MURDOCH RESEARCH REPOSITORY

This is the author’s final version of the work, as accepted for publication following peer review but without the publisher’s layout or pagination.

The definitive version is available at http://dx.doi.org/10.1016/j.micromeso.2015.06.031

Huang, W., Yu, X., Tang, J., Zhu, Y., Zhang, Y. and Li, D. (2015) Enhanced adsorption of phosphate by flower-like mesoporous

silica spheres loaded with lanthanum. Microporous and Mesoporous Materials, 217. pp. 225-232.

http://researchrepository.murdoch.edu.au/28090/

Copyright: © 2015 Elsevier Inc.

It is posted here for your personal use. No further distribution is permitted.

Accepted Manuscript

Enhanced Adsorption of Phosphate by Flower-like Mesoporous Silica SpheresLoaded with Lanthanum

Weiya Huang, Xiang Yu, Jinpeng Tang, Yi Zhu, Yuanming Zhang, Dan Li

PII: S1387-1811(15)00364-9

DOI: 10.1016/j.micromeso.2015.06.031

Reference: MICMAT 7190

To appear in: Microporous and Mesoporous Materials

Received Date: 16 February 2015

Revised Date: 11 June 2015

Accepted Date: 24 June 2015

Please cite this article as: W. Huang, X. Yu, J. Tang, Y. Zhu, Y. Zhang, D. Li, Enhanced Adsorptionof Phosphate by Flower-like Mesoporous Silica Spheres Loaded with Lanthanum, Microporous andMesoporous Materials (2015), doi: 10.1016/j.micromeso.2015.06.031.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1

Enhanced Adsorption of Phosphate by Flower-like Mesoporous Silica

Spheres Loaded with Lanthanum

Weiya Huang a, Xiang Yu c, Jinpeng Tang c, Yi Zhu c, Yuanming Zhang c, and Dan Lib,*

a Department of Materials Science and Engineering, Taizhou University, Linhai, 317000,

China;

b Chemical and Metallurgical Engineering and Chemistry, School of Engineering and

Information Technology, Murdoch University, Murdoch, Western Australia, 6150,

Australia. Corresponding author: Tel: +61 08 9360 2569, Email: l.li@murdoch.edu.au.

c Department of Chemistry, Jinan University, Guangzhou, 510632, China.

Abstract

Monodispersed flower-like mesoporous silica spheres doped with lanthanum were

synthesized via a facile method and explored as novel adsorbents for efficient phosphate

adsorption for the first time. The synthesized adsorbents possessed a flower-like structure,

especially with unique mesoporous channels featuring with small inner pores and wide outer

pores. By raising lanthanum loading, the surface areas and pore volumes of resulting

adsorbents were observed to continuously decrease, accompanied with gradually increasing

pore sizes; that led to a dramatic improvement in phosphate adsorption from water. By taking

cost effectiveness for practical application into consideration, the adsorbent with a theoretical

La/Si molar ratio of 0.1 possessed a promising performance as compared to the literature ever

reported, thanks to the large outer mesopores likely minimizing the diffusion problem caused

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2

by the pore blocking when LaPO4 nanorods formed during adsorption. In the kinetic study, its

phosphate adsorption followed the pseudo-second-order equation better, suggesting

chemisorption. Meanwhile, superior phosphate adsorption capacities were achieved within

the pH range of 3.0 and 6.0; and almost unaffected by the presence of F−, Cl−, NO3−, and

SO42-.

Keywords: flower-like; mesoporous silica; lanthanum; phosphate; adsorption

1. Introduction

Excessive presence of phosphorous in water bodies is one of the major causes to

eutrophication, which leads to the outbreaks of algal bloom, diminishes dissolved oxygen,

lowers water quality, and reduces biodiversity in aquatic ecosystems [1, 2]. Therefore, it is of

vital importance to reduce and control the level of phosphorous in the industrial, urban, and

agricultural wastewater prior to its discharge into environment. In this regard, several

methods, including chemical precipitation, biological treatment, ion exchange and adsorption,

have been investigated to remove phosphorous from aqueous solution [3]. In particular, the

adsorption-based process is considered as one of the promising routes, due to its simple

operation and low sludge production [4, 5]. Until now, various types of new adsorbents have

been developed for phosphorus removal; especially the composites made from metal oxide or

hydroxide depositing on the surface of porous support materials, among which activated

carbon and silica are popular options due to their non-toxic nature [6-8]. In general,

commercial activated carbon is consisting of disordered porous structure and irregular pore

size distribution. The portions of micropores, mesopores, and macropores in activated carbon

are varied largely depending on the raw materials used during synthesis. Fundamentally, the

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3

richness of micropores endows a potential for large surface area of activated carbon, but

lowering mass transfer and in turn adsorption capacity / rate. By contrast, mesoporous silica

has received considerable attention owing to its attractively unique features, such as a highly

ordered structure, ultrahigh surface area, tunable pore size and structure, as well as favourable

surface chemical property for functionalization or modification [9, 10].

Recently, there has been a significant progress made in developing metal-modified

mesoporous silica materials as phosphate adsorbents [11-18]. Until now, a variety of metals,

including Fe, Al, Zr and La, have been successfully impregnated into mesoporous silica, such

as SBA-15 or MCM-41, to form novel adsorbents with tailorable properties [6, 7, 11-18]. In

particular, mesoporous silica materials incorporated with lanthanum (La) have exhibited

several promising advantages in phosphorus adsorption, i.e. a superior adsorption capacity,

wide operating pH range, and high removal rate in low phosphate concentration [13, 17-20].

