Basic Oxygen Furnace steel slag aggregates for phosphorus treatment. Evaluation of its potential use as a substrate in constructed wetlands

Water Research 89 (2016) 355e365

Contents lists available at ScienceDirect

Water Research journal homepage: www.elsevier.com/locate/watres

Basic Oxygen Furnace steel slag aggregates for phosphorus treatment. Evaluation of its potential use as a substrate in constructed wetlands
enz de Miera c, Gemma Ansola d Ivan Blanco a, *, Pascal Molle b, Luis E. Sa
n, Calle La Serna, nº 56, 24071, Leo
n, Spain Instituto de Medio ambiente, Recursos Naturales y Biodiversidad, Universidad de Leo Wastewater Treatment Team, Freshwater System, Ecology and Pollution Research Unit, Irstea, 5 rue de la Doua, CS70077, 69626, Villeurbanne Cedex, France c
n, Campus de Vegazana s/n, 24071, Leo
n, Spain Departamento de Biología Molecular, Universidad de Leo d
n Ambiental, Universidad de Leo
n, Campus de Vegazana s/n, CP: 24071, Leo
n, Spain Departamento de Biodiversidad y Gestio a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 July 2015 Received in revised form 4 November 2015 Accepted 27 November 2015 Available online 2 December 2015

Basic Oxygen Furnace (BOF) steel slag aggregates from NW Spain were tested in batch and column experiments to evaluate its potential use as a substrate in constructed wetlands (CWs). The objectives of this study were to identify the main P removal mechanisms of BOF steel slag and determine its P removal capacity. Also, the results were used to discuss the suitability of this material as a substrate to be used in CWs. Batch experiments with BOF slag aggregates and increasing initial phosphate concentrations showed phosphate removal efficiencies between 84 and 99% and phosphate removal capacities from 0.12 to 8.78 mg P/g slag. A continuous flow column experiment filled with BOF slag aggregates receiving an influent synthetic solution of 15 mg P/L during 213 days showed a removal efficiency greater than 99% and a phosphate removal capacity of 3.1 mg P/g slag. In both experiments the main P removal mechanism was found to be calcium phosphate precipitation which depends on Ca2þ and OH release from the BOF steel slag after dissolution of Ca(OH)2 in water. P saturation of slag was reached within the upper sections of the column which showed phosphate removal capacities between 1.7 and 2.5 mg P/g slag. Once Ca(OH)2 was completely dissolved in these column sections, removal efficiencies declined gradually from 99% until reaching stable outlet concentrations with P removal efficiencies around 7% which depended on influent Ca2þ for limited continuous calcium phosphate precipitation. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Phosphorus Steel slag Batch experiment Column experiment Constructed wetland

1. Introduction Phosphorus (P) removal in constructed wetlands (CW) occurs through a combination of several processes: peat/soil accretion, plant uptake, microbial growth, substrate adsorption and chemical precipitation (Vymazal, 2007). Among these, adsorption and chemical precipitation play the largest role, particularly in saturated subsurface flow CWs where the contact between wastewater and substrate is enhanced (Babatunde et al., 2009; Vymazal, 2005). Sand, gravel and soils have been traditionally used as filter materials in CWs with limited results in terms of P removal (Vohla et al.,

* Corresponding author. E-mail addresses: [email protected] (I. Blanco), [email protected]
enz de Miera), gemma.ansola@ (P. Molle), [email protected] (L.E. Sa unileon.es (G. Ansola). http://dx.doi.org/10.1016/j.watres.2015.11.064 0043-1354/© 2015 Elsevier Ltd. All rights reserved.

2011). For more than two decades research has been directed to selection of alternative materials (Drizo et al., 1999; Yamada et al., 1986). Long-term P removal depends on the physico-chemical characteristics of the material such as Al, Fe and Ca content; specific surface area; porosity; particle size distribution; and hydraulic conductivity because they all affect the P retention capacity (Barca et al., 2012; Drizo et al., 2002; Reddy et al., 1980); whereas its suitability as a substrate is site specific, and depends on its recycling potential, cost and local availability (Drizo et al., 1999). A large number of studies have been conducted on a wide variety of materials for potential use as substrates in CWs or in other small-scale filter systems, including natural materials (minerals, rocks, soils and marine sediments), industrial by-products (from the steel, mining and power plant industry), and man-made products (lightweight aggregates) (Johansson, 2006).

356

I. Blanco et al. / Water Research 89 (2016) 355e365

Among them, slag materials from the steel industry demonstrate promising properties with regard to P sorption capacity (Barca, 2013; Bowden et al., 2009; Drizo et al., 2002; Johansson, 1999; Shilton et al., 2006). In Europe there are three main types of steel industry slag, each named after the process involved in their production: Blast Furnace (BF) iron slag, Basic Oxygen Furnace (BOF) steel slag and Electric Arc Furnace (EAF) steel slag. BF slags are the by-products of iron production, while BOF and EAF slags are by-products of steel production (Proctor et al., 2000). Many studies have been conducted to investigate the potential of slag materials to remove P from wastewater (Bowden et al., 2009; Claveau-Mallet et al., 2013; Drizo et al., 2002, 2006; Xue et al., 2009). Variation of results may be attributed to differences in experimental conditions, physical and chemical compositions of materials, which may vary depending on the ore resource, the process temperature, the factory or even the batch test conditions (Drizo et al., 2002; Xie et al., 2012). However, when comparing steel production slags under the same experimental procedures, BOF slags show greater phosphate removal capacities than EAF slags (Barca et al., 2012). Batch experiments are commonly used to compare P sorption efficiency of different materials due to their simplicity and short time requirements. However, batch experimental parameters are arbitrarily set, and the results are often non-comparable and misleading (Cucarella and Renman, 2009). Moreover, the use of isotherm equations can lead to biased and unrealistic estimates of the adsorption parameters if the experimental data obtained from the batch experiments is intended to be used for the estimation of P retention capacities in field scale applications. As a result, while batch experiments can be useful to compare the P retention capacities of different materials, such tests should be coupled with longer-term column experiments, with similar hydraulic conditions to those present in CWs, for the estimation of longevity, P removal efficiencies and retention capacities of steel slag filters to be used in CWs (Drizo et al., 2002). In this study a locally available BOF slag in NW Spain was submitted to batch and column experiments in order to: (i) determine its P removal capacity, (ii) identify the main P removal mechanisms, and (iii) evaluate its suitability to be used in constructed wetlands. 2. Materials and methods 2.1. BOF slag characterization The BOF slag tested in this study was collected from the Arce~ a de Abajo, Gijo
n, NW lorMittal steel factory LD-II located in Verin Spain. Particle-size distributions were determined using drysieving techniques (European Norm EN, 933-1:1997) to calculate d10, d60 (mesh diameter allowing, respectively, 10 or 60% of the material mass to pass through) and uniformity coefficient (UC ¼ d60/d10). Porosity was determined from the amount of water needed to saturate a known volume of component, and density was measured by the volume of water displaced by a known mass of medium. The specific area of particles was evaluated from the size distribution, assuming grains to be spherical. A representative BOF slag sample (Spanish norm UNE-EN 9322:1999) was crushed for its chemical and mineralogical characterization. Inductively Coupled Plasma Atomic Emission Spectroscopy was used to determine the content of major elements (Ca, Fe, Mg, Al and P) using an Optima 8300 PerkineElmer spectrometer. The following sample digestion procedure was selected after confirming a complete digestion of the sample: 0.5 g of sample were acid digested with a mixture of 10 mL of HNO3 65% and 3 mL of HCl 35%, with a reflux system under atmospheric pressure. The following digestion steps were used: temperature increase to 45  C in 30 min, stable temperature at 45  C for 1 min, temperature increase to 65  C

