. 2007;42:1521– 1529. [19] Yamada H, Kayama M, Saito K, Hara M. A fundamental research on phosphate removal by using slag. Water Res. 1986;20(5):547–557. [20] Lin SH, Juang RS. Heavy metal removal from water by sorption using surfactant-modified montmorillonite. J Hazard Mater. 2002;92:315–326. [21] Wang CC, Juang LC, Lee CK, Hsu TC, Lee JF, Chao HP. Effects of exchanged surfactant cations on the pore structure and adsorption characteristics of montmorillonite. J Colloid Interf Sci. 2004;280:27–35. [22] Krishna BS, Murty DSR, Prakash BSJ. Thermodynamics of chromium(VI) anionic species sorption onto surfactantmodified montmorillonite clay. J Colloid Interf Sci. 2000;229:230–236. [23] Speight JG. Lange’s handbook of chemistry. 16th ed. New York: McGraw-Hill; 2005; p 1.333.
[24] Zhu Y, Zhang X, Chen Y, et al. A comparative study on the dissolution and solubility of hydroxylapatite and fluorapatite at 25°C and 45°C. Chem Geol. 2009;268:89–96. [25] Azam HM, Finneran KT. Fe(III) reduction-mediated phosphate removal as vivianite (Fe3 (PO4 )28 H2 O) in septic system wastewater. Chemosphere. 2014;97:1–9. [26] Speight JG. Lange’s handbook of chemistry. 16th ed. New York: McGraw-Hill; 2005; p 1.335. [27] Naiya TK, Bhattacharya AK, Das SK. Adsorption of Cd(II) and Pb(II) from aqueous solutions on activated alumina. J Colloid Interf Sci. 2009;333:14–26. [28] Shen S, Pan T, Liu X, et al. Adsorption of Rh(III) complexes from chloride solutions obtained by leaching chlorinated spent automotive catalysts on ion-exchange resin Diaion WA21J. J Hazard Mater. 2010;179:104–112. [29] Wu FC, Tseng RL, Juang RS. Initial behavior of intraparticle diffusion model used in the description of adsorption kinetics. Chem Eng J. 2009;153:1–8. [30] McKay G, Otterburn MS, Sweeney AG. The removal of color from effluent using various adsorbents III silica: rate processes. Water Res. 1980;14:15–20. [31] McKay G. The adsorption of dyestuffs from aqueous solutions using activated carbon. III. Intraparticle diffusion process. J Chem Technol Biotechnol A. 1983;33:196–204.
ISSN: 0959-3330 (Print) 1479-487X (Online) Journal homepage: http://www.tandfonline.com/loi/tent20
Effective adsorption of phosphate from wastewaters by big composite pellets made of reduced steel slag and iron ore concentrate Hongjuan Wang, Shaobo Shen, Longhui Liu, Yilong Ji & Fuming Wang To cite this article: Hongjuan Wang, Shaobo Shen, Longhui Liu, Yilong Ji & Fuming Wang (2015) Effective adsorption of phosphate from wastewaters by big composite pellets made of reduced steel slag and iron ore concentrate, Environmental Technology, 36:22, 2835-2846, DOI: 10.1080/09593330.2015.1050069 To link to this article: http://dx.doi.org/10.1080/09593330.2015.1050069
Accepted online: 03 Jun 2015.Published online: 10 Jun 2015.
Submit your article to this journal
Article views: 27
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tent20 Download by: [University of Nebraska, Lincoln]
Date: 26 September 2015, At: 21:27
Environmental Technology, 2015 Vol. 36, No. 22, 2835–2846, http://dx.doi.org/10.1080/09593330.2015.1050069
Effective adsorption of phosphate from wastewaters by big composite pellets made of reduced steel slag and iron ore concentrate Hongjuan Wanga,b , Shaobo Shena,b∗ , Longhui Liua,b , Yilong Jia,b and Fuming Wanga,b a State
Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing, Beijing 100083, People’s Republic of China; b School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing, Beijing 100083, People’s Republic of China
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
(Received 15 November 2014; accepted 21 April 2015 ) In order to remove phosphate from wastewater, a large plastic adsorption column filled with big phosphate-adsorbing pellets with diameters of 10 mm, heated by electromagnetic induction coils, was conceived. It was found that the prepared big pellets, which were made of reduced steel slag and iron ore concentrate, contain magnetic Fe and Fe3 O4 . The thermodynamics and kinetics of adsorption of phosphate from synthetic wastewaters on the pellets were studied in this work. The phosphate adsorption on the pellets followed three models of Freundlich, Langmuir and Dubinin–Kaganer–Radushkevick. The maximum phosphate adsorption capacity Qmax of the pellets were 2.46, 2.74 and 2.77 mg/g for the three temperatures of 20°C, 30°C and 40°C, respectively, based on the Langmuir model. The apparent adsorption energies were − 12.9 kJ/mol for the three temperatures. It implied that ion exchange was the main mechanism involved in the adsorption processes. The adsorbed phosphate existed on the pellet surface mainly in the form of Fe3 (PO4 )2 . A reduction pre-treatment of the pellet precursor with H2 greatly enhanced pellet adsorption for phosphate. The adsorption kinetics is better represented by a pseudo-first-order model. The adsorbed phosphate amounts were similar for both real and synthetic wastewaters under similar adsorption conditions. The percentage of adsorbed phosphate for a real wastewater increased with increasing pellet concentration and reached 99.2% at a pellet concentration of 64 (g/L). Some specific phosphate adsorption mechanisms for the pellets were revealed and the pellets showed the potential to efficiently adsorb phosphate from a huge amount of real wastewaters in an industrial scale. Keywords: phosphate adsorption; big composite pellets; reduced steel slag and iron ore concentrate; thermodynamic and kinetics; adsorption mechanisms
1. Introduction Since the late 1990s, 77% of the lakes and 30.8% of the reservoirs in China have been reported to be polluted with eutrophication.[1] Phosphorus (P) is usually considered to be the limiting nutrient for the eutrophication of water bodies. For this reason, legislation on P rejects for wastewater treatment plants (WWTPs) is becoming stricter around the world.[2] The discharging limit of phosphorus in WWTPs is between 1.0 and 2.0 mg P/L.[3,4] When plants and animals excrete wastes or die, the various forms of organic phosphorus in the decaying matter are mineralized by microorganisms and thus transformed to inorganic phosphorus, mostly in forms of orthophosphate and polyphosphate.[5] Subsurface flow and other treatment wetlands have been used for the purification of industrial and domestic wastewater.[6] They were the main means to remove phosphate from a huge amount of industrial or domestic wastewaters.[6,7] However, phosphate removal with the wetlands is somewhat problematic and phosphate concentration left in the treated wastewater was still high.[7] Thus, some additional separated
*Corresponding author. Email: [email protected] © 2015 Taylor & Francis
filter units containing adsorbents for phosphate are suitable techniques to upgrade phosphate removal by wetlands.[8] Steel slag is molten oxides composite obtained at about 1600°C. It has very high hardness and can resist the corrosion of acidic and basic solutions. Its density is about 3.1–3.6 g/cm3 . It mainly contains about 40wt% of CaO, 10 wt% of SiO2 and 10 wt% of FeO. Many laboratoryscale studies have demonstrated that steel slag is a suitable adsorbent for efficient and economical phosphate removal from wastewater.[9–13] Zeng et al. studied P removal from wastewaters by adsorption using the iron oxide tailing,[9] which had an average particle size of 0.068 mm, derived from a mineral-processing industry and found that the maximum phosphate adsorption capacity Qmax of the iron oxide tailing was 8.21 mg/g at room temperature based on the Langmuir model and P adsorption increased with increasing temperature. The majority of crystalline iron oxides in the tailings was magnetite (Fe3 O4 ). The adsorption mechanism was not disclosed in that study. Jha et al. studied P adsorption using thermally activated steel-making slag under air,[10] which had an average particle size