Since metal oxides which are embedded into mesoporous silica act as active sites specifically

adsorbing phosphate, high loadings of metal oxides are usually desirable to enhance

adsorption capacity. Nevertheless, excessive loading of metals might attack and partially

destruct the ordered porous framework, resulting in a less access to the adsorption sites inside

the ruined pores and in turn a lower adsorption capacity [11, 16, 21, 22]. For example, Shin et

al. reported that SBA-15 modified with a high loading of Al (30%) exhibited lower

phosphate adsorption capacity (19.2 mg P/g) than those of 10% (26.7 mg P/g). This was

attributed to that the pore structure was destructed so that adsorbates could not access the

adsorption sites inside the blocked pores [16]. In Yang’s work by using SBA-15 with

relatively large pores of 8.87 nm, the high loading of La in the silica materials lowered La

usage efficiency [18]. This was possibly explained by the heterogeneous nucleation and

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4

subsequent growth of LaPO4 nanorods during adsorption; which blocked the confined

channels and caused the diffusivity problem [18].

In 2011, Gai et al. reported the synthesis of a family of novel flower-like mesoporous silica

(FMS) spheres for drug delivery; which exhibited unique fine pore systems with narrow inner

channels and wide outer pore openings, offering an easy and high accessibility to a large

number of active sites inside the pores [23, 24]. They are shown as promising supporting

material candidates, potentially leading to the formation of superior phosphate adsorbents

with improved diffusion characteristic and adsorption performance, as compared with the

traditional mesoporous silica, e.g. SBA-15 or MCM-41.

Herein, we designed novel phosphate adsorbents by modifying FMS with lanthanum; to the

best of our knowledge, this was reported for the first time. The structural and morphological

properties of synthesized La-modified FMS, as well as their corresponding phosphate

adsorption behaviours were investigated in detail. In particular, the sample FMS-0.1La

(theoretical La/Si molar ratio = 0.1) exhibited the highest phosphate removal efficiency in

terms of P/La molar ratio; thereby was selected for study on its phosphate capture by varying

time, pH and co-existing ions. Our study suggests the developed flower-like mesoporous

silica spheres loaded with La be a good choice for the adsorption of phosphate from aqueous

environment.

2. Experimental

2.1 Chemicals and materials

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5

All of the chemicals were used as received without further purification. Cyclohexane (≥

99%), 1-pentanol (≥ 99%), cetyltrimethyl-ammonium bromide (CTAB), and tetraethyl

orthosilicate (TEOS, 98%) were purchased from Sigma-Aldrich. Lanthanum nitrate

hexahydrate [La(NO3)3·6H2O] was purchased from Fluka. Potassium dihydrogen phosphate

(KH2PO4), urea, acetone, NaCl, NaSO4, NaNO3, NaF, and Na2CO3 were the products from

Merck Pty Limited, Australia. Other chemicals were obtained from Biolab Ltd, Australia.

2.2 Synthesis and characterization of adsorbents

The flower-like mesoporous silica (FMS) spheres were synthesized as reported [23, 24].

Typically, 1.5 mL of pentanol was dissolved in 30 mL of cyclohexane, followed by adding

2.7 mL (12 mmol) of TEOS to form solution A. 1.87 g (5.14 mmol) of CTAB and 0.6 g of

urea were dissolved into 30 mL deionized (DI) water to get solution B. Subsequently,

solution B was quickly added into solution A under vigorous stirring. After 30 min, the

mixture was transferred into a Teflon-lined autoclave, which was then heated at 120 °C for 4

h. The obtained product was washed with acetone and deionized (DI) water, and dried in

oven. The final product was calcined in air from 25 °C to 550 °C (1 °C min−1), after which

the calcination was maintained at 550 °C for 6 h.

The modification of FMS with lanthanum oxide was carried out as following:

La(NO3)3·6H2O was selected as the lanthanum precursor and incorporated into FMS using

the ethanol evaporation method [18, 19]. Typically, 0.4 g of FMS was added into 100 mL of

ethanol solution containing desired amount of lanthanum precursor. The mixture was stirred

at 60 °C for 24 h, and then the temperature was increased to 80 °C to evaporate the ethanol

completely. The product was calcined at 550 °C for 5 h. The theoretical molar ratios of La

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6

versus Si were varied from 0.02, 0.04, 0.1, to 0.2. In this way, four samples with different

theoretical La/Si molar ratios were obtained and denoted as FMS-xLa, e.g. FMS-0.02La,

FMS-0.04La, FMS-0.1La and FMS-0.2La, respectively. The parent sample was prepared and

referred as FMS for comparison.

Surface morphologies of the samples were examined by scanning electron microscopy (SEM,

JSM-7401F, JEOL Ltd., Japan) and energy dispersive spectroscopy (EDS, Horiba EX-250,

Japan). Transmission electron microscopy (TEM, JEM2010-HR, Japan) was used to

characterize the sample at an accelerating voltage of 200 kV; in which the sample was firstly

dispersed in ethanol via ultrasonic treatment and then collected with a carbon copper grid. X-

ray powder diffraction (XRD) patters were recorded in the 2θ range of 5–80° with a scan

speed of 1°/min by using a diffractometer (Bruker D8 Advance diffractometer, Germany)

with Cu Kα radiation (40 mA, 45 kV). Nitrogen adsorption-desorption isotherms were

measured at 77 K using ASAP 2010 (Micromeritics Inc., USA). Prior to analysis, the samples

were degassed at 120 °C for 12 h under vacuum. The specific surface area, SBET, was

determined from the linear part of the BET plot (P/P0 = 0.05-0.20). The pore size was

calculated from the desorption branch of isotherm by using Barrett-Joyner-Hallenda (BJH).

The total pore volume, Vtotal, was evaluated from the adsorbed nitrogen amount at a relative

pressure of 0.98.

