Investigation of phosphate adsorption onto ferrihydrite by X-ray Photoelectron Spectroscopy

Journal of Colloid and Interface Science 407 (2013) 95–101

Contents lists available at SciVerse ScienceDirect

Journal of Colloid and Interface Science www.elsevier.com/locate/jcis

Investigation of phosphate adsorption onto ferrihydrite by X-ray Photoelectron Spectroscopy M. Mallet ⇑, K. Barthélémy, C. Ruby, A. Renard, S. Naille Laboratoire de Chimie Physique et Microbiologie pour l’Environnement (LCPME), UMR 7564, CNRS – Université de Lorraine, 405 rue de Vandœuvre, F-54600 Villers-lès-Nancy, France

a r t i c l e

i n f o

Article history: Received 18 February 2013 Accepted 14 June 2013 Available online 1 July 2013 Keywords: Ferrihydrite Phosphate Adsorption X-ray Photoelectron Spectroscopy

a b s t r a c t The objective of this study was to characterize phosphate adsorption onto synthetic 2-lines ferrihydrite using surface analysis by X-ray Photoelectron Spectroscopy and batch experiments. Surface analysis of ferrihydrite samples before phosphate sorption gives very reproducible Fe:O surface ratios of (1:3 ± 0.1). Phosphate sorption onto ferrihydrite was investigated by means of pH, initial phosphate concentration, and ionic strength effects. Additionally, potential background electrolyte influence on phosphate adsorption was also determined. Phosphate uptake by ferrihydrite significantly increases with decreasing pH, with a maximum uptake of 104.8 mg PO4 g1 obtained at pH = 4. Phosphate removal increases with the enhancement of ionic strength in agreement with the formation of inner-sphere complexes. The presence of chloride, nitrate, and sulfate showed no competing effect on phosphate removal efficiency. Sorption kinetics follow a pseudo-second order model (R2 > 0.99) and the Freundlich isotherm model adequately describes sorption (R2 = 0.995). The careful examination of high resolution Fe 2p, O 1s, and P 2p spectra before and after phosphate sorption allows the characterization of the modifications occurring onto the ferrihydrite surface. The binding energy of the P 2p peak agrees well with that observed in Fe-PO4 compounds. Additionally, binding energy shifts in the Fe 2p spectra combined to variations in the relative intensity of the components in the high resolution O 1s spectra illustrate well the formation of chemical bonding between iron and phosphate anions at the ferrihydrite surface. Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction Phosphorus is well known to be an essential nutrient for the growth of microorganisms in aquatic environments. However, excessive supply of orthophosphate (denoted phosphate in the present study) to all water bodies through agricultural, industrial, and household activities has been identified to be at the origin of alteration of nutrient ratios and thus eutrophication of lakes, rivers, and sea [1]. For example, dramatic proliferation of blue-green algae in coastal waters is one of the typical manifestations of eutrophication caused by excessive phosphorus concentration [1]. In a general way, the mobility of contaminants in groundwater systems is well known to be strongly dependent on the interaction developed with iron and aluminum based materials which are commonly encountered in soils and waterways. In particular, iron oxides display strong affinity for the sequestration of transition metals [2,3] and anions such as arsenate (AsO2 4 ) [4,5], chromate 3 (CrO2 4 ) [4,6], and phosphate (PO4 ) [7–9]. The environmental abundance of iron oxides combined to their low cost, their easy availability, and the lack of human toxicity make them excellent

⇑ Corresponding author. Fax: +33 3 83275444. E-mail address: [email protected] (M. Mallet). 0021-9797/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2013.06.049

candidates in contaminant removal strategies. Many studies are focused on the goethite mineral (a-FeOOH), the most common and most stable iron oxyhydroxide in natural environments. The main reason is that synthesis of goethite is now perfectly controlled in terms of size particles and geometry which makes this mineral a reference compound for numerous studies. Conversely, ferrihydrite is a poorly ordered iron oxide also commonly encountered in geochemical processes [10,11]. Ferrihydrite is a precursor to other iron minerals such as goethite and hematite (a-Fe2O3) and is thought to play a significant role in geochemical processes [12] in view of its high surface area [10,13]. However, ferrihydrite identification and characterization are difficult in view of its low crystallinity and small particles size [11]. In addition, at the laboratory scale, slightly different products may be obtained depending on the synthesis conditions which may affect the mineral reactivity and make the comparison of the literature data challenging [14]. In any way, the experimental approaches used in the previous studies relative to phosphate sorption by iron oxides involved thermodynamic modeling of sorption data and potentiometric titrations [13], Infrared IR spectroscopic investigations [5,15,16], and X-ray Absorption Near Edge Structure (XANES) investigations [9,17–19]. X-ray Photoelectron Spectroscopy (XPS) is a surface sensitive method that has been widely used to investigate both the electronic structure and the bonding of molecules in a wide range

