9846
Ind. Eng. Chem. Res. 2010, 49, 9846–9856
Development of a New Inorganic-Organic Hybrid Ion-Exchanger of Zirconium(IV)-Propanolamine for Efficient Removal of Fluoride from Drinking Water S. K. Swain,† Sulagna Mishra,‡ Prachi Sharma,‡ Tanushree Patnaik,§ V. K. Singh,† Usha Jha,† R. K. Patel,| and R. K. Dey*,§ Department of Applied Chemistry, Birla Institute of Technology, Mesra-835 215, Ranchi, India, Department of CiVil Engineering, Birla Institute of Technology, Mesra-835 215, Ranchi, India, Post-Graduate Department of Chemistry, RaVenshaw UniVersity, Cuttack-753 003, Orissa, India, and Department of Chemistry, National Institute of Technology, Rourkela-769 008, Orissa, India
This investigation reports the synthesis of a new adsorbent material (ZrPA), from Zr(IV) and propanolamine (PA), synthesized by sol-gel technique at room temperature following the green chemistry principle. The suitability of ZrPA as a potential adsorbent is assessed for the removal of fluoride following the batch mode of operation. The isotherm, kinetics, and thermodynamics of fluoride adsorption on ZrPA have been studied at various experimental conditions (initial fluoride concentration, adsorption time, adsorbent dose, and temperature). The characteristic of the adsorbent, before and after fluoride adsorption, was examined using Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), coupled with energy dispersive spectrum (EDS) techniques. Further measurement of surface area, pore volume, and pore diameter using N2 intrusion automated gas sorption system shows the microporous nature of the prepared material. Adsorption of fluoride was found to be strongly affected by pH. Mathematically, pseudosecond-order kinetic model was found to best describe the reaction rate, which was consistent with the actual measurement. Applications of Freundlich, Langmuir, Temkin, and Dubinin-Radushkevich (D-R) isotherm models for the adsorption process were evaluated, and the data were also compared for six different error functions, that is, the sum of the squares of errors (SSE), sum of the absolute errors (SAE), the average relative error (ARE), the hybrid fractional error function (HYBRID), the Marquardt’s percent standard deviation (MPSD), and regression coefficient (R2), to test the adequacy and accuracy of the model equations. Thermodynamic parameters such as enthalpy, entropy, and free energy were calculated using van’t Hoff equations, which shows that fluoride adsorption on ZrPA indicates the spontaneous and endothermic nature of adsorption. The reusability of the ZrPA adsorbent material was tested up to 10 consecutive cycles for a sustainable commercial application purpose. Quantitative desorption of fluoride from ZrPA was found to be more than 95% at pH 12. To test the efficacy, the performance of the adsorbent material was studied with water samples collected from a fluorosis endemic region. 1. Introduction Fluoride is a naturally occurring element in mineral and geochemical deposits.1 The presence of excess fluoride in drinking water is responsible for pandemic health problems and has attracted the attention of researchers all around the world.2 Long-term ingestion of water with excessive amounts of fluoride can cause dental and skeletal fluorosis.3 The World Health Organization (WHO) guideline suggested a limiting value of 1.5 mg L-1 for fluoride present in drinking water.4 Many techniques such as reverse osmosis, electrodialysis, donnan dialysis, ion-exchange, limestone reactor, and activated alumina column have been reported for the removal of fluoride from drinking water.5,6 However, in all cases, sustainable commercial utilization and viability of a designed process and parameter always depend upon the simplicity of the process, availability of a wide range of chemicals/materials, as well as material characteristics.7 * To whom correspondence should be addressed. Tel.: +91-6712507624. E-mail:
[email protected]. † Department of Applied Chemistry, Birla Institute of Technology. ‡ Department of Civil Engineering, Birla Institute of Technology. § Ravenshaw University. | National Institute of Technology.
