Journal of Environmental Management 151 (2015) 450e460

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

A fixed-bed column for phosphate removal from aqueous solutions using an andosol-bagasse mixture
we
, Daniel Njopwouo Emmanuel Djoufac Woumfo*, Jean Mermoz Sie Laboratoire de Physico-chimie des Mat
eriaux Min
eraux, D
epartement de Chimie Inorganique, Facult
e des Sciences, Universit
e de Yaound
e I, B.P. 812, Yaound
e, Cameroon

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 September 2013 Received in revised form 25 September 2014 Accepted 26 November 2014 Available online 21 January 2015

It is difficult to eliminate phosphate from large volumes of water in batch mode using an adsorbent such as andosol. In a fixed-bed column, andosol has a very low permeability. In this study, andosol was mixed with bagasse to increase permeability. The mixture was then applied for the adsorption of phosphate in a fixed-bed column. Optimum and stable permeability was obtained with a 50/50 mixture of andosol and bagasse. The maximum adsorption capacity obtained was 4.18 mg/g for a column with a bed depth of 1.8 cm and a flow rate of 4 mL/min. The experimental data fit best to Thomas and AdameBohart models. These experimental results were applied in the treatment of natural phosphate-containing water from
Municipal Lake in Cameroon. Column performance increased by 60% due to the presence of Yaounde Ca2þ and Mg2þ in the natural water. These cations form complexes with phosphate at the andosol surface. The standard enthalpy 15.964 kj/mol indicated that phosphate adsorption on andosol-bagasse mixture was an endothermic process. Kinetic experiments demonstrated that phosphate adsorption fitted better with a pseudo-second-order model. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Phosphate Fixed-bed Adsorption Andosol Breakthrough

1. Introduction Phosphate found in wastewater or surface water primarily originates from fertilisers (via runoff waters), human and animal droppings, detergents and other domestic wastes. Within the pH 2 range of 3e11, H2PO 4 and HPO4 are the predominant phosphate species. Phosphate concentrations 0.1 mg/L (Rodier, 1986) favour massive algal growth, which leads to the eutrophication of water bodies (Zhang et al., 2009). Many processes for the elimination of phosphate by physical, chemical or biological treatment have therefore been developed to avoid excessive algal growth. According to Karaça et al. (2004), methods such as inverse osmosis and electro-dialysis are very costly and inefficient because they eliminate only 10% of the total phosphate. Precipitation with lime, iron and aluminium salts is the most commonly used chemical process; however, the process requires a large quantity of these salts. They eliminate approximately 98% of the total phosphate but produce a large quantity of mud (Youcef and Achour, 2005). According to Yeom and Jung (2009), biological processes can

* Corresponding author. E-mail address: [email protected] (E.D. Woumfo). http://dx.doi.org/10.1016/j.jenvman.2014.11.029 0301-4797/© 2014 Elsevier Ltd. All rights reserved.

eliminate up to 97% of phosphate. Notably, various biological methods have been developed but have been found to be inefficient at low phosphate concentrations (Shin et al., 2004). Various studies, including Kuo and Lotse (1974) and Clark and McBride (1984), have instead recommended the elimination of phosphate at low concentrations through adsorption on solid materials such as alumina, activated carbon and clay. According to Yan et al. (2014), adsorption is considered a more effective way to remove pollutants, due to its high removal efficiency, low operational costs and low residue production. Andosol is a dark-coloured soil derived from parent material of volcanic origin, such as volcanic ash, volcanic tuff and pumice. Reports on phosphate adsorption by andosol in the batch mode place the adsorption capacities between 4.15 and 8.37 mg/g of soil (Siewe et al., 2008). This capacity can reach 28 mg/g for allophonerich soils (Parfitt and Hemni, 1980). Andosols used as adsorbents are abundant and low cost. The results obtained experimentally using the batch mode seem to be difficult to apply to the treatment of large volumes of water containing phosphate (e.g., lakes and ponds). Adsorbing the components of a fluid mixture flowing through a packed bed of a porous adsorbent material is the basis of several important applications in chemical engineering. Continuous adsorption experiments are generally performed to

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

Nomenclature A Ad Bg Cad Co Ct EBCT kAB kTh kYN mtotal No

cross-sectional area of the bed, cm2 modified andosol bagasse concentration of phosphate removal, mg/L influent concentration, mg/L effluent concentration, mg/L empty bed contact time, min kinetic constant, L/mg min Thomas model constant, mL/min mg rate constant, min1 total amount of phosphate ion applied to the column, g saturation concentration, mg/L

qo qe qtotal Q t ttotal Uo Veff H P

451

adsorption capacity, mg/g equilibrium phosphate uptake or maximum capacity of column, mg/g the total mass of phosphate adsorbed, mg volumetric flow rate, mL/min flow time, min total flow time, min superficial velocity, cm/min effluent volume, mL bed depth of the fixed-bed column, cm percent removal of phosphate ions, %

Greek symbols time required for 50% adsorbate breakthrough, min

t

investigate the effect of certain parameters, such as flow rate, bed depth and initial concentration, on pollutant removal (Gupta and Babu, 2010). Sugarcane bagasse is the residue left after pressing sugarcane stalks to extract the juice. It is generally employed to generate heat and power to run the sugar milling process. As with many other lignocellulosic materials, there has been an increased interest in developing techniques to convert these materials into environmentally friendly chemicals and biomaterials (Pehlivan et al., 2013). Bagasse is a low density (200 kg/m3) porous structure with great water retention capacity (8.40 g/g) (Brandaohttp:// academic.research.microsoft.com/Author/47799706/poliana-cbrandao et al., 2010). Some laboratory experiments have indicated that during the flow of an effluent in a column containing andosol, fine particles progressively pass through the column and fill the pores through which the effluent percolates. This process leads to a progressive decrease in the permeability of the clay material. A great majority of studies undertaken on phosphate removal have been conducted in batch mode using adsorbents such as andosols, slag, fly ash, powdered aluminium oxides and activated red mud (Lu et al., 2009; Xu et al., 2010; Safaa and Ragheb, 2013). In fixedbed columns, the removal of phosphates was conducted with ion exchange resins (Nur et al., 2013), nanosized FeOOH-modified anion resins (Li et al., 2013), and nanosized magnetic impregnated granular activated carbon (Zach-Maor et al., 2011). These synthetic adsorbents are expensive. Andosol and bagasse are inexpensive local materials. Their use in a fixed-bed column has not yet been studied. The aim of this work was to study the elimination of phosphate in batch and dynamic mode after improving the permeability of andosol with bagasse. The percentage of andosol that enables stable permeability and reproducibility was determined. The effects of several important design parameters such as bed height, flow rate and initial phosphate concentration in solution were investigated. In addition, models were applied to fit the data obtained from the
bed study. A natural phosphate-water sample from Yaounde Municipal Lake was applied to the fixed-bed column.

change of 4.40 (Siewe et al., 2008). Sugarcane pulp was provided
te
Sucrie re du Cameroun” (Cameroon). The pulp confrom “Socie sisted of an average of 45% cellulose, 24% hemicellulose and 25% lignin. The average density was 175 kg/m2.

