Enhanced adsorptive removal of phosphate from aqueous solution by activated sand/metal layered double hydroxides

ABSTRACT In the present study, Mg-Fe and Ni-Fe layered double hydroxides have been examined to enhance the uptake capacity of activated sand as low-cost adsorbents for adsorption of phosphate. The physical and chemical properties of adsorbents were investigated by X-ray diffractometer (XRD), X-ray fluorescence (XRF) spectrum, Fourier transform infrared (FTIR) spectroscopy, scanning electron microscopy (SEM), and Brunauer–Emmett–Teller (BET) analysis. The best efficiency of both adsorbents was obtained at the optimum condition of initial phosphate concentration: 1000 mg/L, adsorbent dose: 1 g, pH: 7–7.5, and a temperature of 298 K. The equilibrium isotherm studies show that the best result was attained in the Langmuir model for phosphate; therefore, homogeneous adsorption is dominated, which was also emphasised by three-parameter-isotherm models. The maximum adsorption capacity for functionalised sand with Mg-Fe and Ni-Fe layered double hydroxides were obtained 69.47 and 66.64 mg/g, respectively. The kinetic data fit well with the pseudo-second-order kinetics and demonstrate that the porous diffusion effectiveness is over film diffusion. Thermodynamic parameters evoke the feasibility of the adsorption process as well as the positive influence of temperature on adsorption.

As mentioned above, it seems that the layered double hydroxides (LDH) (known as synthetic clay) are the most suitable compound for increasing the sand potential for phosphate uptake.The aim of this project is to evaluate six metals layered double hydroxides-sand composites to choose the best ones for enhancement in the amount of phosphate removal.Then, the characterisation of the best adsorbents, the optimal conditions, thermodynamic, and kinetic studies have been investigated to find the mechanism of phosphate uptake.

Materials and physical measurements
The sand sample was collected from the Pakdasht sand mine located 25 km southeast of Tehran Province.It was sieved to the particle size of 106 − 125 µm and used as a basic adsorbent material.Al(NO 3 ) 3 .9H 2 O, Ni(NO 3 ) 3 .6H 2 O, Fe(NO 3 ) 3 .9H 2 O, Mg(NO 3 ) 3 .6H 2 O, ZnSO 4 .7H 2 O and CuCl 2 .2H 2 O salts were applied as modifier agents and NaOH (Merck Company) was applied as a precipitating agent without purification.
X-ray diffractometer (XRD) and X-ray fluorescence (XRF) spectrum were obtained by using a Philips X-ray diffractometer 1140 and a Philips X-ray diffractometer Xunique II, respectively.Fourier transform infrared (FTIR) spectroscopy has been used for characterising the functional groups of raw and modified sands.FTIR spectra from 4000 to 400 cm -1 was recorded on a Shimadzu FTIR instrument, using KBr pellets.Scanning electron microscopy (SEM) micrographs taken from the Philips instrument (XL30), were used to recognise the morphology and size of the raw and modified sands.The BET (Brunauer-Emmett-Teller) method was used to determine the specific surface area of raw and modified sand by BELSORP-MINI II Surface Area Analyser.
The final concentration of phosphate after adsorption was determined by ultraviolet spectroscopy methods using a Unicam UV-visible spectrophotometer, based on the absorption of electromagnetic waves through a spectrophotometer at λ max 420 nm.First, standard solutions of phosphate in concentrations of 5, 10, 15, and 20 ppm were provided and after adding 25 ml Vanadate-molybdate reagent and 5 ml sulphuric acid, the absorption value was measured and the absorption standard diagram was obtained to calculate the phosphate concentration of the unknown sample.