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2836
H. Wang et al.
of 0.015 mm. It was found that the measured phosphate uptake was 261 mg/g with a pre-treatment temperature of 700°C. The majority of crystalline iron oxides in the calcined slag was kirschsteinite (CaFeSiO4 ). The majority of phosphate was adsorbed in the form of calcium phosphate. Drizo et al. studied P removal from wastewaters by column adsorption using the electric arc furnace steel slag,[11] which had a particle size of 2.5–10 mm, and found that the phosphate adsorption was very efficient. Adding limestone to the electric arc furnace steel slag did not significantly improve P removal efficiency, probably because steel slag is already rich in Ca. Kim et al. studied P removal from wastewaters by adsorption using the converter steel slag,[12] which had a particle size of 0.075–0.015 mm, and by precipitation using pH adjustment, under the assumption that phosphates can be removed by producing a hydroxyapatite precipitate, Ca5 (OH)(PO4 )3 . It was found that phosphate from wastewater was removed by adding Ca2+ to the mixture of steel slag and wastewater and P removal increased with increasing temperature. Xiong et al. studied P removal from wastewaters by adsorption using the steel slag,[3] which had particle size less than 2 mm and found that the maximum phosphate adsorption capacity Qmax was 5.3 mg/g at 25°C and P adsorption increased with increasing temperature. Barca et al. studied P removal from wastewaters by column adsorption using the steel slags as received,[2] which had a particle size of 5–12 mm, and found that the phosphate adsorption was very efficient. The main mechanism of P removal was related to CaO dissolution from slag followed by calcium phosphate precipitation and accumulation of the precipitates into the filters. However, large differences in experimental conditions lead to difficulties in comparing the results of the different studies.[13] The steel slag is very cheap and available in large quantity in China and other countries.[1,3,14,15] An attractive feature of the steel slag adsorbent is that the phosphate-loaded steel slag can be used in agriculture as phosphorus fertilizer and soil conditioner.[16] Thus, phosphate adsorption using an adsorption column made of plastic was conceived by us. Initially, a converter slag was used as the adsorbent in the column. It was found by us that phosphate adsorption increased with increasing solution temperature. This phenomenon was also reported by other researchers.[2,3,9,15,16] In cold weather, the flow rate of wastewater through a filter column will be reduced due to increasing water viscosity at low temperature. In order to increase phosphate adsorption and wastewater flowrate in the column during cold weather, the adsorbent and the wastewater passing through the column must be heated. An electromagnetic induction heating coil was preferred for this purpose. Compared to the traditional electrical resistance wire heating, the electromagnetic induction heating technology can save more than 70% of energy and its heating rate can be increased by more than 60%. Additionally, this heating equipment is free of maintenance. In order
to realize electromagnetic induction heating, the adsorbent must contain magnetic Fe or Fe3 O4 . The iron content from steel slags was about 10 wt%, which was too low. Thus, an iron ore concentrate containing 68 wt% of metallic iron was added to the steel slag to increase the iron content. In order to reduce the flowing resistance of wastewater through the filter column, a big cylinder pellet with a diameter of 10 mm was used in this work. Some deionized water was added to a powder mixture composed of steel slag, iron ore concentrate and bentonite. Then part of this mixture was compressed into a big cylinder pellet precursor. After that, this pellet precursor was sintered under air and then reduced with H2 . A reduced composite pellet with iron content of about 48 wt% was thus obtained. Previously, only steel slags as received without chemical H2 reduction pre-treatment were reported to adsorb phosphate from wastewaters. It was first reported in this work that chemical reduction pre-treatment with H2 greatly increased steel slag adsorption for phosphate and the mechanism involved was never reported before. Additionally, composite pellet made of reduced steel slag and iron ore concentrate has also never been studied for phosphate adsorption and the adsorption capacity of the big composite pellet was unknown. The objectives of this study were to investigate the adsorption capacities of the synthetic big composite pellets and study thermodynamic, kinetic and mechanisms of phosphate adsorption on the novel composite pellets.