2. Materials and methods

The dried Andosol-bagasse (AdBg) mixture that was introduced into the column had a total mass of 2.00 g. To obtain a homogeneous mixture, the mixture was stirred for 30 min on a magnetic stirring plate before being introduced into the column. The inner diameter of the glass column was 3.1 cm. Its overall length was 26.0 cm. The column was packed with known quantities of AdBg mixtures within two supporting layers of sponge. Phosphate ions were titrated using a spectrophotometer at a wavelength of 750 nm (Rodier, 1986). Continuous titration at the outlet of the column was carried out by sampling 1 mL of filtrate every 5 min.

2.1. Materials Allophanic andosol was sampled on the southern flank of Mount Meletan at the summit of the Bambouto Mountain (West Cameroon), which has a peak altitude of between 2000 and 2740 m (Tematio et al., 2004). Raw andosol contained 12.5% organic matter. Other characteristics include a cation exchange capacity of 32.20 meq/100 g, a BET surface area of 50.12 m2/g and a zero point

2.2. Elimination of organic matter from andosol In a previous work (Siewe et al., 2008), it was shown that hydrogen peroxide aqueous solution destroys the organic matter and improves phosphate adsorption. The elimination of organic matter was carried out as follows: Raw andosol was introduced into an aqueous solution of 32% H2O2 and stirred overnight. After sedimentation, the sample was collected and washed with distilled water. The resulting andosol was dried at 80  C before being ground and passed through a 160-mm-diameter sieve. This sample was denoted as Ad. 2.3. Purification of the bagasse The sugar cane pulp was air-dried in the laboratory, crushed and passed through a 250-mm-diameter sieve. In order to avoid any interaction due to saccharose, its remaining traces were eliminated by extraction with stirring in hot distilled water. The filtration was repeated several times until the filtrate appeared colourless. This bagasse (Bg) was then dried in an oven at 110  C for 24 h. 2.4. Reagents The reagents used in this study: dibasic sodium phosphate 12€n) and 0.1 M hydrate (PANREAC), 0.1 M soda solution (Riedel-de Hae €n) solution were all of extra pure hydrochloric acid (Riedel-de Hae analytical grade. A synthetic phosphate ion solution was prepared by dissolving the above phosphate salt in distilled water. A natural
Municipal Lake. solution of phosphate was obtained from Yaounde 2.5. Experimental methods

452

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

2.6. Influence of bagasse content on mixture (AdBg) permeability Measurement of AdBg permeability was carried out by determining the flow rate of 100 mL of distilled water through the mixture in the column. For each mixture, the average flow rate was calculated for 10 consecutive trials. The bagasse content was varied between 0 and 75%. The level of influent at the top of the column was kept constant during the experiments (excess influent flowing out of the column was recovered and reused). Consequently, the flow of influent into the column was determined by atmospheric pressure.

Q 1000

t¼total Z

Cad dt

(2)

t¼0

where Cad is the concentration of phosphate removal (mg/L). Equilibrium phosphate uptake qe (mg/g) or the maximum capacity of the column, is calculated as follows:

qe ¼

qtotal m

(3)

where m (g) is the dry weight of adsorbent in the column. The total amount of phosphate ions entering the column (mtotal) is determined using the following equation (Oguz and Ersoy, 2010):

2.7. Fixed-bed column adsorption 2.7.1. Synthetic phosphate solution The fixed-bed column studies were performed using the same glass column (internal diameter of 3.1 cm and length of 26 cm). The column was packed with 2.00, 4.00 and 6.00 g of the AdBg mixture within two supporting layers of sponge. Synthetic solutions of phosphate were pumped into the top of the column using a Masterflex peristaltic pump equipped with a flow rate controller. The AdBg mixture was set at a ratio (determined by the above experiments) that allowed for a constant non-zero permeability over time and a particular adsorbent bed height (equiv. to 1.8, 3.6 and 5.7 cm of bed depth) at room temperature (30 ± 2  C). The flow rate and influent phosphate concentration were kept constant at 4 mL/min and 10 mg/L, respectively. Phosphate solutions of known concentration (10, 15, 20, 25 mg/L) at pH 5.5 were allowed to flow through the column. Effluent samples were collected at regular time intervals to determine the phosphate concentration. The effluent flow was continued until there was no further adsorption, i.e., the phosphate concentration in the influent and effluent remained unchanged. Three parameters were studied using this experimental design: initial concentration of phosphate in mg/L (10, 15, 20 and 25), flow rate in min/mL (4, 6 and 8) and bed depth in cm (1.8, 3.6 and 5.7). 2.7.2. Natural water from Yaound
e Municipal Lake The parameters of the column were as follows: flow rate ¼ 4 mL/ min; bed depth ¼ 1.8 cm; phosphate concentration ¼ 10 mg/L.

mtotal ¼

Co Qttotal 1000

Breakthrough curves generally represent the performance of fixed-bed columns. The time to breakthrough appearance and the shape of the breakthrough curve are very important characteristics for determining the operation and dynamic response of a sorption column (Ahmad and Hameed, 2010; Han et al., 2009a). The effluent concentration (Ct) from the column that reaches approximately 0.1% of the influent concentration (Co) is the breakthrough point. The point where the effluent concentration reaches 98% is usually referred to as the “point of column exhaustion” (Kundu et al., 2004). Breakthrough curves are usually expressed as Ct/Co as a function of time or volume of effluent for a given bed depth (Han et al., 2009b; Kundu et al., 2004). The effluent volume, Veff (mL), can be determined from the following equation (Uddin et al., 2009):

(1)

where Q is the volumetric flow rate (mL/min) and ttotal is the total flow time (min). The value of the total mass of adsorbed phosphate, qtotal (mg), can be calculated from the area under the breakthrough curve (Eq. (2) (Han et al., 2009b)):

(4)

The percent removal of phosphate ions can be obtained from Eq. (5)

  q Y % ¼ total  100 mtotal

(5)

The flow rate represents the empty bed contact time (EBCT) in the column, as described in Eq. (6) (Netpradit et al., 2004): EBCT (min) ¼ bed volume (mL)/flow rate (mL/min)

(6)

2.9. Kinetics of phosphate adsorption Measurements of phosphate uptake using a batch method were conducted by placing 0.2 g of 50/50 andosol/bagasse mix in a flask containing 50 mL (50 mg L1) phosphate solution at pH 5.5. The content of the flask was stirred at different temperatures (20, 30 and 50  C) by using a thermocryostat (MGW LAUDA RM6). Remaining phosphate in suspension was titrated after centrifugation by the colorimetric method (Rodier, 1986). Samples were run in duplicate. The amounts of phosphate adsorbed were calculated by concentration difference using the following formula:

Qt ðmg=gÞ ¼

2.8. Column data analysis

Veff ¼ Q $ttotal

qtotal ¼

ðCo  Ct Þ$V=1000 m

(7)

where Qt is the amount of phosphate ion adsorbed on the adsorbent at time t, Co, the initial concentration of phosphate (mg/L), Ct, the concentration (mg/L) of phosphate ion in solution at time t, V, the volume (mL) of phosphate ion used, and m (g) the weight of the adsorbent used. 2.9.1. Adsorption kinetics A pseudo-first-order kinetics model and a pseudo-second-order kinetics model were used to test the dynamical experimental data. The pseudo-first-order equation:

lnðQe  Qt Þ ¼ k1 t þ ln Qe

(8)

Qe and Qt are the amount of dye (mg/g) adsorbed at equilibrium and at time t, k1 is the rate constant (min). The rate constant, k1 was obtained from slope of the linear plots of ln(Qe  Qt) against t (Bhattacharyya and Gupta, 2008). The pseudo-second-order rate equation:

t 1 1 ¼ þ t Qt k2 Qe2 Qe

(9)

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

k2 is the rate constant of pseudo-order adsorption (g/mg min). The rate constant k2 can be determined experimentally from the slope and intercept of the plot of t/Qt against t. 2.9.2. Thermodynamics of adsorption Thermodynamic parameter can be determined using the equilibrium constant, Kd(Qe/Ce) which depends on temperature and the isotherm model. The change in enthalpy (DH ) and entropy (DS ) associated to the adsorption process were calculated by using following equations (Cottet et al., 2014) 

lnKd ¼

DS DH  R RT



(10)

where R(8.3145 J/mol K) is the ideal gas constant, and T(K) is the temperature. The Gibbs free energy, (DG ) of specific adsorption is calculated from the equation: 



DG ¼ DH  TDS



(11)

According to Eqs. (10) and (11), DH , DS and DG parameters can be calculated.

453

A plot of lnk2 against 1/T provided a linear curve. Using the Arrhenius equation (Eq. (12)) the parameters of activation could be estimated from the slope and the intercept of the straight line passing through the points.

lnk2 ¼ 

Ea þ ln k0 RT

(12)

where Ea is the activation energy (kJ/mol), k2 is the pseudo-second rate constant of adsorption, ko is the pre-exponential factor, R is the ideal gas constant (8.3145 J/mol K) and T is the temperature (K) (Lin et al., 2014). 3. Results and discussion 3.1. Influence of the bagasse percentage on the permeability of the AdBg mixture The permeability of modified andosol (Ad) decreased from 2.4 mL/min to 1 mL/min after 10 consecutive trials (Fig. 1). The addition of 13e35% bagasse also displayed a progressive decrease in permeability during consecutive trials. This finding can be

Fig. 1. Effect of the percentage of bagasse on the permeability of the mixture.

454

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

explained by the migration of fine andosol particles that obstructed the micropores and mesopores as the liquid flowed through the column. However, the average permeability of the mixture improved with increasing bagasse content (Fig. 2). After 10 consecutive trials, the average permeabilities of columns containing 0 and 35% bagasse were 1.9 and 7 mL/min, respectively. The average permeability versus bagasse content was linear in the range of 35e75% bagasse content (Fig. 2). In this range, the average permeability was directly proportional to the bagasse content. AdBg mixtures containing at least 50% bagasse showed constant permeability from the fifth trial onward. A flow rate of 10 mL/min was obtained for a column containing 50% bagasse. The flow rate reached 17 mL/min in a column containing 75% bagasse. Experiments examining the adsorption of phosphate ions using bagasse demonstrate that it is ineffective as an adsorbent. The only role of bagasse was to improve the mixture permeability. Andosol was the only phosphate ion adsorbent in the mixture. The maximum percentage of andosol that enabled stable permeability and reproducibility was 50%. A ratio of 1: 1 (50% Bg: 50% Ad) was thus used for future experiments. 3.2. Effect of flow rate on the adsorption of phosphate ions The effect of flow rate on the adsorption of phosphate ions was examined in the ranges of 4e8 mL/min with an initial concentration of 10 mg/L phosphate and a bed depth of 1.8 cm at room temperature (30 ± 2  C). The experimental breakthrough curves for phosphate adsorption at various flow rates displayed an S shape with two inflection points (Fig. 3). The first inflection point, which is located in the lower part of the curves, corresponds to the breakpoint. This inflection point indicates that the packed bed column started to become saturated (Vassilis and Stavros, 2006). The second inflection point, or the exhaustion point, which is located in the upper part of the curves, indicated that the bed was completely saturated. Breakthrough occurred significantly faster at an increased flow rate. As the flow rate increased from 4 to 8 mL/ min, the EBCT decreased from 2.97 to 1.49 min (Table 1), and the exhaust time (corresponding to 98% of the influent concentration) decreased from 107 to 38 min (Fig. 3). At a higher flow rate, as the external film mass resistance at the surface of the adsorbent tended to decrease, the residence time, and thus the saturation time, decreased, resulting in a lower removal efficiency (Han et al.,

Fig. 3. Effect of flow rates on the experimental breakthrough curves (Initial concentration: 10 mg/L; adsorbent loading: 1 g; bed depth: 1.8 cm; room temperature: 30 ± 2  C; pH ¼ 5.5).

2009b). The breakthrough curve became steeper as the flow rate increased because, at a high flow rate, the influent did not have enough contact time with the andosol. The low contact time resulted in a lower extent of phosphate removal. The influent flow rate also strongly influenced the phosphate uptake capacity. As the flow rate increased from 4 to 8 mL/min, the amount of total phosphate uptake (qe) decreased from 4.18 to 2.82 mg/g (Table 1). Similar results were also previously obtained in other studies (Vinodhini and Das, 2010; Kundu and Gupta, 2005; Maji et al., 2007; Aguayo-Villarreal et al., 2011; Chen et al., 2012). The results obtained indicated that a flow rate of 4 mL/min offered an optimal breakthrough curve. Therefore, subsequent experiments were carried out at 4 mL/min. 3.3. Effect of bed depth on the breakthrough curve The effect of bed depth on the breakthrough curve was investigated by using an initial phosphate concentration of 10 mg/L and an initial flow rate of 4 mL/min at room temperature (30 ± 2  C). As bed depth increased, both exhaustion time and effluent volume (Veff) increased (Fig. 4). The increase in Veff was probably due to the high contact time between phosphate ions and the adsorbent (Baral et al., 2009; Chen et al., 2012). Table 1 indicates an increasing trend in the phosphate ion removal efficiency of the column with increasing adsorbent mass (Sousa et al., 2010). The slope of the breakthrough curve decreased with increasing bed depth, which also resulted in a broadened mass transfer zone (Ahmad and Hameed, 2010; Song et al., 2011). Bed depth strongly influenced the total mass of phosphate adsorbed (Table 1). The increase in the total mass of phosphate adsorbed with increasing bed depth might be attributed to the increased adsorbent surface area, which provided more phosphate binding sites (Gupta and Babu, 2009; Sharma and Singh, 2013). The EBCT also increased from 2.97 min to 9.41 min with an increase in bed depth from 1.8 cm to 5.7 cm (Table 1). These results indicated that a bed depth of 1.8 cm offered the column optimal adsorption capacity (4.18 mg/g). Therefore, subsequent experiments were carried out at a bed depth of 1.8 cm. 4. Effect of the influent phosphate concentration on the breakthrough curve

Fig. 2. Average permeability as a function of bagasse content.

