Environmental Materials Unit, National Environmental Engineering Research Institute, Nehru Marg, Nagpur 400 020, India
a r t i c l e i n f o a b s t r a c t
Article history:Received 24 April 2008Accepted 1 November 2008Available online 31 January 2009
Keywords:DefluoridationChitosanTitanium macrospheresFluoride adsorptionEffect of coexisting anions and pH
In the present study, the metal-binding property of chitosan is used to incorporate titanium metaland applied as an adsorbent for fluoride adsorption. Titanium macrospheres (TM) were synthesizedby a precipitation method and characterized by FTIR, SEM, and XRD. The Langmuir and Freundlichadsorption models were applied to describe the adsorption equilibrium and the adsorption capacitieswere calculated. Thermodynamic parameters of standard free energy change (�G◦), standard enthalpychange (�H◦), and standard entropy change (�S◦) were also calculated. The effects of various physico-chemical parameters such as pH, initial concentration, adsorbent dose, and the presence of coexistinganions were studied. The fluoride uptake was maximum at neutral pH 7 and decreased in acidic andalkaline pH. The presence of coexisting anions has a negative effect on fluoride adsorption. TM wasfound to have very fast kinetics in the first 30 min and then the rate slowed down as equilibrium wasapproached. A comparison of fluoride removal in simulated and field water shows a high adsorptioncapacity in simulated water.
Fluoride is an essential micronutrient present in water for hu-man and animal health. The presence of fluoride in permissiblelimits is beneficial for producing and maintaining healthy bonesespecially in children. The excess intake of fluoride causes severedental or skeletal fluorosis [1,2]. Fluoride contamination in ground-water is a worldwide problem in countries like Africa, Asia, andthe United States [3]. In India almost 17 states have been af-fected by fluoride [4]. The main source of fluoride ingestion inhuman body is drinking water; therefore, it is necessary to treatthe fluoride-contaminated water and bring down the fluoride con-centration within a permissible limit, i.e., 1.5 mg L−1 according toWHO guidelines [5].
There are several existing technologies for fluoride removalfrom water, which include chemical precipitation [6,7], electrodial-ysis [8], ion exchange [9], membrane filtration [10], and adsorption[11]. The Nalgonda technique [6] is widely used for fluoride re-moval in India because of its simplicity, but sludge generationis a major concern in this technique. Adsorption is a widely ac-cepted, ideal, and appropriate technique compared to others notonly at the community level but also at the domestic level. Ac-tivated alumina is a most commonly used adsorbent for defluo-ridation [12] and several other adsorbents reported include metal
Among the different adsorbents reported for defluoridation,rare earth or transition metal-based adsorbents are found to bepotential materials for fluoride removal. Fluoride ion is classifiedas a hard base due to its high electronegativity and small ionicsize and has strong affinity for electropositive multivalent metalions like Al(III), Zr(IV), Ce(IV), Fe(III), and La(III). As the rare earthmetals are expensive, mixing the rare earth metals with cheapermetals or loading metals on supports will enhance the adsorptioncapacity and would be economical. A number of publications havebeen reported mainly on alumina and metals like iron [19] andlanthanum [20], but there are only a few publications in which ti-tanium oxide or hydroxide has been used for fluoride adsorption[21,22]. Ngee et al. [21] used mesoporous titanium oxyhydroxidefor fluoride removal and Lokshin and Belikov [22] used titanium-rich bauxite for fluoride removal.
In our previous publication we have studied defluoridation ofdrinking water using lanthanum-modified chitosan flakes [20]. Inthis study, titanium ion supported on a biopolymer matrix hasbeen investigated for fluoride adsorption.
Chitosan is a biopolymer obtained by deacetylation of naturallyoccurring biopolymer chitin. Chitin is a most abundant biopolymerin the ecosphere after cellulose. Chitosan is a cationic copolymerof 2-glucosamine and N-acetyl-2-glucosamine. Chitosan is able tocoordinate with metals because of high concentrations of amino(–NH2) functional groups. Nitrogen present in the amine groupacts as an electron donor and forms chelates or complexes with
S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290 281
metal ions [23]. Thus, the amino and hydroxyl groups in chitosanact as the active sites for adsorption. Chitosan has been used forremoving metal ions such as chromium [24], cadmium [25], cop-per [26], and arsenic [27]. Chitosan has a tendency to form a gelin acetic acid and this chitosan gel can be precipitated in alkali toobtain spherical beads.
