Preparation of Mesoporous Alumina and In Situ ATR-FTIR Investigation of Phosphate at the Alumina-Water Interface

2008:047

M A S T E R ' S T H E S I S

Preparation of Mesoporous Alumina and In Situ ATR-FTIR Investigation of Phosphate at the Alumina-Water Interface

Tingting Zheng

Luleå University of Technology Master Thesis, Continuation Courses Chemical and Biochemical Engineering Department of Chemical Engineering and Geosciences

Division of Chemistry

Abstract...2

1. Introduction...3

1.1 Background ...3

1.2 ATR-FTIR (Attenuated Total Reflection Fourier Transform Infrared ) spectroscopy...3

2. Materials and methods ...6

2.1 Mesoporous alumina preparation...6

2.2 General characterization of mesoporous alumina powder ...6

2.3 ATR-FTIR experiments...7

2.3.1 Prepare of deposit thin alumina film ...7

2.3.2 Flow cell experiments ...7

2.3.2.1 PH envelope ...7

2.3.2.2 Adsorption isotherms ...8

2.3.2.3 Adsorption kinetics...8

3. Results and discussion ...9

3.1 General characterization...9

3.1.1 XRD pattern ...9

3.1.2 SEM images ...10

3.1.3 Zeta potential charges...10

3.1.4 N

adsorption/desorption analysis...12

3.2 ATR-FTIR spectra ...13

3.2.1 Species distribution and ART-FTIR spectra of aqueous H

PO

at different pH .13 3.2.2 PH envelope spectra...15

3.2.3 Adsorption isotherms ...17

3.2.4 Adsorption kinetics ...21

3.2.5 Curve fitting at different pH and phosphate concentration ...22

3.2.6 Possible molecular configurations ...23

4. Conclusions...24

References...25

Acknowledgement ...26

Abstract

The aim of this study was to synthesize aluminum oxide with high surface area and then investigate phosphate adsorption on the aluminum oxide surface. The method we used to probe phosphate surface complexes at aluminum oxide phase was in situ Attenuated Total Reflection Fourier Transform Infrared (FTIR) spectroscopy, which is very sensitive to the coordination environment and protonation state of phosphate complexes, and therefore a useful tool for characterizing phosphate surface species at the molecular scale.

Mesoporous alumina material had been synthesized using PEG as structure directing reagent and dodecylamine as co-template. XRD, SEM, N

adsorption/desorption technique and zeta potential charge measurement were used to characterize the general properties of the synthesized material. The results showed that the aluminum oxide material is λ-Al

O

with a high surface area viz. 329 m

g

. The deposited alumina on the crystal surface was a film assembled with small alumina particles about 20-30 nm and the thickness of the film was around 3-4 μm.

The zeta potential was affected by electrolyte concentration and the adsorbed phosphate on the surface of the alumina film could significantly influence the charge as well as the PZC value and make PZC to occur at pH values less than that without phosphate.

Phosphate adsorption on alumina surface was investigated by in situ ATR-FTIR

spectroscopy. Spectroscopic studies on these systems have shown that phosphate

forms inner-sphere complexes at the solid–liquid interface. In higher pH value (9.0)

only one complex exist in the water-alumina interface, however, a mixture of different

phosphate complexes formed at lower pH value (4.05 and 4.96). The adsorption

isotherms and kinetics at acid and base environment were also studied. The results

show that the amount of phosphate adsorbed on aluminum oxide increased with

increasing concentration but the increase gradually became smaller at higher

concentrations. The adsorption data were evaluated with the Langmuir and Freundlich

isotherm models. It was indicated that the adsorption data fits the Langmuir isotherm

better than the Freundlich isotherm at lower [PO

] concentration, however, it fits the

Freundlich isotherm better than the Langmuir isotherm at higher [PO

] concentration

both at pH 9.0 and 4.15. Adsorption kinetics show that both adsorption at 9.0 and 4.96

have similar characteristics except at 5μM, with a fast adsorption between t = 0 and

10 min, and a slower adsorption at longer times. Finally, seven possible molecular

symmetries are proposed for this Al-P complex with C

or lower symmetry.

1. Introduction

1.1 Background

Phosphate (P) is of major concern in environmental chemistry. It is essential for plant growth in soils and has been recognized as one of the main nutrients that controls eutrophication in surface water bodies [1, 2]. Its transport and fate in soils and aquifers must be well understood to better evaluate its environmental impact. It is well known that the mobilization of phosphate in the environment is markedly influenced by mineral surfaces. Adsorption at mineral surfaces determines the quantity of phosphate that is retained in the solid phase of soils, groundwaters and surface waters and therefore is one of the primary processes that affect and control the transport and bioavailability of this anion [3]. Various abiotic and biotic factors (pH, redox, ionic strength, adsorbent type, percentage organic matter content, temperature, concentration, competitive adsorbates, solubility product effects, and nonreductive/reductive dissolution of adsorbate) greatly affect the reactivity, speciation, mobility, and bioavailability of P. Because adsorption to mineral surfaces is one of the most important rate-limiting factors controlling P release in subsurface environments, it is vital to study the mechanisms of P adsorption on naturally occurring soil minerals.