Previously, a number of materials such as granular activated carbon, laterites, kaolinite, bentonite, lignite, etc., have been reported as adsorbents for fluoride removal from drinking water.8 Although some of them are natural as well as low cost materials, the efficacy and some other limitations restricted the effective commercial utilization of these materials for effective fluoride removal from drinking water. Investigators9 have also studied the feasibility of many other materials such as hydrous tin oxide, impregnated alumina, impregnated silica gel, resin, granular calcite, fly ash, clays, spent catalyst, ion-exchange resins, activated alumina, mixed adsorbent materials, etc., for the purpose, but the problems associated with respect to regeneration and recovery processes restricted these materials for wider commercial applications. Thus, considering the effort to develop fluoride selective materials, rare earth metal hydroxides/oxides are projected as new potential adsorbents primarily because of their strong affinity toward fluoride.10 The high electronegativity and small ionic size of the fluoride ion favor its strong affinity toward multivalent metal ions including Al(III), Fe(III), and Zr(IV). Preparation of hybrid material is an economical way for the removal of fluoride with high adsorption capacity. The importance of ion-exchange materials based on cation/anion-exchangers including various inorganic cation-exchangers such as silica
10.1021/ie1012536 2010 American Chemical Society Published on Web 09/28/2010
Ind. Eng. Chem. Res., Vol. 49, No. 20, 2010
gel, alumina gel, and chelating resins loaded with high-valence metals such as iron(III), lanthanum(III), cerium(IV), and zirconium(IV) for the preparation of fluoride selective hybrid materials is well documented.11 In this aspect, a recent report shows the development of some new class of ion-exchangers from acidic salt of tetravalent metals and organic monomers such as zirconium(IV) oxide-phenylethylamine and zirconium(IV) oxide-ethanolamine.12 These ion-exchangers showed more selectivity and specificity toward certain ionic species. Crystalline ion-exchangers of zirconium phosphate with layered based flexible structures such as R- and γ-zirconium phosphate have previously been reported by various researchers.13 Here, the tetravalent metal ions of these layeredbased products form interlayer spaces between two sheets, which can be tunable by swelling with polar solvents. Organic bases such as 1,10-phenanthroline could easily be intercalated between the layers of these crystalline materials, and by using such a procedure, a variety of materials can be synthesized with desired functionalities. However, to design an effective adsorption/ separation unit, the evaluation of efficacy of the adsorbent material is a prerequisite. The effectiveness of adsorbent material is normally determined for the evaluation of optimal performance under a variation of a series of experimental parameters. The present Article reports the synthesis of a new ionexchanger, which is based on a combination of zirconium(IV) and propanolamine. The reaction leads to the formation of an amorphous gel kind of material, which was utilized as an anionexchanger specifically for the removal of fluoride from drinking water. A simple synthetic principle was adopted where the tetravalent metal hydrolyzed to produce hydrous zirconium(IV) oxide, and the organic functionality is expected to get adsorbed at the interface of the gel to produce final adsorbent material (ZrPA). The prepared adsorbent material was characterized using advanced instrumentations such as FTIR, TGA/DTA, and XRD techniques. Surface morphology of the prepared material was evaluated using scanning electron micrography (SEM) coupled with energy dispersive spectrum (EDS) technique. The textural properties of the synthesized material were determined by the N2-BET method using a surface area porosimeter analyzer. The investigation primarily focuses on the kinetics and related isotherm studies, and thermodynamic studies where the influences of solution pH, contact time, adsorbent quantity, initial fluoride concentration, and temperature of the medium upon removal of fluoride were evaluated. The experimental results were tested for various kinetic models such as pseudo firstorder, pseudo second-order, first-order reversible, and intraparticle diffusion. Four other different isotherm equations such as Freundlich, Langmuir, Dubinin-Radushkevich (D-R), and Temkin isotherm were also fitted to determine the best-fit equation. Further, six different error functions, that is, SSE, SAE, ARE, HYBRID, MPSD, and R2, were computed to test the adequacy and accuracy of the model equations. Additionally, the suitability of adsorbent material in the presence of co-ions and with respect to elution and regeneration processes was also tested satisfactorily. The efficiency of adsorption of fluoride was investigated with the real groundwater samples collected from a fluoride-affected region of Orissa State, India. 2. Materials and Methods Zirconium oxychloride (Aldrich) and propanolamine (Aldrich) were used for the synthesis of the adsorbent material. The synthesis of Zr(IV)-propanolamine ion-exchanger was carried out by using a 0.1 M aqueous solution of zirconium oxychloride, which was mixed with a 0.1 M aqueous solution of propano-
9847
Table 1. Synthesis of Adsorbent Material with Variation of Ratio of ZrOCl2 · 8H2O and Propanolamine mixing ratio sample
ZrOCl2 · 8H2O
propanolamine
pH
% fluoride removal
ZrPA1 ZrPA2 ZrPA3 ZrPA4
1 1 1 1
1 2 3 4
3 10 10 10
99 69 66 61