96

M. Mallet et al. / Journal of Colloid and Interface Science 407 (2013) 95–101

of iron oxides [20,21]. However, there is only few studies that report the use of XPS to characterize phosphate sorption and bonding onto the surface of iron oxides [22,23]. The aim of the present study is thus to characterize phosphate sorption at the ferrihydrite surface using bulk solution data combined with XPS information. To the best of the authors’ knowledge, such a careful XPS examination of physicochemical modification occurring at the ferrihydrite surface has never been reported in the literature before. 2. Materials and methods 2.1. Ferrihydrite synthesis Two-line ferrihydrite (Fh) samples were synthesized using a slightly modified procedure from that reported by Cornell and Schwertmann [10]. Chloride (FeCl38H2O, Sigma–Aldrich, 98%) was used as the source of ferric salt. Briefly, the ferric solution was prepared by dissolving ferric chloride in demineralized water to obtain a 0.8 M initial concentration. The pH of the solution was raised to 7.5 using 3 M NaOH and maintained to this pH value for 3 h under vigorous stirring. The ferrihydrite precipitates were then centrifuge-washed until the surface chloride atomic concentrations determined by XPS were reduced below 3 at.%. The pastes were then dried at room temperature and crushed in a porcelain mortar to get pulverulent compounds. XRD pattern (not shown) revealed two broad peaks with interplanar spacings of about 0.26 and 0.15 nm which unambiguously confirmed the presence of two-line ferrihydrite particles [10]. The specific surface area of synthetic ferrihydrite determined by single point BET method was 309 m2 g1 in agreement with commonly reported values in the literature for the BET method [10]. 2.2. Phosphate adsorption experiments All glassware and plastic bottles used in the experiments were first cleaned in concentrated HCl and washed with demineralized water and ethanol to avoid any trace analysis. The pH of the suspensions was maintained at a fixed value by the addition of a minimum volume of 0.1–2 M NaOH or HCl at constant time intervals. A 0.1 M NaCl supporting electrolyte was used unless otherwise stated. Samples were stirred on an orbital shaker at room temperature during a 24 h reaction period determined from preliminary experiments and in agreement with literature data [9,13,15]. Experiments were carried out at pH = 7 unless otherwise specified, in agreement with municipal wastewater pH that typically ranges from 6.5 to 7.3. Samples were at different reaction times (0, 10, 20, 35, 50, 70, 105, 135, 180, 240, 360, 500, and-1400 min for kinetics experiments; 0 and 1440 min for isotherms and pH effect experiments), withdrawn from the reaction vessels and filtered through 0.22 lm polypropylene syringe filters before phosphate determination. All experiments were carried out in duplicate, and the mean values are reported. Phosphate adsorption was investigated at different ionic strengths (70, 200, 3000 mg L1 NaCl), using a 1 g L1 ferrihydrite concentration and a 70 mg L1 phosphate concentration (NaH2PO42H2O Aldrich, 99%). The effect of pH was examined using a 200 mg L1 phosphate concentration and a 2 g L1 ferrihydrite concentration while maintaining pH at different values ranging from 4.0 ± 0.1 to 9.0 ± 0.1. The studies of the influence of initial phosphate concentration and the sorption kinetics were carried out at a 1 g L1 ferrihydrite concentration and phosphate concentrations in the range [53–215] mg L1. Phosphate adsorption isotherm experiments were determined at a ferrihydrite concentration of 2 g L1 and initial phosphate concentrations in the range

50–500 mg L1. Finally, the potential competitive effect of background electrolyte anions (NaNO3, Na2SO4, and NaCl solutions) on phosphate adsorption onto ferrihydrite was examined by using initial phosphate, nitrate, and sulfate anions concentrations of 70 mg L1 and chloride anions concentrations in the range of 0– 3000 mg L1 at a ferrihydrite concentration of 1 g L1. 2.3. Ion chromatography The concentrations of phosphate in solution were determined by ion chromatography (IC). A Metrosep A Supp 5-250 column packed with polyvinyl alcohol particles functionalized with quaternary ammonium group (5 lm particles diameter) was used for the separation. The mobile phase consisted of a mixture of 3.2 mM Na2CO3 (Sigma–Aldrich, 99.5+%) and 1 mM NaHCO3 (Sigma–Aldrich, 99.7+%) in ultrapure water (18.2 mX cm at 293 K). The flow rate was 0.7 mL min1 and the sample loop volume was 20 lL. The duplicate experiments demonstrated the high repeatability of the IC method and the experimental error could be controlled within 1–3%. The removal efficiency [PO4ads] (%) was determined by the following equation:

½PO4ads  ð%Þ ¼

ðC i  C t Þ  100 Ci

ð1Þ

where Ci and Ct are the initial and at time t phosphate concentrations (mg L1), respectively. Additionally, phosphate uptake (qe, mg g1) by ferrihydrite at near equilibrium conditions was determined as follows:

qe ¼

ðC i  C e Þ  V m

ð2Þ

where Ce is the equilibrium phosphate concentration (mg L1), V is the solution volume (L), and m is the dried weight (g) of ferrihydrite used. 2.4. X-ray Photoelectron Spectroscopy (XPS) The surface physicochemical properties of ferrihydrite samples were examined by XPS. Spectra were obtained with a KRATOS Axis Ultra X-ray photoelectron spectrometer (Kratos Analytical, Manchester, UK) equipped with a monochromated Al Ka X-ray (hm = 1486.6 eV) operated at 150 W. The samples were pressed onto a Cu tape fixed on a holder and introduced into the spectrometer. The base pressure in the analytical chamber was about 109 mbar. Spectra were collected at normal take-off angle (90°), and the analysis area was 700  300 lm2. Wide scans were recorded using an analyzer pass energy of 160 eV and narrow scans using a pass energy of 20 eV. Ag 3d5/2 full width at half maximum (FHWM) was determined to be 0.75 eV under these recording conditions. Charge correction was carried out using the C 1s core line, setting adventitious carbon signal (H/C signal) to 284.6 eV. Spectra of iron, oxygen, and phosphorus were fitted using a Shirley background and a Gaussian/Lorentzian (70/30) peak model. The core levels are reported from the peak maxima and have an experimental error of ±0.05 eV. 3. Results and discussion 3.1. Removal of phosphate in batch experiments 3.1.1. Ionic strength effect on phosphate adsorption The phosphate adsorption increases with ionic strength under the conditions of the experiments (Fig. 1). The effect of ionic strength is usually correlated with the type of surface complexes that are involved [13,15]. In fact, ions that form outer-sphere com-

97

M. Mallet et al. / Journal of Colloid and Interface Science 407 (2013) 95–101

80

Phosphate removal (%)

70 60 50 40 30 -1

c°(NaCl) = 3133 mg L -1 c°(NaCl) = 201 mg L -1 c°(NaCl) = 70 mg L -1 c°(NaCl) = 0 mg L

20 10 0 0

200

400

600

800

1000

1200

1400

Time (min) Fig. 1. Time dependence of phosphate removal by ferrihydrite for different concentrations of sodium chloride (pH = 7; [PO4] = 70 mg L1; [Fh] = 1 g L1; [NaCl] = 0.1 M).

plexes via electrostatic interaction compete with the supporting electrolyte ions for adsorption on surface sites. Decreasing phosphate adsorption is thus generally observed with increasing ionic strength. On the other hand and depending on the pH of the solution, no or less competition is involved when ions form innersphere complexes that are directly coordinated to surface groups via ligand exchange. As a result, less influence of the anions of the electrolyte support is observed [13,15]. It is important to note that increasing phosphate adsorption with increasing ionic strength has been reported in many cases of inner-sphere coordination [15,24,25]. Such a typical behavior is interpreted with the modification of the electrical potential at the interface, which decreases the electrostatic repulsion between the charged surface and the ion and thus favors the adsorption process. Our results thus probably reflect the formation of inner-sphere complexes in agreement with the literature data on ferrihydrite [13,15]. 3.1.2. Effect of pH Phosphate sorption onto metal oxides occurs via electrostatic attraction and/or ligand exchange and pH is one of the important factors that affect the sorption behavior. A strong pH dependency of phosphate sorption onto ferrihydrite is evidenced in the range