The main objectives of this study were to utilize the metal-binding characteristic of chitosan to incorporate the Ti(III) metaland to gain improved kinetics of fluoride uptake. This new mate-rial has been used for fluoride adsorption from simulated waterand the groundwater obtained from the Dhar district in MadhyaPradesh, India. Effects of pH, initial concentration, adsorbent dose,and existing co-ions were also studied in a batch study.
2. Experimental
2.1. Materials
Chitosan used in this study was obtained from local commercialsources. Ti-isopropoxide was obtained from Across Organics (USA).All other chemicals used in this study were analytical grade ob-tained from E-Merck India Ltd. (Mumbai, India). A stock solutionof fluoride was prepared by dissolving sodium fluoride in deion-ized water and a working solution of 5 mg L−1 was obtained byappropriate dilutions of stock solution.
2.2. Synthesis of Ti macrospheres
Chitosan was dissolved in 5% v/v acetic acid with constantstirring. Titanium solution was prepared by adding 10 ml con-centrated HCl to 8.2 g of Ti-isopropoxide. This titanium solutionwas then added slowly to chitosan gel under vigorous stirring.This Ti-chitosan solution was precipitated in 50% v/v ammoniasolution dropwise with continuous stirring to obtain the spher-ical beads and allowed to stabilize for 1 h. After stabilization,the beads are separated from ammonia solution and washed thor-oughly with distilled water till removal of ammonia. The sphericalbeads were dried in an oven at 80 ◦C. Finally these Ti macrosphereswere washed with distilled water before being used for defluorida-tion experiments. Similarly Ti macrospheres (TM) with 10, 15, and20 wt% Ti loading were synthesized.
2.3. Adsorption experiment
The adsorption studies for fluoride removal form drinking waterby TM were carried out in a batch process. A 50-ml fluoride solu-tion of desired concentration was taken into a PVC conical flaskand a known weight of adsorbent was added into it and shakenfor 24 h on a horizontal rotary shaker (Model No. CIS-24, RemiInstruments, Mumbai, India) to attain the equilibrium. Adsorbentwas then separated using Whatman filter paper No. 42 and thefiltrate was analyzed for residue using an ion-selective electrode.All adsorption experiments were carried out at room temperature(30 ± 2 ◦C). The specific amount of fluoride adsorbed was calcu-lated using the equation
qe = (C0 − Ce) × V
W, (1)
where qe is the adsorption capacity (mg g−1) in the solid at equi-librium; C0 and Ce are initial and equilibrium concentrations offluoride (mg L−1), respectively; V is volume of the aqueous solu-tion, and W is the mass (g) of adsorbent used in the experiments.The effect of pH on fluoride removal was studied by adjusting thepH of the solution using 0.1 N HCl and 0.1 NaOH. In the study ofeffect of co-anions present, the pH of the medium was not con-trolled.
2.4. Kinetic study
In order to estimate equilibrium adsorption rate for the up-take of fluoride by TM, time-dependent sorption studies were con-ducted in a three-necked glass vessel having a capacity of 1000 ml.The particular concentration of fluoride solution was transferredinto the vessel and a known weight (1 g L−1) of the adsorbent wasadded to it. The suspension was stirred using a four-blade, pitchedturbine impeller with the stirring speed 500 rpm. Samples werewithdrawn from the vessel at frequent time intervals and analyzedfor the concentration of fluoride using an ion-selective electrode.
2.5. Method of analysis
Fluoride concentrations in the experimental samples were an-alyzed using a fluoride ion-selective electrode (Orion 9490 on aSargent Welch Model PAX 900 pH/activity meter). The solutionpH was measured using the same ion meter coupled with pHelectrode. The titanium and other heavy metals if leached fromadsorbent were determined using inductively coupled plasma–atomic emission spectroscopy (ICP-AES, Model OPTIMA 4100DV).The physico-chemical parameters of treated and untreated waterwere determined using standard methods of water and wastewateranalysis [28]. Carbonate and bicarbonate alkalinity, total hardness,and chlorides were determined by a titrimetric method, whereasthe concentrations of nitrate and sulfate were determined spec-trophotometrically using a UV visible spectrophotometer (Chemito,Model UV-2100). Most of the experiments were repeated twicefor better accuracy and blank experiments were also performedthroughout the studies. The experimental error was observed tobe within ±3%. Similar experimental procedures were followed forthe effects of coexisting ions, pH, initial concentration, etc.