Synthesising of metal layered double hydroxides onto the sand
Layered double hydroxides (LDH) as inorganic double metal hydroxide compounds have attracted many researchers' attention to remove anionic pollution from aqueous solutions [45,47,49,53].Their general formula is as: Where M 2+ and M 3+ are the divalent and trivalent cations in octahedral and tetrahedral positions within the hydroxide layers, respectively, and A n-is an exchangeable anion [54].
In the first step, sand was treated by H 2 SO 4 to eliminate trace organic compounds and calcite minerals that caused increased porosity of the surface.LDH of Ni-Fe, Ni-Al, Mg-Al, Zn-Al, Mg-Fe, and Cu-Fe in ratio mole 3:1 of M 2+ :M 3+ were synthesised into activated sand (treated by H 2 SO 4 ) by co-precipitation method.The solution of NaOH (0.1 M) was added slowly to the solution containing M 2+ (Ni, Mg, Zn and Cu), M 3+ salts (Al, Fe), and activated sand under intense stirring for 4 h at room temperature.Then, the solution was maintained at 70°C for 4 h in the oven, and thereafter it was filtered and dried at room temperature.Concerning ratios M 3+ / M 2+ +M 3+ in all synthesised compounds were 0.25.This means that no formation of unwanted phases was not observed since the ratio is between 0.2 and 0.33 [55].

Adsorption experiments
Initial experiments were conducted under identical conditions to select two adsorbents with a maximum adsorption capacity.The equilibrium phosphate concentration after adsorption (q e ) and adsorption efficiency (R%) was calculated by Eq. ( 2) and (3).
Where V is the volume of the sulphate solution (L), C 0 and C e are the initial and final concentrations (mg/L), and m is the amount of adsorbent (g).

Equilibrium studies
To identify the adsorption process, the experimental data of phosphate adsorption on the surface of Mg-Fe and Ni-Fe sand composites were evaluated using two-parameter (Langmuir, Freundlich, Temkin, D-R) and three-parameter (Khan and Toth) isotherm models.The non-linear equations of these models were represented in Table S1 [56].
For this purpose, 1 g of Ni-Fe and Mg-Fe sand composite (particle size of 106-125 µm) was stirred at 600 rpm in 100 mL of phosphate solution with different initial concentrations (400-1500 mg/L) for 30 min at a temperature of 298, 308, 323, and 343 K.

Kinetic studies
The mechanism of phosphate adsorption to Ni-Fe and Mg-Fe sand composites has been investigated using reaction-based kinetic models (pseudo-first-order and pseudo-secondorder) and diffusion-based kinetic models (Boyd, intra-particle diffusion, and mass transfer).The linear form of the equations of these models is demonstrated in Table S2 [57][58][59][60][61][62].For kinetic studies, 1 g of Ni-Fe and Mg-Fe sand composites (particle size of 106-125 µm) were added to 100 mL of phosphate solution with an initial concentration of 400-1500 mg/L at different contact times (5 to 90 min) with a shaking speed of 600 rpm.

Thermodynamic studies
For the thermodynamic investigation, 1 g of Ni-Fe and Mg-Fe sand composites (particle size of 106-125 µm) was added to 100 ml of phosphate solution at concentrations of 400 to 1500 mg/L, and the amount of removed phosphate was determined at various temperatures of 298, 308, 323, and 343 K. Thermodynamic parameters (ΔS, ΔH and ΔG) were calculated by equations represented in Table S3 [63].

Primary adsorption experimental
To select the best adsorbents having the highest amount of phosphate adsorption capacity, experiments were conducted with the same conditions for 1 g adsorbent of Ni-Fe, Ni-Al, Mg-Al, Zn-Al, Mg-Fe, and Cu-Fe sand composites to remove 1000 mg/L initial phosphate concentration from a 100 ml aqueous solution in 30 minutes.According to Table 1, all the metal layer double hydroxides functionalised activated sand cause to increase phosphate uptake in comparison with raw sand and activated sand.Meanwhile, Ni-Fe and Mg-Fe sand composites are more efficient to remove phosphate; consequently, further studies have been carried out to adsorb phosphate onto these adsorbents.

X-ray fluorescence (XRF)
The elemental analysis results (Table 2) indicate that silicon and calcium are the main elements and magnesium, iron, and aluminium are trace elements of raw sand.Acidic activation of sand (sulphuric acid) reduced the weight percentage of CaO, Al 2 O 3 and SiO 2 and significantly caused the increase in SO 3 and LOI, 1 which implies the elimination of some sand components and the formation of new insoluble sulphate on the sand surface (Table 2).Significant weight percentages of iron-nickel oxides in Ni-Fe and magnesium-iron oxides in Mg-Fe sand composites indicate that metal double hydroxides were synthesised on the surface of the sands.