of 4–9 (Fig. 2). Indeed, the adsorbed amount continuously decreases with increasing pH. Such a behavior is well known for the adsorption of anions onto iron oxides and in particular is well documented for hematite [26], ferrihydrite [13], goethite [27], and ferric green rust [28]. In fact, the surface charge becomes more negative as the pH increases, thus increasing the repulsion between the phosphate anions and the surface sites. The fresh ferrihydrite usually exhibits a point of zero charge (pzc) around 8.0 ± 0.2 [13]. It should be noted here that the phosphate removal (Fig. 2) remains high at pH values close and above the pzc where the surface becomes negatively charged. Thus, the adsorption process may not be due to electrostatic attraction only and requires the implication of inner-sphere surface complexes between phosphate species and surface groups. From Fig. 2, a maximum sorption uptake of 104.8 mg PO4 g1 is thus obtained at pH = 4. 3.1.3. Adsorption kinetics Kinetic study was carried out to highlight the effect of various initial concentrations on the phosphate removal rate. Pseudo-first order, pseudo-second order, and Elovich models were tested. The pseudo-first order and Elovich models did not adequately fit the adsorption of phosphate onto the ferrihydrite samples (data not shown). The equation of the pseudo-second order model is given by Ho and McKay [29]:

dqt ¼ kðqe  qt Þ2 dt

ð3Þ

where qt (mg g1) is the amount of phosphate adsorbed at time t (min) and k is the apparent rate constant of sorption (g mg1 min1). The integration of Eq. (3) leads to:

qt ¼

t

ð4Þ

2

1=kqe þ t=qe

which can be expressed in the linearized form (5) as:

t 1 1 ¼ þ t qt kq2e qe

ð5Þ

The kinetic data were fitted according to Eq. (5), using the linear least squares fitting technique. The corresponding plot of t/qtvs t is shown in Fig. 3, and the calculated kinetic parameters are given in Table 1. Correlation coefficients are above 0.99 for every initial phosphate concentration studied. These results indicated that the adsorption system belongs to the pseudo-second order kinetic

35

100 30 25

-1

q e (mg g )

80

t/qt

60

20 15

40 10 -1

20

c°(phosphate) = 53 mg L -1 c°(phosphate) = 72 mg L -1 c°(phosphate) = 215 mg L

5 0

0 4

5

6

7

8

9

pH Fig. 2. Effect of pH on phosphate adsorption onto ferrihydrite ([Fh] = 2 g L1; [PO4] = 200 mg L1 [NaCl] = 0.1 M).

0

200

400

600

800

1000

1200

1400

Time (min) Fig. 3. Plot of t/qt vs t for adsorption of phosphate onto ferrihydrite (pH = 7; [Fh] = 2 g L1, [NaCl] = 0.1 M).

98

M. Mallet et al. / Journal of Colloid and Interface Science 407 (2013) 95–101

Table 1 Kinetic parameters for phosphate adsorption onto ferrihydrite (pH = 7; [Fh] = 1 g L1). Phosphate initial concentration (mg L1)

53 72 215

Pseudo-second order kinetic parameters k (g mg1 s1)

qe,cal (mg g1)

R2

4.0  104 5.5  104 2.3  104

40.3 41.4 67.1

0.999 0.999 0.999

qe,exp (mg g1)

38.9 40.5 64.6

model. The values of k are in the same order of magnitude except for an initial phosphate concentration of 215 mg L1. The lower obtained value may reflect a decrease in the sorption rate. Calculated and experimental qe values agree well with a maximum value of 69 mg g1 obtained for the highest initial phosphate concentration (Table 1). According to the pseudo-second order model, and as stated by Ho and McKay [29], the phosphate adsorption onto ferrihydrite leads to a chemisorption process involving valency forces through sharing or exchange of electrons between sorbent and sorbate. 3.1.4. Adsorption isotherm Sorption isotherm was determined to assess phosphate distribution between solid and aqueous phases as a function of adsorbed concentration. Only the Langmuir and Freundlich isotherm models were tested for their applicability, those being commonly used to describe phosphate equilibrium sorption onto solids. Langmuir model did not adequately fit the experimental data (not shown). Freundlich equation is expressed as follows:

qe ¼ KC ne

which reflects a better adsorption capacity. The value of the n exponent, around 0.2, indicates that the adsorption is favorable for these experimental conditions as it is generally stated in the literature for n values in the range 0.1–0.5 [31]. 3.1.5. Effect of competitive anions The interference of coexisting anions onto the sorption process was evaluated by conducting experiments with various competing 2 anions, i.e., Cl, NO 3 and SO4 . From Table 2, it appears that the presence of coexisting chloride at the same level than phosphate (experiment 2) exerts no influence onto phosphate sorption. Equimolar concentrations of Cl, 2 NO 3 and SO4 as that of the phosphate solution (experiment 3) also do not show any influence onto phosphate removed from solution.