2. Experimental 2.1. Materials 2.1.1. Preparation of composite pellets A steel slag (converter steel slag) and an iron ore concentrate from two metallurgical plants in China were used in this work. They were dried, ground and sieved. The samples of steel slag and iron ore concentrate with a particle size fraction of 74 − 106 μm were used in this work. The samples were dried at 105°C for 2 h and kept in a desiccator. The chemical compositions of the samples are listed in Table 1. The contents of iron and phosphate from the steel slag were 10.48 and 0.48 wt%, respectively. The contents of iron and phosphate from the iron ore concentrate were 68.78 and 0.00 wt%, respectively. In order to make pellets, about 10 g of deionized water was added into a powder mixture composed of 50 g of converter slag, 50 g of iron ore concentrate and 1 g of bentonite. The mixture thus obtained was mixed sufficiently to make a precursor sample. After that, about 1.5 g of precursor sample was taken from the mixture and pressed under a pressure of 40 MPa into a cylinder pellet with a diameter of 10 mm and a height of 7 mm. This pellet was placed in a quartz boat. The quartz boat was put at the centre of a horizontal tube furnace. The furnace temperature was first raised to 150°C and kept at this temperature for 30 min. Then, the furnace temperature was gradually raised to 600°C.
Environmental Technology
2837
Table 1. Chemical compositions of the samples analysed by XRF (wt%). Samples A B C
Fe
O
Ca
P
Mg
Al
Si
S
Ti
Mn
10.48 68.78 48.34
45.83 24.47 26.7
29.88 0.81 12.88
0.48 0 0.34
2.04 0.78 1.16
1.17 0.34 0.88
6.47 4.57 7.48
0 0.24 0.13
0.68 – 0.33
2.60 – 1.17
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Notes: A, steel slag; B, concentrated iron ore; C, pellet before adsorption.
After that, H2 was passed through the quartz tube with a flowrate of 150 mL/min; meanwhile, the furnace temperature was kept at 600°C. After 2 h, the furnace power was shut down and the furnace temperature was decreased naturally to room temperature. Then, the pellet in the boat was taken out and put in a plastic bag, which was then placed in a desiccator. A magnetic pellet composed of reduced steel slag and iron ore concentrate was thus synthesized. The pellet weight was about 1.2 g. This pellet was used in the adsorption experiments. In another experiment regarding the effects of temperature and time of H2 reduction pre-treatment on phosphate adsorption, some bigger pellets with a diameter of 30 mm and a weight of about 20 g were prepared in a very similar way. 2.1.2.
Powder of adsorbed substances
After adsorption at 20–40°C, some white powder was found to be deposited on the pellet surface. The white powder on the pellet surface was scrapped off and collected. The collected powder was characterized with X-ray diffractometer (XRD) and scanning electron microscope (SEM). 2.1.3.
Synthetic phosphate-containing wastewaters
Some synthetic phosphate-containing wastewaters used in this work were prepared by dissolving NaH2 PO4 ·2H2 O in deionized water. The pH of wastewaters were adjusted to 7.0 by HCl or NaOH solution. 2.1.4. Real phosphate-containing wastewater A real phosphate-containing wastewater from a secondary activated sludge process of a WWTP in Beijing was used in this work. The wastewater pH was 7.01, the concentration of phosphate in this wastewater was 15.64 mg/L. 2.2. Adsorption thermodynamics experiment One hundred milliliter of synthetic phosphate-containing wastewaters of different phosphate concentrations (18.34, 38.19, 55.22, 72.25, 92.09, 112.87, 132.09, and 154.75 ppm) was placed in each of 8 plastic bottles of 250 mL with tight screw lids. One pellet of about 1.2 g with a diameter of 10 mm as described before was placed in each of the 8 bottles. Then, the 8 bottles were put sequentially in an incubator (BS-1E, China) with a time interval of 0.5 h
agitated at 150 rpm and controlled at a preset temperature. After 24 h, the 8 bottles were taken out of the incubator sequentially and filtrated immediately on a Whatman GFA membrane. The filtrate thus obtained was analysed for phosphate concentration. The thermodynamic experiments of phosphate adsorption on big pellets were carried out at 20°C, 30°C and 40°C, respectively. The adsorbed phosphate amount, Qt in mg/g, and adsorbed phosphate percentage, At in %, were calculated by Equations (1) and (2), respectively (C0 V0 − Ct Vt ) × 0.001, W (C0 V0 − Ct Vt ) At = × 100, C0 V0
Qt =
(1) (2)
where C0 and Ct are the phosphate concentrations in an initial phosphate-containing synthetic wastewater and the filtrate in mg/L, respectively; V0 and Vt are the volumes of the initial synthetic wastewater and the filtrate in mL, respectively; W is the weight of the pellet in g. The distribution coefficient of phosphate ions, K d (L solution/kg pellet), is defined as the ratio of the phosphate concentration in the solid pellet (C1 ) to that in the solution phase (C2 ) and is calculated by the following equation: Kd =
C1 × 103 , C2
(3)
where C1 is the phosphate ions mass (mg) adsorbed in 1 g of the solid pellet (mg/g) and C2 is the phosphate ions mass (mg) left in 1 l of solution after adsorption (mg/L). 2.3. Adsorption kinetics experiment One hundred milliliter of synthetic phosphate-containing wastewater with a phosphate concentration of 39.09 ppm was placed in each of 11 plastic bottles of 250 mL with tight screw lids. One pellet of about 1.2 g with a diameter of 10 mm as described before was placed in each of the 11 bottles. Then, the 11 bottles were put simultaneously in an incubator (BS-1E, China) agitated at 150 rpm and controlled at a preset temperature with a variation of ± 1°C. After a certain time (0.5, 1, 2, 3, 4, 6, 8, 9, 10, 12 and 24 h), one of the bottles was taken out of the incubator and filtrated immediately on a Whatman GF-A membrane. The filtrate thus obtained was analysed for phosphate concentration. The adsorption kinetics experiments were carried out at 20°C, 30°C and 40°C.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2838
H. Wang et al.
2.4. Phosphate adsorption in a real wastewater One hundred milliliter real wastewater from a WWTP in Beijing as described previously was put in each of 5 plastic bottles of 250 mL with tight screw lids. Then, one, two, three, four and five weighed pellets (about 1.2 g with a diameter of 10 mm for each pellet) were placed in each of the 5 bottles. After that, the 5 bottles were put sequentially in an incubator (BS-1E, China) with a time interval of 0.5 h agitated at 150 rpm and controlled at 30°C. After 24 h, the 5 bottles were taken out of the incubator sequentially and filtrated immediately on a Whatman GFA membrane. The filtrate thus obtained was analysed for phosphate concentration.
adsorption temperature (Figure 2(b)). Thus, a higher temperature facilitates the phosphate adsorption on the pellets.