X-ray diffraction (XRD)
Mineralogical studies (Figure 1a) suggest that the main minerals of the sand samples include quartz (SiO 2 ) and calcite (CaCO 3 ).Quartz is the major mineral in the raw and modified sands.The main diffract of quartz mineral in all samples is attributed to plans (100) (101), and (102) [64,65].The principal diffracts of calcite in the XRD pattern of raw sand appear at 2ɵ = 29.74 which is related to plan (104) [66,67].After the acid treatment (washing by H 2 SO 4 ), the diffraction plane of calcite mineral has disappeared and new characteristic reflection peaks which are related to plans (020), (021), and (041) of gypsum mineral (Figure 1b) are observed at 2ɵ = 11.74,20.94, and 29.38, respectively [68,69].As shown in Table 3, these planes (gypsums) are also observed in diffraction patterns of Ni-Fe and Mg-Fe sand composites in 2θ = 11.72,20.88, 28.17 and 2θ = 11.73,20.91, 29.37, respectively.In other words, gypsum (CaSO 4 ) has been synthesised on the surface of sand by reaction between sulphuric acid (H 2 SO 4 ) and calcite (CaCO 3 ).
According to Table 3, the basal spacing of different planes (d hkl ) of quartz and gypsum minerals have not been changed after functionalising the surface of the sand with metal double hydroxides.On the other hand, the metal layered double hydroxides functionalised activated sand do not have a significant effect on the interlayer space of quartz and gypsum planes.These phenomena are similar to other adsorbents like the metal layered double hydroxide functionalised iron ore waste and oxalate loaded haematite [65,71].

Fourier transform infrared (FTIR)
Figure 2 illustrates the FTIR spectrum of raw sand and synthesised composites.In the case of raw sand, the observed broad band at the wave number 3430 cm −1 is related to the stretching vibration bond between hydrogen and oxygen of different water molecules in the sand sample so that the broad band at the 3400 cm −1 area in FTIR spectrum is related to H 2 O molecules [71,78].The C-O stretching and bending vibrations of CO 3 can be seen at wavenumbers of 1429 and 874 cm −1 , respectively, which indicate the presence of calcite in the structure of raw sand.The shoulder vibration bands at 1089 and 1035 cm −1 are attributed to stretching of symmetric and asymmetric Si-O bonds.In addition, bending vibration of O-Si-O was found at 465 cm −1 which belongs to quartz component [67,71,79].A small broad band is observed at a wavenumber of 2515 cm −1 due to the symmetric stretching of the C-H bond of CH 2 group, which is probably due to the presence of impurities or organic compounds in the sample [80,81].Concerning the FTIR spectrum of activated sand, activation of the sand (by H 2 SO 4 ) causes the disappearance of C-O vibration bands related to calcite at 1429 and 874 cm −1 .The vibration bands at 668 and 602 cm −1 are attributed to vibrational symmetry and antisymmetry of the S-O bond in sulphate molecule proving the formation of gypsum mineral [82,83].
According to Figure 2 the wave numbers of the stretching and bending vibrations of the O-H bond belong to Si-OH.In addition, the Si-O-Si shoulder starching bands in Ni-Fe and Mg-Fe sand composites were not significantly shifted compared to the activated sand.The observation of shoulder vibration bands at 1383 cm −1 can be assigned to the nitrate in the interlayer space of metal hydroxides layer (Ni(NO 3 ) 3 .6H 2 O and Mg(NO 3 ) 3 .6H 2 O) [47,[84][85][86][87].It should be noted that because of the overlapping of vibration bands in the region below 1000 cm −1 , the vibration bands of metal hydroxides are not well identified.The waveforms and shapes of vibrating bands in the FTIR spectra of activated sand, Mg-Fe and Ni-Fe sand composites are approximately the same.

SEM Morphology and Specific Surface Area (BET analysis)
The SEM images of the raw and synthesised adsorbents in Figure 3 illustrate that the raw sand has a coarse structure with a large surface.After activating, particles of the sand samples are crushed into sheets and layers and the porosity is increased.It is noticeable that this morphology is similar to other reports about acid washing and modified sand [88,89].In synthesised samples, metal double hydroxide particles cover the sand surface, causing the particle size of sand to increase and turn it into plate-shaped.In addition, the specific surface area of raw sand, Mg-Fe and Ni-Fe sand composites were measured to be 11.53, 19.11, and 21.43 m 2 /g, respectively (Table 4).As Table 4 expresses, the modified sands have a high specific surface area and total pore volume (0.054 cm 3 /g); thus, these abilities make them potential adsorbents towards contaminants [90,91].