in 25 min, stable temperature at 65  C for 5 min, temperature increase to 100  C in 15 min, and stable temperature at 100  C for 120 min. X-ray diffraction (XRD) was used to study the mineralogical composition of the representative BOF slag sample. XRD data were recorded at ambient temperature on a Bruker D8 Advance powder diffractometer, in the angular range 5e70 (2q), collected with a step of 0.02 and a step time of 8 s. Before and after column experiments larger BOF slag particles were gold coated (Balzers SCD 004) prior to observation using a JSM-6480 LV JEOL scanning electron microscope (SEM). The surface composition of the BOF slag particles before and after column experiments was assessed on carbon coated BOF slag particles with an EDX microanalysis instrument (Energy Dispersive X-ray System) coupled to the SEM. 2.2. Kinetic batch experiments Initial PO4-P concentration of 30 mg/L similar to concentrations present in dairy parlors were prepared in 500 mL Erlenmeyer flasks using KH2PO4 in distilled water, and adjusted to pH 7 using NaOH before each test. Then, 20 g of slag were added to each flask, leading to a slag to solution mass ratio of 1:25 (adapted from Nair et al. (1984)). Flasks were covered and kinetic batch experiments were carried out with three replicates on an orbital shaker at 120 rpm under controlled temperature conditions (25  C). A slag granular size of <10 mm was used to perform the batch experiments since it has been suggested that sieving of unsorted fines would not be necessary to increase reactivity and efficacy in full-scale treatment systems (Bowden et al., 2009). Solutions were measured for pH, P and Ca2þ concentrations at 0.5, 1, 2, 3.5, 5, 10, 24 and 48 h. Potential Ca2þ release capacity was calculated as:

Qt ¼

Cat Vt M

(1)

where Qt is the potential Ca2þ release capacity at time t (mg/g), Vt is the volume of the solution at time t (L), M is the mass of slag (g), and Cat is the Ca2þ concentration of the solution at time t (mg/L). Then, the potential Ca2þ release capacity was plotted according to the pseudo-first order kinetic equation of Lagergren (1898):

lnðQ e  Q t Þ ¼ ln Q e  k1 t

(2)

where Qe is the potential Ca2þ release capacity at equilibrium (mg/ g), Qt is the potential Ca2þ release capacity at time t (mg/g), t is the time (h) and k1 is the rate constant of pseudo-first order release (1/ h). If pseudo-first order kinetics are applicable, this suggests that one of the reactants is present in great excess over the other reactants in the reaction mixture. Qe must be known to exploit equation (2) with experimental data. Qe was considered an adjustable parameter whose value was estimated by trial and error. The experimental potential Ca2þ release capacities observed after 48 h were used as initial Qe-trial in order to calculate Qe from the intercept of the plot of ln(Qe-trial  Qt) against t. Then the Qe-trial was adjusted until the difference between Qe-trial and Qe was less than 0.1% of the value of Qe-trial. The relationship between Ca2þ and OH was evaluated through linear regression analysis. 2.3. Isotherm batch experiments Isotherm batch experiments tested a broad range of initial PO4-P concentrations (5, 10, 25, 50, 100, 200, 300 and 400 mg/L). Initial P

I. Blanco et al. / Water Research 89 (2016) 355e365

concentrations (5e50 mg/L) simulate values that may be observed in real environments such as agricultural field drainage, sewage treatment works, and effluent from dairy parlor. Higher concentrations (100e400 mg/L) were useful to determine the maximum phosphate removal capacities of the BOF slag. Solutions with different initial PO4-P concentrations were prepared in 250 mL Erlenmeyer flasks using KH2PO4 in distilled water, and were adjusted to pH 7 using NaOH before each test. Then, 10 g of slag was added to each flask, leading to the slag to solution mass ratio of 1:25. Flasks were covered and tests were carried out with three replicates on an orbital shaker at 120 rpm under controlled temperature conditions (25  C). After 48 h, solutions were measured for pH, P and Ca2þ concentrations. Phosphate removal capacities (PRCs, mg P/g slag) were calculated as:

PRC ¼

ðPin  PÞV M

(3)

357

filled with 5 cm of inert gravel (0.45 porosity) material at the bottom, 20 cm of BOF slag (0.49 porosity), and 5 cm of gravel at the top; received a top to bottom flow and were maintained under hydraulically saturated conditions at room temperature (20 ± 3  C). A flow rate of 3.4 mL/min that resulted in a P loading rate of 73.5 mg P/day for the Experimental column was set up to achieve a theoretical Hydraulic Retention Time (HRT) of 8 h. The selection of a Hydraulic Retention Time (HRT) is crucial for the long-term operation of a steel slag filter. A 8 h HRT is long enough to allow the dissolution of CaO or Ca(OH)2 from the slag and kinetically raise pH to allow phosphorus precipitation and crystallization (ClaveauMallet et al., 2012). It is also greater than other HRTs that result in loose crystal organization and poorer retention capacities (Claveau-Mallet et al., 2012), and low enough to minimize the precipitation of carbonates that takes place at longer HRTs (Liira et al., 2009). Phosphorus removal efficiencies (P removal efficiencies, %) were calculated as:

where V is the volume of the solution (L), M is the mass of slag (g), Pin is the initial PO4-P concentration (mg P/L) and P is the residual PO4-P concentration of the solution after 48 h (mg P/L).

P removal efficiency ¼

2.4. Column experiments

where Pin is the inlet PO4-P concentration (mg P/L) and Pout is the outlet PO4-P concentration of the solution at the specific sampling point (mg P/L).

Two column experiments were designed to evaluate: (i) Ca2þ release from BOF slag, and (ii) P removal from a synthetic solution. The Ca2þ release column experiment (Control column) was designed as a short experiment which exclusive purpose was to understand the mechanisms for Ca2þ release and pH increase in the absence of P which might affect Ca2þ and OH concentrations in solution due to calcium phosphate precipitation. It was fed with tap water and monitored at the inlet and outlet for Ca2þ concentration during the first 13 days of experimentation, whereas pH and EC were monitored during the whole experiment (a total of 67 days). The results were used to evaluate the relationship between Ca2þ and OH through linear regression analysis. The P removal column experiment (Experimental column) was fed with a synthetic solution prepared using KH2PO4 in tap water with a concentration of 15 mg P/L, similar to those likely to be encountered in real future applications; and was adjusted to a conductivity of 1000 ± 100 mS/cm with NaCl, which is comparable with the effluent from a first vertical stage in constructed wetlands (Molle, 2003). The Experimental column was monitored during 213 days for pH, EC, and P concentrations at three vertical sampling points (at a distance of 5, 10, and 15 cm from the top of the BOF slag volume), the inlet and the outlet of the column. Both columns (Fig. 1) (diameter ¼ 12 cm; height ¼ 30 cm) were

Fig. 1. Schematic layout of the column experiments.