2.6. Thermodynamics
Thermodynamic parameters of adsorption standard free energychange (�G◦), standard enthalpy change (�H◦), and standard en-tropy change (�S◦) were calculated at 303, 313, and 323 K.
2.7. Physical characterization
XRD patterns of chitosan and TM were recorded on Rigaku X-ray diffractometer. The sample was scanned for 2θ range from 10to 60◦ . The FTIR spectra of chitosan and TM (before and after flu-oride adsorption) in KBr pellets were recorded on a Bruker ModelVertex 70 spectrometer. The SEM analysis of chitosan and TM wascarried out using a Jeol JXA-840 A electron probe microanalyzer(Japan) with different magnifications. Surface area of TM has beenmeasured using the Brunauer Emmett Teller (BET) method.
3. Results and discussion
3.1. Characterization of TM
3.1.1. FTIR analysisThe functional groups of chitosan such as amino and hydroxyl
groups are very important for adsorption. The FTIR spectra of chi-tosan flakes and TM are given in Fig. 1. The band at 3294 cm−1 inchitosan is attributed to stretching vibration of the hydroxyl groupwhich is shifted to 3305 cm−1 in TM and the peak at 3695 cm−1 isdue to stretching vibration of the N–H group of chitosan which isshifted 3660 cm−1 in TM. This band shifting may be due to forma-tion of weak intermolecular hydrogen bonding between the aminogroup and the hydroxide group of chitosan and incorporation ofTi–OH. Pauline et al. [29] have reported that in chitosan the bandat 1571 cm−1 is more intense than at 1676 cm−1, which suggests
282 S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290
Fig. 1. FTIR spectra of (a) chitosan flakes, (b) TM before adsorption, (c) TM after adsorption.
effective deacetylation. However, in TM the band at 1676 cm−1
in chitosan has shifted to 1681 cm−1 due to incorporation of Ti–OH. The bands at 2924 and 1376 cm−1 in chitosan and 2889 and1361 cm−1 in TM are attributed to C–H stretching vibrations inpolymeric backbone and C–H bending, respectively. The significantdecrease in the band intensity is due to the interaction betweenchitosan and Ti. The bands at 1155 and 1075 cm−1 in chitosan maybe due to stretching vibration of C–O. In the case of TM the inten-sity of these bands has decreased drastically, suggesting that thereis an interaction between the C–O group of chitosan and Ti–OH.FTIR spectra of TM before and after adsorption were obtained andshown in Fig. 1. The bands at 1681 and 1571 cm−1 in TM beforeadsorption are shifted to 1701 cm−1 after fluoride adsorption, indi-cating weak electrostatic interaction between chitosan and fluorideions. From the FTIR spectra of TM before and after adsorption itcan be noted that there are no new peaks appearing after the ad-sorption process.
3.1.2. Scanning electron microscopy of TMThe SEM analysis was performed to understand the morphology
of chitosan and TM. Chitosan flakes exhibit rough surface morphol-ogy. The lack of uniformity in geometry of these flakes is due touneven sizes of the flakes. SEM of TM shows that it has a denseand firm structure with minimum porosity (Fig. 2). TM is found tohave a particle size in the range of 1114.76 to 1154.81 μm. The sur-face area of TM is increased more than that of the chitosan flakes.The BET surface areas of chitosan and TM were found to be 2.29and 3.18 m2/g, respectively. TM exhibit spherical geometry withsmooth and folded surface morphology more accurately describedas surface depression. This may be due to titanium hydroxide thatgives a uniform surface property. It was observed that the mor-phology of TM does not change after fluoride adsorption.
3.1.3. X-ray diffraction of TMThe X-ray diffraction patterns of chitosan and TM are shown in
Fig. 3. The XRD patterns of chitosan present three peaks at 2θ = 8,20, and 29◦ , respectively. However TM presents two major peaks
at 2θ = 20 and 25◦ . The peak at 2θ = 25◦ in TM is due to thepresence of titanium hydroxide. The peaks in TM are weak, whichindicates that the degree of crystallinity of TM is lower than thatof chitosan.