The effect of influent phosphate concentration on the breakthrough curve was examined in the range of 10e25 mg/L at a flow rate 4 mL/min and a bed depth of 1.8 cm at room temperature (30 ± 2  C). The adsorption process reached saturation faster, and

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

455

Table 1 Column data parameters obtained at different inlet phosphate concentrations, bed heights and flow rates. Exp. parameters

Co (mg/L)

Q (mL/min)

H (cm)

tp (min)

ttotal (min)

mtotal (mg)

qtotal (mg)

qe (mg/g)

Veff (mL)

P (%)

EBCT (min)

Flow rate

10 10 10 10 10 10 10 15 20 25

4 6 8 4 4 4 4 4 4 4

1.8 1.8 1.8 1.8 3.6 5.7 1.8 1.8 1.8 1.8

60 33 17 60 80 120 60 40 25 20

120 70 40 120 180 285 120 105 85 75

4.8 4.2 3.2 4.8 7.2 11.4 4.8 6.3 6.8 7.5

4.18 3.53 2.82 4.18 6.47 10.37 4.18 5.42 5.64 5.94

4.18 3.53 2.82 4.18 3.24 3.46 4.18 5.42 5.64 5.94

428 336 304 428 592 908 428 364 296 264

87 84 82 87 90 91 87 86 83 79

2.97 1.98 1.49 2.97 5.94 9.41 2.97 2.97 2.97 2.97

Bed depth

Co

the breakthrough time decreased with increasing influent phosphate concentration (Fig. 5). A decrease in phosphate concentration gave a later breakthrough curve, and the treated volume was greatest at the lowest influent concentration. As expected, the adsorption capacity increased with an increasing influent phosphate concentration. This result can be explained by the fact that a larger concentration gradient caused faster transport due to an increased diffusion coefficient (Uddin et al., 2009; Chen et al., 2011). With an increase in influent phosphate concentration from 10 to 25 mg/L, the uptake and corresponding total influent adsorbed were found to increase from 4.18 to 5.94 mg/g and 4.80e7.50 mg, respectively (Table 1). This might be attributed to the high influent phosphate concentration, which provided a greater driving force for the transfer process to overcome mass transfer resistance (Baral et al., 2009). As the influent phosphate concentration increased from 10 to 25 mg/L, the exhaust time for andosol decreased from 120 to 75 min. These results demonstrated that a higher initial influent concentration led to a higher driving force for mass transfer; hence the adsorbent achieved saturation faster, resulting in a decrease in the exhaust time and adsorption zone length (Malkoc et al., 2006; Baral et al., 2009). Fig. 5 shows that the breakthrough curves are sharper with increasing influent phosphate concentration, indicating a relatively small mass transfer zone. These results also suggest that intra-particle diffusion controlled the sorption process (Baral et al., 2009). 4.1. Breakthrough curve modelling Successful design of a column adsorption process requires prediction of the breakthrough curve for the effluent (Han et al.,

Fig. 4. Effect of bed depth on the experimental breakthrough curves (initial concentration: 10 mg/L; flow rate: 4 mL/min; room temperature: 30 ± 2  C; pH ¼ 5.5).

2009b). Several simple mathematical models have been developed to describe and analyse laboratory-scale column studies for industrial applications (Kumar and Chakraborty, 2009; Han et al., 2009b; Vinodhini and Das, 2010). In this study, AdamseBohart, Thomas and YooneNelson models were used to determine the best model for predicting the dynamic behaviour of the column. 4.1.1. AdamseBohart model The AdamseBohart model (Bohart and Adams, 1920) is based on a fundamental equation established by surface reaction theory, which describes the relationship between Ct/Co and t in a continuous system. This model assumes that equilibrium is not instantaneous. The model is used to describe the initial phase of the breakthrough curve. The expression is as follows:

    Ct H ¼ kAB Co t  kAB No ln Uo Co

(13)

where Co and Ct are the influent and effluent concentrations (mg/L), respectively; kAB is the kinetic constant (L/mg min); No is the saturation concentration (mg/L); H is the bed depth of the fix-bed column (cm); and Uo is the superficial velocity (cm/min), which is defined as the ratio of the volumetric flow rate Q (cm3/min) to the cross-sectional area of column A (cm2). The range of t taken into consideration included the entire breakthrough region. The parameters kAB and No can be calculated from the linear plot of ln(Co/ Ct) against t (diagram not shown). For all breakthrough curves using linear regression analysis, the respective values of kAB and No were calculated (Table 2) along with the correlation coefficients (R2). The R2 values from the AdamseBohart model ranged between 0.856 and 0.958. The values of kAB decreased with increasing bed depth and influent

Fig. 5. Effect of influent concentration on the experimental breakthrough curves (flow rate: 4 mL/min; bed depth: 1.8 cm; adsorbent loading: 1 g; room temperature: 30 ± 2  C; pH ¼ 5.5).

456

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

Table 2 Parameters of AdamseBohart, Thomas and YooneNelson model under different conditions using linear regression analysis. Exp. parameters

Flow rate

Bed depth

Co

AdamseBohart

Thomas

YooneNelson

Co (mg/L)

Q (mL/min)

H (cm)

kAB  103 (L/mg min)

No (mg/L)

R2

kTh  103 (L/mg min)

qo (mg/g)

R2

kYN  102 (min1)

t (min)

R2

10 10 10 10 10 10 10 15 20 25

4 6 8 4 4 4 4 4 4 4

1.8 1.8 1.8 1.8 3.6 5.7 1.8 1.8 1.8 1.8

3.30 4.70 8.20 3.30 2.60 2.50 3.30 2.20 1.94 1.64

1049.91 954.75 691.95 1049.91 3072.41 7066.25 1049.91 1301.69 1393.25 1489.65

0.931 0.889 0.915 0.931 0.958 0.856 0.931 0.914 0.914 0.891

6.30 9.50 18.20 6.30 4.60 4.10 6.30 4.93 4.35 3.92

2895.87 2523.16 1751.21 2895.87 8716.52 20320.98 2895.87 3231.89 3414.25 3531.63

0.899 0.839 0.917 0.899 0.969 0.984 0.899 0.955 0.978 0.992

6.30 9.50 18.20 6.30 4.60 4.10 6.30 7.40 8.70 9.80

72.40 42.05 21.15 72.40 108.96 169.34 72.40 53.87 42.68 35.32

0.889 0.839 0.917 0.889 0.969 0.984 0.889 0.955 0.978 0.992

concentration but increased with increasing flow rate. The value of No decreased with increasing flow rate but increased with increasing bed depth and influent concentration.

obtained from the Thomas or YooneNelson models (0.839e0.992) demonstrated that the process of adsorption onto the column was more closely described by these two models.