lamine in different volume ratios (v/v) at pH 3.0 and 10.0 (Table 1). The pH of the solution was adjusted using 0.1 M HCl/NaOH solution. The white colored gel so obtained was allowed to stand for 24 h at room temperature to ensure complete precipitation. The gel was separated from the mother liquor by decantation and washed with distilled water a number of times to ensure a pH 7 of the effluent wash. The material was dried at approximately 50 °C in an oven for 24 h. It is noted that preliminary investigation with 10 mg L-1 fluoride solution with fixed quantity of adsorbent, at natural pH and ambient temperature (25 °C) of the medium, showed comparatively better fluoride removal capacity for which the material ZrPA1 was selected for further detail studies. The infrared spectra of all of the sample materials were recorded at a resolution of 4 cm-1 and 64 scans using Shimadzu IR Prestige-21 FTIR instrument. X-ray diffraction (XRD) data were obtained by a diffractometer (Shimadzu XRD-6000 diffractometer) using Cu KR radiation. Surface morphology of the prepared adsorbent material was examined using a scanning electron microscope (Jeol, JSM 6390 LV) coupled with energy dispersive spectrum (EDS) technique under vacuum, at an accelerating voltage of 20 kV. Surface area and pore volume were determined using a Quantachrome Autosorb Automated gas sorption system. The zeta potential of adsorbent material in solution was determined by electrophoretic studies using a Malvern Zeta meter (model Nano ZS). The thermogravimetric analysis was performed on a Shimadzu instrument (DTG 60) in nitrogen atmosphere under a flow of 30 mL min-1 and heating rate of 10 °C min-1 varying the temperature from 25 to 1000 °C. The extent of fluoride removal was studied using the batch experiment technique, which was carried out by mixing adsorbent material with fixed volume of fluoride solution. The mixture was agitated at 200 rpm in a mechanical shaker (Remi Equipments, Mumbai, India) placed in an incubator, until equilibrium was attained. The solution was filtered using a microfilter, and the fluoride concentration in the filtrate was measured using an ion selective electrode (Orion 720 A+ ion analyzer). The experiments were conducted at variable dose, initial concentration, agitation time, pH, and temperature where the optimum experimental conditions were ascertained. Fluoride uptake by the ion-exchange material was calculated using the following equation: Qe ) (Co- Ce)V/W
(1)
where Qe is equilibrium concentration (mg g-1). V is the volume of solution (L). Co (mg L-1) and Ce (mg L-1) are the initial concentrations of fluoride and equilibrium concentration at time t. W is the mass of the adsorbent material (g). 3. Results and Discussion 3.1. Characterization. The characterization of sample materials was carried out using various advance instrumentation techniques. The Fourier transform infrared (FTIR) spectrum of
9848
Ind. Eng. Chem. Res., Vol. 49, No. 20, 2010
Figure 3. Scanning electron micrograph (SEM) embedded with EDS showing the surface morphology of ZrPA. Figure 1. FTIR spectra of ZrPA and ZrPAF.
Figure 2. Thermal analysis showing TGA and DTA of adsorbent (ZrPA).
the adsorbent material before and after adsorption of fluoride is shown in Figure 1 as ZrPA and ZrPAF, respectively. In the spectra, the peaks in the range 3500-3200 cm-1 show the presence of -OH group. The hydrogen-bonded NH group is observed at 3500 cm-1 as a weaker band. Further, the presence of a peak at 1610 cm-1 in ZrPA was attributed to the -NHbending vibration band, which noticed a shift as well as a slight broadening in case of ZrPAF that may be taken as indicative of bonding between the protonated nitrogen and the fluoride.14 In addition, the -OH group in ZrPAF also noticed slight changes in feature (broadening of band) at 3420 cm-1, indicating the possible electrostatic attraction between the functional group and fluoride. The peak observed at 500 cm-1 in ZrPA is assigned to the metal-oxygen bonding, which also noticed a shift to 534 cm-1 with decreased intensity in case of ZrPAF. Analysis of the spectra thus provides evidence toward the participation of nitrogen and hydroxyl groups as possible coordinating sites for bonding with fluoride in the solution.11 The thermal analysis of the sample material ZrPA is furnished in Figure 2. The thermogravimetric analysis (TGA) curve shows rapid weight loss up to a temperature 450 °C. However, on increasing the temperature, a sharp deflection in the TGA curve was noticed at 480 °C, which is attributed to the decomposition of organic molecule present in the matrix at higher temperature. Correlating the data with differential thermal analysis (DTA),
Figure 4. Scanning electron micrograph (SEM) embedded with EDS showing the surface morphology of ZrPAF.
an exotherm was also noticed at the same temperature, that is, at 480 °C, which further supported the evidence.12 An endothermic peak noticed at 100 °C in the DTA curve indicates the loss of water molecules present as free/bonded form in the core of the matrix. Thermogravimetric analysis, therefore, strongly indicates that the organic molecule propanolamine could be present in the matrix or it may also remain adsorbed on the surface of the hydrous zirconium oxide gel. The scanning electron micrograph embedded with the EDS spectrum of adsorbent material (ZrPA) and fluoride adsorbed material (ZrPAF) are furnished in Figures 3 and 4, respectively. From the micrograph, the surface texture of adsorbent materials could be easily evaluated. The micrograph in Figure 4 shows whitish spots distributed over the surface, which could result from the adsorption of fluoride on the adsorbent. Further, the EDS spectrum clearly shows the presence of elements such as C, O, Zr, and F in the fluoride adsorbed material ZrPAF. The XRD spectrum, in Figure 5, of the adsorbent material shows the amorphous characteristic. The specific surface area, micropore volume, and average micropore diameter of the adsorbent material were determined to be 201.62 m2 g-1, 1.038 cm3 g-1, and 2.43 nm, respectively. 3.2. Sorption Studies. 3.2.1. Variation of Adsorbent Dose. Adsorption of fluoride onto adsorbent material is dependent on various parameters such as contact time, adsorbent
Ind. Eng. Chem. Res., Vol. 49, No. 20, 2010
9849
Figure 5. X-ray diffractogram of ZrPA showing the amorphous characteristics. Figure 7. Measurement of the zeta potential of ZrPA as a function of pH using 0.01 M NaCl and 0.001 M NaCl as supporting electrolyte.