2.3 Phosphate adsorption experiments

A series of batch tests were conducted to investigate the phosphate adsorption performances

of the adsorbents. The solution of different phosphate concentrations was prepared by

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7

dissolving desired amount of KH2PO4 in DI water. The phosphate concentration was analysed

by Autoanalyzer 3 (Bran and Luebbe Inc., Germany).

In equilibrium experiments, 0.025 g of the adsorbent was added into 50.0 mL of phosphate

solution with varous initial concentrations in polypropylene bottles. The pH of the solution

was adjusted to 5.0 by 0.10 M HCl or 0.10 M NaOH. The samples were placed in the shaker

bath at 25 °C for 48 h. The phosphate concentration was analyzed after contrifugation for 10

min at 12,000 rpm. The amount of phosphate adsorbed onto the sample at the equilibrium (qe)

was calculated by Eq. (1),

q� =(�����)

� …………… (1)

where C0 and Ce are the initial and equilibrium phosphate concentrations in solution (mg

P/L), respectively; V is the volume of solution (L) and m is the mass of adsorbent (g).

The equilibrium data were fitted to the Langmuir and Freundlich isotherm models, and their

equations were shown as Eqs. (2) and (3) [25, 26], respectively,

Langmuir model: q� q�K���

������ …………… (2)

Freundlich model:q� = K�C��/� …………… (3)

where qm (mg P/g) is the amount of adsorption in Langmuir adsorption model corresponding

to the maximum adsorption capacity, and KL (L/mg P) is the constant in the Langmuir

isotherm model related to energy or net enthalpy of adsorption; KF is the constant in the

Freundlich isotherm model which indicates the relative adsorption capacity and n is the

Freundlich coefficient which measures the adsorption intensity.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8

The essential characteristics of the Langmuir equation can be expressed in a

dimensionless separation factor or equilibrium parameter, RL, as defined in Eq. (4), to

determine whether the adsorption is favourable or not:

R� =�

������ …………… (4)

where C0 is the highest initial phosphate concentration (mg P/L) and KL is the Langmuir’s

adsorption constant (L/mg P).

Adsorption kinetic experiments were conducted as follows: 0.10 g of the adsorbent was

added in 100.0 mL of phosphate solution with an initial concentration of 50 mg P/L and 2 mg

P/L, respectively. The sealed polypropylene bottle was then placed in the shaker bath for 48

h. 2.0 mL of suspension was taken out of the bottle over a given period of time and the

phosphate concentration was analysed after contrifugation. The amount of phosphate

adsorbed onto the sample at different time periods (qt) was calculated by Eq. (5),

q� =(�����)

� …………… (5)

where C0 and Ct are the phosphate concentrations (mg P/L) in solution at initial and different

time periods, respectively.

In order to analyse the kinetic mechanism of adsorption process, the experimental data were

fitted in the pseudo-first-order and pseudo-second-order models, which are described as Eqs.

(6) and (7) [27-30]:

Pseudo-first-order: log(q� − q�) = logq� −� �

!.#$# …………….(6)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9

Pseudo-second-order: �

%�=

�

�&%�& +

�

%� …………… (7)

where qt and qe are the amounts of phosphate adsorbed over a given period of time t (mg P/g)

and at equilibrium (mg P/g), respectively; t is the adsorption time (h); k1 (1/h) and k2 (g/mg

P/h) are the adsorption rate constants of the pseudo-first-order and the pseudo-second-order

adsorption, respectively.

To investigate the effect of pH on the phosphate adsorption, 0.025 g of the adsorbent was

added in 25.0 mL of 50 mg P/L phosphate solution of different initial pH values, ranging

from 3.0 to 11.0. The initial pH of phosphate solution was adjusted with 0.10 M NaOH and

HCl solution. To study the effect of coexisting anions on the phosphate removal rates, 0.025

g of the adsorbent was added into the phosphate solution containing 0.01 M coexisting ion,

which was prepared by dissolving sodium salt form of F−, Cl−, NO3−, SO4

2−, or CO32− into 50

mg P/L phosphate solution.

3. Results and discussion

3.1 Morphological and structural properties

Fig. 1 exhibits the SEM images and XRD patterns of the as-prepared parent FMS and FMS-

xLa materials (x = 0.02, 0.04, 0.1, 0.2) with different lanthanum loadings. SEM images (Fig.

1a-e) show that all of the samples are consisting of uniform and monodispersed spheres with

an average particle diameter of approximately 300 nm. The possible formation mechanism of

the series of FMS and FMS-xLa samples is spectulated via a self-assembly process as shown

in Fig. S1 [23, 24]. In the synthetic micro-emulsion system, CTAB micelles firstly assemble

into flower-like template molecules (Fig. S1a). Simultaneously, the negatively charged

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10

silicate molecules, produced after the hydrolysis and ionization of TEOS with urea, are

arranged within the space among the CTAB template molecules (Fig. S1b); the silicates are

then condensed at high temperature to form silica material. The following calcination at 550

⁰C removes the CTAB template, thus generating flower-like ordered porous silica spheres

(FMS) (Fig. S1c). The impregnation of La3+ (Fig. S1d) and subsequent calcination (Fig. S1e)

result in the generation of active sites to phosphate anions; which endows P adsorption

capacity to the La-loaded FMS mateirals (Fig. S1f). In particular, the shperes exhibit unique

hierachially flower-like porous structure (see Fig. 1), enabling the high surface area and easy

phosphate accessiblity. There is no significant difference observed in the morphologies

between the parent FMS and La-functionalized FMS-xLa materials (Fig. 1a-e), i.e. FMS-

0.02La, FMS-0.04La, FMS-0.1La, and FMS-0.2La; that indicates FMS is stable and can

totally maintain its interesting flower-like structure during the lanthanum impregnation

process. Moreover, the images suggest that all lanthanum (III) have been loaded into the

porous FMS; this is further supported by the EDX analysis shown in Table 1.