3.2. Effect of titanium loading
To study the effect of Ti loading on fluoride removal, Ti loadingwas varied from 5 to 50 wt%. The effect Ti loading on adsorptionof fluoride is shown in Fig. 4a. It was observed that an increase intitanium loading from 5 to 15% has improved fluoride removal 61to 89%. A 15% Ti loading was optimum as it shows the highest flu-oride removal and reduces the fluoride level to permissible limits.Further increasing titanium loading has a negative effect on fluo-ride removal and this may be due to overlapping of active sites.Hence further studies are carried out for 15% Ti-loaded macro-spheres.
3.3. Preliminary adsorption experiment
A preliminary adsorption experiment was carried out using chi-tosan and TM at initial F− concentration 5.2 mg L−1, contact time24 h, to check the fluoride removal efficiency of both adsorbents.Fig. 4b shows the plot of Ce versus qe for TM. The modified chi-tosan shows much higher fluoride removal efficiency compared tountreated chitosan. Therefore further adsorption studies for fluo-ride removal were carried using modified chitosan. The optimizedTi3+ loading was found to be 15% Ti-loaded TM as shown inFig. 4a.
3.4. Effect of adsorbent dose
The effects of adsorbent dose at fixed initial fluoride concentra-tion 5 mg L−1, pH 7.03, shaking speed 150 rpm, and contact time24 h are shown in Fig. 5. The percentage removal fluoride distinctlyincreases from 19 to 81.98% with the increase in adsorbent dosefrom 0.2 to 1 g L−1 and this is due to high adsorbent/fluoride ra-tio. After a dose of 1 g L−1 there is no significant improvement
S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290 283
Fig. 2. SEM of (a) chitosan flakes, (b) TM before F− adsorption, (c) TM after F− adsorption.
in fluoride removal. However, the fluoride concentration has beenreduced to 1.27 mg L−1, which is below the permissible limit of1.5 mg L−1 [5]. This may be because at lower adsorbent dose, thedriving force (concentration gradient) responsible for adsorptionfluoride is high and hence adsorption of fluoride is high and viceversa [20].
3.5. Effect of initial concentration
Effect of initial concentration on the percentage removal offluoride was studied at different initial fluoride concentrationsby keeping other parameters constant such as adsorbent dose1 mg L−1, pH 7.03, shaking speed 150 rpm, and contact time 24 h.The effect of initial concentration on removal of fluoride is shownin Fig. 6. It was observed that with increase in initial concentrationof fluoride, the percentage removal of fluoride decreases, while ad-sorption capacity increases. This may be because the amount offluoride ions available for adsorption increases with the increasein concentration.
3.6. Effect of pH
The pH of the medium plays a significant role in fluoride ad-sorption. The effect of pH on fluoride removal by Ti macrospheres
was studied over a broad pH range of 3–11 with adsorbent dose1 g L−1, initial concentration 5.1 mg L−1, shaking speed 150 rpm,and contact time 24 h. The effect of pH on fluoride adsorptionis given in Fig. 7. Maximum fluoride adsorption was observed atneutral pH 7 while in acidic and alkaline pH fluoride removal de-creases. Under acidic conditions, the amount of fluoride adsorbedis slightly decreased and this can be attributed to the formation ofweak hydrofluoric acid. The decrease in fluoride removal under al-kaline conditions may be due to competition of excess of hydroxylions with fluoride ions.
3.7. Effect of co-ions
Fluoride-contaminated groundwater contains several other ionswhich can compete with fluoride in the adsorption process. Tostudy the effect of interfering ions, adsorption studies were car-ried out in the presence of 0.01 M sodium salt solution of Cl−,SO2
4, NO3−, CO2−3 , and HCO3−, separately. The effect of coexisting
ions on fluoride removal is shown in Fig. 8. It was observed thatbicarbonate, carbonate, chloride, sulfate, and nitrate were havingnegative effects on fluoride adsorption. In presence of bicarbonateand carbonate, there was practically no fluoride removal. This maybe due to competing effects of these anions as well as change inpH. The pHs of these solutions were 7.9, 7.0, 6.97, 8.26, and 10.9,
284 S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290
(a)
(b)
Fig. 3. XRD patterns of (a) chitosan flakes, (b) TM.
respectively, for chloride, sulfate, nitrate, carbonate, and bicarbon-ate, while the pH of the solution without addition of salts/ions was7.03. It was also confirmed from our experiment (Section 3.6) of ef-fect of pH that fluoride adsorption decreases in alkaline pH as alsoexplained.