ðPin  Pout Þ  100 Pin

(4)

2.5. Geochemical modeling Geochemical modeling of the speciation of the initial solutions in the batch isotherm and column experiments were performed to calculate the saturation indexes (SI) of the minerals considered using the geochemical software PHREEQC v.3 (Parkhurst and Appelo, 2013) with the wateq4f database modified taking into account the Ksp of the different minerals according to Stumm and Morgan (1996) at the temperature at which the tests were conducted. Batch isotherm experiments and column inlet water was measured for OxidationeReduction Potential (ORP). In batch isotherm experiments, Kþ and Naþ concentrations were calculated from the initial added solution concentrations and Naþ was used in the modeling program to charge balance the simulations. pH, Total alkalinity, major cations (Ca2þ, Mg2þ, Naþ, Kþ, Total Fe, Al3þ) and 2  major anions (F, Cl, SO2 4 , PO4 , NO3 ) were measured from the Experimental column inlet water. 2.6. Water analyses In the batch and column experiments pH and EC were measured with portable probes (WTW) calibrated periodically according to manufacturer's procedures. Samples were immediately centrifuged after collection (5 min, 1000 rpm), and the supernatant solution was acidified with HCl to pH 2 to avoid calcium carbonate precipitation, and stored at 4  C until analysis. Dissolved P and Ca concentrations were determined through ICP-AES analysis using an Optima 8300 PerkineElmer spectrometer. For geochemical modeling purposes ORP from initial solutions was measured with a portable probe (WTW). Total alkalinity was measured according to method 2320B (APHA-AWWA-WEF, 2005). Major cation concentrations were determined through ICP-AES analysis using an Optima 8300 PerkineElmer spectrometer, whereas major anion concentrations were determined through ionic chromatography using a Metrohm 881 Compact IC pro with conductivity detection and a Metrosep A Supp 7 column with Na2CO3 3.6 mM as mobile phase at 45  C.

358

I. Blanco et al. / Water Research 89 (2016) 355e365

3. Results and discussion 3.1. BOF slag physical, mineralogical and chemical characterization The particle size distribution of the BOF slag (Table 1) resulted in d10 and d60 values consistent with those reported by Proctor et al. (2000) from 17 BOF slag composite samples. The high UC value informs of a wide range of particle diameters ranging from fine to coarse aggregates (from 0.1 to 10 mm). The XRD pattern of the BOF slag was very complex, with several overlapping peaks resulting from the many minerals present in the sample (Fig. 2). The mineral phases identified were defined as major or minor phases depending on the intensity of the peaks. Some of the overlapping mineral phases that could not be determined with certainty were identified as probable. A very strong diffraction peak appears at 2q ¼ 18 , which represents portlandite (Ca(OH)2), the major mineral phase of the sample. The presence of this mineral is characteristic of weathered BOF slag after the hydration of lime (CaO) in the presence of moisture (Mahieux et al., 2009; Yildirim and Prezzi, 2011). Larnite (b-Ca2SiO4) and corundum chromian (Al1.54O3Cr0.46) are minor phases observed with a second and third strongest diffraction peak. Steel slags can be regarded as weak Portland cement clinkers due to larnite cementitious properties (Shen et al., 2009; Tsakiridis et al., 2008). Additional minor mineral phases include wustite (FeO), merwinite (Ca3Mg(SiO4)2), quartz (a-SiO2) and calcite (CaCO3), as a result of Ca(OH)2 carbonation with atmospheric CO2 (Morone et al., 2014; Ruiz-Agudo et al., 2013). Iron oxides are frequent in the sample. In addition to FeO, the XRD pattern informs of the probable presence of hematite (Fe2O3), brownmillerite (Ca2FeAlO5), magnetite (Fe3O4) and magnesioferrite (MgFe2þ3O4). The presence of Fe3O4 and MgFe2þ3O4 would explain the magnetic behavior of the BOF slag (Belhadj et al., 2012). The BOF slag surface was examined by SEM and EDX before the experiments. The BOF slag surface was characterized by cloudy textures on the surface of the BOF slag particles which can be attributed to carbonation of Ca(OH)2 with atmospheric CO2 (Fig. 3a and b), similar to those observed by Yildirim and Prezzi (2009). The EDX spectrum (Fig. 3c) identifies O and Ca as the main elements present on the surface of the BOF slag (Table 1), which is consistent with the presence of Ca(OH)2 as the main mineral phase identified in the XRD analysis. Although likely, the presence of CaCO3 on the surface of the BOF slag cannot be confirmed through the EDX analysis. Carbon was used to coat the sample in order to create the conductive layer, so this element was not considered in the EDX analysis. The chemical analysis identifies Ca and Fe as the main components of the slag, which add up to 60% of sample mass (Table 1), confirming the presence of Fe mineral phases within the BOF slag. Comparison between the surface EDX analysis and the ICP-AES chemical analysis on the crushed BOF slag sample indicate that

Table 1 Physical properties and chemical composition of the BOF slag. Physical properties