Figure 6. Variation of pH. Concentration of fluoride, 10 mg L-1; time of contact, 60 min; adsorbent dose, 0.1 g.
dose, concentrations of fluoride, pH, and temperature of the medium. In the present investigation, the adsorbent dose was varied from 0.0125 to 0.225 g for a selected contact time period of 60 min, keeping the initial solute concentration at 10 mg L-1. It was observed that with increase in adsorbent dose the extent of fluoride adsorbed also increases, and the value reaches a maximum corresponding to a feeding dose of 0.1 g. The observed result could be attributed to the availability of more adsorbent particles, hence more adsorption sites at the solid adsorbent phase, and in the present case only 0.1 g of the solid adsorbent is just enough to remove almost 98-99% of the fluoride from the solution. 3.2.2. Variation of pH. The solution pH is one of the important parameters in determining the sorption potential of a sorbent material. Adsorption of fluoride by the prepared adsorbent was studied in the range of pH 2-12. The graphical representation, in Figure 6, shows that the adsorption of fluoride increases with increase in pH of the medium. However, beyond pH 7, the adsorption characteristics noticed a decreasing trend, and the optimum removal of fluoride was found to be at a pH of 7.0. It is interesting to note that the material also exhibit fluoride uptake behavior in acidic pH of the medium. For example, at pH 2, nearly 85% of the fluoride could be easily
removed. A progressive decrease in fluoride removal was observed in alkaline range (pH > 7), the reason for which may be ascribed to the fact that in a solution phase the competition of the hydroxyl ions with fluoride ion does exist and subsequently with increase in pH value, the concentration and hence mobility of hydroxyl ion also increase. To consolidate the result obtained in pH studies, the isoelectric point (IEP) of the adsorbent material was determined using electrophoretic mobility in electrolytic medium of two different ionic strengths, that is, 0.01 and 0.001 M NaCl. The surface charge assessed by IEP is defined as the point at which the zeta potential was found to be zero. When pH < pHIEP, the surface charge is positive, and when pH > pHIEP, the surface charge is negative, and when the pH ) pHIEP, it indicates neutral surface.15 The electrophoretic studies, shown graphically as a function of pH versus zeta potential (mV), in Figure 7, show that the IEP values of adsorbent is 7.1 and 7.5, respectively, in two different electrolytes 0.01 and 0.001 M NaCl solution. With increasing ionic strength, the IEP values were found to be decreased, which could be attributed to the adsorption of background electrolyte, that is, NaCl. Apart from this, it could be noted that the anions could also be adsorbed on the surface of the adsorbent materials, neutralizing positive charge of the surface to some extent and resulting in decreasing IEP values. 3.2.3. Variation of Contact Time Period. The effect of variation in contact time on adsorption of fluoride at a fixed initial concentration of 10 mg L-1 indicates that with increasing time of contact, the amount of uptake of fluoride by the adsorbent material also increases, and equilibrium could be reached in a 16 min time period. Thus, a maximum 99% of fluoride could be successfully adsorbed within a contact time of 16 min from a given 10 mg L-1 solution at natural pH of the medium with an adsorbent dose of 0.1 g. Further, the efficacy of fluoride adsorption was computed using various kinetics expressions such as pseudo first-order, pseudo second-order, first-order reversible, and intraparticle diffusion model. Mathematical expressions for the abovementioned model equations and related discussions are presented as follows: The integrated form of the pseudo first-order rate equation16 can be represented as: log(qe- qt) ) log qe- Kft/2.303
(2)
9850
Ind. Eng. Chem. Res., Vol. 49, No. 20, 2010
Table 2. Kinetic Parameters for Fluoride Adsorption onto ZrPA at pH 7.0a pseudo first-orderb qexp
Kf
qe,cal
9.9
0.211
5.0691
a
pseudo second-orderb
reversible first-orderb
R2
Ks
h
qe,cal
R2
0.982
0.0669
7.575
10.638
0.999
-1
-1 b
Adsorbent dose, 0.1 g L ; temp, 298 K; initial fluoride concentration, 10 mg L . min-1); KB, min-1; Ki, C, mg (g min1/2)-1.
KB
R2
0.212
0.982
intraparticle diffusionb Ki
C
R2
1.151
5.459
0.978
-1
-1
The units for qexp, qe,cal, mg g ; Kf, L min ; Ks, h, (mg g-1
where qe and qt denote the amount of fluoride adsorbed (mg g-1) at equilibrium and at time t, respectively. Kf (min-1) is the rate constant of pseudo first-order adsorption reaction. A linear nature of plot of log(q e- qt) versus t suggests the applicability of the kinetic models where both qe and Kf can be determined from the intercept and slope of the curve, respectively. Similarly, the linear form of the pseudo second-order kinetic model16 can be expressed as: t/qt ) 1/Ks(qe)2+ t/qe
(3)
where Ks is the rate constant for the pseudo second-order reaction (g mg-1 min-1). qe and qt are the amounts of solute sorbed at equilibrium and at any time t (mg g-1), respectively. The straight line plot of t/qt versus t for the kinetic data gives the values for qe and Ks, respectively, from the slope and intercept. On further simplification, eq 3 can be rewritten as: h ) Ksq2e
(4)
where h is the initial sorption rate (mg g-1 min-1). To represent the first-order reversible equation, the integrated form17 of the equation is represented as: ln[1 - U(t)] ) KBt
Figure 8. Pseudo second-order kinetic plot for fluoride adsorption on ZrPA. Temp, (25 ( 2) °C; pH, 7.0; fluoride concentration, 10 mg L-1; adsorbent dose, 0.1 g L-1.