In the XRD patterns (Fig. 1f) of FMS and FMS-xLa (x = 0.02, 0.04, 0.1, and 0.2) samples,

noted that three peaks, which are marked by black arrows, arise from the Al2O3 background

plate. In the XRD pattern of FMS, a broad band centered at 2θ value of 23 ⁰ (indicated by a

solid circle) is assigned to the amorphous mesoporous SiO2 (JCPDS No. 29-0085). By

increasing La loading, the SiO2 characterisitic broad band becomes weaker in Fig. 1f; this is

probably due to the coverage of lanthanum onto the mesoporous surface. Meanwhile, several

new weak peaks (as indicated by the open triangles) are observed in FMSM-xLa (x = 0.02,

0.04, 0.1 and 0.2). They can be indexed to lanthanum oxide (PDF No. 83-1943), revealing the

presence of La2O3 in the samples [19].

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

11

N2 adsorption-desorption isotherms and BJH distributions of FMS and FMS-xLa (x = 0.02,

0.04, 0.1, and 0.2) are shown in Fig. 2a and b; the BET surface areas (SBET), pore diameters

(DBJH) and total pore volumes (Vtotal) are summarized in Table 1. The parent sample FMS

synthesized without lanthanum modification shows a typical IV isotherm with a H2-

hysteresis loop, suggesting the existence of irregular mesopores [31, 32]. Seen from the BJH

distribution (Fig. 2b), the FMS sample exhibits a multimodal pore size distribution, centering

at 3.4 nm and 13.5 nm, respectively. The intensity of former one (in the range of 2-5 nm) is

significantly greater, assigned to the small mesopores located inside the sample. On the other

side, the latter one with a relatively low pore density and wide pore distribution in the range

of 6-30 nm corresponds to the presence of large pores on the particle surfaces. This result

reveals that the FMS sample has a unique porous structure, which is consisting of externally

large mesopores and internally small mesopores. The FMS-xLa samples show similar typical

IV isotherms and H2-hysteresis loops (Fig. 2a); whilst their SBET and Vtotal decrease

continuously with increasing x values (Table 1). In comparison with the high BET surface

area (451.6 m2/g) of FMS, the BET surfaces for FMS-xLa samples are markedly reduced

from 278.2 m2/g to 67.4 m2/g by increasing x from 0.02 to 0.2. Similarly, the pore volumes

continuously drop from 0.64 cm3/g (FMS-0.02La), 0.54 cm3/g (FMS-0.04La), 0.32 cm3/g

(FMS-0.1La), to 0.27 cm3/g (FMS-0.2La), respectively. Meanwhile, the hysteresis loops

gradually shift from P/P0 = 0.4 to 0.8 by changing x values from 0.02 to 0.2, indicating the

increasing pore sizes. The corresponding porosimetry results also confirm this increasing

tendency of mesopore diameters from 7.95 nm to 15.02 nm at greater x values (Fig. 2b and

Table 1), all of which are larger than the pore size of FMS (6.36 nm). These changes in

textural properties can be explained by the increasing loading of lanthanum components.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

12

When x is changed from 0.02 to 0.2, the actual amount of lanthanum in the corresponding

sample is increased from 4.25 wt% to 30.02 wt%; which is very close to the theoretical

loading (Table 1). The presence of lanthanum components, likely in the form of lanthanum

oxide (characterized by XRD, Fig.1f), in the unique mesoporous channels of flower-like

silica spheres leads to the decreases of SBET and Vtotal at higher x values. Meanwhile, when

introducing La2O3, it may firstly occupy the small inner mesoporous channels and then fill up

the relatively large outer mesopores. Therefore, it is not surprising to see the increase of BJH

pore diameters from 7.95 nm for FMS-0.02La up to 15.02 nm for FMS-0.2La (Table 1).

These observed changes in textural properties further confirm the successful incorporation of

lanthanum components into the mesoporous channels of FMSM-xLa, which are expected to

act as reactive sites to phosphate anions, leading to excellent phosphorus adsorption

performance.

3.2 Phosphate adsorption

Fig. 3 shows the Langmuir and Freundlich isotherm fitting plots of FMS-xLa (x = 0.02, 0.04,

0.1, and 0.2) at various initial P concentrations. Their corresponding isotherm parameters are

summarized in Table 2. The adsorption performance of FMS which was synthesized without

lanthanum impregnation was also carried out for comparison. The adsorption capacity is

found to be nearly zero, which indicates that the negligible amount of phosphate can be

removed by FMS. In contrast, the FMS-xLa samples have distinct amounts of adsorption,

which are enhanced at greater x (Table 2). Both Langmuir and Freundlich models are suitable

to describe the phosphorus sorption isotherms of the FMS-xLa materials. However, the

Langmuir equation gives a better fit than the Freundlich equation according to their

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

13

correlation coefficients (R2 > 99%, Table 2), indicating that the adsorption process occurs on

the surface of FMS-xLa via monolayer adsorption. Similar findings have also been reported

elsewhere by using lanthanum-modified bentonite [33], lanthanum-doped vesuvianite [34]

and lanthanum-loaded orange waste [35]. The maximum adsorption capacities (qm) of FMS-

xLa estimated by the Langmuir model (Table 2) are 9.72 mg P/g, 14.74 mg P/g, 42.76 mg P/g

and 44.82 mg P/g, corresponding to x = 0.02, 0.04, 0.1 and 0.2, respectively. Moreover, in

Table 2, the factor RL, which is within the range of 0 and 1.0, suggests the favourability of

phosphate adsorption onto all of the lanthanum modified-FMS samples [36].