d10 d60 UC Porosity, % Density, g cm3 Specific area, m1

Chemical composition, %

0.3 4.7 15.4 49 2.94 218.5

Element

EDX analysis

O Mg Al Si Ca Mn Fe P

53.7 1.8 0.9 4.4 33.5 1.5 4.2

ICP-AES analysis 3.3 1.1 33.9 26.2 0.5

Ca content was homogeneous within the slag, whereas Fe was predominantly present within the BOF slag. The low presence of Fe on the slag surface may be related to the carbonation of the slag with atmospheric CO2 during weathering. The cloudy textures observed through SEM analysis are likely covering the Fe oxides on the surface of the steel slag. 3.2. Kinetic batch experiments The experiments showed a quick reduction of the initial PO4-P concentrations from 30 mg/L to levels lower than 1 mg/L during the first half hour of the batch experiments, altogether with increasing pH values and Ca2þ concentrations with time (Fig. 4a and b). The high correlation coefficients obtained from the application of the pseudo-first order equation to the experimental potential Ca2þ release capacity through time indicated that this model described Ca2þ release well (Table 2). This suggests that one of the reactants (Ca(OH)2) was present in great excess over the other reactants in the reaction mixture, and demonstrates that the dissolution of Ca(OH)2 is the primary reaction explaining the increase in Ca2þ. Moreover, increasing OH concentrations calculated from pH are strongly related to Ca2þ concentrations (R2 ¼ 0.96), confirming that pH increases according to the stoichiometric dissolution of Ca(OH)2 in water. Although the potential Ca2þ release capacity and Ca2þ to OH ratio are useful to explain the increase of Ca2þ and OH concentrations in solution with respect to the slag composition, it is worth highlighting that these calculations may be affected by calcium phosphate and (CaCO3) precipitation. The formation of calcium phosphate precipitates, encouraged by the combination of elevated equilibrium pH and increased Ca2þ concentration in solution (Arias et al., 2001; Johansson and Gustafsson, 2000), is responsible for the reduction of PO4-P and Ca2þ concentrations during the batch experiments. whereas dissolution of atmospheric CO2 and CaCO3 precipitation, that may have occurred within the kinetic batch tests due to frequent sampling, also affect Ca2þ concentrations, influencing the calculation of kinetic rates and Ca2þ to OH ratios. It is suggested that future kinetic batch tests should use greater initial PO4-P concentrations in order to properly assess P removal kinetics of BOF slag. Also, Ca2þ release kinetics should be assessed during longer times since similar studies suggest that a pseudoequilibrium can be reached for pH and Ca2þ concentrations after 7 days of agitation (Barca et al., 2012). 3.3. Isotherm experiments As shown in Fig. 5a, tests with initial PO4-P concentrations ranging between 5 and 300 mg/L showed residual PO4-P concentrations in solution lower than 1 mg/L which resulted in PO4-P removal performances greater than 99%. Results differed at the initial 400 mg/L PO4-P concentration tests, where residual PO4-P concentrations increased to an average value of 62.8 mg/L and removal efficiencies decreased to an average 84%. PRCs increased with increasing initial PO4-P concentrations with values between 0.12 and 1.20 mg P/g slag for lower initial PO4P concentrations (5e50 mg/L), which represent levels most likely to be encountered in wastewaters ranging from typical domestic to parlor waste. At greater initial PO4-P concentrations (100e400 mg/ L) PRCs continued to increase, reaching its maximum (8.04 mg P/g slag) at the initial 400 mg/L PO4-P concentration. This increase of PRC with increasing initial PO4-P concentrations in batch experiments has been previously reported in other studies (Barca et al., 2012; Bowden et al., 2009; Xue et al., 2009). Although the results from isotherm batch experiments can lead to biased and unrealistic estimates if used for field scale

I. Blanco et al. / Water Research 89 (2016) 355e365

359

Fig. 2. XRD pattern of BOF slag before experiments.

Fig. 3. SEM observations (a, b) and EDX spectrum (c) of the BOF slag surface before experiments.

2+

2+

(b)

PO4-P

Ca

25

300

20 15

200

10 100

5

0

0 1 6

12

24

36

48

Time, h

12 11 pH

Ca concentration, mg/L

400

PO4-P concentration, mg/L

(a)

10 9 8 7 1 6

12

24

36

48

Time, h

Fig. 4. Kinetics of Ca2þ concentration release and PO4-P concentration removal (a), and pH (b). Bars indicate standard deviation.

Table 2 Correlation coefficients and rate constants for the pseudo-first order kinetic model describing Ca2þ release. Kinetic batch replicates

Qe, mg Ca/g

k1, 1/h

R2

Kinetic batch 1 Kinetic batch 2 Kinetic batch 3

5.36 6.70 7.10

0.166 0.223 0.286

0.972 0.973 0.971

applications, they are useful to compare the PRCs of different materials. PRC results were generally lower than those reported by other authors studying BOF slags in phosphate removal batch experiments (Table 3). These differences may be attributed to different particle size and slag to solution ratio experimental conditions, since the removal efficiency of a material decreases with increasing particle size and slag to solution ratios (Cucarella and

I. Blanco et al. / Water Research 89 (2016) 355e365

60

300

40

200 20

100 0

12 11 pH

400

(b)

PO4-P

2+

Ca

PO4-P concentration, mg/L

500

2+

(a) Ca concentration, mg/L

360

0 5

100

200

300

10 9 8 7

400

5

Initial PO4-P concentration, mg/L

100

200

300

400

Initial PO4-P concentration, mg/L

Fig. 5. Representation at different initial PO4-P concentrations of Ca2þ and PO4-P concentration in solution (a), and pH (b). Bars indicate standard deviation.

Table 3 Experimental conditions and PRCs observed in phosphate removal BOF slag batch experiments. Reference

Particle size, mm

Amount of slag, g

Slag to solution ratio

Initial concentrations, mg P/L

Agitation, rpm

Contact time, h

PRC, mg P/g

Jha et al. (2008) Bowden et al. (2009) Xue et al. (2009) Barca et al. (2012) Han et al. (2015) This study

<0.02 <0.3 <0.6 5e10 <0.18 <10 <10

0.1 0.05 1 40 0e2 10 10

1:500 N.A. 1:100 1:25 >1:125 1:25 1:25

310 1e1000 10e500 5e100 50e125 5e100 200e400

N.A. 150 N.A. 125 N.A. 120 120

24 24 3 168 4 48 48

78.9 89.97 0.82e43.1 0.03e2.49 11.4e20.3 0.12e2.47 4.96e8.04

Renman, 2009). The results were similar between the maximum PRC results obtained by Barca et al. (2012) at the 100 mg/L concentration (2.49 versus 2.47 mg P/g slag obtained in the present study) what may be attributed to similar experimental conditions between both studies (Table 3). The results from isotherm batch experiments are useful to understand the P removal mechanisms. It is more likely that PO4-P removal occurred via precipitation rather than adsorption since conditions were favorable for calcium phosphate precipitation; white precipitates where clearly identifiable at the end of the batch experiments; and P adsorption to metal oxide surfaces decreases with increasing pH (McBride, 1994). Chemical precipitation of calcium phosphate minerals occurs spontaneously at pH 8 (Khelifi et al., 2002; Søvik and Kløve, 2005), and it is encouraged by the combination of elevated equilibrium pH and increased Ca2þ concentration in solution (Arias et al., 2001; Johansson and Gustafsson, 2000). Dissolution of Ca(OH)2 in water increased pH values and Ca2þ concentrations in solution, resulting in PO4-P removal through calcium phosphate precipitation. The decline in PO4-P removal efficiency at the initial 400 mg/L PO4-P concentration may then be attributed to limited Ca2þ release from Ca(OH)2 dissolution in water as indicated by low Ca2þ residual concentrations in solution and increased PO4-P residual concentrations. Geochemical modeling with PHREEQC software was used to simulate the effect of Ca(OH)2 dissolution considering the original quality of the isotherm batch solutions at the starting point of each test. Different calcium phosphate precipitates may form in solution depending on the pH values and the Ca/P molar ratio such as amorphous calcium phosphates (ACPs), dicalcium phosphate (DCP), tricalcium phosphate (TCP), octocalcium phosphate (OCP) and hydroxyapatite (HAP) (Table 4). In all tests DCP was never supersaturated whereas TCP, OCP and HAP were systematically supersaturated in the bulk, and SI values for all the phases increased gradually with increasing initial PO4-P concentrations (Table 5). It is likely that HAP has been the main compound formed in the

solutions at lower initial PO4-P concentrations (5e100 mg/L), when the Ca/P molar ratios, calculated from consumption of Ca2þ and P, are over the theoretical mineral ratio for HAP (Table 5). This is also indicated by the HAP strongly positive saturation index (SI > 19.83), the pH values, Ca2þ and PO4-P concentrations that were in the range of values that support its formation (Kim et al., 2006; Stumm and Morgan, 1996; Valsami-Jones, 2001). At higher initial PO4-P concentrations (200e400 mg/L) SI values indicate greater supersaturation of TCP, OCP and HAP, while Ca/P molar ratios indicate more favorable conditions for TCP, OCP and ACP formation. It is proposed that co-precipitation of all these calcium phosphates occurred during these tests, with TCP, OCP and ACP acting as precursors for HAP crystallization. Substantial HAP formation is consistent with the decrease of pH in solution (Fig. 5b) as suggested by Barca et al. (2012), due to OH consumption during HAP formation:

10Ca2þ þ 6PO4 3 þ 2OH /Ca10 ðPO4 Þ6 ðOHÞ2 Y

(5)

Direct HAP precipitation and HAP crystallization from precursors require OH consumption that results in a gradual pH decrease. Lower pH and Ca/P molar ratio conditions are favorable for TCP, OCP and ACP precipitation that promote further HAP crystallization and pH decrease in solution. 3.4. Control column Outlet pH and Ca2þ concentrations decreased asymptotically and were always greater than inlet concentrations (data not shown). During the first 13 days of monitoring, average inlet Ca2þ concentrations were 27.7 ± 0.8 mg/L while outlet Ca2þ concentrations decreased from 548.9 to 84.7 mg/L. Although carbonates from inlet water alkalinity and atmospheric CO2 dissolution may have resulted in calcium carbonate precipitation, affecting the Ca2þ to OH ratio, the good correlation coefficient (R2 ¼ 0.97) between outlet Ca2þ concentrations and outlet OH concentrations

I. Blanco et al. / Water Research 89 (2016) 355e365

361

Table 4 Calcium to phosphate (Ca/P) molar ratio and solubility products (KPS) of different calcium phosphates (Stumm and Morgan, 1996; Valsami-Jones, 2001). Name

Formula

Ca/P molar ratio

KPS (25  C)

Amorphous calcium phosphate (ACP) Dicalcium phosphate (DCP) Tricalcium phosphate (TCP) Octocalcium phosphate (OCP) Hydroxyapatite (HAP)

Ca3(HPO4)2 CaHPO4 Ca3(PO4)2 Ca4H(PO4)3 Ca10(PO4)6(OH)2

~1.5 1 1.5 1.33 1.67

Variable, more soluble than the rest 106.6 1026 1046.9 10114

Table 5 Ca/P molar ratios and saturation indexes (SI) with respect to different minerals at the beginning of the isotherm batch experiments calculated with PHREEQC at T ¼ 25  C. Initial PO4-P concentration, mg/L

Ca/P molar ratio

DCP

TCP

OCP

HAP

5 10 25 50 100 200 300 400

17.05 20.26 12.65 7.37 3.37 1.56 1.49 1.29

4.29 3.99 3.60 3.30 3.00 2.70 2.53 2.40

2.20 2.80 3.59 4.19 4.79 5.39 5.73 5.98

2.86 3.76 4.95 5.85 6.75 7.65 8.16 8.53

19.83 20.74 21.93 22.83 23.73 22.68 25.14 25.50

calculated from pH values indicates that Ca2þ is released according to the stoichiometric dissolution of Ca(OH)2. 3.5. Experimental column PO4-P concentrations at the inlet, vertical sampling points (5, 10 and 15 cm), and the outlet of the Experimental column are plotted in Fig. 6. Initial PO4-P concentrations at the vertical sampling points were close to 0 mg/L resulting in PO4-P removal rates greater than 95% at the beginning of the experiment. PO4-P concentrations started to increase on days 9, 29 and 45 at the vertical sampling points 5, 10 and 15 cm respectively. This resulted in a gradual decrease of P removal rates until reaching a more stable level of P removal at each column section (Table 6). Experimental column outlet PO4-P concentrations were always close to 0 mg/L resulting in P removal rates greater than 95% during the whole operational period (Table 6). Geochemical modeling with PHREEQC software was used to simulate the effect of Ca(OH)2 dissolution on the column inlet chemical composition in equilibrium with atmospheric CO2 at the top of the Experimental column. The saturation indexes indicate

Table 6 Column section PRC and average P removal rate at the end of the experiment (n ¼ 20). Column section

Average P removal rate, %

Inlet e 5 cm 5e10 cm 10e15 cm 15 cm e Outlet

7.7 6.6 7.3 99.5

± ± ± ±

4.5 3.3 3.8 1.4

PRC, mg P/g slag 1.8 2.5 1.7 6.5

the co-precipitation of calcite (CaCO3, SI ¼ 6.36) and calcium phosphates in solution. CaCO3 precipitation is to be expected since the Experimental columns have carbonate sources in the inlet tap water alkalinity and the atmospheric dissolution of CO2. Its precipitation is relevant since it may compete with calcium phosphates for consumption of Ca2þ ions. TCP (SI ¼ 4.29), OCP (SI ¼ 6.09) and HAP (SI ¼ 18.25) were also supersaturated in solution, indicating that the initial high P removal rates may be attributed to calcium phosphate precipitation. Ca(OH)2 dissolution in water results in high pH and Ca2þ concentrations in solution, ideal conditions for calcium phosphate

Fig. 6. P concentrations at the inlet, 5 cm, 10 cm, 15 cm sampling points, and outlet of the column.

362

I. Blanco et al. / Water Research 89 (2016) 355e365

precipitation. As a consequence of continuous dissolution, the amount of Ca(OH)2 was gradually decreased in the BOF slag, resulting in lower Ca2þ and OH release in solution with time, increase of limiting conditions for calcium phosphate precipitation, gradual increase of PO4-P concentration in solution and gradual decrease in P removal. Once maximum Ca2þ release from BOF slag dissolution was reached, only the Ca2þ present in the influent solution was available for reaction with PO4-P, explaining the low stable levels of P removal (Bowden et al., 2009). In the Experimental column, the PRC values were in the range of values reported in the literature for BOF slag column experiments (Table 7). PRC values varied between 1.7 and 2.5 mg P/g slag within the first three sections of the column, reached a maximum of 6.5 mg P/g slag in column section 15 cm e Outlet, and showed an average of 3.1 mg P/g slag for the whole column. The greater PRC and P removal rates of section 15 cm e Outlet with respect to the rest of the column (Table 6) may be attributed to (i) Ca(OH)2 solubility within the Experimental column and (ii) the accumulation of small diameter BOF slag particulates due to sedimentation from upper sections of the column. Ca(OH)2 dissolution in the upper sections of the column increases Ca2þ and OH concentrations in solution. After calcium phosphate and CaCO3 precipitation, remaining Ca2þ and OH concentrations will control the solubility of Ca(OH)2, a slightly soluble ionic compound, delaying its dissolution in the lower sections of the column. The accumulation of small diameter BOF slag may have affected the P removal efficiency of BOF slag since it strongly depends on its particle size, and it has been shown to decrease with an increase in particle size (Cucarella and Renman, 2009). Most probably, the smaller size slag accumulated at the bottom of the column favored PO4-P removal due to a greater specific surface available for Ca(OH)2 dissolution. The PO4-P removal capacity of the upper sections of the column would have been reduced in comparison since they retained the larger size slag particles with a lower specific surface available for Ca(OH)2 dissolution. This would explain similarities with results from the literature. The first three column sections, with a greater particle size, are more similar to the PRC values achieved by Barca et al. (2014) using columns with similar inlet P concentrations and greater slag particle size. On the other hand, the PRC results from the last section are similar to those obtained by Bowden et al. (2009) using columns with lower particle size and greater specific area. Variation of results is to be expected when comparing studies with other different experimental conditions in addition to particle size. P removal rates decrease with increasing inlet P concentrations (Bowden et al., 2009), and PRC values may increase with the duration of the experiment. Also, P removal may improve with increasing CaO-slag content, and with increasing HRT due to prolonged contact time between slag and solution, favoring CaO dissolution and Ca phosphate precipitation (Barca et al., 2014).