(5)
where U(t) ) (Cao - Ca)/(Cao - Cae), and Cae is the equilibrium solute concentration. KB ) -k#, and k# is denoted as the overall rate constant. KB can be calculated from the slope of the linear plot of ln[1 - U(t)] versus t. The intraparticle diffusion model16 is expressed as: qt ) Kit1/2 + C
(6)
where Ki is the intraparticle diffusion constant (mg g-1 min-1/2) and C is the intercept. The value of Ki is determined from the slope of the plot qt versus t1/2. The value of intercept C provides information about the thickness of boundary layer, that is, the resistance to the external mass transfer.15 In the context of application of the above-mentioned model equations via experimental conditions mentioned in the present investigation, the kinetic behavior was studied for an initial fluoride concentration of 10 mg L-1 with adsorbent dose of 1.0 g L-1, keeping the pH fixed at 7.0 at room temperature to find out the minimum time of contact needed to attain maximum removal. Table 2 lists the detail results of rate constants obtained from the graphical representation of slope and intercept of various linear plots (Figures 8 and 9). Analyzing the data, it can be mentioned that the value of correlation coefficient (R2) for the pseudo second-order kinetic model is found to be very high (0.999), for which the values of qe,exp and qe,cal are also found to be very close to each other. Therefore, it can be mentioned that for the present investigation the sorption kinetics can be conveniently represented by the pseudo second-order model for the adsorption fluoride onto the adsorbent material. From the Weber-Morris intraparticle diffusion plot, the value of intercept C provides information about the thickness of
Figure 9. Weber-Morris intraparticle diffusion plot for fluoride adsorption on ZrPA. Temp, (25 ( 2) °C; pH, 7.0; fluoride concentration, 10 mg L-1; adsorbent dose, 0.1 g L.-1.
boundary layer, that is, the resistance to the external mass transfer. The larger is the value of the intercept, the higher is the external resistant.16 The deviation of straight line from the origin, as shown in Figure 9, may be due to the difference between the rate of mass transfer in the initial and final stages of adsorption, which indicates that the pore diffusion is not the sole rate-controlling step. 3.2.4. Variation of Initial Fluoride Concentration. The effect of variation in fluoride concentration upon adsorption process was also investigated within the concentration range of 2-50 mg L-1. In this case, a maximum contact time period of 30 min and adsorbent dose of 0.1 g L-1 with volume of fluoride solution 100 mL was maintained as the experimental
Ind. Eng. Chem. Res., Vol. 49, No. 20, 2010 Table 3. Linear and Nonlinear Freundlich, Langmuir, Temkin, and D-R Isotherm Constants Related to the Sorption of Fluoride onto ZrPA model/parameters
linear method Freundlich
KF 1/n
29.784 0.562
29.538 0.481
Langmuir KL qm
6.2 32.258
b A
283.604 39.112
1.170 59.119
Temkin 296.189 45.885
D-R qm K
0.1488 0.005
0.149 0.005
condition. It was observed that with progressive increase in initial fluoride concentration, the percentage/amount of removal of fluoride decreases. This could be attributed to the fact that increases in concentration of fluoride lead to the saturation of available coordination sites; hence, the binding capacity of the adsorbent approaches a saturation value resulting in overall decrease in removal percentage. When the initial fluoride concentration was low, almost 100% removal of fluoride from the solution could be achieved. Further, the relationship between the equilibrium amount of fluoride on the adsorbent and the solute concentration in solution was verified using various isotherm equations. To model the present experimental investigation, the experimentally obtained equilibrium data were fitted to four different isotherm equations, Langmuir, Freundlich, Temkin, and D-R. The significance of various isotherm equations in adsorption process has been highlighted in a number of previously published research articles.15,16 It can be mentioned that the correlation of equilibrium data, either theoretical or empirical equations, is essential for practical design and operation of adsorption process. The data obtained for the present investigation are presented in Table 3, where both linear and nonlinear regression methods of analysis were used to evaluate various adsorption parameters. The theoretically predicted isotherm data were determined using a Microsoft excel for linear analysis and a SPSS 13.0 statistic software was used for nonlinear assessment. The isotherm equations are discussed as follows. The linear form of the Freundlich equation15 is expressed as: ln qe ) ln KF+ 1/n ln Ce
(7)
where the Freundlich constants KF and 1/n described the adsorption capacity and adsorption intensity, respectively. The higher is the value of 1/n, the higher will be the affinity between the adsorbate-adsorbent and the heterogeneity of the adsorbent sites. As shown in Table 3, for the present investigation, the calculated values of 1/n were found to be 0.56 (linear method), indicating that nearly 56% of active adsorption sites contain equal energy for adsorption process. The linear form of Langmuir adsorption isotherm15 equation is expressed as: 1/qe ) (1/KLqm)(1/Ce) + 1/qm