In addition, the molar ratio of P adsorbed versus La (defined as P/La) has been considered as

an important indicator showing lanthanum usage efficiency of the adsorbents [19]. Based on

the values of P/La in Table 2, although FMS-0.2La has the largest maximum adsorption (qm),

its P/La is the lowest, indicating that the further increasing loading of lanthanum components

in FMS-xLa materials reduces the lanthanum usage efficiency. Therefore, taking cost

effectiveness into consideration for practical application, the highest P/La value for the

sample FMS-0.1La suggests the best lanthanum usage efficiency, thus it was chosen for

studying the effects of time, pH and co-existing ions on its adsorption.

Table 3 shows the comparison between our FMS-0.1La and other reported La-modified

mesoporous silica adsorbents in terms of the maximum adsorption capacity and P/L ratio.

Obviously, the maximum adsorption capacity of FMS-0.1La, 42.82 mg P/g, is superior to

those of La-doped MCM-41 (7.18 mg P/g) [17], La-coordinated amino-functionalized MCM-

41 (17.72 mg P/g) [13], La-doped SBA-15 (42.16 mg P/g) [18], and La-doped mesoporous

SiO2 (23.25 mg P/g) [37]. On the other side, FMS-0.1La possesses an improved La utilization

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

14

efficiency with a comparable adsorption capacity, in relative to La100SBA-15 and HMS-1/5

[8, 18]. This would thanks to the unique pore structures of FMS, narrow inner channels and

wide outer pores. Since La is less expensive than other rare earth elements (REEs) [20, 34],

taken account of the excellent La utilization efficiency and high adsorption capacity, our La-

doped mesoporous silica adsorbent shows a promising potential in application.

To reveal the structures of FMS-0.1La after adsorption, XRD, TEM and EDX analysis were

carried out (Fig. S2 and Fig. 4). Fig. S2 shows that FMS-0.1La is consisting of LaPO4 species

after adsorption test. In Fig. 4a and b, the flower-like structures of spheres are still retained

after adsorption. Some rod-like nanocrystals with a thickness of around 4 nm (as pointed by c

in Fig. 5b) are observed to grow from the mesopores of flower-like silica spheres. The HR-

TEM image indicates that these rod-like nanocrystals are LaPO4 species with a d120 spacing

of 0.31 nm (Fig. 4c) [38]. This is further supported by the EDX analysis of rod-like

nanocrystal (Fig. 4e), which indicates a La wt% of 64.16% being close to that of pure LaPO4

(59.39%). Similar rod-like LaPO4 crystals have also been observed by using La2O3-modified

SBA-15 in phosphate removal, where the chemical reaction between P and La led to a

heterogeneous nucleation of LaPO4 species in the confined nanopores [18]. Since our flower-

like silica spheres have a unique porous structure with small internal mesopores and wide

mesopore openings. It is reasonable to believe that those rod-like LaPO4 crystals grow from

the internal mesopores of spheres, due to the chemical reaction between adsorbed phosphate

and La active sites on the inner channels. Meanwhile, due to the large outer mesopores of

FMS-0.1La, the diffusion problem caused by the blocking of pore channels with the formed

LaPO4 nanorods could be minimized. In addition to the rod-like LaPO4 crystals, some

particles attached to the flower-like surface of sphere are observed with a dot-like

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

15

morphology, as pointed by (d) in Fig. 4b. The HR-TEM image (Fig. 4d) indicates that such

dot-like species are also LaPO4 (a spacing of 0.26 nm corresponding to (202) plane of

LaPO4). This is further affirmed by the EDX result (Fig. 4e) showing that they are composed

of La, P, Si, and O with an equal La/P molar ratio. However, the mass amount of Si is 47.10

% for the dot-like particle, which is ten times more than that is found in the rod-like crystal.

We suspect the dot-like LaPO4 species may result from the heterogeneous nucleation of

LaPO4 on the large outer mesopore surface of FMS-0.1La; the product appearing alike in

morphology is also reported in the phosphate adsorption by using lanthanum-doped

macroporous silica foams [19].

Fig. 5 shows the impacts of time, pH and co-existing ions on the adsorption of our selected

FMS-0.1La sample. The kinetics of phosphate adsorption onto FMS-0.1La has been

evaluated, as shown in Fig. 5a and Table S1. The equilibrium in the solution with an initial

concentration of 50 ppm can reach after 24 h. Fig. S3 shows the kinetics of adsorption by

using FMS-0.1La in 2 mg P/L solution, which is finished within 2 h. This fast adsorption in a

solution with a low phosphorus concentration proves the potential use of FMS-0.1La in

polishing secondary effluent to meet stringent standard of discharge [8]. To further

understand the adsorption process, the kinetic curves were fitted in the pseudo-first-order and

pseudo-second-order models (Table S1). Relatively higher correlation coefficients when

using the pseudo-second-order model indicate that the adsorption process of phosphate onto

FMS-0.1La is more likely to be chemisorption.

Fig. 5b exhibits the phosphate removal rates of FMS-0.1La by varying initial pH values in the

range of 3.0–11.0 (the corresponding equilibrium pH values are shown in Fig. S4). Fig. S5

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

16

suggests a negligible La3+ concentration observed in the solution while the adsorbent can

largely maintain its morphology within the pH of 3.0 – 11.0. The adsorption process of

phosphate is strongly dependent on the initial pH value of solution. High phosphate

adsorption capacities of around 37 mg/P g are observed from the pH = 3.0 to 6.0. As the pH

is further increased from 6.0 to 11.0, the phosphate adsorption capacities decrease by 34.7%,

gradually from 37.02 mg P/g to 24.17 mg P/g. As known, phosphate may exist in different

ionic species, e.g. H2PO4−, HPO4

2−, and PO43−, depending on the pH of solution [39]. When

the pH value is between 2.13 and 7.20, the main species is monovalent H2PO4−. The high

adsorption capacities observed between pH 3.0 and 6.0, indicating that FMS-0.1La provides a

great affinity to the single charged phosphate species (H2PO4−). The decrease of adsorption

capacities in the pH range of 6.0 to 11.0 may be explained by the ligand-exchange

mechanism. Similarly to other metal oxides, the embedded La2O3 in FMS-0.1La forms

hydroxide when in aqueous solution [6]; specifically attracting phosphate via the ligand-

exchange mechanism, as shown in the following Eqs:

−−

−−

+−↔≡+−≡

+−↔≡+−≡

OHHPOLaHPOOHLa

OHPOHLaPOHOHLa

2)()(2

)(

422

4

4242

The ligand-exchange mechanism involved in the phosphate adsorption process by using

FMS-0.1La can be further confirmed by monitoring the pH change during the adsorption

process, as shown in Fig. S6. The initial pH of 50 mg P/L solution is 5.07. After adding 50

mg FMS-0.1La, the pH value is increased to 5.95 in one hour and then gradually up to 8.95 at

24 h, where the pH value finally reaches the equilibrium. In addition, this observation well

supports the decrease of phosphate uptake observed in the pH range of 6.0 and 11.0 (Fig. 5b),

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

17

because high concentration of OH– is thermodynamically unfavourable for the uptake of

phosphate via the ligand-exchange mechanism [19].

Fig. 5c shows the effects of coexisting anions, including F−, Cl−, NO3−, SO4

2−, and CO32−, on

the phosphate adsorption capacity of FMS-0.1La. The phosphate adsorption capacity (qe) of

FMS-0.1La in the absence of competitive anions is 40.25 mg P/g. When introducing 0.01 M

CO32−, its phosphate adsorption capacity is reduced by 27.6%. Similar finding was also

reported in other studies [18, 19]. It could be explained by the lower Ksp of La2(CO3)3

(3.98×10−34), as compared with that of LaPO4 (3.7×10−23); that favours the displacement of

adsorbed PO43− on FMS-0.1La by CO3

2− ions and the subsequent transformation of formed

LaPO4 to La2(CO3)3, thus deteriorating the phosphate adsorption capacity [18, 19]. Moreover,

the initial pH of the solution with 0.01 M CO32− is 10.63; under that a large amount of OH− is

present. OH− may compete with PO43− for adsorption active sites, thus hindering ligand-

exchange mechanism and lowering phosphate adsorption. However, the presence of other

0.01 M coexisting anion, i.e. F−, Cl−, NO3−, or SO4

2−, exerts a negligible effect on adsorption

capacity, suggesting that the sample FMS-0.1La possess a high selectivity to phosphate

anions

4. Conclusions

A series of flower-like mesoporous silica spheres doped with lanthanum (FMS-xLa) utilized

for phosphate removal were reported for the first time. By increasing the theoretical La/Si

molar ratio from 0.02 to 0.2, the resulting adsorbent possessed improved phosphate

adsorption capacity. Nevertheless, the FMS-0.1La adsorbent prepared with a theoretical La/Si

molar ratio of 0.1 exhibited the highest lanthanum usage efficiency, in the view of P/La molar

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

18

ratio. For FMS-0.1La, its adsorption kinetic data were better described by the pseudo-second-

order model. The optimal pH for phosphate adsorption by using FMS-0.1La was between 3.0

and 6.0. The co-existing ions, except CO32-, had negligible effects on the phosphate

adsorption, suggesting that the FMS-0.1La sample possess a high selectivity to phosphate

anions. In relative to currently reported La-modified mesoporous silica phosphate adsorbents,

high phosphorous removal capacity and superior La utilization efficiency have been observed

by using our as-prepared material, thus holding out a promising prospect of practical use.

Acknowledgements

The part of this research was supported by Zhejiang Provincial Natural Science Foundation

of China under Grant No. LQ14B070005. Dr D. Li acknowledges the support from Murdoch

SEIT & University Small Grants. Dr W. Huang thanks the financial support from Zhejiang

Provincial Key Discipline of Applied Chemistry and the program of China Scholarship

Council (No. 201206780010). Appreciation is also expressed to Dr Yuzhou Wu, from

Monash University (Australia), for providing SEM analysis and images in this manuscript.

Appendix A. Supplementary data

Supplementary data are available: scheme for formation of FMS and FMS-xLa and

illustration of P adsorption; equilibrium pH of solution versus initial pH of solution by

using FMS-0.1La as the adsorbent; the change of solution pH values during the

adsorption; XRD pattern of FMS-0.1La after the adsorption; the residual La

concentrations (with SEM images of FMS-0.1La as insets) in solutions of different initial

pH values; the parameters derived from the pseudo-first-order and pseudo-second-order

fitting plots of FMS-0.1La adsorption in 50 ppm and 2 ppm solution, respectively.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

19

References

[1] K. G. Sellner, G. J. Doucette, G. J. Kirkpatrick, J. Ind. Microbiol. Biot. 30 (2003)

383-406.

[2] G. Gao, B. Qin, R. Sommaruga, R. Psenner, Hydrobiologia 581(2007) 177-188.

[3] G. K. Morse, S. W. Brett, J. A. Guy, J. N. Lester, Sci. Total Environ. 212(1998) 69-

81.

[4] S. K. Kang, K. H. Choo, K. H. Lim, Sep. Sci. Technol. 38(2003) 3853-3874.

[5] J. Yang, L. Zhou, J. Zhang, J. Zou, Z. Yuan, C. Yu, Chem. Eur. J. 19(2013) 5578-

5585.

[6] Z. L. Shi, F. M. Liu and S. H. Yao, New Carbon Mater., 26(2011) 299-306.

[7] L. Zhang, L. Wan, N. Chang, J. Liu, C. Duan, Q. Zhou, X. Li and X. Wang, J.

Hazard. Mater., 190(2011), 848-855.