Table 2 Influence of coexisting anions on the percentage of phosphate removal (pH = 7; [Fh] = 1 g L1). Anions initial concentrations (mg L1)

Experiment 1 Experiment 2 Experiment 3 Experiment 4 Experiment 5

Phosphate removal (%)

Phosphate

Chloride

Nitrate

Sulfate

70

None

None

None

65.0

70

70

None

None

63.0

70

70

70

70

68.0

70

160

None

None

71.4

70

3000

None

None

80.4

ð6Þ

The Freundlich linear form of this equation is therefore: 35

logðqe Þ ¼ logðKÞ þ n logðC e Þ

ð7Þ 30

1

Chloride atomic percentage

n1

where K represents the Freundlich affinity coefficient (mg n Ln g1 ) and n is the Freundlich linearity constant. Fig. 4 shows that phosphate sorption onto ferrihydrite sample is well-described using the Freundlich model (R2 > 0.99). The adsorption process thus corresponds to a heterogeneous system which is not restricted to the formation of a monolayer. The value of K, 1 n1 26:9 mg n Ln g1 , is higher than the values obtained for goethite 1 1 n1 n1 (K ¼ 17:3 mg n Ln g1 ) and akaganeite (K ¼ 10 mg n Ln g1 ) [30],

25 20 15 10 5 0 0

5

10

15

20

Number of cycles Fig. 5. Influence of the number of washing/centrifugation cycles on the percentage of chloride at the ferrihydrite surface as determined by XPS.

Table 3 Surface composition (at.%) of phosphate-reacted ferrihydrite and corresponding Fe:O ratios. Experimental conditions Initial phosphate concentration (mg L1)

Fig. 4. (a) Sorption isotherm of phosphate anions on ferrihydrite (b) relationship between log(qe) vs log(Ce) to verify the Freundlich model. Symbols: experimental data; lines: Freundlich isotherms model fitting (pH = 7; [Fh] = 2 g L1; [PO4] = 50– 500 mg L1; [NaCl] = 0.1 M).

Pristine surface 60 200 500 2000

Atomic percent of the elements (%) pH

Fe

O

Cl

P

Fe:O

7 7 7 4

23.6 23.0 21.3 19.9 16.3

75.6 73.7 74.1 75.2 75.8

0.8 0.6 0.8 0.4 n.d.

n.d. 2.7 3.8 4.5 8.0

1:3.2 1:3.2 1:3.5 1:3.8 1:4.7

99

M. Mallet et al. / Journal of Colloid and Interface Science 407 (2013) 95–101

(a)

Fe 2p 3

6,0x10

3

4,0x10

3

2,0x10

740

730

720

710

700

Binding energy (eV)

(b)

3

8,0x10

O 1s

2

9x10

(c)

P 2p 2

3

8x10

3

7x10

3

6x10

6,0x10

2

4,0x10

2

2,0x10

0,0

2

540

535

530

525

5x10

140

138

Binding energy (eV)

136

134

132

130

Binding energy (eV)

Fig. 6. (a) Fe 2p, (b) O 1s, and (c) P (2p) spectra of ferrihydrite samples reacted with 0 (black crosses), 60 (blue crosses) and 200 (red crosses) mg L1 phosphate solutions (pH = 7; [Fh] = 1 g L1). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

However, if the chloride concentrations were about 2.3 (experiment 4) and about 41.7 (experiment 5) times of the phosphate one, a marked increase in the phosphate adsorption by ferrihydrite is evidenced.

reproducibility of the Fe:O ratio of 1:3(± 0.1) obtained in our study. Additionally, the O2/OH surface ratio of initial ferrihydrite samples was always remarkably constant and equals to 3:2. However, this ratio does not correspond to those expected from the above formula of ferrihydrite [32–34].

3.2. Surface analysis by XPS 7000

710.9 eV

6000 5000 4000 3000 2000 1000

(a) 740

730

720

710

7000

Intensity (counts)