pH values and PO4-P concentrations during the operational period at the 5, 10 and 15 cm sampling points are shown in Fig. 7. Maximum pH values at the beginning of the experiment decreased gradually through time. High initial pH values are to be expected since BOF slags in solution may increase pH values up to 13 units (Mayes et al., 2008), and have a strong tendency to produce high conductivities and Ca2þ concentrations (Barca, 2013), whereas the reduction through time may be attributed to depletion of Ca(OH)2 content in the BOF slag due to continuous dissolution in water. The increase in PO4-P concentrations on days 9, 29 and 45 for the 5, 10 and 15 cm vertical sampling points coincides in time with a drop of pH to values close to 10 and a drop of EC to concentrations similar to those registered at the inlet. These results suggest that the Ca content in solution has decreased, limiting calcium phosphate precipitation, and indicating that the Ca slag content of the corresponding column section is getting exhausted. This behavior is not so clear for sampling point 15 cm; PO4-P concentrations start to increase on day 45 before the previous section is exhausted, suggesting the presence of preferential pathways in this section of the column. In view of these results, pH and EC concentrations could be used as simple indicators of Ca(OH)2 depletion in the BOF slag column. The use of pH as an indicator of Ca(OH)2 depletion helps also explain the great removal capacity at the bottom of the column where pH outlet values remained in the range 10.7e11.3 at the end of the experiment, indicating that dissolution of Ca(OH)2 is still taking place. Ca2þ concentrations were estimated from pH values following the stoichiometric relationship observed in the Control column. The Ca/P molar ratio was greater than 1.67 until the increase of PO4P concentrations (days 9, 29 and 45), indicating that conditions were favorable for direct HAP formation in the column (House, 1999; Johansson and Gustafsson, 2000; Stumm and Morgan, 1996). Additional PO4-P removal after these days may have been caused by the combination of adsorption processes on the slag surface and the precipitation of Ca phosphates with a Ca/P molar ratio lower than 1.67 (Barca et al., 2012). It is worth highlighting that the use of estimated Ca2þ concentrations is used for illustration purposes of the possible PO4-P removal mechanisms, and does not pretend to be an accurate estimate of concentrations since the molar ratio of Ca2þ to OH in solution may have been affected due to calcite and calcium phosphate precipitation. White precipitates were clearly identifiable on the top of the column during the operational period. Microscopic observations by SEM after the column experiment showed a slag surface covered with a crystalline layer (Fig. 8a and b). The EDX analysis (Fig. 8c) revealed that the crystalline surface consisted predominantly of Ca, P and O, with a Ca/P molar ratio of 3.76. A Ca/P molar ratio >1.67 (the theoretical Ca/P molar ratio for HAP) may be a result from an increment of Ca2þ composition due to presence of CaCO3 which substitutes the sites for HAP crystallization (Kim et al., 2006). A

Table 7 Experimental conditions and PRCs observed in phosphate removal BOF slag column experiments. Reference

Particle size, mm

Ca content, %

Main Ca phase

Flow direction

HRT, h

Operation, days

Initial P concentration, mg/L

PRC, mg P/g

Maximum PRC, %

P saturation

Cha et al. (2006) Bowden et al. (2009)

1e2 <20 <20 6e12 20e50 <10

32.1 42e44 42e44 44.1 44.1 33.9

CaO CaO CaO CaO CaO Ca(OH)2

Upflow N.A. N.A. Horizontal Horizontal Downflow

42a 4e22 8 24 24 8

263 406 306 364 364 213

2.3 1e50 100e300 10.8 10.8 15

N.A. N.A. 8.39 1.05 1.01 3.1

100 62 74 >99 95 >99

No N.A. Yes No No Yes

Barca et al. (2014) This study a

Calculated from experimental conditions.

I. Blanco et al. / Water Research 89 (2016) 355e365

363

Fig. 7. Inlet PO4-P concentration (circles) and outlet pH values (squares) at the (a) 5 cm, (b) 10 cm and (c) 15 cm sampling points.

Fig. 8. SEM observations (a, b) and EDX spectrum (c) of the BOF slag surface after the column experiment.

white precipitate leaking from the column was observed at the outlet and identified as CaCO3 after XRD analysis (Fig. 9). The presence of this precipitate confirms the formation of CaCO3 crystals within the column as a result of atmospheric CO2 diffusion and inlet water alkalinity, and supports the possibility of CaCO3 co-

precipitation with calcium phosphates (House, 1999; Kapolos and Koutsoukos, 1999). Its presence is significant as calcite has been recognized to provide sorption sites for P leading to nucleation and precipitation of calcium phosphates (Freeman and Rowell, 1981; Kim et al., 2006; Liira et al., 2009).

Fig. 9. XRD pattern of precipitate at the outlet of the Experimental column.

364

I. Blanco et al. / Water Research 89 (2016) 355e365

3.6. The potential use of BOF slag in constructed wetlands Batch and column experiments are useful to understand the P removal mechanisms of BOF slag and illustrate the potential PRC of this material. However, full-scale supplementary research is required to accurately forecast the performance of filter materials under field conditions. Field scale systems are exposed to environmental conditions such as hydrology that may affect CaO and/or Ca(OH)2 dissolution rates and influence the P removal efficiency (Barca, 2013), or temperature oscillations which may reduce P treatment under low temperature conditions (Barca, 2013; Shilton et al., 2006). Wastewater composition has also a significant effect on P retention capacity: the proportion of phosphate to non-phosphate P (polyphosphates, organic phosphorus, etc.) may affect P treatment (Stumm and Morgan, 1996), whereas other wastewater constituents may compete for Ca (CO2, carbonates) (Liira et al., 2009) or theoretically inhibit HAP crystallization (such as Mg2þ, humic, fulvic and tannic acids), although other phases (OCP, DCP) may be less affected (House, 1999; Valsami-Jones, 2001). Accumulation of total suspended solids (TSS) and development of biofilm due to the presence of organic matter (BOD, COD) and nutrients in wastewater may clog the filter or create preferential pathways (Drizo et al., 2002). Pretreatment of these pollutants is recommended in order to prolong the longevity of the filter (Weber et al., 2007). The use of BOF slag filters generates leachates with high EC concentrations and pH values greater than 12 which may have detrimental effects in the environment, and need to be reduced to levels lower than 9, the legal discharge limit in many European countries (EPA, 2001). In constructed wetlands (CWs) high pH values may impede macrophyte growth and development. In that respect, Drizo et al. (2002) recommend the design of CWs using several separate units in series. The first one would act as a conventional CW for TSS and organic matter treatment and plant growth, and it would be followed by an unplanted steel slag unit to act as a secondary treatment unit for P removal. A final third unit would be used for final effluent polishing and pH reduction. Recent research is focusing on ways to reduce pH from slags and other alkaline filter media. The potential for wetlands and CWs to buffer alkaline lime spoil and steel slag leachates as a passive approach to remediation has been highlighted by Mayes et al. (2009), whereas Drizo and Picard (2014) have developed a steel slag filter where a pH adjustment unit is an integral component of the filter. Since BOF slag field studies are still scarce (Barca, 2013; Shilton et al., 2006), future research should focus on the combination of CWs with BOF slag filters and pH adjustment units. 4. Conclusions Batch and column experiments showed that a locally available BOF slag from NW Spain is a material able to remove P effectively from synthetic solutions. Batch experiments showed phosphate removal efficiencies between 84 and 99% and phosphate removal capacities from 0.12 to 8.78 mg P/g slag with increasing initial phosphate concentrations. Under continuous flow conditions over a period of 213 days of operation, the column experiment showed a removal efficiency greater than 95% and a phosphate removal capacity of 3.1 mg P/g slag. P saturation of slag was reached within the upper sections of the column which showed phosphate removal capacities between 1.7 and 2.5 mg P/g slag, and a gradual decline in removal efficiencies from 99% until reaching stable outlet concentrations with efficiencies around 7%. The main P removal mechanism is calcium phosphate precipitation which depends on Ca2þ and OH release after dissolution of Ca(OH)2 in water. Once Ca(OH)2 is completely dissolved, P removal