Alumina is an important material used widely in a number of industrial applications, particularly as an adsorbent. The predominant crystal phase of alumina is believed to be γ phase. γ-alumina is a stable phase and usually shows a high surface area and make it highly efficient as adsorbent. In addition, the mesoporous structure of the material we used in this study can efficiently promote the adsorption of phosphate in the water-alumina interface due to the large surface area and organized pore channels.

1.2 ATR-FTIR (Attenuated Total Reflection Fourier Transform Infrared) spectroscopy

Vibrational spectroscopy is ideally suited to probe interfaces. A vibrational

spectrum contains detailed structural information of the adsorbate layer such as

interaction mode between surface and adsorbate, orientation of the adsorbate and intermolecular interactions within the adsorbate layer. Furthermore, infrared radiation is noninvasive. The advances in vibrational spectroscopy of interfaces in the past few years went along with progress in instrumentation and data processing. Furthermore, the information contained in the measured spectra can now be read in more detail through comparison with calculations. The spectroscopic method most commonly used to probe phosphate surface complexes at these mineral phases is infrared (IR) spectroscopy, which is very sensitive to the coordination environment and protonation state of phosphate complexes, and therefore a useful tool for characterizing phosphate surface species at the molecular level. Much research has been done on the uptake of phosphate by metal oxide minerals such as goethite, ferrihydrite, boehmite and aluminum oxides using in situ Fourier transform infrared (FTIR) spectroscopy.

Spectroscopic studies on these systems have shown that phosphate forms inner-sphere complexes at the solid–liquid interface of these substrates, and have revealed that the mechanism of phosphate complexation may vary with pH and surface coverage.[4-10]

The principle of ATR-IR spectroscopy is illustrated in Fig. 1-2 for multiple internal reflections and for a thin deposited film. In contrast to transmission IR (T-FTIR) spectroscopy where the IR beam passes directly through the sample, in the ATR mode the IR radiation is reflecting through the internal reflection element (IRE), an IR transparent crystal of high refractive index (n

) in contact with the sample (n

, with n

>n

). The IR radiation propagates through the IRE at an angle of incidence (θ) larger than the critical angle, such that total reflection occurs at the IRE-sample interface. An evanescent electromagnetic field is generated that penetrates into the sample and is attenuated by the sample, thus producing an IR spectrum. The amplitude of the electric field decays exponentially with the distance from the IRE.

The penetration depth (dp) is the distance from the interface where the intensity of the electric field falls to 1/e of its original value at the interface:

1

2 2

2 sin

d

n λ

π θ

= −

where λ

= λ/n

is the wavelength in the denser medium, λ the wavelength of the incoming radiation and n

= n

/n

. The above equation holds for a two-phase system (IRE/sample). Typically, d

is on the order of 1 μm. For bulk materials, the degree of coupling between the evanescent field and the absorbing sample is given by the effective thickness

2

21 0

2cos

p e

n E d

d ⁼ θ

where E

is the amplitude of the electric field at the interface. The effective thickness

expresses the equivalent path length in a hypothetical transmission measurement,

which yields the same absorption as in an ATR experiment. The very short path

length used in ATR-IR spectroscopy, implicit in d

, makes this technique surface

sensitive and, hence, suitable for the in situ characterization of heterogeneous

catalysts. The sensitivity can be enhanced by using multiple reflection elements. The

effective thickness depends on the refractive indices of IRE and sample. For example, by increasing or decreasing n

at constant λ

and n

, d

decreases or increases, respectively. Hence, by changing from Ge (n

=4.0) to ZnSe (n

=2.4), d

increases, i.e., more sample is probed by the IR radiation.