3.2.1. Surface properties of ferrihydrite sorbent XPS was considered as a powerful tool to characterize the surface properties of ferrihydrite. Survey spectra (not shown) indicated not surprisingly the presence of Fe, O, adventitious carbon and variable amounts of chloride, depending on the number of washing/centrifugation cycles. Fig. 5 clearly shows that chloride becomes very limited at the ferrihydrite surface after three washing/centrifugation cycles. However it was not possible to definitively eliminate chloride from the surface even after twenty washing/centrifugation cycles. Additionally, a remarkably constant Fe:Ototal ratio (at%) of 1:(3.0 ± 0.1) was determined from the analysis of five ferrihydrite samples. It is important to recall here that the composition and the structure of ferrihydrite are still under discussion. In fact, several formulae have been proposed in the past: 5Fe2O39H2O [32], Fe5HO84H2O [33], Fe2O3 2FeOOH2.6H2O [34], with 5Fe2O39H2O being the most generally reported formula in the literature. However, the presence of structural OH in the ferrihydrite structure was clearly evidenced by IR spectroscopy [34]. In addition, a goethite-like structure is now accepted in the literature [11] which makes the commonly Fe2O3 formula reported for ferrihydrite inappropriate unless for natural samples containing an important amount of hematite. In any way, all these ferrihydrite compositions correspond to a bulk Fe:Ototal ratio of 1:2.4. The experimental Fe:Ototal ratio derived from our experiments is thus only slightly lower than that expected from the above formula. More recently, bulk composition of the 2-line form of ferrihydrite prepared from nitrate solutions has been reported either as Fe4(O,OH,H2O)12 or as Fe6(O,OH,H2O)12 with an Fe:Ooxide ratio ranging from 1:3 to 1:2 [11] which may be rewritten as Fe:Ototal ratios ranging from 1:9 to 1:6; the surface ratios determined in the present study are thus significantly different. This finding may neither be attributed to the nature of the salt used for ferrihydrite synthesis nor to some surface contamination of our samples in view of the

711.4 eV

6000 5000 4000 3000 2000

(b)

1000 740

730

720

710

7000

712.2 eV

6000 5000 4000 3000 2000 1000 740

(c) 730

720

710

Binding energy (eV) Fig. 7. Fe 2p spectra of (a) initial ferrihydrite, (b) ferrihydrite reacted with a 2000 mg L1 phosphate solution at pH = 4, and (c) FePO42H2O reference sample.

M. Mallet et al. / Journal of Colloid and Interface Science 407 (2013) 95–101

3.2.2. Modification at the ferrihydrite surface after reaction with phosphate solutions Elemental composition of ferrihydrite samples reacted with phosphate solutions at 0, 60, 200, and 500 mg L1 at pH 7 is reported in Table 3. Because batch experiments (Fig. 2) clearly revealed that maximum sorption uptake occurs at pH = 4, additional data obtained at pH = 4 and for a high initial phosphate concentration are also reported in this table. It is important to note that, although these conditions are not commonly encountered in environmental conditions, they make the modification at the ferrihydrite surface easier to observe as it will be discussed later. Data from Table 3 were recalculated to a contaminant-free basis. Residual minor amounts of chloride are observed in all experiments as discussed in the previous section. The amount of phosphorus slightly increases with the initial phosphate solution concentration at pH = 7 and corresponding Fe:O ratios decrease from 1:3.2 to 1:3.8, thus reflecting the sorption of phosphate anion. In agreement with batch experiments, a higher amount of phosphorus is detected at the sample surface at pH = 4, and the corresponding Fe:O ratio is significantly lowered compared to values obtained at pH = 7. Characteristic Fe 2p, O 1s, and P 2p high resolution spectra obtained for ferrihydrite before and after phosphate adsorption at pH = 7 are reported in Fig. 6. Fe 2p spectra display a maximum at 711.2 eV for 2p3/2 core level and a broad satellite peak at 719 eV (Fig. 6a) as is commonly observed for Fe(III) species in iron (oxyhydr)oxides [20]. The careful examination of the position of the 2p3/2 peak seems to indicate a very slight shift of this peak toward higher binding energy with increasing initial phosphate concentration. This assumption will be confirmed later with the

(a)

133.6 eV

900 800 700 600 500 400 140 4000 3500 3000

138

136

134

(b)

132

130

128

130

128

133.6 eV

2500 2000 1500 1000 500 140 1800 1600 1400 1200 1000 800 600 400 200 0 140

138

136

134

(c)

132

132.9 eV

138

136

134

132

130

128

Binding energy (eV)

529.7 eV

531.1 eV

533 eV

531.3 eV

(b) Intensity (counts)

1000

Fig. 9. P 2p spectra of (a) ferrihydrite sample reacted with a 2000 mg L1 phosphate solution at pH = 4, (b) FePO42H2O reference sample, and (c) NaH2PO42H2O reference sample.