4.1.2. Thomas model The Thomas model (Thomas, 1944) assumes plug flow behaviour in the column bed. The linearised form of this model can be described by the following expression:

4.2. Application to the treatment of natural water from Yaound
e Municipal Lake

  Co k qo m  kTh Co t ln  1 ¼ Th Q Ct

(14)

where kTh is the Thomas model constant (mL/min mg), qo is the adsorption capacity (mg/g), and t stands for the total flow time (min). The values of kTh and qo can be determined from the linear plot of ln[(Co/Ct)  1] against t (diagram not shown). Relative constants and coefficients were obtained using linear regression analysis according to Eq. (14) (Table 2). Table 2 shows that kTh values decreased with increasing bed depth. As the flow rate increased, the values of kTh increased, while the values of qo decreased. The values of kTh decreased with increasing influent phosphate concentration. These values can be attributed to the driving force for adsorption due to the concentration gradient. The R2 values from the Thomas model ranged from 0.839 to 0.992, which indicated that this model provided a better fit than the AdamseBohart model. 4.1.3. YooneNelson model Yoon and Nelson (Yoon and Nelson, 1984) developed a model to investigate the breakthrough behaviour of adsorbate gases on activated charcoal. The linearised YooneNelson model for a single component system can be expressed as follows:

 ln

Ct Co  Ct

The treatment of natural water was investigated at a flow rate of 4 mL/min, a bed depth of 1.8 cm under natural pH and a temperature of 30 ± 2  C. The adsorption of phosphate ions was carried out for two water samples containing phosphate. The first sample was a synthetic sample containing 10 mg/L of phosphate ions. The second
Municipal Lake was a natural water sample obtained from Yaounde (Cameroon) whose initial phosphate concentration (2.1 mg/L) was adjusted to 10 mg/L by dissolving an appropriate amount of the synthetic reagent (NaH2PO412H2O) used previously. The breakthrough time of the column for phosphate ion adsorption in the natural milieu was greater than in the synthetic sample (Fig. 6). The adsorption capacity of phosphate ions increased from 4.18 mg/g in the synthetic medium to 6.69 mg/g in the natural medium, representing a 60% improvement (Table 3). This adsorption capacity was low compared to 35.20 mg/g recorded by Li et al. (2014). These authors removed phosphate from aqueous solution using nanostructure iron (III)-copper (II) binary oxides. However, AdBg mixture is available and remains cheapest than the binary oxide cited above. The treated effluent volume before the breakthrough of the column (Veff) increased from 428 to 760 mL, which was a 77.6% increase. These results demonstrated that other

 ¼ kYN t  tkYN

(15)

where kYN is the rate constant (min1) and t is the time required for 50% adsorbate breakthrough (min). A linear plot of ln[Ct/(CoeCt)] against t was used to determine the values of kYN and t from the intercept and the slope of the plot (diagram not shown). Statistical parameters of the YooneNelson model were calculated (Table 2). The kYN values and the 50% breakthrough time t both increased with increasing bed depth. The values of t significantly decreased as the influent phosphate concentration increased because column saturation occurred more rapidly (Calero et al., 2009). However, the values of kYN increased and the values of t decreased with an increase in the flow rate. The adsorption of phosphate ions on the column was not accurately described by the AdameBohart model. The R2 values

Fig. 6. Effect of the natural effluent on the experimental breakthrough curves (flow rate: 4 mL/min; bed depth: 1.8 cm; adsorbent loading: 1 g; room temperature: 30 ± 2  C).

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

457

Table 3 Column parameters for synthetic and natural phosphate adsorption by Ad-M/Bag mixture. Exp. parameters

pH

Co (mg/L)

Q (mL/min)

H (cm)

D (cm)

tp (min)

ttotal (min)

mtotal (mg)

qtotal (mg)

qe (mg/g)

Veff (mL)

P (%)

Synthetical Natural

5.5 6.8

10 10

4 4

1.8 1.8

3.1 3.1

60 96

120 190

4.80 7.60

4.18 6.69

4.18 6.69

428 760

87 88

Table 4
Municipal Lake) before and after passage through the fixed-bed column. Chemical analysis of natural sample (Yaounde Cations

Anions

Samples

pH

Kþ

Naþ

Ca2þ

Mg2þ

Fe

Al3þ

HCO 3

PO3 4

SO2 4

Cl

NO 3

NO 2

Crude (mg/L) Treated (mg/L)

6.8 7.2

13.6 12.1

15.3 15.1

20 15

7.2 1.8

0 0

0.3 0

12.8 12.2

10 1.2

11.8 11.6

12.9 13.1

23.8 23.7

5.1 5.2

ionic species contained in the natural effluent considerably improved the adsorption of phosphate onto the column. Chemical analysis of the natural sample before and after flow through the column (Table 4) indicated low variation in HCO 3,    3þ þ þ SO2 4 , Cl , NO3 and NO2 concentrations. The Al , K and Na concentrations decreased (0.3 mg/L), while the concentrations of Ca2þ and Mg2þ decreased by 5.0 and 5.4 mg/L, respectively. These two cations probably participated in the elimination of phosphate ions through the formation of calcium- or magnesiumeorthophosphate complexes at the surface of the colloidal material (Helyar et al., 1976; Eze and Loganathan, 1990). The phosphate (HPO2 4 ) concentration decreased from 10 to 1.2 mg/L, which corresponds to an 88% elimination. A second passage through a similar column approached 100% phosphate removal from the effluent. 4.3. Adsorption kinetics of phosphate on AdBg mix Kinetic experiments were performed to determine the adsorption rate of phosphate on AdBg mix. The change of adsorbed phosphate as a function of adsorption time was shown in Fig. 7. It was found that the adsorption process could be divided into two steps, a quick step and a slow one (Li et al., 2014). In the first step, the adsorption was very fast and appropriate 90% of the equilibrium adsorption capacity was achieved within the first 6 h. In the second step, the adsorption slowed down. The maximum adsorption had taken place within 20 h. To ensure complete adsorption, the

adsorption time was therefore extended to 24 h for all batch experiments (Doula et al., 1996). The amount of phosphate adsorbed increased with the temperature. Higher temperature should be preferred in the phosphate removal and the process is endothermic. Similar results were reported in a previous study where coal fly ash was used as an adsorbent for phosphate removal (Yeom and Jung, 2009). According to Table 5, kinetic data fitted better with the pseudosecond-order model (R2 ¼ 0.999). This indicates that the adsorption process might be chemisorption (Long et al., 2011).