Fig. 1. Overview of the ATR-FTIR set-up.[11]

Fig. 2. An attenuated total reflection setup allowing multiple internal reflections is

shown. The IR beam propagates through the coated trapezoidal IRE and is totally

reflected at the thin film-liquid interface. The electric field vectors for the parallel

(E

┴) and perpendicular (E

) polarized radiation and a Cartesian axes system on top

of the thin film are also shown.[12]

2. Materials and methods

2.1 Mesoporous alumina preparation

The following general procedure shown in Fig. 3 was used to prepare the mesostructured alumina described in this study. The hydrous mesoporous alumina was prepared by adding NH

·H

O solution (2.5%) slowly to a rapidly stirred AlCl

·6H

O solution (1 mol L

). PEG (molecular weight is 4000) solution was used as a structure-directing reagent and as a dispersant to prevent the solid particles from aggregation dodecylamine was used as co-template . The suspension was centrifuged after being intensively stirred for 1 h at pH 9. The solid parts were then rinsed with distilled water and ethanol in order to remove extra PEG and other impurities. The gel was dissolved in ethanol solution and then distilled at 100

C to remove the alcohol. In the last step, the aluminum hydroxide powder obtained was calcined at 650

C for 6 h.

Fig. 3. Procedure used to prepare the meso-structured alumina

2.2 General characterization of mesoporous alumina powder

SEM. Alumina dispersions (0.5 g L

) were prepared by dispersing solid alumina in

a 0.01 M NaCl solution. The resulting dispersion was shaken during one hour. 2 ml

(≈1 mg) of this dispersion was placed on top of a SEM glass (the glass area is same as

the 45

ZnSe ATR crystal) to make the thickness same as deposited alumina thin film

on the ZnSe ATR crystal. Then, the film was allowed to dry in air overnight in order

to form a dry alumina film. The film was then rinsed with water to eliminate the

excess of alumina particles that did not adhere well to the glass and put in an oven at

60ºC overnight. The images were obtained by a Philips XL30 (detector: SE) scanning

electron microscope.

XRD. Powder X-ray diffraction (XRD) results of the samples were collected on a Rigagu D/MAX-γA diffractometer equipped with a rotating anode and CuKα radiation. Counts were accumulated every 0.02

(2θ ) at a scan speed of 2

(2θ)/min.

N

adsorption–desorption analysis. N

adsorption–desorption isotherms were obtained at 77 K on a micrometrics ASAP 2000 surface and pore-size analyzer using a static adsorption procedure. Samples were degassed at 150

C in vacuum below 10

Torr for 16 h prior to the measurements. The specific surface area was calculated using the BET equation for data in a (P/P

) range between 0.05 and 0.3. The pore-size distributions were calculated from the data of the desorption branch of the isotherm using the Barrett– Joyner–Halenda (BJH) method.

Zeta potential. The zeta potential of alumina powder in solutions with/without phosphate was measured by means of an electrophoretic method and in NaCl electrolyte solution.

2.3 ATR-FTIR experiments

2.3.1 Preparation of deposited thin alumina films

A stock alumina dispersion (0.5 g L

) was prepared by dispersing solid alumina in a 0.01 M NaCl solution. The resulting dispersion was shaken during one hour. 2 ml (≈1 mg) of this dispersion was placed on top of the 45

ZnSe ATR crystal and allowed to dry in air overnight in order to form a dry alumina film. The film was then rinsed with water to eliminate the excess of alumina particles that did not adhere well to the crystal.

2.3.2 Flow cell experiments 2.3.2.1 PH envelope

A Bruker IFS 66 v/S spectrometer equipped with a DTGS detector was used in the collection of the infrared spectra. Spectra recorded during the adsorption process were obtained by taking an average of 128 scans at a resolution of 4 cm

. PH envelopes were measured using the flow cell technique described by Hug and Sulzberger [13], Hug [14], and Peak et al. [15,16]. The flow cell was placed on the horizontal ATR sample stage inside the IR spectrometer and connected to a reaction vessel containing 0.1 L of background electrolyte solution (0.01 M NaCl in distilled H

O), which was stirred with a magnetic bar, and adjusted to the desired pH. A peristaltic pump was used to pass solute from the reaction vessel through the flow cell at a known flow rate. The pH of the solution in the main reaction vessel was continuously monitored, and (re)adjusted if necessary. The experimental flow-cell setup is described in more detail in Peak et al. [15] and Wijnja and Schulthess [17].

Experiments were started by pumping the background electrolyte through the

flow cell at a flow rate of ≈1 mLmin

, and allowing the alumina deposit to

equilibrate with the background solution. Background spectra, consisting of the

absorbance of the ZnSe crystal, the alumina deposit and the electrolyte solution, were

collected regularly during this equilibration period; typically after about 1 h,

successive background spectra showed no further changes, indicating that the alumina deposit had equilibrated with the background solute. The final background spectrum was collected at this time as the average of 128 scans at a 4 cm

resolution, and phosphate was injected into the reaction vessel to start the adsorption experiment. All successive spectra were ratioed to this background spectrum. For the pH envelope experiment, the initial 0.01 M NaCl background solution pumped through the flow cell had been adjusted to pH 9.0. Following equilibration of the alumina deposit with this solution and collection of the background spectrum, phosphate was added at a concentration of 50 μM. Phosphate adsorption on the alumina deposit was monitored by collecting IR spectra every 10 min. When there were no differences between successive spectra, the final spectrum of sorbed phosphate was collected as the average of 128 scans, and the pH of the solution in the reaction vessel was lowered to pH 7.32 by addition of 0.1 M HCl. Phosphate adsorption onto the alumina deposit in the flow cell was allowed to reach a new equilibrium. Then, the final IR spectrum was collected and pH was lowered to the next level. This procedure was repeated down to pH 4.05.