(a)

529.9 eV

533.1 eV

(c)

531.5 eV

532.7 eV

540

1100

Intensity (counts)

100

535

530

525

Binding energy (eV) Fig. 8. O(1s) spectra of (a) initial ferrihydrite, (b) reacted with a 2000 mg L1 phosphate solution at pH = 4, and (c) FePO42H2O reference sample.

results obtained at pH = 4. O 1s spectra (Fig. 6b) clearly display a decrease in the peak intensity at 529.9 eV with concomitant peak intensity increase at 531.2 eV at increasing initial phosphate concentrations. Finally, the P 2p spectrum (Fig. 6c) displays a peak maximum at 133.6 eV, the signal intensity increasing with initial phosphate concentration. As just previously mentioned, the signal features observed at pH = 7 (Fig. 6) are expected to be enhanced at pH = 4, and the modifications at the ferrihydrite surface may be easier to interpret. Note that it was checked that the same conclusions apply to these two pH values. Figs. 7 and 8 thus compare the Fe 2p and O 1s high resolution spectra of ferrihydrite before and after phosphate adsorption at pH = 4 (Table 3) to those obtained from a ferric phosphate FePO42H2O reference sample (Sigma–Aldrich). Additionally, the P 2p high resolution spectrum for the chloride ferrihydrite sample obtained at pH = 4 is compared to that obtained from (i) a FePO42H2O reference sample and (ii) the NaH2PO42H2O salt initially used to prepare phosphate solutions (Fig. 9). It is obviously not possible to discriminate iron in ferrihydrite (Fig. 7a) from iron in a phosphate complex (Fig. 7c), in view of the small amount of phosphate adsorbed combined to close binding energies. However, the assumption according to which the 2p3/ 2 core level shifts toward higher binding energies after phosphate sorption is clearly confirmed here (Fig. 7b), and this shift agrees well with the formation of Fe–O–P bonding (Fig. 7c). O 1s spectra of ferrihydrite samples were fitted with 3 peaks at 529.9 eV, 531.3 eV, and 533.1 eV corresponding to O–Fe bonding, OH or OH + O–P bonding, and adsorbed water, respectively. Note that the binding energies of structural OH in ferrihydrite (Fig. 8a) and O–P in iron phosphate (Fig. 8c) are too close to separate their relative contribution. The component at 531.3 eV (Fig. 8b) reflects both the presence of structural OH in ferrihydrite and the O–P bonding. The increase in this component thus undoubtedly reflects

M. Mallet et al. / Journal of Colloid and Interface Science 407 (2013) 95–101

€ et al. phosphate sorption at the sample surface (Figs. 6 and 8). Lu [22] made the assumption according to which the stoichiometric ratio of surface hydroxyl between the original adsorbent and the phosphate-loaded adsorbent allows to investigate the nature of the inner-sphere complex involved in the sorption of phosphate onto metal oxides. In fact, ratios close to 0.5 and 2 are expected for monodentate and bidentate complex, respectively. Unfortunately, overlapping binding energies for O–P and O–H components (Fig. 8) prevented the reliable determination of this ratio in the present study. Finally, the binding energy of the P 2p peak located at 133.6 eV in the ferrihydrite samples reacted with the phosphate solution (Fig. 6c and Fig. 9a) is very close to that observed for the FePO42H2O reference sample (Fig. 9b). Of particular interest is that the binding energy of the P 2p peak in NaH2PO42H2O reference salt is located at a significantly lower binding energy, i.e., 132.9 eV (Fig. 9c). The simultaneous analysis of P 2p, O 1s, and Fe 2p spectra thus gives spectroscopic evidence for phosphate sorption onto the sample ferrihydrite surface and highlights the formation of an iron phosphate complex. 4. Conclusions Phosphate adsorption mechanisms onto ferrihydrite were examined from the combination of bulk and surface analyses. The adsorption process is enhanced by increasing ionic strength and decreasing pH. A maximum phosphate uptake of around 105 mg PO4 g1 is thus determined at pH = 4. Of particular interest is that it remains high at neutral pH, i.e., 77.8 mg PO4 g1 at pH = 7. In fact, pH = 7 is more representative of the water treatment conditions, municipal wastewater pH normally ranging from 6.5 to 7.3. The modification of the surface properties of ferrihydrite after phosphate adsorption is well illustrated from XPS analyses from the interpretation of high resolution Fe 2p, P 2p, and O 1s spectra. The O 1s spectrum of ferrihydrite and in a general way of iron oxides is very sensitive to the adsorption of oxygen rich species. O 1s spectrum thus consists in an effective molecular probe for sorption processes. Additionally, the P 2p spectrum gives evidence for the formation of Fe–O–P bonding after phosphate sorption. To the best of our knowledge, such an in depth study using XPS has never been reported before in the literature. Acknowledgments The Agence Nationale de la Recherche (ANR program, ECOTECH2009 – No. 0994C0103) is gratefully acknowledged for