4.4. Thermodynamic parameters of phosphate adsorption The calculated values for the standard enthalpy (DH ) and entropy change (DS ) were obtained using the Van'tHoff equation. The plot of lnKd against 1/T gave a straight line of slope e DH /R and intercepts DS /R. The Gibbs free energy change (DG ) values were obtained by using Eq. (11). The thermodynamic parameters for the adsorption of phosphate were given in Table 6. The positive value of standard enthalpy 15.964 kJ/mol, indicated that the adsorption of phosphate on AdBg was an endothermic processes (Zhang et al., 2009; Cottet et al., 2014). The positive values of entropy (DS ), 34.486 J mol1 K1 indicate that the adsorption of phosphate on AdBg mix increased the random in the system (solution/adsorbent) during the adsorption process. The Gibbs free energy (DG ), listed in Table 6, as calculated for adsorption of phosphate on AdBg mix were 5.863, 5.518 and 4.828 kJ/mol at different temperatures. This suggested that the adsorption of phosphate on AdBg mix was feasible and spontaneous thermodynamically. The magnitude of the activation energy (Ea) can indicate if the adsorption mechanism is physisorption (5e40 kJ mol1) or chemisorption (40e800 kJ mol1) (Cottet et al., 2014). The 41.185 kJ mol1 value for the activation energy indicated that the adsorption of phosphate onto AdBg mix had a potential barrier corresponding to a chemisorption process.

Table 5 Kinetic parameters of adsorption for phosphate on AdBg mix at different temperatures. T ( C)

Fig. 7. Effect of temperature on adsorption kinetics of phosphate by AdBg mix: 0.2 g/ 50 mL, initial concentration: 50 mg/L; pH: 5.5.

20 30 50

Pseudo-first-order model

Pseudo-second-order model

Qexp (mg g1)

Qe (mg g1)

k1 (h1)

R2

Qe (mg g1)

k2 (g mg1 h1)

R2

3.281 3.922 4.960

1.666 1.915 2.104

0.201 0.165 0.118

0.884 0.957 0.969

2.847 3.552 5.145

0.215 0.482 1.080

0.999 0.999 0.999

458

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

Table 6 Thermodynamic parameters values for adsorption of phosphate onto AdBg mix sample. R2

DS DG (kJ mol1) Ea R2 1 (kJ mol1) (J mol1 K1) (kJ mol ) 20  C 30  C 50  C DH

AdBg 0.997 15.964

34.486

5.863 5.518 4.828 41.185

0.973

4.5. FTIR analysis Fig. 8 presented the superposition of six FTIR spectra of samples used in this study: the Phosphate salt NaH2PO4 12H20 (P), Bagasse (Bg), Andosol (Ad), Andosol and Bagasse mixture before (AdBg) and after phosphate adsorption in solutions of two different concentrations: 0.3 g/L (AdBgP0.3) and 3.0 g/L (AdBgP3.0). The samples were characterised by Fourier Transform Infrared (FTIR) spectroscopy (Bruker Alpha-p) in the spectral range of 4000e400 cm1.

Table 7 Band assignments for infrared spectra of NaH2PO4 12H20 (P), Bagasse (Bg), Andosol (Ad), Andosol Bagasse mixture (AdBg), Andosol Bagasse mixture after phosphate adsorption in a 0.3 g/L aqueous solutions (AdBgP0.3) and Andosol Bagasse mixture after phosphate adsorption in a 3.0 g/L aqueous solutions (AdBgP3.0). IR frequencies (cm1)

Band assignments

P

Bg

Ad

AdBg AdBgP0.3 AdBgP3.0

/ 1616 / / / / 1186 1054 / / / 860 /

1724 / 1604 1514 1373 1240 / / 1033 / / / /

/ 1624 / / / / / / / 1010 913 / 795

1722 / 1604 1511 1374 1238 / / 1030 / 913 / 796

/ 1648 / / / / / / 1028 / / 858 797

/ 1652 / / / / 1145 1048 / / / 855 /

C]O bending vibration HeOeH stretching vibration C]O bending vibration C]C bending vibration eCH2e bending vibration CeO bending vibration AlePO4eAl bending vibration PeO bending vibration SieO, AleO stretching vibration SieO bending vibration AleOH bending vibration OeP]O bending vibration AleO stretching vibration

Table 7 presented the band assignment for the six samples cited above. Vibrational bands were identified in relation to the crystal structure in terms of fundament vibrating units namely H2PO and H2O for NaH2PO4 12H20, which were assigned according to the literatures (Unuabonah et al., 2007; Zhang et al., 2009; Costa et al., 2013; Li et al., 2014). AdBg IR spectrum showed bands which could be recognised on Ad and Bg spectra. For example, the band at 1240 cm1 on Bg (CeO bending vibration) was absent on Ad spectrum but present on AdBg one at 1238 cm1. The bands at 795 cm1 (AleO stretching vibration) on Ad was absent on Bg but present on AdBg at 796 cm1. After the AdBg mix reaction with phosphate solution, the AleOH bending bands (913 and 795 cm1) from andosol completely disappeared on AdBgP3.0 spectrum and the news bands at 1145, 1048 and 855 cm1 appeared. They were assigned to the bending vibration of PeO, OeP]O and OePeO (Table 7). After phosphate adsorption, the band at 858 cm1 appeared on AdBgP0.3, and its intensity increase obviously with phosphate concentration on AdBgP3.0 at 855 cm1. These two bands could be attributed to OeP]O bending vibration band from phosphate spectrum at 860 cm1. The presence of a strong band at 1048 cm1 on AdBgP3.0 sample, was attributed to the asymmetric vibration of PeO (Zhang et al., 2009; Li et al., 2014). These results confirmed the phosphate adsorption onto AdBg surface through ligand exchange between phosphate anions and eOH. Eqs. (16) and (17) described this ligand 2 exchange. H2PO 4 and HPO4 anions can also react with SieOH or AleOH via hydrogen bonding according to follows Eqs. (18)e(21) (Sabah and Majdan, 2009; Yan et al., 2014):

Fig. 8. Superposition of FTIR spectra from six samples used in this study: Phosphate salt NaH2PO4$12H20 (P), Bagasse (Bg), Andosol (Ad), Andosol and Bagasse mixture (AdBg) before and after phosphate adsorption in solutions of two different concentrations: 0.3 g/L (AdBgP0.3) and 3.0 g/L (AdBgP3.0).

þ 2^SeOH þ HPO2 4 þ 2H / ^S2O2PO2H þ 2H2O

(16)

þ  ^SeOH þ HPO2 4 þ H / ^SeOPO3H þ H2O

(17)

þ 2^SeOH þ H2PO 4 þ 2H / ^S2O2PO2H þ 2H2O

(18)

þ ^SeOH þ H2PO 4 þ H / ^SeOPO3H2 þ H2O

(19)

 ^SeOH þ H2PO 4 / ^SeOH$$$H2PO4

(20)

2 ^SeOH þ HPO2 4 / ^SeOH$$$HPO4

(21)

where ^S represent the active surface functional groups including the silanol (SieOH) and aluminols (AleOH).