2.3.2.2 Adsorption isotherms

The phosphate adsorption isotherm measurements were performed at fixed pH values of 9.01 and 4.05. For these experiments, the alumina deposit was first equilibrated with the 0.01M NaCl background electrolyte adjusted to the pH of interest, and next phosphate was injected at a concentration of 50 μM. Phosphate adsorption onto the alumina coating inside the flow cell was allowed to reach equilibrium; then, the final IR spectrum of sorbed phosphate was collected, and the phosphate concentration in the reaction vessel was raised for new equilibration. This procedure was repeated until the aqueous phosphate concentrations in the main reaction vessel reached 2500 (2000) μM. Preliminary experiments using a noncoated ZnSe crystal indicated that contributions from aqueous phosphate in the IR spectrum become visible above noise level at solution concentrations of approximately 1000 μM. The ATR-FTIR spectra collected in the pH envelope and phosphate isotherm experiments are therefore dominated by absorbance from alumina-sorbed phosphate complexes, and contributions from aqueous phosphate are negligible.

2.3.2.3 Adsorption kinetics

Adsorption kinetics was measured at different phosphate concentration at pH

9.00 and 4.94, respectively. For these experiments, the alumina deposit was first

equilibrated with the 0.01M NaCl background electrolyte adjusted to pH 9.00, and

then phosphate was injected at a concentration of 5 μM. This time was set as initial

time of the adsorption reaction. Spectra were then recorded as a function of time

every ten minutes. Then phosphate concentration was increased to desired value to

record the spectra every ten minutes. Subsequently, pH was fixed at 4.94 and the

procedure above was repeated.

3. Results and discussion

3.1 General characterization

3.1.1 XRD pattern

Fig. 4 XRD diagram of mesoporous alumina

XRD pattern of mesoporous alumina is shown in Fig. 4. It shows typical γ-Al

O

which have been reported by many groups. The diffraction peaks are relatively broad

which imply a small particle size and poor crystallinity.

3.1.2 SEM images

Fig. 5 SEM image of deposited alumina

Fig. 5 shows a thin film of deposited alumina on the ZnSe crystal surface and the SEM images in horizontal and vertical direction, respectively. It was indicated by the magnified horizontal image that the alumina film is to some extent assembled from small alumina particles, about 20-30nm. From the vertical image it is evident that the thickness of the deposited alumina film is around 3-4 μm.

3.1.3 Zeta potential measurements.

The results from z-potential measurements are shown in figures 6 and 7:

Fig. 6. The figure shows the zeta potential of mesoporous alumina at different electrolyte concentration.

Fig. 7. The figure shows the zeta potential of mesoporous alumina at different [PO

] concentrations.

The zeta potential values of the alumina particles are shown in Fig. 6 as a function of pH and at different phosphate concentration. As shown in Fig. 6, the PZC of alumina occurred at a pH value of 8.9-9.5 in the absence of phosphate using NaCl electrolyte solutions with the concentrations 0.001M, 0.01M, and 0.1M. The concentration of electrolyte solution hardly affect the zeta potential at pH lower than PZC but significantly at pH higher than PZC.

With the presence of phosphate shown in Fig. 7, the zeta potential decreased more

rapidly with pH than the z-potential without phosphate. At a phosphate concentration

of 5μM, the zeta potential was negative already at pH values above 6.5, whereas a

still higher phosphate concentration (50μM) resulted in a PZC value of 5.3. This is

because of the adsorption of phosphate at the water-alumina interface.

3.1.4 N ₂ adsorption/desorption analysis.

The following figures show results from BET measurements:

Fig. 8 N

adsorption/desorption isotherms and corresponding pore size distribution of

mesoporous alumina

Table 1 Parameters of N

adsorption/desorption analysis

The nitrogen adsorption/desorption isotherms and corresponding pore size distribution are given in Fig. 7-8. The BET surface area, pore size, pore volume of the samples are listed in Table 1. Testing results show that the synthesized alumina material displayed a high surface area of 329 m

g

and a pore size centered at 9.3nm with a narrow distribution.