which is otherwise known as the separation factor or equilibrium parameter RL. The expression for RL is: RL ) 1/(1 + KLCo)
nonlinear method
(8)
where the values of Langmuir constants qm and KL relate to the capacity and energy of the adsorption process. The isotherm criterion can be described by another dimensionless constant,
9851
(9)
where Co is the initial concentration of fluoride (mg L-1), and KL is the Langmuir isotherm constant. The feasibility of isotherm criterion can be evaluated from the values of RL, where RL < 1 represents favorable adsorption and RL > 1 represents unfavorable adsorption processes. The calculated value of RL for the initial fluoride concentration of 10 mg L-1 was found to be 0.015 and 0.078 by linear and nonlinear regression analysis, respectively, indicating a favorable condition for the adsorption of fluoride. Application of the Temkin isotherm model started basically with the assumption of a linear relationship that exists between adsorbent-adsorbate interactions with a decrease in heat of adsorption for all the molecules covered in a layer. Further, the model also emphasizes upon the adsorption process, which is characterized by uniform distribution of binding energies. The linearized form of the Temkin equation15,16 can be written as: qe ) B1 ln A + B1 ln Ce
(10)
The value of the constant B1 is related to the heat of adsorption, and A is related to the equilibrium binding constant (L mol-1) corresponding to the maximum binding energy. A plot of qe versus ln Ce enables the constants B1 and A to be determined from the slope and intercept, respectively. The value of heat of adsorption, B1, in the present investigation, was estimated to be 8.376 kJ mol-1. Information regarding adsorption mechanism related to the material-anion interaction can be obtained from the DR isotherm model. The linearized form of eq 15 can be written as: ln qe ) ln qm-Kε2
(11)
where ε2 is the Polanyi potential equal to RT ln(1 + 1/Ce). Both qe and qm related to the amount of adsorbate adsorbed at equilibrium per unit of adsorbent (g g-1) and the theoretical saturation capacity (g g-1), respectively. Ce is the equilibrium solid concentration (g L-1), and K is a constant related to adsorption energy. Most importantly, K gives the value of mean free energy of adsorption. The relationship between K and the energy (E) per mole of the adsorbate (energy required to transfer one mole of fluoride to surface of adsorbent material) can be computed using the following equation: E ) (-2k)-0.5
(12)
The magnitude of E is quite useful for estimating the type of adsorption, and if this value is between 8 and 16 kJ mol-1, the adsorption type can be described as ion-exchange.15 For the values of E < 8 kJ mol-1, the adsorption process is said to be physical in nature. The values of E in this study were found to be 10.0 kJ mol-1, which suggests that the adsorption proceeded by the ion-exchange process. Both values of K and qm, respectively, can be obtained from the slope and intercept of the plot qe versus ε2. The fitting presentations for all the isotherm models are presented in Figures 10 and 11, respectively, for linear and nonlinear regression analysis. For the present investigation, analysis of the results indicated that the DR isotherm model could be the best-fit model to describe the material-anion interaction process.
9852
Ind. Eng. Chem. Res., Vol. 49, No. 20, 2010 Table 4. Isotherm Error Deviation Data Related to the Sorption of Fluoride onto CIC Using Six Commonly Used Functions R2
error functions
SSE
SAE
ARE
HYBRID
MPSD
44.517 25.816 65.969 21.684
56.600 32.020 205.947 32.038
13.751 19.753 56.054 4.386
38.091 40.701 167.913 6.991
Linear Approach Freundlich Langmuir Temkin D-R
0.987 2691.118 124.651 60.559 0.994 604.409 60.031 35.764 0.911 214.362 39.852 105.413 0.997 131.456 34.624 34.135 Nonlinear Approach
Freundlich Langmuir Temkin D-R
0.995 0.985 0.912 0.997
11.320 35.072 209.088 6.237
9.243 16.272 39.851 6.982
22.371 31.411 89.040 6.876
Table 5. Thermodynamic Parameters for the Sorption of Fluoride on ZrPA
Figure 10. Adsorption isotherm modeling of fluoride removal using ZrPA using linear regression analysis. Adsorbent dose, 0.1 g L-1; pH, 7.0; temp, (25 ( 2) °C; contact time, 30 min.
temp (°C)
∆H° (kJ mol-1)
∆S° (kJ mol-1 K-1)
∆G° (kJ mol-1)
Ea (kJ mol-1)
S*
10 20 30 40 50
84.852
0.3757
-21.47 -110.08 -113.84 -117.59 -121.35
83.148
0
percentage of uptake of fluoride by the adsorbent material increases with increase in temperature of the medium. The thermodynamic feasibility of the adsorbent-fluoride interaction process can be represented as:18 ∆Go ) -RT ln KD
(13)
where ∆G° is the change in free energy, T is the absolute temperature, R is the universal gas constant (8.314 J mol-1 K-1), and KD is the equilibrium constant. Negative values of ∆G° at all temperatures, as indicated in Table 5, show the spontaneous nature of the adsorption process, which is associated with increased temperature. Further, the values of the enthalpy change (∆H°) and the entropy change (∆S°) associated with the processes were evaluated using the relationship: ln KD ) ∆So /R - ∆Ho /RT Figure 11. Adsorption isotherm modeling of fluoride removal using ZrPA using nonlinear regression analysis. Adsorbent dose, 0.1 g L-1; pH, 7.0; temp, (25 ( 2) °C; contact time, 30 min.