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

5. Conclusion Andosol is an effective fixed bed adsorbent for phosphate removal, but its application is limited by low permeability. Bagasse in its raw state did not adsorb phosphate ions. However, when bagasse was mixed with 50% andosol, the permeability of the entire mixture stabilised and improved significantly from 1 mL/min to 10 mL/min. Phosphate removal from an effluent could then be conducted in a fixed-bed column with a reasonable flow rate. Optimal adsorption capacity of the AdBg mixture (4.18 mg/g) was obtained with a bed depth of 1.8 cm and a flow rate of 4 mL/min. Influent concentrations equal to 10 and 15 mg/L gave elimination percentages of 87 and 86%, respectively. Column data obtained under different conditions were described using the Thomas, YooneNelson and AdameBohart models. The correlation coefficients indicated a better fit of these data to the models of Thomas and YooneNelson. Adsorption of phosphate ions in a fixed-bed column was applied to two samples having the same concentration of phosphate ions. The first sample was synthetic, while the second one was natural. The presence of Ca2þ and Mg2þ ions in natural waters improved the performance of the column by approximately 60%; believe due to complex formation at the andosol surface. Kinetic experiments demonstrated that phosphate adsorption onto AdBg mix fitted better with a pseudo-second-order model and it was a chemisorption process. FTIR analysis indicated that phosphate anions may form inner-sphere surface complexes at the water/oxide interface, which is the main phosphate removal mechanism. References
ndez-Montoya, V., MontesAguayo-Villarreal, I.A., Bonilla-Petriciolet, A., Herna
n, M.A., Reynel-Avila, H.E., 2011. Batch and column studies of Zn2þ removal Mora from aqueous solution using chicken feathers as sorbents. Chem. Eng. J. 167, 67e76. Ahmad, A.A., Hameed, B.H., 2010. Fixed-bed adsorption of reactive azo dye onto granular activated carbon prepared from waste. J. Hazard. Mater. 175, 298e303. Baral, S.S., Das, N., Ramulu, T.S., Sahoo, S.K., Das, S.N., Chaudhury, G.R., 2009. Removal of Cr (VI) by thermally activated weed Salvinia cucullata in a fixed-bed column. J. Hazard. Mater. 161, 1427e1435. Bhattacharyya, K., Gupta, S., 2008. Influence of acid activation on adsorption of Ni (II) and Cu (II) on kaolinite and montmorillonite: kinetic and thermodynamic study. Chem. Eng. J. 136, 1e13. Bohart, G.S., Adams, E.Q., 1920. Behavior of charcoal towards chlorine. J. Chem. Soc. 42, 523e529. ~o, P.C., Souza, T.C., Ferreira, C.A., Hori, C.E., Romanielo, L.L., 2010. Removal of Branda petroleum hydrocarbons from aqueous solutions using sugarcane bagasse as adsorbent. J. Hazard. Mater. 175, 1106e1112.
zquez, G., Tenorio, G., Martín-Lara, M.A., 2009. Study of Calero, M., Hern
ainz, F., Bla Cr(III) biosorption in a fixed-bed column. J. Hazard. Mater. 171, 886e893. Chen, N., Zhang, Z.Y., Feng, C.P., Li, M., Chen, R.Z., Sugiura, N., 2011. Investigations on the batch and fixed-bed column performance of fluoride adsorption by Kanuma mud. Desalination 268, 76e82. Chen, S., Yue, Q., Gao, B., Li, Q., Xu, X., Fu, K., 2012. Adsorption of hexavalent chromium from aqueous solution by modified corn stalks: a fixed-bed column study. Bioresour. Technol. 113, 114e120. Clark, C.J., McBride, M.B., 1984. Chemisorption of Cu(II) and Co(II) on allophane and imogolite. Clays Clay Miner. 32, 300e310. Costa, T.C.de Carvalho, Diniz Melo, J.D., Paskocimas, C.A., 2013. Thermal and chemical treatments of montmorillonite clay. Ceram. Int. 39, 5036e5067. Cottet, L., Almeida, C.A.P., Naidek, N., Viante, M.F., Lopes, M.C., Debacher, N.A., 2014. Adsorption characteristics of montmorillonite clay modified with iron oxide with respect to methylene blue in aqueous media. Appl. Clay Sci. 95, 25e31. Doula, M., Ioannou, A., Dimirkou, A., 1996. Thermodynamics of phosphate adsorption-desorption by alfisols, entisols, vertisol, and inceptisols. Commun. Soil Sci. Plant Anal. 27, 1749e1764. Eze, O.C., Loganathan, P., 1990. Effects of pH on phosphate sorption of some Paleudults of southern Nigeria. Soil Sci. 150, 613e621. Gupta, S., Babu, B.V., 2009. Modeling, simulation, and experimental validation for continuous Cr(VI) removal from aqueous solutions using sawdust as an adsorbent. Bioresour. Technol. 100, 5633e5640. Gupta, S., Babu, B.V., 2010. Experimental investigations and theoretical modeling aspects in column studies for removal of Cr(VI) from aqueous solutions using activated tamarind seeds. J. Water Resour. Prot. 2, 706e716.