3.2 ATR-FTIR spectra

3.2.1 Species distribution and ART-FTIR spectra of aqueous H _n PO ₄ ^3-n at different pH.

Fig. 9 Species distribution of phosphate solution at different pH (ploted by Medusa

software)

Fig. 10 ART-FTIR spectra of aqueous H

PO

phosphate species at different pH Based on the dissociation constants of phosphate, the distribution of species in the phosphate solution at different pH can be determined, as illustrated in Fig. 9. The sensitivity of IR spectroscopy toward the protonation of phosphate is well illustrated by the differences between the IR spectra of PO

(aq), HPO

(aq), H

PO

(aq), and H

PO

(aq) as shown in Fig. 10. The spectra of these aqueous phosphate species have also been reported by Tejedor-Tejedor and Anderson [8], and are useful for demonstrating the use of symmetry arguments in interpreting IR spectra, as outlined by Nakamoto [18]. The use of attenuated total reflectance Fourier-transform infrared (ATR-FTIR) spectroscopy for aqueous media allows for characterization of two phosphate vibrations: the nondegenerate symmetric stretching ν

, and the triply degenerate symmetric stretching ν

. The nonprotonated PO

anion has tetrahedral symmetry and belongs to the point group T

. This species has one active ν

band, centered at 1006 cm−1, whereas the ν

vibration is inactive (spectrum a, Fig. 10).

Protonation of this complex to HPO

(aq) leads to a symmetry reduction from Td to C

, and as a result, the triply degenerate ν

vibration splits into two separate ν

bands located at 1078 and 987 cm

, and the ν

band is activated and can be seen at 848 cm

(spectrum b, Fig. 10). The addition of another proton to form H

PO

(aq) further reduces the symmetry to C

. This symmetry reduction leads to splitting of the ν

vibration into three ν

bands at 1159, 1074 and 939 cm

, and the ν

band remains active (at 877 cm

), so that a total of four bands are present for this phosphate species (spectrum c, Fig. 10). Formation of H

PO

(aq) increases the symmetry back up to C

, and as a result, two ν

bands (at 1172 and 1006 cm

) and one ν

band (at 887 cm

) are seen for this phosphate complex (spectrum d, Fig. 10). The spectra of the H

PO

(aq) and H

PO

(aq) complexes also contain a broad band at 1220–1240 cm

, which has been assigned to the δ(POH) bending mode [4].

The relation between the number of IR active bands and the symmetry of the

phosphate anion illustrated in Fig. 10 can be used to interpret phosphate bonding

configurations at mineral surfaces. For instance, when phosphate forms a

nonprotonated monodentate inner-sphere complex, the symmetry of the phosphate

molecule is C

, and two ν

bands are expected to be present in addition to the active ν

band. When phosphate forms a monodentate binuclear (i.e., bridging) or a bidentate mononuclear (i.e., edge-sharing) inner-sphere complex, the symmetry of the surface complex is C

so that three active ν

bands should be present in addition to the ν

band. Surface complexes having symmetry lower than C

(i.e., C

complexes) also will have three active ν

bands and one ν

band. If phosphate coordinates as an outer-sphere complex, a slight shift in the ν

vibration is expected relative to those of the phosphate ions in solution due to distortion resulting from the near-surface electrical field, but the number of ν

bands is not expected to change.

3.2.2 PH envelope spectra

The following spectra show results from the adsorption of phosphate at different pH values:

Fig. 11 ATR-FTIR spectra resulting from flow-cell adsorption pH envelope

experiment, conducted at [PO

]=50 μM. Spectra were collected at ph 9.00, 7.32, 6.97,

6.16, 4.96, and 4.05, respectively.

Fig. 12. Results from the flow-cell pH envelope experiment, conducted at [PO4]=50uM. The spectra were collected at pH (a)7.32, (b)6.96, (c)6.16, (d)4.96, and (e)4.05 (normalized spectra resulting from Fig. 11).

Fig. 13. Difference spectra of selected spectra from Fig. 11 (The spectra 4.96-9.0, 4.96-7.32, 4.96-6.97, and 4.96-6.16 are obtained by subtracting spectra at pH 9.0, 7.32, 6.97, 6.16 from the spectrum at pH4.96. Spectra are shown in Fig. 11.

Fig. 11 shows the ATR-FTIR spectra of phosphate complexes forming at the

alumina–water interface in the pH range 4.05–9.00 at an aqueous phosphate

concentration of 50 μM. The increase in intensity of the IR absorbance with

decreasing pH indicates that phosphate sorption increases when pH is lowered. In

addition to intensity changes, significant changes in spectrum shape and position are

observed as a function of pH as well (Fig. 12) . Spectra are dominated by three IR

frequencies, centered at approximately 1088, 1065 and 1016 cm

. When pH is lowered, the band at 1088 cm

gradually shifts to higher wavenumbers. Spectra recorded at pH 4.05 and 4.96 show evidence for the presence of at least four bands，

which indicate that at least two different phosphate complexes are present at the alumina surface in the pH range 4.05 - 4.96.