Further, to consolidate the results obtained using both linear and nonlinear regression analysis and for the evaluation of the most suitable isotherm model to represent the experimental data, six different error functions16 were also used to calculate the error deviation between the experimental and predicted equilibrium data. The different well-known error functions used to calculate the error deviation can be listed as (i) the sum of the squares of errors (SSE), (ii) the sum of absolute errors (SAE), (iii) the average relative errors (ARE), (iv) the hybrid fractional error function (HYBRID), (v) the Marquardt’s percent standard deviation (MPSD), and (vi) the regression coefficient (R2). The calculated values of the error functions are listed in Table 4. Compilation of the results shows that the DR isotherm model is the most suitable model to satisfactorily describe the adsorption phenomenon based on the highest R2 value and lowest SSE, SAE, ARE, HYBRID, and MPSD values. 3.2.5. Variation of Temperature. The effect of temperature upon the adsorption process was studied with variation of temperature of the medium from 10 to 50 °C, and the related thermodynamic parameters were also calculated. The experimental observation showed that, within the range of study, the
(14)
Here, the plot of R ln KD versus 1/T yielded a straight line (Figure 12), from which the ∆H° and ∆S° values were evaluated from the slope and intercept, respectively. The results are presented in Table 5, where the positive value of ∆H° indicates
Figure 12. Plot of ln KD versus 1/T for ZrPA.
Ind. Eng. Chem. Res., Vol. 49, No. 20, 2010 Scheme 1. Proton Shifting Mechanism in the Formation of ZrPA
+
-
MOH + H3O + F f
9853
MOH+ 2 ----F +
H2O (electrostatic attraction)
and (iii) ion-exchange reaction between positively charged metal center and fluoride: MOH + H3O++ F- f M+---F-+ 2H2O (ion-exchange) MOH + H3O++ F- f M+---F-+ 2H2O (ion-exchange) Further, when the pH of the medium remains relatively in a neutral range, that is, at pH 7.00, the fluoride adsorption onto the neutral solid surface can be described by a ligand or ionexchange reaction mechanism, which is represented as: that the process is endothermic in nature. The value of ∆S° was found to be 0.375 kJ mol-1 K-1, indicating a relatively ordered state rather than a chaotic distribution. Further, we compute the factor sticking probability (S*), which is related to the surface coverage (θ) by the adsorbed molecule and associated activation energy, for which eq 17 can be written as: S* ) (1 - θ) exp(-Ea /RT)
(15)
where θ denotes the surface coverage, which can be determined using the following equation: θ ) 1 - Ce/Co
(16)
where Co and Ce represent the initial and equilibrium fluoride ion concentrations, respectively. Both values of Ea and θ can be calculated, respectively, from the slope and intercept of the plot of ln(1 - θ) versus 1/T. The positive values of ∆H° and activation energy Ea confirm the endothermic nature of the sorption process. The value of S* is found to be 0, clearly indicating that the adsorption follows a chemisorption process.15,16 3.2.6. Mechanism of Fluoride Adsorption. Understanding the mechanism of any adsorption process plays a major role in understanding the material characteristics and design of the new adsorbent for future applications purpose. Thus, considering the results obtained from the experimental investigations and computing the obtained results using mathematical/theoretical models, a mechanism for the adsorption of fluoride by the newly synthesized ZrPA can be proposed. On the basis of zeta potential measurement, the uptake of fluoride by ZrPA could be explained on the basis of a proton shifting mechanism, which occurs from the acidic site of the gel to the adsorbed organic molecule to form a protonated amine (Scheme 1). At lower pH of the medium, when pH < pHIEP, surface sites are positively charged and, therefore, attract negatively charged fluoride by an electrostatic interaction process. Upon hydration, the zirconium surface completes the coordination shells with the available OH group, and depending on the variation of pH, these surfaceactive OH groups may further bind or release H+ where the surface remains positive due to the reaction: MOH + H3O+ f MOH+ 2 + H2O Thus, when pH < 7.00 (pHzpc ) 7.0-7.5), the overall mechanism of fluoride adsorption can be represented in three different forms, which are outlined as follows: (i) electrostatic interaction between positively charged center (nitrogen) and negatively charged fluoride molecule in solution, (ii) electrostatic attraction between positively charged surface hydroxyl group and fluoride,