4. Equilibrium modeling
The Langmuir and Freundlich models are most commonly usedisotherms to represent the equilibrium distribution of adsorbatefrom a liquid phase onto a solid phase. To determine the equi-
S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290 285
(a)
(b)
Fig. 4. (a) Effect of Ti loading in chitosan for fluoride adsorption. (b) Comparison ofchitosan and Ti microspheres for fluoride removal.
Fig. 5. Effect of adsorbent dose for fluoride adsorption.
librium isotherms, adsorbent doses were varied, while the initialconcentration of fluoride was kept constant. The Freundlich model,which is indicative of surface heterogeneity of the sorbent, is givenby the following linearized equation:
log(qe) = log K F + 1/n log(Ce). (2)
Fig. 6. Effect of concentration on fluoride adsorption.
Fig. 7. Effect of pH on fluoride adsorption.
Fig. 8. Effect of presence of co-ions on fluoride adsorption.
Here K F and 1/n are Freundlich constants, related to adsorptioncapacity and adsorption intensity (heterogeneity factor), respec-tively. The values of K F and 1/n were obtained from the slope andintercept of the linear Freundlich plot of log qe vs log Ce shown inFig. 12 and were found to be 1.66 mg−1 g and 0.4161, respectively,with regression coefficient (R2) of 0.95. Since the value of adsorp-tion intensity (heterogeneity factor) is less than unity, it indicatesthat the system shows favorable adsorption.
286 S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290
(a) (b)
(c)
Fig. 9. (a) Effect of time on fluoride adsorption, (b) Lagergren plot, (c) intraparticle mass transfer curve for fluoride adsorption.
The Langmuir equation, which is valid for monolayer sorptiononto a surface, is
1
qe= 1
qmax K× 1
Ce+ 1
qmax, (3)
where qmax is the maximum amount of the fluoride ion per unitweight of chitin to form a complete monolayer on the surfacebound at high Ce and K is a constant related to the affinity of thebinding sites. qe represents a particle limiting adsorption capacitywhen the surface is fully covered with solute and assists in thecomparison of adsorption performance, particularly in cases wherethe sorbent did not reach its full saturation in experiments. Thevalues of Langmuir parameters qmax and K were calculated fromthe slope and intercept of the linear plots of 1/qe vs 1/Ce shownin Fig. 13 and were found to be 7.20 mg g−1 and 0.7863 L mg−1,respectively, with regression coefficient (R2) of 0.98. The values ofLangmuir and Freundlich constants were given in Table 2. In orderto predict the adsorption efficiency of the process, the dimension-less quantity (r) was calculated by using the equation
r = 1
1 + K C0, (4)
where C0 and K are the initial concentration of fluoride andLangmuir isotherm constant. If the value of r < 1, it representsfavorable adsorption and greater than 1.0 represents unfavorableadsorption. The value of r for an initial fluoride concentration of
5 mg L−1 was found to be 0.202. It indicates that our system isfavorable for adsorption.
5. Adsorption kinetics
The kinetics of the adsorption process was studied to explainthe fluoride uptake mechanism in TM. From Fig. 9a it was observedthat fluoride uptake was very rapid in the first 30 min and thenthe rate slowed down as the equilibrium was approached. The re-sults obtained from the experiment were used to determine therate-limiting step. Kinetic models are used to examine the rate ofthe adsorption process and potential rate-controlling step. The ca-pability of the pseudo-first-order kinetic model was examined inthis study. The pseudo-first-order equation of Lagergren is gener-ally expressed as [30]
log(qe − q) = log qe − Kad
2.303t, (5)
where qe and q (both in mg g−1) are the amount of fluoride ad-sorbed per unit mass of adsorbent at equilibrium and time t , re-spectively. The adsorption rate constant (Kad) for fluoride sorptionwas calculated from the slope of the linear plot log(qe −q) vs time(t) as shown in Fig. 9b. The values of adsorption rate constants(Kad) were found to be 0.0343, 0.040, and 0.051 min−1 for initialfluoride concentrations of 5.1, 10, and 15.1 mg L−1, respectively. Inthe case of strict surface adsorption, a variation of adsorption rate
S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290 287
Scheme 1. Metal chelating with chitosan microspheres.