2.5. Analysis and characterization The phosphate concentrations in the solutions were analysed with inductively coupled plasma-atomic emission spectroscopy (SPECTRO ARCOS EOP, SPECTRO Analytical Instruments GmbH). After adsorption, the white substances coated on the surface of several pellets (Figure 1(a)) were scrapped off. The white powder substances thus collected were used in the characterization of adsorbed substances with XRD and SEM. The SEM samples were prepared by dispersing the ground white powder on one side of double-sided carbon conductive tape. The other side of the tape was supported on a pure copper block. SEM and energy dispersive spectrometer (EDS) observation was performed with Zeiss Ultra 55. XRD patterns of the samples were recorded with a Rigaku D/max-TTRIII X-ray diffractometer equipped with a Cu Kα radiation source (λ = 0.15405 nm). The sample patterns of XRD were analysed using the software MDI Jade5.0 with the aid of the JCPDF database.
3.3.1.
3.
Results and discussion
3.1. Effect of the contact time The distribution coefficients (K d ) of phosphate ions reached plateaus for 24 h at 20°C, 30°C and 40°C (Figure 2(a)). It implies that the adsorption equilibrium time for phosphate ions was 24 h at 20°C, 30°C and 40°C under the adsorption conditions. Based on these findings, an agitation time of 24 h was used in all thermodynamic experiments.
3.3. Thermodynamic studies of phosphate adsorption It is important to evaluate the most appropriate correlations for equilibrium data to optimize the design of an adsorption system. Langmuir, Freundlich, Dubinin– Kaganer–Radushkevich (DKR) isotherm models were used to describe the adsorption equilibrium. Experimental isotherm data were obtained at an equilibrium time of 24 h for different initial concentrations of phosphate ions in the solution. Langmuir adsorption model
The Langmuir adsorption isotherm is based on the assumption that all sites possess equal affinity for the adsorbate and monolayer of the adsorbate on the adsorbent.[17] The Langmuir isotherm is expressed by the following equation; Ceq Qeq
=
1 bQmax
Ceq , Qmax
(4)
where b (L/mg) and Qmax (mg/g) are the Langmuir constant and the theoretical monolayer saturation capacity of the pellet, respectively. High b values indicate high adsorption affinity. The equilibrium data were fitted to the Langmuir equation. The plot for this is shown in Figure 2(c). The linear plots of Ceq /Qt versus Ceq (Figure 2(c)) with regression correlation coefficients greater than 0.95 indicated the applicability of the Langmuir adsorption isotherm for the cases at three temperatures. The values of b and Qmax calculated from the intercept and slope of the linear plots are shown in Table 2. The values of Qmax were 2.46, 2.73 and 2.77 mg/g for the temperatures 20°C, 30°C and 40°C, respectively (Table 2). It implied that the saturation adsorption capacity of phosphate ions Qmax increased with an increase in the temperature. The values of b were 0.05, 0.06 and 0.11 for the temperatures 20°C, 30°C and 40°C, respectively (Table 2). It indicated that the bonding strength of phosphate ions to the pellet increased with an increase in the temperature. In order to predict the adsorption efficiency of the adsorption process, the dimensionless equilibrium parameter RL was determined by using the following equation [18]: RL =
3.2. Variation of Qeq with Ceq Qeq (mg/g) and Ceq (mg/L) are the amount of adsorbed phosphate ions per unit weight of adsorbent and unabsorbed phosphate ions concentration in solution at equilibrium, respectively. The variation of Qeq with Ceq is shown in Figure 2(b). Qeq increased with increasing Ceq . Ceq was proportional to the initial phosphate concentration C0 . Thus, Qeq increased with increasing initial phosphate concentration C0 . Additionally, Qeq increased with increasing
+
1 , 1+bC0
(5)
where C0 is the initial concentration of phosphate ions and b is the Langmuir constant. The values of RL < 1 represent favourable adsorption and the values greater than 1.0 represent unfavourable adsorption.[18] Also, RL values equal to 0 indicate irreversible adsorption, From this study, RL values for phosphate ions adsorption ranged from 0.056 to 0.52 for the three temperatures (Figure 4(b)). This is for initial phosphate concentrations ranging from 18.34 to 154.75 mg/L. Therefore, the adsorption process is favourable.
2839
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Environmental Technology
Figure 1. (a) Photos of the pellets before and after adsorption; (b) SEM image of the white powder (a) scrapped from the surface of the pellets after adsorption; (c) EDS of the red rectangle frame region indicated in (b); (d) elemental contents obtained according to the EDS analysis of (e) XRD of the white powder (a) scrapped from surface of the pellets after adsorption; and (f) variation of adsorbed phosphate amount and percentage with pellet concentration in a secondary wastewater from a WWTP in Beijing (Wastewater = 100 mL; initial phosphate concentration = 15.6 ppm; pH = 7.01; 30°C; 150 rpm; 24 h).
Moreover, RL values decreased with an increase in the initial phosphate concentration. A higher initial phosphate concentration was more favourable for the adsorption of phosphate on the pellet. So, the pellet can be used as a potential source for adsorption of phosphate ions from the phosphate-containing wastewater.
3.3.2. Freundlich adsorption model The empirical Freundlich equation is based on adsorption on a heterogeneous surface and is derived from the assumption that the adsorption sites are distributed exponentially with respect to the heat of adsorption.[17,18] The logarithmic linear form of the Freundlich equation is
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2840
H. Wang et al.
(a)
(b)
(c)
(d)
(e)
(f)
Figure 2. (a) Variation of distribution coefficient K d of phosphate with time (phosphate solution = 100 mL; pH = 7.0; pellet = 1.2 g; 150 rpm); (b) variation of equilibrium adsorption amount Qeq with equilibrium concentration of phosphate Ceq ; (c) Langmuir plot for phosphate adsorption on pellets; (d) variation of equilibrium parameter RL with initial concentration of phosphate; (e) Freundlich plot for phosphate adsorption on pellets and (f) DKR plot for phosphate adsorption on pellets (phosphate solution = 100 mL; pH = 7.0; pellet = 1.2 g; 150 rpm; 24 h).