Adsorption characteristics
To find the optimal conditions for phosphate adsorption on Ni-Fe and Mg-Fe sand composites, the effects of phosphate initial concentration, pH, and number of composites on the adsorption efficiency were investigated as follows.

The effect of Mg-Fe and Ni-Fe sand composites quantity
To specify the optimal number of Mg-Fe and Ni-Fe sand composites, the experiments were carried out in various quantities of modified sands (1-5 g) under the same conditions.Figure 4(a, b) present the results of phosphate adsorption with different adsorbent values, as phosphate uptake (phosphate concentration of 1000 mg/L) increases in the range of 20-75% with the increasing mass of both adsorbents from 1 to 3 g.The enhancement of   phosphate uptake can be due to increasing available adsorption sites on the surface of Mg-Fe and Ni-Fe sand composites.Similar adsorptive removal behaviour has been observed in the phosphate adsorption process with other adsorbents [20,21,[92][93][94].Nevertheless, the increase in the amount of adsorbents (from 3 to 5 g) did not cause significant changes in the phosphate uptake percentage because most of the phosphate ions in the solution are adsorbed and a very negligible amount of phosphate has remained in the solution.
As Figure 4(a, b) illustrates the amounts of adsorbed phosphate (q e ) declines with an increasing number of Mg-Fe and Ni-Fe sand composites.On the other hand, decreasing the adsorbed phosphate (q e ) and increasing the dosage of adsorbent causes the distribution coefficient (K d ) to decrease.Considering the adsorption process that has occurred in a fixed volume, enhancing the amount of Mg-Fe and Ni-Fe sand composites makes competition among available functional groups on the surface of modified sands, which causes decreasing their activities [95,96].

The effect of phosphate concentration
Adsorption studies were performed on different initial concentrations of phosphates from 400 to 1500 mg/L at constant weights of Mg-Fe and Ni-Fe sand composites.Figure 5(a, b) illustrate that by increasing the initial phosphate concentration, the removal percentage of phosphate decreases (from 75% to 42%) and the equilibrium phosphate concentration after adsorption increases (from 28 to 67 mg/g) for both adsorbents.At higher phosphate concentrations, the available active sites of the adsorbent surface are reduced because of saturation, as a result, phosphate adsorption percentage decreases [21,97].In the case of low pH, phosphorus species is H 3 PO 4 which could weakly bound to the adsorption sites of sand composites due to lack of electrostatic force and ion exchange; hence, the low phosphate uptake in low pH is quite reasonable but with increasing pH up to 7, the percentage of phosphorus species HPO 4 2-and H 2 PO 4 − increases and the surface of adsorbents are positively charged by protonation so that at pH = 7 the maximum adsorption value is observed by involvement of ion exchange [17].

The effect of pH on adsorption mechanism
The drift method was applied to measure the pH PZC value and the resulting pH PZC of 7.7 for Mg-Fe sand composites and 7.4 for Ni-Fe sand composite (Figure 7); that is, the surface charge of Mg-Fe and Ni-Fe sand composites at pH greater than pH PZC has changed from positive to negative.As can be seen in Figure 6(a, b), even though the concentration of PO 4 3-has been increased in the pH >7, the amount of adsorbed phosphate has decreased.It can be described so as the concentration of OH − increased, electrostatic attraction between the phosphate and the Mg-Fe and Ni-Fe sand composites became negligible due to deprotonation of the adsorbent surface and the strong competition of hydroxide with phosphate; consequently, the adsorption gradually decreases due to electrostatic repulsions between negative surface charges of adsorbents and phosphate species [17,[99][100][101][102][103][104][105][106][107].