The gradual shifting of bands with pH in the pH range 9.00-4.05 suggests that the phosphate IR spectra observed between pH 9.00 and 4.05 consist of the combined absorbance of at least two different phosphate surface complexes: the phosphate complex that dominates at high pH, and a second surface complex that grows as the pH is lowered. To investigate this option, we took the difference spectra of the data collected at pH 9.00, 7.32, 6.96, 6.16, and 4.96 in order to isolate the IR spectra of the (additional) phosphate species coordinating to the alumina surface when pH is lowered. The results are shown in Fig. 13. The difference spectra in all cases show one strong vibrational band at approximately 1130 cm

except that the spectrum at pH 4.96 contains additional peak frequency at 1010 cm

corresponding to the IR spectrum of phosphate sorbed at lower pH.. The difference spectra shown in Fig. 13 indicate that lowering pH leads to the formation of a phosphate surface complex that has an IR spectrum very different from that of the phosphate species dominating the surface speciation at high pH.

3.2.3 Adsorption isotherms

Fig. 14. ATR-FTIR spectra resulting from flow-cell adsorption isotherm experiments

conducted at pH 9.01. Spectra from adsorption equilibria were collected at aqueous

phosphate concentrations of 5, 10, 25, 50, 200, 500, 1000, 1500, 2000, and 2500 μM.

Fig. 15. Normalized spectra as obtained from Fig. 14

The results of the adsorption isotherm experiment performed at pH 9.01 are presented in Fig. 14 for aqueous phosphate concentrations between 5 and 2500 μM.

The IR absorbance intensities increase with increasing phosphate concentrations, indicating that the amount of phosphate sorbed increases as the phosphate concentration is raised. This indicates that alumina surface site saturation is not achieved in the phosphate concentration range applied here. Although the spectra of the sorption complexes collected at the various phosphate concentrations overall look quite similar at pH 9.01 (Fig. 14), a subtle change seems to occur in the IR spectra as the phosphate concentration is raised. The change in the normalized IR spectra shown in Fig. 15 suggests that at higher phosphate concentrations, i.e. at higher surface coverage, there is a shift to higher wavenumbers.

Fig. 16 shows the results from measurements of the adsorption equilibrium at different concentrations, which is the basis of the adsorption isotherm.

Fig. 16. Two isotherms for phosphate adsorbed on mesoporous alumina at pH 9.01.

The solid and dotted lines are the results of curve fitting, using Freundlich and Langmuir adsorption isotherms, respectively.

It was found that the amount of phosphate adsorbed on aluminum oxide increased with increasing concentration but the increase gradually became smaller at higher concentrations. The adsorption data were evaluated using the Langmuir and Freundlich isotherm models, as demonstrated in Fig. 16. Obviously it was indicated that the adsorption data fits the Langmuir isotherm better than the Freundlich isotherm at lower [PO

] concentration. However, it fits the Freundlich isotherm better than the Langmuir isotherm at higher [PO

] concentration at least at pH 9.0.

Fig. 17. ATR-FTIR spectra resulting from flow-cell adsorption isotherm experiments conducted at pH 4.15. Spectra were collected at aqueous phosphate concentration of 5, 10, 25, 50, 200, 500, 1000, and 2000 μM.

Fig. 18. Normalized spectra from Fig. 17.

The results from the isotherm experiment performed at pH 4.15 are shown in Fig.

17, which presents the IR spectra of phosphate–aulmina sorption complexes forming at aqueous phosphate concentrations ranging from 5 to 2000 μM. The IR absorbance intensities increase with increasing phosphate concentration, indicating that the amount of phosphate sorbed increases as the phosphate concentration is raised. This indicates that alumina surface site saturation is not achieved in the phosphate concentration range applied here.

Consistent with the results from the pH envelope experiment described in the previous section, the IR spectra of sorbed phosphate at pH 4.15 (Fig. 17) and normalized absorbance spectra in Fig. 18, appear to consist of at least two different phosphate sorption complexes. There is a little bit of change in the shape of spectrum, however, as the phosphate concentration increases, indicating that the phosphate surface speciation changes with surface loading.

Fig. 19. The figure shows the adsorption isotherm of phosphate at pH 4.15 when phosphate is adsorbed on a mesoporous alumina surface. The solid and dotted lines are the results of curve fitting with Langmuir and Freundlich adsorption isotherms, respectively.