will depend on influent Ca2þ for limited calcium phosphate precipitation. The use of BOF steel slag as a substrate in constructed wetlands would produce effluents with high pH values and EC concentrations that need to be considered in the constructed wetland design, and need to be corrected before discharge on water bodies. Further research on pilot scale studies of BOF slag filters with real wastewater under environmental conditions is required in order to properly characterize its suitability as a substrate in constructed wetlands. Acknowledgments
n Blanco Rubio received a FPI fellowship mutually funded Dr. Iva
n of the Junta de Castilla y Leo
n by the Consejería de Educacio (Orden EDU/1817/2004, de 29 de noviembre) and the European Social Fund. The authors would like to thank ArcelorMittal for providing the BOF steel slag, Dr. Harouiya and the staff from
ez and Mr. Hector Cemagref for their technical support, Mr. Luis Pela
rraga for their contribution during the development of Luis Astia this work. The authors also thank the reviewers for their valuable contributions. References APHA-AWWA-WEF, 2005. Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington, D.C., USA, p. 1496. Arias, C.A., Del Bubba, M., Brix, H., 2001. Phosphorus removal by sands for use as media in subsurface flow constructed reed beds. Water Res. 35 (5), 1159e1168. Babatunde, A., Zhao, Y., Burke, A., Morris, M., Hanrahan, J., 2009. Characterization of aluminium-based water treatment residual for potential phosphorus removal in engineered wetlands. Environ. Pollut. 157 (0), 2830e2836. Barca, C., 2013. Steel Slag Filters to Upgrade Phosphorus Removal in Small Waste
Nantes Angers Le Mans, water Treatment Plants (Ph.D. thesis). L'Universite Nantes, France.
rente, C., Meyer, D., Chazarenc, F., Andres, Y., 2012. Phosphate removal Barca, C., Ge from synthetic and real wastewater using steel slags produced in Europe. Water Res. 46 (7), 2376e2384. Barca, C., Meyer, D., Liira, M., Drissen, P., Comeau, Y., Andres, Y., Chazarenc, F., 2014. Steel slag filters to upgrade phosphorus removal in small wastewater treatment plants: removal mechanisms and performance. Ecol. Eng. 68 (0), 214e222. Belhadj, E., Diliberto, C., Lecomte, A., 2012. Characterization and activation of Basic Oxygen Furnace slag. Cem. Concr. Compos. 34 (0), 34e40. Bowden, L.I., Jarvis, A.P., Younger, P.L., Johnson, K.L., 2009. Phosphorus removal from waste waters using basic oxygen steel slag. Environ. Sci. Technol. 43 (7), 2476e2481. Cha, W., Kim, J., Choi, H., 2006. Evaluation of steel slag for organic and inorganic removals in soil aquifer treatment. Water Res. 40 (5), 1034e1042. Claveau-Mallet, D., Wallace, S., Comeau, Y., 2012. Model of phosphorus precipitation and crystal formation in electric arc furnace steel slag filters. Environ. Sci. Technol. 46 (3), 1465e1470. Claveau-Mallet, D., Wallace, S., Comeau, Y., 2013. Removal of phosphorus, fluoride and metals from a gypsum mining leachate using steel slag filters. Water Res. 47 (4), 1512e1520. Cucarella, V., Renman, G., 2009. Phosphorus sorption capacity of filter materials used for on-site wastewater treatment determined in batch experiments e a comparative study. J. Environ. Qual. 38 (0), 381e392. Drizo, A., Comeau, Y., Forget, C., Chapuis, R.P., 2002. Phosphorus saturation potential: a parameter for estimating the longevity of constructed wetland systems. Environ. Sci. Technol. 36 (21), 4642e4648. Drizo, A., Forget, C., Chapuis, R.P., Comeau, Y., 2006. Phosphorus removal by electric arc furnace steel slag and serpentinite. Water Res. 40 (8), 1547e1554. Drizo, A., Frost, C.A., Grace, J., Smith, K.A., 1999. Physico-chemical screening of phosphate-removing substrates for use in constructed wetland systems. Water Res. 33 (17), 3595e3602. Drizo, A., Picard, H., 2014. System for removing phosphorus from wastewater. US Patent 8721885 B2. Filed August 30 2010, and issued May 14 2014. EPA, 2001. Parameters of Water Quality e Interpretation and Standards. Environmental Protection Agency, Wexford, Ireland, p. 133. European Norm EN 933-1:1997. Tests for geometrical properties of aggregates. Determination of particle size distribution. Sieving method. Asociacion Espanola de Normalizacion y Certificacion. Madrid, Spain. Freeman, J.S., Rowell, D.L., 1981. The adsorption and precipitation of phosphate onto calcite. J. Soil Sci. 32 (1), 75e84. Han, C., Wang, Z., Yang, H., Xue, X., 2015. Removal kinetics of phosphorus from synthetic wastewater using basic oxygen furnace slag. J. Environ. Sci. 30 (0),