459

Han, R., Wang, Y., Zhao, X., Wang, Y.F., Xie, F.L., Cheng, J.M., Tang, M.S., 2009a. Adsorption of methylene blue by phoenix tree leaf powder in a fixed-bed column: experiments and prediction of breakthrough curves. Desalination 245, 284e297. Han, R., Zou, L., Zhao, X., Xu, Y., Xu, F., Li, Y., Wang, Y., 2009b. Characterization and properties of iron oxide-coated zeolite as adsorbent for removal of copper(II) from solution in fixed bed column. Chem. Eng. J. 149, 123e131. Helyar, K.R., Munns, D.N., Burau, R.G., 1976. Adsorption of phosphate by gibbsite. II. Formation of a surface complex involving divalent cations. J. Soil Sci. 27, 315e323. Karaça, S., Gurses, A., Ejder, M., Açıkyıldız, M., 2004. Kinetic modeling of liquidphase adsorption of phosphate on dolomite. J. Colloid Interface Sci. 277, 257e263. Kumar, P.A., Chakraborty, S., 2009. Fixed-bed column study for hexavalent chromium removal and recovery by short-chain polyaniline synthesized on jute fiber. J. Hazard. Mater. 162, 1086e1098. Kundu, S., Gupta, A.K., 2005. Analysis and modeling of fixed bed column operations on As(V) removal by adsorption onto iron oxide-coated cement (IOCC). J. Colloid Interface Sci. 290, 52e60. Kundu, S., Kavalakatt, S.S., Pal, A., Ghosh, S.K., Mandal, M., Pal, T., 2004. Removal of arsenic using hardened paste of Portland cement: batch adsorption and column study. Water Res. 38, 3780e3790. Kuo, S., Lotse, E.G., 1974. Kinetics of phosphate adsorption and desorption by lake sediments. Soil Sci. Soc. Am. Proc. 38, 50e54. Li, G., Gao, S., Zhang, G., Zhang, X., 2014. Enhanced adsorption of phosphate from aqueous solutions by nanostructured iron(III)-copper(II) binary oxides. Chem. Eng. J. 235, 124e131. Lin, Y., Chen, Zh, Chen, Z., Megharaj, M., Naidu, R., 2014. Decoloration of acid violet red B by bentonite-supported nanoscale zero-valent iron: reactivity, characterization, kinetics and reaction pathway. Appl. Clay Sci. 93-94, 56e61. Li, N., Ren, J., Zhao, L., Wang, Z.L., 2013. Fixed-bed adsorption study on phosphate removal using nanosized FeOOH anion resin. Hindawi Publ. Corp. J. Nanomater. 2013, 1e5. Long, F., Gong, J.L., Zeng, G.M., Chen, L., Wang, X.Y., Deng, J.H., Niu, Q.Y., Zhang, H.Y., Zhang, X.R., 2011. Removal of phosphate from aqueous solution by magnetic FeeZr binary oxide. Chem. Eng. J. 171, 448e455. Lu, S.G., Bai, S.Q., Zhu, L., Shan, H.D., 2009. Removal mechanism of phosphate from aqueous solution by fly ash. J. Hazard. Mater. 161, 95e101. Maji, S.K., Pal, A., Pal, T., Adak, A., 2007. Modeling and fixed bed column adsorption of As(III) on laterite soil. Sep. Purif. Technol. 56, 284e290. Malkoc, E., Nuhoglu, Y., Dundar, M., 2006. Adsorption of chromium (VI) on pomacean olive oil industry waste: batch and column studies. J. Hazard. Mater. B138, 142e151. Netpradit, S., Thiravetyan, P., Towprayoon, S., 2004. Evaluation of metal hydroxide sludge for reactive dye adsorption in a fixed-bed column system. Water Res. 38, 71e78. Nur, T., Shim, W.G., Jobir, M.A.H., Vignesvaran, S., Kandasamy, J., 2013. Modelling of phosphorus removal by ion exchange resin (Purolite FerrIX A33E) in fixed-bed column experiments. Desalin. Water Treat. 52 (4e6), 784e790. Oguz, E., Ersoy, M., 2010. Removal of Cu2þ from aqueous solution by adsorption in a fixed bed column and neural network modelling. Chem. Eng. J. 164, 56e62. Parfitt, R.L., Hemni, T., 1980. Structure of some allophanes from New Zealand. Clays Clay Miner. 28, 285e294.
draogo, W.K.I., Schmidt, C., Zachmann, D., Bahadir, M., Pehlivan, E., Tran, H.T., Oue 2013. Sugarcane bagasse treated with hydrous ferric oxide as a potential adsorbent for the removal of As(V) from aqueous solutions. Food Chem. 138 (1), 133e138. me e
dition. Rodier, J., 1986. L’analyse de l’eau. Dunod, Paris, p. 1383, 8 e Sabah, E., Majdan, M., 2009. Removal of phosphorus from vegetable oil by acidactivated sepiolite. J. Food Eng. 91 (3), 423e427. Safaa, M., Ragheb, 2013. Phosphate removal from aqueous solution using slag and fly ash. HBRC J. 9, 270e275. Sharma, R., Singh, B., 2013. Removal of Ni(II) ions from aqueous solution using modified rice straw in a fixed-bed column. Bioresour. Technol. 146, 519e524. Shin, E.W., Han, J.S., Jang, M., Min, S.H., Park, J.K., Rowell, R.M., 2004. Phosphate adsorption on aluminium-impregnated mesoporous silicates: surface structure and behavior of adsorbents. Environ. Sci. Technol. 38, 912e917. Siewe, J.M., Djoufac, W.E., Bitom, D., Figueras, F., Djomgue, P., Njopwouo, D., Azinwi, P.T., 2008. Andosols of the Bambouto Mountains (West Cameroon): characteristics, superficial properties e study of the phosphate ions adsorption. Open Inorg. Cem. J. 2, 106e115. Song, J.Y., Zou, W.H., Bian, Y.Y., Su, F.Y., Han, R.P., 2011. Adsorption characteristics of methylene blue by peanut husk in batch and column mode. Desalination 265, 119e125. Sousa, F.W., Oliveira, A.G., Ribeiro, J.P., Rosa, M.F., Keukeleire, D., Nascimento, R.F., 2010. Green coconut shells applied as adsorbent for removal of toxic metal ions using fixed-bed column technology. J. Environ. Manage. 91, 1634e1640. Tematio, P., Kengni, L., Bitom, D., Hodson, M.E., Fopoussi, J.C., Leumbe, O., Mpakam, H.G., Tsozue, D., 2004. Soils and their distribution on Bambouto volcanic mountain, West Cameroon highland, Central Africa. J. Afr. Earth Sci. 39, 447e475. Thomas, H.C., 1944. Heterogeneous ion exchange in a flowing system. J. Am. Chem. Soc. 66, 1466e1664.

460

E.D. Woumfo et al. / Journal of Environmental Management 151 (2015) 450e460

Uddin, M.T., Rukanuzzaman, M., Khan, M.M.R., Islam, M.A., 2009. Adsorption of methylene blue from aqueous solution by jackfruit (Artocarpus heteropyllus) leaf powder: a fixed-bed column study. J. Environ. Manage. 90, 3443e3450. Unuabonah, E.I., Olu-Owolab, B.I., Adebowale, K.O., Ofomaja, A.E., 2007. Adsorption of lead and cadmium ions from aqueous solutions by tripolyphosphateimpregnated Kaolinite clay. Colloids Surf. A Physicochem. Eng. Asp. 292, 202e211. Vassilis, J.I., Stavros, G.P., 2006. Adsorption, Ion Exchange and Catalysis. Elsevier Press, Athens, p. 274. Vinodhini, V., Das, N., 2010. Packed bed column studies on Cr (VI) removal from tannery wastewater by neem sawdust. Desalination 264, 9e14. Xu, K., Deng, T., Lin, J., Peng, W., 2010. Study on the phosphate removal from aqueous solution using modified fly ash. Fuel 89 (12), 3668e3674.

Yan, Y., Sun, X., Ma, F., Li, J., Shen, J., Han, W., Liu, X., Wang, L., 2014. Removal of phosphate from wastewater using alkaline residue. J. Environ. Sci. 26, 970e980. Yeom, S.H., Jung, K.Y., 2009. Recycling wasted scallop shell as an adsorbent for removal of phosphate. J. Ind. Eng. Chem. 15, 40e44. Yoon, Y.H., Nelson, J.H., 1984. Application of gas adsorption kinetics. I. A theoretical model for respirator cartridge service life. Am. Ind. Hyg. Assoc. J. 45, 509e516.
de
s physicoYoucef, L., Achour, S., 2005. Elimination des phosphate par les proce chimiques. Larhyss J. 4, 129e140. Zach-Maor, A., Semiat, R., Shemer, H., 2011. Fixed-bed phosphate adsorption by immobilized nano-magnetic matrix: experimental and a new modeling approach. Adsorption 17, 929e936. Zhang, G., Liu, H., Liu, R., Qu, J., 2009. Removal of phosphates from water by a FeeMn binary oxide adsorbent. J. Colloid Interface Sci. 335, 168e174.