given below: ln Qeq = ln KF +
1 ln Ceq , n
(6)
where Qeq (mg/g) and Ceq (mg/L) are the amount of adsorbed phosphate ions per unit weight of adsorbent
and unabsorbed phosphate ions concentration in solution at equilibrium, respectively, and K F and 1/n the Freundlich constants representing the adsorption capacity (mg1−n g−1 Ln ) and intensity (dimensionless) of the adsorbent, respectively. The values of 1/n less than 1 represent a favourable adsorption.[19]
Environmental Technology Table 2. Constants of adsorption isotherms of various models for phosphate on pellet (phosphate solution = 100 mL; pellet = 1.2 g; 150 rpm; 24 h). Temperature (°C)
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Model
Parameter
20
30
40
Freundlich K F (mg1−n g−1 Ln ) 0.4821 0.7046 0.9655 1/n 0.2010 0.2616 0.3125 R2 0.9561 0.9720 0.9786 Langmuir Qmax (mg/g) 2.46 2.73 2.77 b (L/mg) 0.050 0.061 0.107 R2 0.9548 0.9792 0.9844 DKR X m (mg/g) 4.84 5.21 5.23 B (mol2 /kJ2 ) 3 × 10−9 3 × 10−9 3 × 10−9 E ad (kJ/mol) 12.9 12.9 12.9 R2 0.9620 0.9597 0.9794
The equilibrium data were fitted to the Freundlich equation. The plot for this is shown in Figure 2(e). The linear plots of ln Qcq versus ln Ceq with regression correlation coefficients greater than 0.95 indicated the applicability of the Freundlich adsorption isotherm for the cases at the three temperatures. The values of K F and 1/n were calculated from the intercept and slope of the linear plots between ln Qeq and ln Ceq and are shown in Table 2. The values of K F were 0.48, 0.70 and 0.96 for the temperatures 20°C, 30°C and 40°C, respectively (Table 2). The values of 1/n were 0.20, 0.26 and 0.31 for the temperatures 20°C, 30°C and 40°C, respectively (Table 2), which indicated that the phosphate ions adsorption on the pellet was favourable for the three temperatures. 3.3.3. DKR model Langmuir and Freundlich isotherms do not give any idea about the adsorption mechanism. In order to distinguish between physical and chemical adsorption, the data were simulated with the DKR isotherm model.[20–22] The DKR equation is expressed as ln Qeq = ln Xm − Bε2 ,
(7)
where ε (Polanyi Potential) is [RT ln(1 + (1/Ceq )] (kJ/mol), Ceq are the unabsorbed phosphate ions concentration in solution at equilibrium (mol/L), X m is the DKR constant related to the maximum adsorption capacity of the pellet with respect to phosphate ions (mg/g) and B is the DKR constant related to the free energy of adsorption per mole of the adsorbate as it migrates to the surface of the adsorbent from an infinite distance in the solution (mol2 /kJ2 ). The plots of ln Qeq versus ε2 (Figure 2(f)) yielded straight lines with regression correlation coefficients greater than 0.98, which indicated the applicability of the DKR adsorption isotherm. The values of X m and B were calculated from the intercept and slope of the linear
2841
plots are shown in Table 2. The values of X m calculated from DKR adsorption isotherms were 4.84, 5.21 and 5.23 mg/g for the temperatures 20°C, 30°C and 40°C, respectively, which implied that the adsorption capacity (mg/g) of pellet for phosphate ions increased with an increase in the temperature. This finding was consistent with that obtained from the Langmuir adsorption isotherm (Table 2). The apparent adsorption energy (E ad ) from the DKR isotherm,[20–22] which was calculated using the following equation: −1 Ead = √ . 2B
(8)
The calculated values of E ad were − 12.9 kJ/mol for the temperatures of 20°C, 30°C and 40°C (Table 2). It is implied that the adsorption was an exothermic process. It is known that the magnitude of apparent adsorption energy E ad is useful for estimating the type of adsorption and if this value is below 8 kJ/mol the adsorption type can be explained by physical adsorption, between 8 and 16 kJ/mol the adsorption type can be explained by ion exchange, and over 16 kJ/mol the adsorption type can be explained by chemical adsorption.[20–22] The absolute values of E ad found in this study were 12.9 kJ/mol (Table 2), which implied that ion exchange was the main mechanism involved in the adsorption processes for the three temperatures. The effects of H2 reduction temperature and reduction time of the pellets on phosphate adsorption are presented in Figure 3. Both adsorbed phosphate amount and adsorbed phosphate percentage increased significantly with increasing reduction temperature and time (Figure 3). Thus, H2 reduction pre-treatment greatly enhanced the pellet adsorption capacity for phosphate. After adsorption, some white powder was left on the pellet surface (Figure 1(a)). The SEM and EDS of the white powder scrapped from the pellet surface are shown in Figure 1(b) and 1(c), respectively. The phosphate content in the pellet before adsorption was 0.34 wt% (Table 1). After adsorption, this value changed to 0.86 wt% (Figure 1(d)). Thus, at least part of phosphate ions were adsorbed on the pellet surface and existed in the white powder. The XRD of the white powder (Figure 1(a)) is shown in Figure 1(e). The main phases existing in the white powder were A-CaO (PDF no. 17-912), B-CaSiO3 (PDF no. 2-689), C,D,F,I,K-Fe3 O4 (PDF no. 75-1609), N-Al2 O3 (PDF no. 4-877), E,J,M-FeO (PDF no. 74-1886), G-Fe (PDF no. 87-722), H-Fe3 (PO4 )2 (PDF no. 48-1880) and L-Ca2 Fe9 O13 (PDF no. 83-1902). Therefore, most of the phosphates were adsorbed on the pellet surface probably in the form of Fe3 (PO4 )2 . Obviously, the scraped white powder contained both the adsorbed substances and the pellet surface layer. Thus, the ion exchange occurred on the pellet surface during the phosphate adsorption can be represented
2842
H. Wang et al.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
(a)
(b)
Figure 3. Effect of (a) H2 reduction temperature and (b) H2 reduction time of the pellets on phosphate adsorption (phosphate solution = 100 mL; pellet = 20 g; pellet diameter = 30 mm; initial phosphate concentration = 15.55 ppm; pH = 7.00; 30°C; 150 rpm; 4 h).