Similar to the result obtained at pH 9.01, the amount of phosphate adsorbed on aluminum oxide increased with increasing concentration but the increase gradually became smaller at higher concentrations. The adsorption data were evaluated according to the Langmuir and the Freundlich isotherm models, as shown in Fig. 19.

The adsorption data fits the Langmuir isotherm better than the Freundlich isotherm at

lower [PO

] concentration. However, it fits the Freundlich isotherm better than the

Langmuir isotherm at higher [PO

] concentration and pH 4.15.

3.2.4 Adsorption kinetics.

Figures 20 and 21 show results from adsorbtion rate experiments:

Fig. 20. The integrated absorbance of phosphate at the mesoporous alumina-water interface at different [PO

] concentrations and constant pH 9.0, as a function of time.

Fig. 21. Integrated absorbance of [PO

] on mesoporous alumina-water interface at different [PO

] concentrations and constant pH 4.94, as a function of time

Fig. 20-21 shows the phosphate adsorption kinetics at pH 9.0 and 4.94. The different curves

with doted symbols represent different initial phosphate concentrations. All curves have similar

characteristics showing a fast adsorption between t = 0 and 10 min, and a slower adsorption at

longer times, except that at 5μM. Although adsorption seems to reach completion at around 120

min for adsorption at pH 9.0 and 300 min at pH 4.94, some long-term kinetic experiments showed

that adsorption continued after this period of time, but very slowly. Data in Fig. 20-21 are very

similar to those reported in several other articles and, as it has been proposed by several authors,

we also suggest that adsorption takes place in at least two interconnected processes: a very fast

initial process, which seems to take place in 5 min or less, followed by a slower process that takes place in hours or even days. [19]

3.2.5 Curve fitting at different pH and phosphate concentration.

The spectral band shapes of the originating from adsorbed phosphate species were subjected to curve fitting using the Opus software, as shown in figures 22 and 23:

Fig. 22 Curve fitting of spectra recorded at pH 4.05(a), 7.32 (b) and 9.0(c), after first having normalized the spectra.

Fig. 22 shows curve fitted spectra recorded at pH 4.05(a), 7.32 (b) and 9.0(c) after

first having normalized each spectrum. The broad band is composed of four (a) or

three (b and c) ν

vibrations, indicating C

or lower symmetry. Spectra collected at

high pH value could be fitted with three ν

bands, as illustrated in Fig. 22 b-c for the

spectra collected at pH 7.32 and 9.0. The fitting results were different with that at

lower pH, however. The three ν

bands are in the range between 780-1060 cm

. In

addition, different surface complexes are formed at varying pH from 4.05 to 9.0. At

the lowest pH value (pH 4.05), four bands are distinguishable in the 1200–900 cm

region. As discussed in the introduction, the maximum number of phosphate ν

bands

expected, based on symmetry arguments, is three. This means that at least two

different phosphate complexes are present at the alumina surface at pH 4.05.

Fig. 23 Curve fitted spectra collected at [PO

] = 5 and 500 μM. The pH of the solution was 9.0.

Fig. 23 shows curve fitted spectra collected at [PO

]= 5 and 500 μM (pH 9.0). The broad band is composed of three ν

vibrations positioned at 1069, 1009(1012), and 969(967)cm

, indicating C

or lower symmetry. The fitted curves at different phosphate concentration are quite similar in peak positions. However, some difference concerning the intensity of the fitted curves is evident.

3.2.6 Possible surface complexes

Fig. 24. This figure shows possible molecular symmetries (C

or lower) of protonated

Phosphate inner-sphere complexes at the alumina-water interface: a. monoprotonated monodentate mononuclear, b. diprotonated monodentate mononuclear, c.

monoprotonated bidentate mononuclear, d. diprotonated bidentate mononuclear, e.

nonprotonated monodentate mononuclear with hydrogen bonded with the hydroxyl group of Al-H, f. monoprotonated monodentate mononuclear with hydrogen bonded with the hydroxyl group of Al-H, g. diprotonated monodentate mononuclear with hydrogen bonded with the hydroxyl group of Al-H.[20]

As shown in Fig. 24, several molecular symmetries are possible for the Al–P complex with C

or lower symmetry.

4. Conclusions

Mesoporous alumina materials have been synthesized using PEG as structure directing reagent and dodecylamine as co-template. XRD, SEM, N

adsorption/desorption technique and zeta potential measurements were used to characterize the general properties of the synthesized material. The results show that the aluminum oxide material is λ-Al

O

having a high surface area viz. 329 m

g

. The deposited alumina on the crystal surface is a film composed of small alumina particles, about 20-30nm in diameter, and the thickness of the film is between 3 and 4 μm. The zeta potential was affected by electrolyte concentration and the adsorbed phosphate on the surface of the alumina film could significantly influence the surface charge as well as the PZC value and make PZC to occur at pH values less than that without phosphate.