[8] W. Huang, Y. Zhu, J. Tang, X. Yu, X. Wang, D. Li, Y. Zhang, J. Mater. Chem. A

2(2014) 8839-8848.

[9] C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli, J. S. Beck, Nature

359(1992) 710-712.

[10] D. Y. Zhao, J. L. Feng, Q. S. Huo, N. Melosh, G. H. Fredrickson, B. F. Chmelka, G.

D. Stucky, Science 279(1998) 548-552.

[11] P. Delaney, C. McManamon, J. P. Hanrahan, M. P. Copley, J. D. Holmes and M. A.

Morris, J Hazard. Mater. 185(2011) 382-391.

[12] W. Chouyyok, R. J. Wiacek, K. Pattamakomsan, T. Sangvanich, R. M. Grudzien, G.

E. Fryxell, Environ. Sci. Technol. 44(2010) 3073-3078.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

20

[13] J. Zhang, Z. Shen, W. Shan, Z. Mei, W. Wang, J. Hazard. Mater. (2011) 18676-

18683.

[14] J. Zhang, Z. Shen, Z. Mei, S. Li, We. Wang, J. Environ. Sci. 23(2011) 199-205.

[15] Y. Tang, E. Zong, H. Wan, Z. Xu, S. Zheng, D. Zhu, Micropor. Mesopor. Mat.

155(2012) 192-200.

[16] E. W. Shin, J. S. Han, M. Jang, S. H. Min, J. K. Park, R. M. Rowell, Environ. Sci.

Technol. 38(2004) 912-917.

[17] J. Zhang, Z. Shen, W. Shan, Z. Chen, Z. Mei, Y. Lei, W. Wang, J. Environ. Sci.

22(2010) 507-511.

[18] J. Yang, L. Zhou, L. Zhao, H. Zhang, J. Yin, G. Wei, K. Qian, Y. Wang, C. Yu, J.

Mater. Chem. 21(2011) 2489-2494.

[19] J. Yang, P. Yuan, H. Y. Chen, J. Zou, Z. G. Yuan, C. Z. Yu, J. Mater. Chem.

22(2012) 9983-9990.

[20] E. W. Shin, K. G. Karthikeyan, M. A. Tshabalala, Environ. Sci. Technol. 39(2005)

6273-6279.

[21] D. Y. Zhao, Q. S. Huo, J. L. Feng, B. F. Chmelka, G. D. Stucky, J. Am. Chem. Soc.

120(1998) 6024-6036.

[22] S. Garg, K. Soni, G. M. Kumaran, R. Bal, K. Gora-Marek, J. K. Gupta, L. D.

Sharma, G. M. Dhar, Catal. Today., 141(2009) 125-129.

[23] S. Gai, P. Yang, P. Ma, D. Wang, C. Li, X. Li, N. Niu, J. Lin, J. Mater. Chem.

21(2011) 16420-16426.

[24] S. Gai, P. Yang, P. Ma, L. Wang, C. Li, M. Zhang, L. Jun, Dalton Trans. 41(2012)

4511-4516.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

21

[25] I. Langmuir, J. Am. Chem. Soc., 38(1916) 2221-2295.

[26] H. M. F. Freundlich, J. Phys. Chem. 57(1906) 385-471.

[27] Y. S. Ho, J. Hazard. Mater. 136(2006) 681-689.

[28] Y. S. Ho, G. McKay, Process Biochem. 34(1999) 451-465.

[29] S. Lagergren, Kungliga Svenska Vetenskapsakademiens Handlingar 24(1898) 1-39.

[30] Y. S. Ho, G. McKay, Chem. Eng. J. 70(1998) 115-124.

[31] Z. Sun, Y. Deng, J. Wei, D. Gu, B. Tu, D. Zhao, Chem. Mater. 23(2011) 2176-2184.

[32] T. Horikawa, D. D. Do, D. Nicholson, Adv. Colloid Interface Sci. 169(2011) 40-58.

[33] F. Haghseresht, S. Wang, D. D. Do, Appl. Clay Sci. 46(2009) 369-375.

[34] H. Li, J. Ru, W. Yin, X. Liu, J. Wang, W. Zhang, J. Hazard. Mater. 168(2009), 326-

330.

[35] B. K. Biswas, K. Inoue, K. N. Ghimire, S. Ohta, H. Harada, K. Ohto, H. Kawakita,

J. Colloid Interface Sci. 312(2007) 214-223.

[36] Y. Al-Degs, M. A. M. Khraisheh, S. J. Allen, M. N. Ahmad, Water Res. 34(2000)

927-935.

[37] E. Ou, J. Zhou, S. Mao, J. Wang, F. Xia, L. Min, Colloids Surf. A 308(2007) 47-53.

[38] Z. Chai, L. Gao, C. Wang, H. Zhang, R. Zheng, P. A. Webley, H. Wang, New J.

Chem. 33(2009) 1657-1662.

[39] D. D. Perrin, B. Dempsey, Buffers for pH and Metal Ion Control, Chapman & Hall,

London, 1974.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

22

Figure Captions:

Fig. 1. SEM images (a-e) and XRD patterns (f) of the samples: FMS (a), FMS-0.02La (b),

FMS-0.04La (c), FMS-0.1La (d), and FMS-0.2La (e), respectively. In Fig. 1 (f), the peaks

indicated by black arrows are from the Al2O3 background plates, whilst the solid circle refers

to the diffraction of amorphous SiO2 and the open triangles refer to the diffractions of La2O3.

Fig. 2. N2 adsorption-desorption isotherms (a) and BJH distributions (b) of FMS, FMS-

0.02La, FMS-0.04La, FMS-0.1La, and FMS-0.2La, respectively.

Fig. 3. The Langmuir adsorption isotherm (a) and Freundlich adsorption isotherm (b) fitting

plots of FMS-xLa (x = 0.02, 0.04, 0.1, 0.2) samples.