Based on the literature,[26] K sp = 9.91 × 10−16 , the −8 M. calculated soluble [PO3− 4 ] = 3.15 × 10 Thus, the formation of phosphate precipitate has the following decreasing order based on the thermodynamic calculations:
by the following equation:
(9) It means that Fe2+ on the pellet surface adsorbed phos3+ in the pellet precursor was phate (PO3− 4 ). When Fe reduced by H2 to a larger extent, more Fe2+ can be produced. This explained why a reduction pre-treatment of the pellet precursor by H2 greatly enhanced phosphate adsorption on the pellet as shown in Figure 3. Fe2+ on the pellet surface originated from Fe3 O4 (PDF no. 75-1609) and FeO (PDF no. 74-1886) on the pellet surface. Some studies indicated that most of the phosphates from the wastewaters was removed using steel slags in the form of calcium phosphates.[10,12] However, Drizo et al. found that adding Ca2+ to the wastewaters containing steel slags did not enhanced phosphate removal.[11] It implied that the formation of calcium phosphates probably did not completely govern the removal of phosphate. The solubility product constants (K sp ) of some common phosphate mineral precipitates are listed as follows: Ca3 (PO4 )2 → 3Ca2+ + 2PO3− 4 .
(10) −29
Based on the literature,[23] K sp = 2.07 × 10 −6 M calculated soluble [PO3− 4 ] = 1.44 × 10 Ca5 (OH)(PO4 )3 → 5Ca
2+
−
+ OH +
3PO3− 4 .
For the composite pellet used in this work, the atomic percentages of Fe and Ca were 15.06% and 10.47%. It implied that the Fe atom number was larger than that of Ca in the composite pellet. Thus, FePO4 · 2H2 O and Fe3 (PO4 )2 · 8H2 O were more readily to be formed than Ca3 (PO4 )2 or Ca5 (OH)(PO4 )3 for this kind of composite pellet. In addition, the number of Fe2+ from the composite pellet was likely much larger than that of Fe3+ due to H2 reduction pre-treatment based on XRD characterization (Figure 2(e)). This explained why only Fe3 (PO4 )2 phase from the pellet surface was detected by XRD (Figure 2 (e)). It is noteworthy that the Fe3 (PO4 )2 precipitate can be formed in the neutral solution,[25] although most of the − phosphates probably exist in the form of HPO2− 4 or H2 PO4 at this pH.
3.4. (11)
(12)
Based on the literature,[25] K sp = 1 × 10−35.77 , the −8 M calculated soluble [PO3− 4 ] = 5.50 × 10 FePO4 · 2H2 O → Fe3+ + PO3− 4 + 2H2 O.
Ca5 (OH)(PO4 )3 .
, the
Based on the literature,[24] K sp = 1 × 10−53.28 , the −5 M calculated soluble [PO3− 4 ] = 1.32 × 10 Fe3 (PO4 )2 · 8H2 O → 3Fe2+ + 2PO3− 4 + 8H2 O.
FePO4 · 2H2 O ≈ Fe3 (PO4 )2 · 8H2 O > Ca3 (PO4 )2 >
(13)
Kinetic studies of phosphate ions adsorption
Kinetic adsorption experiments were carried out in this work to investigate the adsorption rate and adsorption mechanism of phosphate ions. 3.4.1. Variation of Qt with t The variation of phosphate adsorption amount Qt with adsorption time t is shown in Figure 4(a). Qt increased significantly with t before 9 h. After that, Qt increased slowly with t between 9 and 24 h.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Environmental Technology (a)
(b)
(c)
(d)
(e)
(f)
2843
Figure 4. (a) Variation of Qt with t; (b) pseudo-second-order kinetics plot for phosphate adsorption on the pellets; (c) pseudo-first-order kinetics plot for phosphate adsorption on the pellets; (d) Arrhenius equation plot for pseudo-first-order adsorption of phosphate on the pellet; (e) external diffusion model for phosphate adsorption on the pellets and (f) interparticle diffusion model for phosphate adsorption on the pellets (phosphate solution = 100 mL; initial phosphate concentration = 39.1 ppm; pH = 7.0; pellet = 1.2 g; 150 rpm).
3.4.2. Pseudo-first and pseudo-second-order models The pseudo-first-order rate expression based on the solid capacity is generally expressed as follows [27]: dQ = k1 (Q − Qeq ), dt Qt = −k1 t. ln 1 − Qeq
The pseudo-second-order equation is also based on the adsorption capacity of the solid phase. It is expressed as follows [27]: dQ = k2 (Q − Qeq )2 , dt
(14)
1 = 1 + Qeq k2 t, 1 − (Qt /Qeq )
(15)