should be proportional to the first power of concentration [12].But when intraparticle diffusion limits the adsorption process, therelationship between the initial solute concentration and the rateof adsorption is no longer linear. So, the possibility was studiedin terms of a graphical relationship between the amount of F ad-sorbed and the square root of time as shown in Fig. 9c.
In order to test the contribution of intraparticle diffusion on theadsorption process, the rate constant for intraparticle diffusion wasobtained by using the following equation:
q = K pt0.5. (6)
For calculating the intraparticle diffusion rate constant K p (mg g−1
min−1/2), the amount of F− adsorbed per unit mass of adsorbent,q at any time t , was plotted as a function of square root of timet1/2 (Fig. 9c). The K p values were obtained from the slope of thelinear portions of the curves and were found to be 0.931, 0.679,and 0.317 mg g−1 min−1/2 for the initial F− concentration of 5.1,
10, and 15.1 mg L−1, respectively. This shows that the adsorptionis governed by intraparticle diffusion. The plot for intraparticlediffusion shows that the initially curved portion reflects film orboundary layer diffusion effects and the subsequent linear por-tion is attributed to the intraparticle diffusion effect. The linearportions of the curves do not pass through the origin (Fig. 9c),indicating that the mechanism of fluoride removal on TM is com-plex and both the surface adsorption and the intraparticle diffusioncontribute to the rate-determining step [31].
6. Mechanism of fluoride adsorption on TM
It has been reported by earlier work [11] that incorporation ofelectropositive metal enhances selectivity and adsorptive proper-ties of the material for fluoride. This concept has been used indesigning and modifying chitosan for enhancing its selectivity to-ward fluoride. Chitosan forms metal chelates or complexes with
288 S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290
Fig. 10. Comparison of fluoride adsorption in simulated and field water.
Table 1Detailed characteristics of field water before and after treatment.
Parameters Unit Value of field waterbefore treatment
Value of field waterafter treatment
Color – Colorless ColorlessOdor – Odorless OdorlessTurbidity in NTU NTU <1 0.5Total hardness as CaCO3 mg L−1 60 64Total iron as Fe mg L−1 <0.01Chloride as Cl mg L−1 292 4.86Fluoride as F mg L−1 10.2 1.06Nitrate as NO3 mg L−1 <0.01 <0.01Sulfate as SO4 mg L−1 101 96.9Alkalinity as CaCO3 mg L−1 37 96Total dissolved solids mg L−1 477 478pH – 7.4 6.8Cadmium mg L−1 <0.002 <0.002Chromium mg L−1 <0.002 <0.002Copper mg L−1 <0.01 <0.01Lead mg L−1 <0.002 <0.002Arsenic mg L−1 0.008 0.007Titanium mg L−1 <0.001 <0.001
metal ions and is well documented. The nitrogen from the aminogroup (–NH2) of chitosan acts as an electron donor and the metalions are coordinated with chitosan. The FTIR confirms the aminogroups (RNH2, RNH3), hydroxyl groups (OH–), and Ti–OH. The ca-pacity for adsorption of anions on chitosan depends on the aminogroups [32] and in turn on incorporation of metal hydroxides andoxides [11] on chitosan. Ti undergoes hydroxylation to form Ti–OH. The OH− is then replaced with fluoride ions. The Ti may beresponsible for connecting fluoride to the surface. Fluoride is ad-sorbed on TM mainly due to –NH (RNH2, RNH3) and Ti–OH. Over-all, the major mechanism of fluoride adsorption on TM is seen inScheme 1.