Equilibrium of adsorption process
Regarding the values of the correlation coefficients of the two-parameter isotherm models (Table 5 and Table 6), the Langmuir model has the highest correlation coefficient (R [2]>0.99)at 298-343 K in comparison with other models.In other words, the Langmuir model can describe phosphate adsorption better than the other models; therefore, the phosphate adsorption process was performed homogeneously for monolayer coverage on the surface of Mg-Fe and Ni-Fe sand composites.Based on the Langmuir R L values (0 < R L < 1), it can be concluded that the phosphate adsorption process on the Mg-Fe and Ni-Fe sand composites has favourable adsorption at different temperatures [45].This is similar to the removal of phosphate by Mg-Al layered double hydroxide [47], lime   modified zeolite [13], Zn-Al layered double hydroxides [52], marble waste [108], dolomitemodified biochar [20], multifunctional biochar [35], nanoscale zerovalent iron [109], and Mg-Al layered double hydroxide [22] ) between phosphate ions and adsorbent; as the temperature rises, the chemical process tends to increase.
In order to determine the adsorption process more accurately in equilibrium, the threeparameter models (Toth and Khan) are investigated.According to Table 5 and Table 6, the values of correlation coefficients (R 2 ) and error functions (MPSD, RMSE and APE) at all temperatures imply that the experimental data fits well with the Toth model.Since the value of parameter 't' at all temperatures is close to number one; thus, the Toth model tends to be Langmuir model, which confirms the homogeneity of the phosphate adsorption process onto Mg-Fe and Ni-Fe sand composites.

Kinetics of adsorption process
The mechanism of phosphate adsorption onto Mg-Fe and Ni-Fe sand composites has been investigated using kinetic models based on surface reactions (pseudo first and pseudo second-order) and particle penetration (intra-particle, Boyd and mass transfer).As shown in Figure 8(a, b), an increase in the initial concentration of phosphate causes the uptake of phosphate to be enhanced.The amount of adsorbed phosphate by Mg-Fe and Ni-Fe sand composites at all initial concentrations increase with increasing contact time, so that in the first 5 min beginning of adsorption, the amount of removed phosphate from the solution reaches over 65.7 and 63.4 mg/g at initial concentration 1000 mg/L for Mg-Fe and Ni-Fe sand composites, respectively; however, after about 5 min, the amount of adsorbed phosphate increases at very low rates until it reaches an equilibrium state at 60 min of contact time.In other words, the phosphate adsorption process for both adsorbents includes two steps; the first step is rapid adsorption due to the electrostatic interaction between the adsorbent molecules (phosphate) and the adsorbent surface (Mg-Fe and Ni-Fe sand composites), and the second step is gradual adsorption with a low rate due to the propagation of the particles dissolved into the Mg-Fe and Ni-Fe sand composites pores to the extent that all active sites occupy.This behaviour is common in mineral adsorbents [110].
The different parameter values of reaction-based kinetic models to find out the adsorption mechanism are represented in Table 7.The pseudo-second order model describes the adsorption process very well.This was clarified by the high correlation coefficient (R 2 > 0.99) for both adsorbents and a good agreement between calculated (q e(cal) ) and experimental (q e(exp) ) adsorptive capacity values.This means that the phosphate adsorption onto Mg-Fe and Ni-Fe sand composites is controlled by the chemisorption interaction [35,52,92,111,112]; accordingly, the initial phosphate concentration and the number of active sites of Mg-Fe and Ni-Fe sand composites surface affect the adsorption process.With the increasing initial concentration of phosphate from 400 to 1500 mg/L, for both adsorbents, the rate constant of pseudo-second order kinetics (K 2 ) is reduced.
To determine the speed controller step in the phosphate adsorption onto Mg-Fe and Ni-Fe sand composites, the kinetic experimental data were fitted by Boyd, intraparticle diffusion, and mass transfer models.
With reference to Table 8, the intra-particle diffusion plot is not linear and does not pass through the origin for both adsorbents, implying that some other mechanisms are involved with intraparticle diffusion.This could be because of the wide distribution of pore size for adsorbents as well [113].To accurately determine the speed controller step in the adsorption process, the kinetic model of Boyd was studied.The plot of B t (Boyd constant) against t (contact time) is not linear and does not pass through the origin but has an intercept for both adsorbents; consequently, the film diffusion can be effective in the adsorption process.
To find the speed controller step of the adsorption process precisely, the determination of the mass transfer resistance was performed.The mass transfer resistance can depend on film diffusion, porous diffusion, or both [58].The variation of the global mass transfer factor [k L a] g , film mass transfer factor [k L a] f , and internal diffusion factor [k L a] d versus C e /C 0 are shown in Figure 9(a-c).With increasing C e /C 0 , the internal diffusion factor and global mass transfer factor decreased while the film mass transfer factor increased.In the process of adsorbing phosphate on Mg-Fe and Ni-Fe sand composites, the internal mass transfer coefficient is greater than the film mass transfer ([k L a] d > [k L a] f ), which indicates the effectiveness of porous diffusion (intra-particle diffusion).