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2844
H. Wang et al.
where Qeq and Qt (both in mg/g) are the amount of phosphate ions adsorbed per unit mass of the resin at equilibrium and time t (h), respectively, and k 1 (1/h) and k2 (g/mg h) are the rate constants of the pseudo-first-order adsorption and pseudo-second-order adsorption, respectively. The kinetic experimental data were simulated using pseudo-second-order and pseudo-first-order models and the simulating curves are shown in Figure 4 (b) and 4(c), respectively. It is observed that for the pseudo-secondorder model there were not linear relationships between 1/(1 − Qt /Qeq ) and adsorption time t at the three temperatures (Figure 4(b)). However, such linear relationships occurred for the pseudo-first-order model (Figure 4(c)). The linear plots of ln(1 − (Qt /Qeq )) versus time t with regression correlation coefficients greater than 0.98 indicated the applicability of pseudo-first-order model at the three temperatures. Thus, it indicated that the adsorption kinetics data are better represented by the pseudo-firstorder model. The values for k1 for the three temperatures are calculated from the curve slopes shown in Figure 4(c). The adsorption rate constant k1 were 0.1947, 0.2166 and 0.2351 h−1 for temperatures 20°C, 30°C and 40°C, respectively (Figure 4(c)). The linear plot of lnk1 versus (103 /T) (K−1 ) (Figure 4(d)) with regression correlation coefficient 0.9968 indicated the applicability of the following Arrhenius equation: ln k1 = ln A −
Ea,1 , RT
(16)
where k 1 is the rate constant of pseudo-first-order reaction (1/h), A is the pre-exponential factor, E a,1 is the apparent adsorption activation energy (J/mol), R is the gas constant (8.314 J K−1 mol−1 ) and T is the temperature (K). The apparent adsorption activation energy E a,1 for the pseudo-first-order model was calculated from the slope of the linear plot of ln k 1 versus (103 /T) and found to be 7.2 kJ/mol (Figure 4(d)). 3.4.3. External diffusion model At early times of contact (e.g. between 0 and 10 min of contact) the system could be simplified by assuming that the concentration of phosphate ions on the pellet surface tended towards zero (pure external diffusion) and the internal diffusion was negligible. Thus, Ficks’ laws may be applied to describe the mass-transfer rate of external diffusion. The diffusion flux J of Ficks’ law could be expressed as follows [28]: J = kf Cb = −
dnb VdCb =− , Adt Adt
(17)
where Cb is the bulk concentration of phosphate ions (mol/m3 ), kf is the external mass-transfer coefficient of phosphate ions (m/h), A is the pellet surface (m2 ), V is the
volume of bulk solution (m3 ) and t is the diffusion time (h)
t
− 0
A kf dt = V ln
Ct
d ln Cb , C0
A C0 = kf t, Ct V
(18)
where C0 and Ct are the bulk concentrations of phosphate ions (mol/m3 ) at times of 0 and t, respectively. If external diffusion was a rate-controlling step at the beginning of phosphate adsorption process, there was a linear relationship between ln C0 /Ct and t as described by Equation (18) and the line pass through the original at a certain temperature. Such linear relationships at the time range between 0 and 8 h existed for the 3 temperatures (Figure 4(e)). However, after 8 h such linear relationships did not exist and the lines also did not pass through the original (Figure 4(e)). Thus external diffusion was the rate-controlling step only before 8 h. 3.4.4. Interparticle diffusion model Besides for phosphate adsorption on the external surface of adsorbent pellet, there is also a possibility of transporting adsorbate phosphate from the solution to the internal pores of the adsorbent pellet (intraparticle diffusion). The most commonly used technique for identifying the mechanism involved in the internal pore diffusion process is by using the intraparticle diffusion model [29–31] given by: Qt = kid t(1/2) + c,
(19)
where Qt (mg/g) is the amount of Pt ions adsorbed per unit mass of the resin at time t (h), kid is the intraparticle diffusion rate constant (mg g−1 h−1/2 ), and c is the intercept. For the intercept c, McKay et al. [30] have indicated that ‘extrapolation of the linear portion of the plot back to the y-axis provides intercepts which are proportional to the extent of the boundary layer thickness, that is, the larger the intercept the greater the boundary layer effect’. Plots of Qt versus t1/2 for the three temperatures are shown in Figure 4(f). From the figure, it is observed that there are two linear regions. One was between 0 and 8 h and the other was between 8 and 12 h. The intercepts of first linear portion were negative. Similar phenomena were also observed by McKay et al. [30,31]. They believed that the boundary layer thickness retarded intraparticle diffusion and led to negative intercepts.[30,31] The intercepts of second linear portion were positive, which indicated that intraparticle diffusion was a rate-controlling step between 8 and 12 h. In summary, the external diffusion was the main rate-controlling step of phosphate adsorption before 8 h and internal diffusion became the rate-controlling step between 8 and 12 h.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
Environmental Technology 3.5. Adsorption of phosphate from real wastewater In order to investigate the adsorption performance of phosphate from a real wastewater on the pellets, the effect of pellet concentration in the wastewater on phosphate adsorption was studied and the results are presented in Figure 1(f). The adsorbed phosphate amount increased with increasing pellet concentration and reached a plateau value of 1.55 mg/g at a pellet concentration of 50 (g/L) (Figure 1(f)). The adsorbed phosphate amount was 0.95 (mg/g) for the real wastewater with an initial phosphate concentration of 15.6 ppm and a pellet concentration of 12 (g/L) at 30°C (Figure 1(f)). The adsorbed phosphate amount Qeq at equilibrium was 1.44 (mg/g) for the synthetic phosphate-containing solution with an initial phosphate concentration 18.3 ppm and a pellet concentration of 12 (g/L) at 30°C (Figure 2(b)). Thus, adsorbed phosphate amounts were similar for both synthetic and real phosphate-containing wastewaters. The adsorbed phosphate percentage also increased with increasing pellet concentration and reached 99.2% at pellet concentration of 64 (g/L) (Figure 1(f)). Thus, the pellets prepared in this work have the potential to be used as adsorption media filled in an adsorption column to remove phosphate from real wastewaters. 4.
Conclusions
The phosphate adsorption on the pellets made of the composite composed of reduced steel slag and concentrated iron ore by H2 was studied. It was found that the adsorption equilibrium time was about 24 h. Phosphate adsorption increased with increasing temperature. The phosphate adsorption on the pellets followed the three models of Freundlich, Langmuir and DKR. The maximum phosphate adsorption capacity Qmax of the pellets were 2.46, 2.74 and 2.77 mg/g for the temperatures 20°C, 30°C and 40°C, respectively, based on the Langmuir model. The apparent adsorption energy (E ad ) was − 12.9 kJ/mol for the 3 temperatures based on the DKR model. It implied that ion exchange was the mechanism involved in the adsorption processes. The adsorbed phosphate existed on the pellet surface mainly in the form of Fe3 (PO4 )2 . The adsorption kinetics data are better represented by the pseudo-firstorder model. The external diffusion was the main ratecontrolling step of phosphate adsorption before 8 h and internal diffusion became the rate-controlling step between 8 and 12 h. The adsorbed phosphate amounts were similar for both real and synthetic wastewaters under similar adsorption conditions. The percentage of adsorbed phosphate for a real wastewater increased with increasing pellet concentration and reached 99.2% at a pellet concentration of 64 (g/L). Some specific phosphate adsorption mechanisms for the composite pellets were revealed and the pellets showed the potential to efficiently adsorb phosphate from a huge amount of real wastewaters in an industrial scale.
2845
Disclosure statement No potential conflict of interest was reported by the authors.