Phosphate adsorption on alumina surfaces was investigated by in situ ATR-FTIR

spectroscopy. Spectroscopic studies on these systems have shown that phosphate

forms inner-sphere complexes at the solid–liquid interface. At higher pH value (9.0)

only one complex exists at the water-alumina interface, but a mixture of different

phosphate complexes formed at lower pH values (4.05 and 4.96). The adsorption

isotherms and kinetics at acid and base environment were also studied. The results

showed that the amount of phosphate adsorbed on aluminum oxide increased with

increasing concentration but the increase gradually became smaller at higher

concentrations. The adsorption data were evaluated using the Langmuir and

Freundlich isotherm models. It was shown that the adsorption data fits the Langmuir

isotherm better than the Freundlich isotherm at lower [PO

] concentration. However,

it fitted the Freundlich isotherm better than the Langmuir isotherm at higher [PO

]

concentration both at pH 9.0 and pH 4.15. Adsorption kinetics showed that adsorption

both at pH 9.0 and pH 4.96 had similar characteristics except at the lowest

concentration studied (5μM), with a fast adsorption between t = 0 and 10 min, and a

slower adsorption at longer reaction times. Finally, seven possible surface complexes

are proposed for this Al-P complex, having C

or lower symmetry.

References

[1] H. Klapper, Control of Eutrophication in Inland Water, Ellis Horwood, Chichester, 1991.

[2] Q. Zhou, C.E. Gibson, Y. Zhu, Chemosphere 42 (2001) 221.

[3] D.L. Sparks, Environmental Soil Chemistry, Elsevier Science, New York, 2003

[4] R.J. Atkinson, R.L. Parfitt, R.S.C. Smart, J. Chem. Soc. Faraday Trans. 1 70 (1974) 1472.

[5] R.L. Parfitt, J.D. Russell, V.C. Farmer, J. Chem. Soc. Faraday Trans. 1 72 (1976) 1082.

[6] R.L. Parfitt, R.J. Atkinson, Nature 264 (1976) 740.

[7] R.L. Parfitt, R.J. Atkinson, R.S.C. Smart, Soil Sci. Soc. Am. Proc. 39 (1975) 837.

[8] M.I. Tejedor-Tejedor, M.A. Anderson, Langmuir 6 (1990) 602.

[9] P. Persson, N. Nielsson, S. Sjöberg, J. Colloid Interface Sci. 177 (1996) 263.

[10] Y. Arai, D.L. Sparks, J. Colloid Interface Sci. 241 (2001) 317.

[11] Fredriksson A, Holmgren A. An in situ ATR-FTIR study of the adsorption kinetics of xanthate on germanium.

Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2007, 302(1-3):96-101 [12] D. Ferri, T. Bürgi and A. Baiker, Helv. Chim. Acta 85 2002 3639

[13] S.J. Hug, B. Sulzberger, Langmuir 10 (1994) 3587.

[14] S.J. Hug, J. Colloid Interface Sci. 188 (1997) 415.

[15] D. Peak, R.G. Ford, D.L. Sparks, J. Colloid Interface Sci. 218 (1999) 289.

[16] D. Peak, E.J. Elzinga, D.L. Sparks, in: H.M. Selim, D.L. Sparks (Eds.), Heavy Metals Release in Soils, Lewis Publishers, Boca Raton, FL, 2001.

[17] H. Wijnja, C.P. Schulthess, Soil Sci. Soc. Am. J. 65 (2001) 324.

[18] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, Wiley, New York, 1997.

[19] R. Strauss, G.W. Brümmer, N.J. Barrow, Eur. J. Soil Sci. 48 (1997) 101..

[20] Yuji Arai, D. L. Sparks. Journal of Colloid and Interface Science, 241,317-326 (2006)

Acknowledgements

I would like to thank my supervisor Professor Zhongxi Sun. Thank you for being inspiring, sharing, patient, and understanding and because of you I feel incredibly fortunate to have the chance to work in Sweden.

I would like to thank my supervisor Professor Allan Holmgren for giving me this opportunity to work for half year at Luleå university of technology as an exam worker.

Thank you for your valuable suggestions and encouraging words in my research work.

Also you have helped me a lot in my daily life and English speaking ability.

Xiaofang, I really appreciate your advice and help in my research work. In addition, you also are my friend in my life. Also, I would like to thank all my colleagues in chemistry division especially Payman and Yu for the help in chemicals needed and instrument instructions.

Last but definitely not least, I want to express my great appreciation to my family.

To my parents, I cherish your devoted affection. Without you, I would not be the person I am.

I am always feeling warm and happy in this cold winter because of all of you.