Thermodynamic of adsorption process
Thermodynamic parameters for phosphate adsorption onto Mg-Fe and Ni-Fe sand composites are given in Table 9.Based on the equilibrium constant K d , the values of ΔG were calculated using at 298, 308, 323, and 343 K.The positive amount of enthalpy adsorption changes (ΔH) for both adsorbents indicates that the process of phosphate adsorption onto Mg-Fe and Ni-Fe sand composites are endothermic; hence, increasing the temperature leads to an increase in phosphate adsorption.It is noticeable that with increasing initial concentration of phosphate from 400 to 1500 mg/L, the values of enthalpy adsorption (ΔH) are declined and enhanced for Mg-Fe and Ni-Fe sand composites, respectively.The positive entropy (ΔS) for both adsorbents indicates that the adsorption process is dissociative (non-associative), which could be due to the release of new ions in the solution during phosphate adsorption.Of course, similar to enthalpy changes, with increasing the initial concentrations of phosphate, the values of ΔS for Mg-Fe and Ni-Fe sand composites reduced and increased, respectively.
As seen in Table 9, for Mg-Fe sand composite adsorbent, the values of Gibbs free energy values (ΔG) in the initial concentrations of phosphate 400 and 600 mg/L at all temperatures except 298 K are negative, which means that the phosphate adsorption process is spontaneous in the conditions mentioned above.Whilst by increasing the concentration of phosphate greater than 800 mg/L, the values of ΔG changed completely and the phosphate adsorption process was non-spontaneous.
Moreover, the positive values of ΔG for Ni-Fe sand composite at the initial concentrations of phosphate 400 and 600 mg/L at all temperatures except 343 K demonstrate the nonspontaneous nature of phosphate adsorption.A similar result for adsorption of phosphate onto a chitosan-polymer composite was reported [114].

Evaluation of amounts of adsorbed phosphate and mechanism of the adsorption process
Experimental data showed that the functionalization of activated sand by Mg-Fe and Ni-Fe layered double hydroxides resulted in a significant increase of about 44% and 53% in the phosphate uptake, respectively.As noted in the article, before the functionalization, the sand is activated by acid washing to remove impurities, especially calcite.This treatment creates gypsum (CaSO 4 ) mineral, which also has a good ability to adsorb phosphate [115].Gypsum is irregularly formed on the surface of the sand, as proven by the SEM analysis.The functionalization of activated sand is a noncovalent interaction, and it is mainly based on van der Waals forces.
There are two types of interaction between composite and phosphate: (a) Electro-static interaction between phosphate and gypsum, so that phosphate is trapped by gypsum [124].Phosphor gypsum may also form under certain conditions [125], which it was not observed in these adsorption processes (b) Hydrogen bonding between hydroxyl groups of Mg-Fe LDH and Ni-Fe LDH with oxygen atoms of phosphate molecules.
The involvement of these two interactions in the phosphate adsorption process can vary according to the ionic strength of the solution.

Conclusion
The main objective of this paper is to demonstrate the potential of sand as suitable bedding for loading of metal layered double hydroxide (LDH) for removal of the environmental contamination of soluble phosphate.The maximum adsorption capacity of phosphate by adsorbents functionalised with LDH is related to Mg-Fe and Ni-Fe sand composites.This considerable improvement in phosphate adsorption can be attributed to two main reasons a) Mg-Fe and Ni-Fe layered double hydroxide b) acid washing which can cause Gypsum (CaSO 4 ) mineralisation in the sand.
The investigation of optimum conditions implied that the amount of adsorbent and the pH of the solution are the most important variables affecting the uptake of phosphate.The surface charge of Mg-Fe and Ni-Fe sand composites at pH >7 has changed from positive to negative, thereby reducing the amount of adsorbed phosphate with increasing pH of the solution.Equilibrium studies demonstrate that the phosphate adsorption process occurs homogeneously and single-layer with a uniform distribution of heat adsorption.The phosphate uptake for both adsorbents is controlled by chemisorption interaction.
According to reaction-based kinetic models, the phosphate initial concentration and number of the active sites of adsorbents surface have affected the adsorption process.For diffusion-based models (intra-particle, Boyd, and mass transfer), the effectiveness of porous diffusion is greater than that of film diffusion.Thermodynamic studies illustrate that the temperature increase has a positive influence on phosphate uptake by Mg-Fe and Ni-Fe sand composites; furthermore, the positive values of ΔS indicate that the adsorption of phosphate is non-associative.