7. Comparison of fluoride water in simulated and field water
TM was tested in fluoride-contaminated groundwater samplescollected from Dhar, Madhya Pradesh, India, and simulated water(prepared by dissolving NaF in double-distilled water). A compar-ison of fluoride uptake in simulated and field water is given inFig. 10. The detailed characteristic parameters of field water be-fore and after treatment are given in Table 1. It was observed thatthe percentage removal is very high in simulated water as com-pared to distilled water. This may be due to two reasons: (i) thepresence of interfering ions in field water, and (ii) alkaline pH ofthe field water. Table 1 also shows that field water contains differ-ent interfering anions, which compete with fluoride in adsorption,
Fig. 11. Reusability of adsorbent for fluoride uptake.
Fig. 12. Freundlich isotherm plot for fluoride removal using Ti microspheres.
Fig. 13. Langmuir isotherm plot for fluoride removal using Ti microspheres.
and reduces the fluoride uptake capacity of TM. From the resultsof effect of pH in Section 3.6, it has been discussed that fluorideadsorption is low in alkaline pH. Therefore, fluoride removal is lowin field water as compared to simulated water.
S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290 289
Table 2Langmuir and Freundlich constants for fluoride removal using Ti microspheres.
Langmuir constant Freundlich constant
qmax (mg/g) K (ml/g) R2 K F (mg/g) 1/n R2
TM 7.21 0.79 0.98 1.66 0.416 0.95
Fig. 14. The plot of ln(qe/Ce) against Ce for thermodynamic parameter.
8. Regeneration and reuse
A fluoride adsorption experiment (initial fluoride concentration10.2 mg L−1, dose 1 g L−1, contact time 24 h) was performed andthe TM separated by filtration was dried in an oven to determineits reusability. From Fig. 11 it is seen that the used adsorbentshows a lower adsorption capacity than the fresh TM. The per-centage fluoride removal decreases from 90 to 15% in fresh use tothird reuse. It is possible to regenerate the adsorbent by an alumsolution treatment. Further studies for optimization of regenera-tion process are in progress.
9. Thermodynamic parameters
Thermodynamic parameters of adsorption standard free energychange (�G◦), standard enthalpy change (�H◦), and standard en-tropy change (�S◦) were calculated using the following equations.
Standard free energy change (�G◦) is given by the equation
�G◦ = −RT ln K0, (7)
where �G◦ is standard free energy change of sorption (kJ/mol), Tis the temperature in Kelvin, and R is universal gas constant (8.314J/mol/K) and K0 is the sorption equilibrium constant determinedfrom the slope of ln(qe/Ce) against Ce (Fig. 14) at different tem-peratures and extrapolating to zero Ce as the method suggested byKhan and Singh [33].
The standard enthalpy change (�H◦) and standard entropychange (�S◦) were calculated using the equation
ln K0 = �H◦
RT+ �S◦
RT, (8)
where (�H◦) is standard enthalpy change (kJ/mol) and (�S◦) isstandard entropy change (kJ/mol K). The values of �H◦ and �S◦were obtained from the slope and intercept of ln K0 against 1/T .The values of K0, �G◦ , �H◦ , and �S◦ are given in Table 3.
The negative value of standard free energy change and positivevalue of entropy change indicate that fluoride adsorption is a spon-
Table 3Thermodynamic parameters for fluoride adsorption by Ti microspheres.
taneous reaction. The positive value enthalpy change indicates thatthe reaction is endothermic.
10. Conclusion
TM shows excellent fluoride removal capacity, which is veryhigh compared to chitosan. The major advantage of using TM forremoval of fluoride over other adsorbents is that the TM is in theform of beads and has good stability and settling properties sothat it can be easily separated from water. The experimental dataare well fitted in Langmuir and Freundlich isotherms. From thethermodynamic study, the process of fluoride adsorption by TMis spontaneous and endothermic. The adsorption of fluoride on TMmainly depends on the pH and the existing co-ions present in thewater. In alkaline pH (above 7) fluoride uptake is very low com-pared to acidic pH. The presence of other co-anions, particularlycarbonates and bicarbonate, in water has negative effects on fluo-ride uptake.
Acknowledgments
We thankfully acknowledge the constant support and guidanceof Dr. Sukumar Devotta, Director NEERI. Financial assistance from,UNIECF Bhopal, India for this work is gratefully acknowledged.