I. Blanco et al. / Water Research 89 (2016) 355e365 21e29. House, W.A., 1999. The physico-chemical conditions for the precipitation of phosphate with calcium. Environ. Technol. 20 (7), 727e733. Jha, V.K., Kameshima, Y., Nakajima, A., Okada, K., 2008. Utilization of steel-making slag for the uptake of ammonium and phosphate ions from aqueous solution. J. Hazard. Mater. 156 (1e3), 156e162. Johansson, L., 1999. Blast furnace slag as phosphorus sorbents e column studies. Sci. Total Environ. 229 (0), 89e97. Johansson, L., 2006. Substrates for phosphorus removal e potential benefits for onsite wastewater treatment? Water Res. 40 (0), 23e36. Johansson, L., Gustafsson, J.P., 2000. Phosphate removal using blast furnace slags and opoka-mechanisms. Water Res. 34 (1), 259e265. Kapolos, J., Koutsoukos, P.G., 1999. Formation of calcium phosphates in aqueous solutions in the presence of carbonate ions. Langmuir 15 (19), 6557e6562. Khelifi, O., Kozuki, Y., Murakami, H., Kurata, K., 2002. Nutrients adsorption from seawater by new porous carrier made from zeolitized fly ash and slag. Mar. Pollut. Bull. 45 (0), 311e315. Kim, E.H., Yim, S.B., Jung, H.C., Lee, E.J., 2006. Hydroxyapatite crystallization from a highly concentrated phosphate solution using powdered converter slag as a seed material. J. Hazard. Mater. B136 (0), 690e697. Lagergren, S., 1898. About the theory of so-called adsorption of soluble substances. K. Sven. Vetensk. Handl. 24 (4), 1e39. €t, K., 2009. Active Liira, M., Koiv, M., Mander, U., Motlep, R., Vohla, C., Kirsima filtration of phosphorus on Ca-rich hydrated oil shale ash: does longer retention time improve the process? Environ. Sci. Technol. 43 (10), 3809e3814. Mahieux, P.Y., Aubert, J.E., Escadeillas, G., 2009. Utilization of weathered basic oxygen furnace slag in the production of hydraulic road binders. Constr. Build. Mater. 23 (0), 742e747. Mayes, W.M., Batty, L.C., Younger, P.L., Jarvis, A.P., Koiv, M., Vohla, C., Mander, U., 2009. Wetland treatment at extremes of pH: a review. Sci. Total Environ. 407 (13), 3944e3957. Mayes, W.M., Younger, P.L., Aumonier, J., 2008. Hydrogeochemistry of alkaline steel slag leachates in the UK. Water Air Soil Pollut. 195 (0), 35e50. McBride, M.B., 1994. Environmental Chemistry of Soils. Oxford University Press, Oxford, UK, p. 406.
s de roseaux: limites hydrauliques et re
tention du Molle, P., 2003. Filtres plante
Montpellier II Sciences et Techniques du phosphore (Ph.D. thesis). Universite Languedoc, Montpellier, France. Morone, M., Costa, G., Polettini, A., Pomi, R., Baciocchi, R., 2014. Valorization of steel slag by a combined carbonation and granulation treatment. Miner. Eng. 59 (0), 82e90. Nair, P.S., Logan, T.J., Sharpley, A.N., Sommers, L.E., Tabatabai, M.A., Yuan, T.L., 1984. Interlaboratory comparison of a standardized phosphorus adsorption procedure. J. Environ. Qual. 13 (4), 591e595. Parkhurst, D.L., Appelo, C.A.J., 2013. Description of Input and Examples for PHREEQC Version 3-A Computer Program for Speciation, Batch-reaction, One-dimensional Transport, and Inverse Geochemical Calculations: U.S. Geological Survey Techniques and Methods. Book 6, Chap. A43, 497 p., Available only at: http:// pubs.usgs.gov/tm/06/a43/.

365

Proctor, D.M., Fehling, K.A., Shay, E.C., Wittenborn, J.L., Green, J.J., Avent, C., Bigham, R.D., Connolly, M., Lee, B., Shepker, T.O., Zak, M.A., 2000. Physical and chemical characteristics of blast furnace, basic oxygen furnace, and electric arc furnace steel Industry slags. Environ. Sci. Technol. 34 (0), 1576e1582. Reddy, K.R., Overcash, M.R., Khaleel, R., Westerman, P.W., 1980. Phosphorus adsorption-desorption characteristics of two soils utilized for disposal of animal wastes. J. Environ. Qual. 9 (1), 86e92. Ruiz-Agudo, E., Kudlacz, K., Putnis, C.V., Putnis, A., Rodriguez-Navarro, C., 2013. Dissolution and carbonation of portlandite [Ca(OH)2] single crystals. Environ. Sci. Technol. 47 (19), 11342e11349. Shen, W., Zhou, M., Ma, W., Hu, J., Cai, Z., 2009. Investigation on the application of steel slag-fly ash-phosphogypsum solidified material as road base material. J. Hazard. Mater. 164 (1), 99e104. Shilton, A.N., Elmetri, I., Drizo, A., Pratt, S., Haverkamp, R.G., Bilby, S.C., 2006. Phosphorus removal by an ‘active’ slag filter e a decade of full scale experience. Water Res. 40 (0), 113e118. Søvik, A.K., Kløve, B., 2005. Phosphorus retention processes in shell sand filter systems treating municipal wastewater. Ecol. Eng. 25 (0), 168e182. Spanish norm UNE-EN 932-2:1999. Ensayos para determinar las propiedades
ridos. Parte 2: Me
todos para la reduccio
n de muestras de generales de los a laboratorio. Asociacion Espanola de Normalizacion y Certificacion. Madrid, Spain. Stumm, W., Morgan, J.J., 1996. Aquatic Chemistry. Wiley-Interscience, Nueva York, NY, USA, p. 1022. Tsakiridis, P., Papadimitriou, G., Tsivilis, S., Koroneos, C., 2008. Utilization of steel slag for Portland cement clinker production. J. Hazard. Mater. 152 (2), 805e811. Valsami-Jones, E., 2001. Mineralogical controls on phosphorus recovery from wastewaters. Mineral. Mag. 65 (5), 611e620. Vohla, C., Koiv, M., Bavor, H.J., Chazarenc, F., Mander, U., 2011. Filter materials for phosphorus removal from wastewater in treatment wetlands e a review. Ecol. Eng. 37 (0), 70e89. Vymazal, J., 2005. Horizontal sub-surface flow and hybrid constructed wetlands systems for wastewater treatment. Ecol. Eng. 25 (0), 478e490. Vymazal, J., 2007. Removal of nutrients in various types of constructed wetlands. Sci. Total Environ. 380 (0), 48e65. Weber, D., Drizo, A., Twohig, E., Bird, S., Ross, D., 2007. Upgrading constructed wetlands phosphorus reduction from a dairy effluent using electric arc furnace steel slag filters. Water Sci. Technol. 56 (3), 135e143. Xie, J., Wu, S., Lin, J., Cai, J., Chen, Z., Wei, W., 2012. Recycling of basic oxygen furnace slag in asphalt mixture: material characterization & moisture damage investigation. Constr. Build. Mater. 36 (0), 467e474. Xue, Y., Hou, H., Zhu, S., 2009. Characteristics and mechanisms of phosphate adsorption onto basic oxygen furnace slag. J. Hazard. Mater. 162 (0), 973e980. Yamada, H., Kayama, M., Saito, K., Hara, M., 1986. A fundamental research on phosphate removal by using slag. Water Res. 20 (5), 547e557. Yildirim, I.Z., Prezzi, M., 2009. Use of Steel Slag in Subgrade Applications. Indiana Department of Transportation and Purdue University, West Lafayette, IN, USA. Yildirim, I.Z., Prezzi, M., 2011. Chemical, mineralogical, and morphological properties of steel slag. Adv. Civ. Eng. 2011 (0), 1e13.