Figure 1 .
Figure 1.XRD pattern of raw sand, activated sand, Mg-Fe sand composite and Ni-Fe sand composite.

Figure 2 .
Figure 2. FTIR analysis of raw sand, activated sand, Mg-Fe sand composite and Ni-Fe sand composite.

Figure 6 (
Figure 6(a, b) illustrates the effect of the pH range of 2-12 on the capacity of Mg-Fe and Ni-Fe sand composites to remove phosphate from aqueous solution.The phosphate adsorption capacity increases for both adsorbents with increasing pH up to about seven.In this pH range, the maximum amount of adsorbed phosphate is obtained by both adsorbents.The pH changes for these two adsorbents resulted from the competing surface charge of adsorbents and the dominant phosphorus species PO 4 3-, HPO 4 2-, H 2 PO 4 − , and H 3 PO 4 .According to a literature survey, phosphorus appears in various forms at different pH ranges of the solution.So that at pH >12 PO 4 3-form; at pH = 7-12 a mixture of HPO 4 2-and PO 4 3-; and in the range of pH = 2-7 HPO 4 2-and H 2 PO 4 − compounds are predominant.In addition, H 3 PO 4 is individually observed at pH <2 [17,20,98].In the case of low pH, phosphorus species is H 3 PO 4 which could weakly bound to the adsorption sites of sand composites due to lack of electrostatic force and ion exchange; hence, the low phosphate uptake in low pH is quite reasonable but with increasing pH up to 7, the percentage of phosphorus species HPO 4 2-and H 2 PO 4 − increases and the surface

Figure 7 .
Figure 7. Plots of the pH pzc results for Mg-Fe and Ni-Fe sand composites.

Figure 9 .
Figure 9. Variation of mass transfer factor: (a) [k L a] g , (b) [k L a] f and (c) [k L a] d for phosphate adsorption onto Mg-Fe and Ni-Fe sand composites.

Table 2 .
Chemical analyses of raw and modified sand (XRF)

Table 3 .
The results of X-ray diffraction of raw and modified sand

Table 4 .
Parameters of BET analyses for raw sand, Mg-Fe and Ni-Fe sand composites.

Table 5 .
Isotherm model constants for phosphate adsorption onto the Mg-Fe sand composite

Table 6 .
Isotherm model constants for phosphate adsorption onto Ni-Fe sand composite . As the temperature increases from 298 to 343 K, the amount of adsorbed phosphate increases from 69.47 to 86.75 mg/g and 66.6 to 84.2 mg/g for Mg-Fe and Ni-Fe sand composites, respectively.The parameter b (Langmuir constant) increases from 0.014 L/mg at 298 K to 0.11 L/mg at 343 K, and 0.019 L/mg at 298 K to 0.444 L/mg at 343 K for Mg-Fe and Ni-Fe sand composites, respectively, indicating the positive effect of temperature on phosphate adsorption onto adsorbents.Based on the D-R isotherm model, the amount of free energy (E) increases from 14.3 KJ/mol at 298 K to 15.9 KJ/mol at 343 K for Mg-Fe sand composite, and from 10.63 KJ/mol at 298 K to 15.19 KJ/mol at 343 K for Ni-Fe composites.It denotes that phosphate adsorption process for both adsorbents as the chemical interaction (D-R model: E > 8 kJ mol _1

Table 7 .
Reaction-based model parameters for phosphate adsorption onto Mg-Fe and Ni-Fe sand composite

Table 8 .
Intra-particle diffusion and Boyd kinetic models for phosphate adsorption onto Mg-Fe and Ni-Fe sand composite

Table 9 .
Values of thermodynamic parameters for phosphate removal onto Mg-Fe and Ni-Fe sand composite