101

financial support. K. Barthélémy expresses his sincere gratitude to SAUR for the Ph.D. position grant. The authors also acknowledge Dr. C. Despas (LCPME, UMR 7564 CNRS – Université de Lorraine) for helping in the ionic chromatography. References [1] C. Trepanier, S. Parent, Y. Comeau, J. Bouvrette, Water Res. 36 (2002) 1007. [2] S.L.S. Stipp, M. Hansen, R. Kristensen, M.F. Hochella, L. Bennedsen, K. Dideriksen, T. Balic-Zunic, D. Leonard, H.J. Mathieu, Chem. Geol. 190 (2002) 321. [3] P.J. Swedlund, J.G. Webster, G.M. Miskelly, Appl. Geochem. 18 (2003) 1671. [4] S. Fendorf, M.J. Eick, P. Grossl, D.L. Sparks, Environ. Sci. Technol. 31 (1997) 315. [5] Y.F. Jia, L.Y. Xu, X. Wang, G.P. Demopoulos, Geochim. Cosmochim. Acta 71 (2007) 1643. [6] L. Legrand, A. El Figuigui, F. Mercier, A. Chausse, Environ. Sci. Technol. 38 (2004) 4587. [7] O. Bastin, F. Janssens, J. Dufey, A. Peeters, Ecol. Eng. 12 (1999) 339. [8] N. Boujelben, J. Bouzid, Z. Elouear, A. Feki, F. Jamoussi, A. Montiel, J. Hazard. Mater. 151 (2008) 103. [9] K.C. Ruttenberg, D.J. Sulak, Geochim. Cosmochim. Acta 75 (2011) 4095. [10] R.M. Cornell, U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrences and Uses, Wiley-VCH GmbH & Co. KGaA, Weinheim, 2003. [11] J.L. Jambor, J.E. Dutrizac, Chem. Rev. 98 (1998) 2549. [12] V. Barron, J. Torrent, E. de Grave, Am. Mineral. 88 (2003) 1679. [13] J. Antelo, S. Fiol, C. Perez, S. Marino, F. Arce, D. Gondar, R. Lopez, J. Colloid Interface Sci. 347 (2010) 112. [14] A. Navrotsky, L. Mazeina, J. Majzlan, Science 319 (2008) 1635. [15] Y. Arai, D.L. Sparks, J. Colloid Interface Sci. 241 (2001) 317. [16] P. Persson, N. Nilsson, S. Sjoberg, J. Colloid Interface Sci. 177 (1996) 263. [17] N. Khare, D. Hesterberg, J.D. Martin, Environ. Sci. Technol. 39 (2005) 2152. [18] N. Khare, J.D. Martin, D. Hesterberg, Geochim. Cosmochim. Acta 71 (2007) 4405. [19] Y.T. Liu, D. Hesterberg, Environ. Sci. Technol. 45 (2011) 6283. [20] A.P. Grosvenor, B.A. Kobe, N.S. McIntyre, S. Tougaard, W.N. Lennard, Surf. Interface Anal. 36 (2004) 632. [21] M. Mullet, Y. Guillemin, C. Ruby, J. Solid State Chem. 181 (2008) 81. € , H. Liu, R. Liu, X. Zhao, L. Sun, J. Qu, Powder Technol. 233 (2013) 146. [22] J. Lu [23] A. Zach-Maor, R. Semiat, H. Shemer, J. Colloid Interface Sci. 363 (2011) 608. [24] J. Antelo, M. Avena, S. Fiol, R. Lopez, F. Arce, J. Colloid Interface Sci. 285 (2005) 476. [25] R. Rahnemaie, T. Hiemstra, W.H. van Riemsdijk, Langmuir 23 (2007) 3680. [26] G. Horányi, P. Joó, J. Colloid Interface Sci. 247 (2002) 12. [27] C. Luengo, M. Brigante, J. Antelo, M. Avena, J. Colloid Interface Sci. 300 (2006) 511. [28] K. Barthélémy, S. Naille, C. Despas, C. Ruby, M. Mallet, J. Colloid Interface Sci. 384 (2012) 121. [29] Y.S. Ho, G. McKay, Process Biochem. 34 (1999) 451. [30] R. Chitrakar, S. Tezuka, A. Sonoda, K. Sakane, K. Ooi, T. Hirotsu, J. Colloid Interface Sci. 298 (2006) 602. [31] R.E. Treybal, Mass-transfer Operations, third ed., McGraw-Hill Book Company, New York, 1980. [32] M. Fleischer, G.Y. Chao, L.J. Cabri, Am. Mineral. 60 (1975) 736. [33] K.M. Towe, W.F. Bradley, J. Colloid Interface Sci. 24 (1967) 384. [34] J.D. Russell, Clay Miner. 14 (1979) 109.