MOH + F- f M+---F-+ OHThis proposed mechanism is supported with the evidence of observed notable increase of equilibrium solution pH. The maximum adsorption of fluoride at neutral pH range is supported by the obtained values of isoelectric points, which were found to be 7.1 and 7.5 for two different electrolytic mediums. Here, the mechanism of fluoride removal process has been outlined similar to that discussed for metal oxide by Tor19 and Sarkar et al.8 where the exchange between available surface OH group and fluoride occurs due to the fact that both F- and OH- are isoelectronic having comparable ionic radius. With an increase in pH of the medium, the ZrPA surface could become more negative, for which the fluoride adsorption capacity reduces due to the columbic repulsion force where like charges repel each other. However, the low adsorption capacity obtained at higher pH could be explained on the basis of the mechanism of adsorption of available Na+ in solution in the first sphere following competition of fluoride ion with hydroxyl ion at the secondary adsorption sphere at the solid adsorbent phase. MOH + Na++ F-+ OH- f MO-----Na+---F-+ H2O It is to be noted that, although here we have tried to outline a possible mechanism for the newly synthesized adsorbent ZrPA, the modeling of the specific adsorption of fluoride on any material surface depends on a number of external factors such as temperature, pH, initial fluoride concentration, as well as the density of surface functional groups available for coordination. In light of the above-mentioned mechanism of adsorption, it may be further noted that ZrPA showed adsorption capacity at a wide pH range, which could be useful for commercial exploitation purpose. Here, we can mention the main role of propanolamine, which is basically 2-fold. First, the attachment of organic molecule (propanolamine) to the structure develops nucleophilicity. Second, the attachment of the organic molecule could be responsible for the development of porosity in the structure and for imparting a large specific surface area after the process of drying. Further, the organic molecule may also combine with the zirconium to form a layer where the neighboring particles could be interconnected. 3.2.7. Effect of Coexisting Ions. Drinking water contains many other ions such as sulfate, chloride, nitrate, etc., along with fluoride, which may compete with fluoride for active sorption sites.20 The effect of various diverse ions/competing co-ions upon adsorption of fluoride was investigated using sulfate, nitrate, chloride, bicarbonate, and phosphate ions. The experimental condition for initial fluoride concentration was kept at 10 mg L-1, varying the initial concentration of co-ions from
9854
Ind. Eng. Chem. Res., Vol. 49, No. 20, 2010
Figure 13. Effect of co-ions upon removal of fluoride from solution. Amount of adsorbent, 0.1 g L-1; concentration of fluoride, 10 mg L-1; time of contact, 30 min; temperature, (25 ( 2) °C.
0 to 600 mg L-1. The results of the observation showed graphically (Figure 13) indicated that NO3-, Cl-, and SO4-2 had little effect upon fluoride removal processes. However, the experimental result showed that the presence of PO4-3 and HCO3- significantly affected the fluoride removal process. Fluoride sorption capacity was reduced by nearly 50-70% for an initial phosphate and bicarbonate concentration of 300 mg L-1. Maliyekkal et al.21 have also reported a similar kind of observation for the removal of fluoride using a new adsorbent material, magnesia amended activated alumina. It is expected that phosphate and bicarbonate may compete for the active sorption sites along with the fluoride during the adsorption process, which resulted in a decrease in adsorption capacity. Apart from this, the effect of these anions toward adsorption may be due to their affinity toward the adsorbent material. The defluoridation capacity of adsorbent material was tested in groundwater samples collected from Orissa, India. The synthetic water sample was prepared by adding sodium fluoride to deionized water. The results of such investigations are presented in Figure 14. It was observed that more fluoride removal could be possible from the synthetic water sample in comparison to the collected groundwater sample. The reason could be attributed to the fact that the real groundwater sample is always associated with a number of cations and anions, which could well interfere with the adsorption process. A summary of the characteristics such as pH, total hardness, alkalinity, presence of various metal ions, etc., of the field water samples collected from different places is given in Table 6. It is to be noted here that the presence of various cations and hence the total concentration of salts can also play a vital role in the determination of adsorption of fluoride. 3.2.8. Reuse and Regeneration of Adsorbent. Regeneration and reuse of the adsorbent material carries utmost importance, which directly affects the cost factor and hence its utility in continuous batch adsorption processes. Only the adsorbent materials that can be reused have practical value in real system. The reusability capacity of adsorbent was performed with dried adsorbent. As shown in Figure 15, the percentage of adsorption of fluoride by adsorbent was found to be reduced from 98% to 80% following a sequence of the first to tenth cycle of batch operation study. Desorption studies were carried out with variation of concentration of NaOH solution. It was observed that leaching of
Figure 14. Comparison of adsorption of fluoride in synthetic water and collected groundwater sample. Amount of adsorbent, 0.1 g L-1; time of contact, 30 min; temperature, (25 ( 2) °C. Table 6. Characteristics of Field Water Samples parameters
valuesa
turbidity (NTU) total hardness as CaCO3 (mg L-1) total iron as Fe (mg L-1) chloride as Cl (mg L-1) fluoride as F (mg L-1) nitrate as NO3 (mg L-1) sulfate as SO42- (mg L-1) alkalinity as CaCO3 (mg L-1) Na (mg L-1) K (mg L-1) pH conductivity (νS) Mn (mg L-1)
0.6-5.0 75-466