Funding This work was supported by the grants from the National Natural Science Foundation of China [grant number 50874011] and the Programs for Innovative Research Team of undergraduate students at University of Science and Technology Beijing [grant numbers 14020118 and 14020367].
References [1] Ma J, Li H. Preliminary discussion on eutrophication status of lakes, reservoirs and rivers in China and overseas. Resour Environ Yangtze Basin. 2002;11:575–577 (in Chinese). [2] Barca C, Meyer D, Liira M, et al. Steel slag filters to upgrade phosphorus removal in small wastewater treatment plants: Removal mechanisms and performance. Ecol Eng. 2014;68:214–222. [3] Xiong J, He Z, Mahmood Q, Liu D, Yang X, Islam E. Phosphate removal from solution using steel slag through magnetic separation. J Hazard Mater. 2008;152:211–215 (in Chinese). [4] Tanada S, Kabayama M, Kawasaki N, et al. Removal of phosphate by aluminum oxide hydroxide. J Colloid Interf Sci. 2003;257:135–140. [5] Lee SH, Vigneswaran S, Moon H. Adsorption of phosphorus in saturated slag media columns. Sep Purif Technol. 1997;12:109–118. [6] Kadlec RH, Knight RL. Treatment wetlands. Boca Raton (FL): Lewis, CRC Press; 1996; p. 893. [7] Brix H, Arias CA, Del Bubba M. Media selection for sustainable phosphorus removal in subsurface flow constructed wetlands. Water Sci Technol. 2001;44:47–54. [8] Vohla C, Kõiv M, Bavor HJ, Chazarenc F, Mander Ü. Filter materials for phosphorus removal from wastewater in treatment wetlands– a review. Ecol Eng. 2011;37:70–89. [9] Zeng L, Li X, Liu J. Adsorptive removal of phosphate from aqueous solutions using iron oxide tailings. Water Res. 2004;38:1318–1326. [10] Jha VK, Kameshima Y, Nakajima A, Okada K. Hazardous ions uptake behavior of thermally activated steel-making slag. J Hazard Mater. 2004;B114:139–144. [11] Drizo A, Forget C, Chapuis RP, Comeau Y. Phosphorus removal by electricarc furnace steel slag and serpentinite. Water Res. 2006;40:1547–1554. [12] Kim EH, Lee DW, Hwang HK, Yim S. Recovery of phosphates from wastewater using converter slag: kinetics analysis of a completely mixed phosphorus crystallization process. Chemosphere. 2006;63:192–201. [13] Chazarenc F, Brisson J, Comeau Y. Slag columns for upgrading phosphorus removal from constructed wetland effluents. Water Sci Technol. 2007;56:109–115. [14] Bowden LI, Jarvis AP, Younger PL, Johnson KL. Phosphorus removal from wastewaters using basic oxygen steel slag. Environ Sci Technol. 2009;43:2476–2481. [15] Claveau-Mallet D, Wallace S, Comeau Y. Model of phosphorus precipitation and crystal formation in electric arc furnace steel slag filters. Environ Sci Technol. 2012;46: 1465–1470. [16] Johansson-Westholm L. Substrates for phosphorus removal –potential benefits for on-site wastewater treatment? Water Res. 2006;40:23–36.
Downloaded by [University of Nebraska, Lincoln] at 21:27 26 September 2015
2846
H. Wang et al.
[17] Saeed MM, Bajwa SZ, Ansari MS. Investigation of the removal of lead by adsorption onto 1-(2-thiazolylazo)-2naphthol (TAN) imbedded polyurethane foam from aqueous solution. J Chin Chem Soc. 2007;54:173–183. [18] Basha S, Murthy ZVP. Kinetic and equilibrium models for biosorption of Cr(VI) on chemically modified seaweed, Cystoseira indica. Process Biochem. 2007;42:1521– 1529. [19] Yamada H, Kayama M, Saito K, Hara M. A fundamental research on phosphate removal by using slag. Water Res. 1986;20(5):547–557. [20] Lin SH, Juang RS. Heavy metal removal from water by sorption using surfactant-modified montmorillonite. J Hazard Mater. 2002;92:315–326. [21] Wang CC, Juang LC, Lee CK, Hsu TC, Lee JF, Chao HP. Effects of exchanged surfactant cations on the pore structure and adsorption characteristics of montmorillonite. J Colloid Interf Sci. 2004;280:27–35. [22] Krishna BS, Murty DSR, Prakash BSJ. Thermodynamics of chromium(VI) anionic species sorption onto surfactantmodified montmorillonite clay. J Colloid Interf Sci. 2000;229:230–236. [23] Speight JG. Lange’s handbook of chemistry. 16th ed. New York: McGraw-Hill; 2005; p 1.333.
[24] Zhu Y, Zhang X, Chen Y, et al. A comparative study on the dissolution and solubility of hydroxylapatite and fluorapatite at 25°C and 45°C. Chem Geol. 2009;268:89–96. [25] Azam HM, Finneran KT. Fe(III) reduction-mediated phosphate removal as vivianite (Fe3 (PO4 )28 H2 O) in septic system wastewater. Chemosphere. 2014;97:1–9. [26] Speight JG. Lange’s handbook of chemistry. 16th ed. New York: McGraw-Hill; 2005; p 1.335. [27] Naiya TK, Bhattacharya AK, Das SK. Adsorption of Cd(II) and Pb(II) from aqueous solutions on activated alumina. J Colloid Interf Sci. 2009;333:14–26. [28] Shen S, Pan T, Liu X, et al. Adsorption of Rh(III) complexes from chloride solutions obtained by leaching chlorinated spent automotive catalysts on ion-exchange resin Diaion WA21J. J Hazard Mater. 2010;179:104–112. [29] Wu FC, Tseng RL, Juang RS. Initial behavior of intraparticle diffusion model used in the description of adsorption kinetics. Chem Eng J. 2009;153:1–8. [30] McKay G, Otterburn MS, Sweeney AG. The removal of color from effluent using various adsorbents III silica: rate processes. Water Res. 1980;14:15–20. [31] McKay G. The adsorption of dyestuffs from aqueous solutions using activated carbon. III. Intraparticle diffusion process. J Chem Technol Biotechnol A. 1983;33:196–204.