Fig. 4. TEM images of FMS-0.1La before adsorption (a) and after adsorption (b); (c) and (d)

HR-TEM images of rod-like crystal (pointed by c in Fig. 4b) grown from the mesopores and

dot-like crystal (pointed by d in Fig. 4b) in FMS-0.1 La; (e) EDX data for (c) and (d).

Fig. 5. (a) The pseudo-second-order fitting plot of phosphate adsorption onto FMSM-0.1La;

(b) effect of pH on the phosphate removal capacities of FMS-0.1La; (c) effect of co-existing

anions on the phosphate adsorption capacities of FMS-0.1La.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

23

Table Captions:

Table 1. Structural characteristics of FMS and FMS-xLa (x = 0.02, 0.04, 0.1, and 0.2).

Table 2. Langmuir and Freundlich isotherm parameters in the phosphate adsorption by using

FMS-xLa (x = 0.02, 0.04, 0.1, and 0.2).

Table 3. Comparison of phosphate adsorption capacities and P/La between FMS-0.1La and

other lanthanum-modified mesoporous silica in literature.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

a is the mean actual La wt%, calculated by averaging three EDS results; b is the theoretical La wt%.

Samples SBET (m2/g) DBJH (nm) Vtotal (cm3/g) Laa (wt%) Lab (wt%)

FMS 451.6 6.36 0.84 0 0

FMS-0.02La 278.2 7.95 0.64 4.25 4.40

FMS-0.04La 231.2 8.30 0.54 6.70 8.39

FMS-0.1La 102.6 13.02 0.32 16.56 18.30

FMS-0.2La 67.4 15.02 0.27 30.02 30.53

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

a is calculated by molar ratio of P adsorbed (calculated from qm) versus La (calculated from

actual La loading wt% in Table 1).

Samples (P/La)a Langmuir Freundlich

qm (mg P/g) KL (L/mg P) RL R2 n KF R2

FMS-0.2La 0.67 44.82 0.386 0.0314 0.994 2.963 13.53 0.967

FMS-0.1La 1.16 42.76 0.350 0.0345 0.993 3.545 14.24 0.988

FMS-0.04La 0.99 14.74 0.317 0.0379 0.996 6.945 7.37 0.940

FMS-0.02La 1.03 9.72 0.465 0.0262 0.999 5.401 4.70 0.981

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Samples Adsorption

capacity (mg P/g)

P/La References

La10-meso-SiO2 23.10 / [37]

La-coordinated amino-functionalized MCM-41 17.72 / [13]

La-doped MCM-41(La25M41) 7.18 0.43 [17]

La40MCM-41 23.77 0.96 [18]

La80SBA-15 40.52 0.92 [18]

La100SBA-15 45.63 0.88 [18]

HMS-1/5 47.89 0.96 [8]

FMS-0.1La 42.76 1.16 Present work

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Research highlights

• Flower-like mesoporous silica loaded with La was used for P adsorption.

• It showed a maximum adsorption capacity of 44.8 mg P/g at 25 °C.

• Its adsorption was pH-dependent and little affected by competitive ions.

• It exhibited superior La utilization efficiency compared to literature.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1

Supplementary Data for

Enhanced Adsorption of Phosphate by Flower-like Mesoporous Silica

Spheres Loaded with Lanthanum

Wei-Ya Huang, a Xiang Yu,

c Jinpeng Tang,

c Yi Zhu,

c Yuan Ming Zhang

c and Dan Li

b*

aDepartment of Materials Science and Engineering, Taizhou University, Linhai, 317000,

China

bChemical and Metallurgical Engineering and Chemistry, School of Engineering and

Information Technology, Murdoch University Murdoch, Western Australia, 6150, Australia;

cDepartment of Chemistry, Jinan University, Guangzhou, 510632, China

*Corresponding author. Tel: +61 08 9360 2569, Email: l.li@murdoch.edu.au (D. Li).

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2

Fig. S1. Formation of FMS and FMS-xLa (x = 0.02, 0.04, 0.1, 0.2) and illustration of P

adsorption [1, 2].

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3

Fig. S2. XRD pattern of FMS-0.1La after the adsorption; the standard XRD pattern of LaPO4

is shown at the bottom.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4

Fig. S3. Kinetics of adsorption by using FMS-0.1La in the solution with an initial

concentration of 2 mg P/L.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5

Fig. S4. Equilibrium pH of solution versus initial pH of solution by using FMS-0.1La as the

adsorbent.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6

Fig. S5. The residual La concentrations in the solutions of different pH values; the insets are

SEM images of FMS-0.1La in the solution of pH = 3.0 and 11.0.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7

Fig. S6. The change of solution pH values during the adsorption in the solution with an initial

concentration of 50 mg P/L.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8

Table S1. Parameters derived from the pseudo-first-order and pseudo-second-order fitting

plots of FMS-0.1La adsorption in the solution with an initial concentration of 50 mg P/L and

2 mg P/L.

First-order kinetics Second-order kinetics

C0 (mg P/L) k1 qe (cal) (mg P/g) R2 k2 qe (cal) (mg P/g) R

2

50 0.1556

(1/h) 39.6 0.9738

0.00403

(g/mg P/h) 46.1 0.9924

2 1.0363

(1/min) 1.8 0.8906

0.097

(g/mg P/min) 2.0 0.9999

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9

References:

[1] S. Gai, P. Yang, P. Ma, D. Wang, C. Li, X. Li, N. Niu, J. Lin, J. Mater. Chem.

21(2011) 16420-16426.

[2] S. Gai, P. Yang, P. Ma, L. Wang, C. Li, M. Zhang, L. Jun, Dalton Trans. 41(2012)

4511-4516.