Appendix A. Nomenclature
qe Equilibrium adsorbate capacity (mg g−1)C0 Initial concentrations of fluoride (mg L−1)Ce Equilibrium concentrations of fluoride (mg L−1)V Volume of the aqueous solution (L)W Mass of adsorbent (g)K F Freundlich constants related to adsorption capacity
Kad Rate constant (min−1)q Amount of fluoride adsorbed per unit mass of adsorbent
at time t (mg g−1)K p Intraparticle diffusion rate constant (mg g−1 min−1/2)r Dimensionless quantity
References
[1] S. Ayoob, A.K. Gupta, Crit. Rev. Environ. Sci. Technol. 36 (2006) 433–487.[2] R.C. Meenakshi, Maheshwari, J. Hazard. Mater. 137 (2006) 456–463.[3] N. Mameri, A.R. Yeddou, H. Lounici, D. Belhocine, H. Grib, B. Bariou, Water
Res. 32 (1998) 1604–1612.[4] S.M. Maliyekkal, A.K. Sharma, L. Philip, Water Res. 40 (2006) 3497–3506.[5] WHO, Guidelines for Drinking-Water Quality, vol. 1, first addendum to third
edition, 2006.[6] W.G. Nawlakhe, D.N. Kulkarni, B.N. Pathak, K.R. Bulusu, Indian J. Environ.
Health 17 (1975) 26–65.[7] H. Mjengera, G. Mkongo, in: 3rd Water Net Symposium: Water Demand Man-
agement for Sustainable Development, 2002.[8] S. Annouar, M. Mountadar, A. Soufiane, A. Elmidaoui, M.A. Menkouchi Sahli,
290 S. Jagtap et al. / Journal of Colloid and Interface Science 332 (2009) 280–290
[13] M.S. Onyango, Y. Kojima, O. Aoyi, E.C. Bernardo, H. Matsuda, J. Colloid InterfaceSci. 279 (2004) 341–350.
[14] A. Tor, Desalination 201 (2006) 267–276.[15] P.P. Coetzee, L.L. Coetzee, R. Puka, S. Mubenga, Water SA 29 (2003) 331–338.[16] S. Kagne, S. Jagtap, P. Dhawade, S.P. Kamble, S. Devotta, S.S. Rayalu, J. Hazard.
Mater. 154 (2008) 88–95.[17] R.L. Ramos, J. Ovalle-Turrubiartes, M.A. Sanchez-Castillo, Carbon 37 (1999) 609–
617.[18] S. Chidambaram, A.L. Ramanathan, S. Vasudevan, Water SA 29 (2003) 339–344.[19] W. Ma, Y. Fei-Qun, M. Han, R. Wang, J. Hazard. Mater. 143 (2007) 296–302.[20] S. Kamble, S. Jagtap, N. Labhsetwar, D. Thakre, S. Godfrey, S. Devotta, S.S. Ray-
alu, Chem. Eng. J. 129 (2007) 173–180.[21] Li Ngee Ho, T. Ishihara, S. Ueshima, H. Nishiguchi, Y. Takita, J. Colloid Interface
Sci. 272 (2004) 399–403.[22] E.P. Lokshin, M.L. Belikov, Russian J. Appl. Chem. 76 (9) (2003) 1466–1471.[23] W.S. Wan, S. Ngah, A. Ghani, A. Kamari, Bioresour. Technol. 96 (2005) 443–450.
[24] G. Rojas, J. Silva, J.A. Flores, A. Rodriguez, M. Ly, H. Maldonado, Sep. Purif. Tech-nol. 44 (2005) 31–36.
[25] G.L. Rorrer, T.Y. Hsien, J.D. Way, Ind. Eng. Chem. Res. 32 (1993) 2170–2178.[26] L. Nan, R. Bai, Sep. Purif. Technol. 42 (2005) 237–247.[27] L. Dambies, E. Guibal, A. Roze, Colloids Surf. A 170 (2000) 19–31.[28] APHA, AWWA, WPCF, Standard Methods for the Examination of Water and
Wastewater, Am. Public Health, 1989.[29] A.T. Pauline, J.I. Simionato, J.C. Garcia, J. Nozaki, Carbohydr. Polym. 64 (2006)
98–103.[30] D.P. Das, J. Das, K. Parida, J. Colloid Interface Sci. 261 (2003) 213–220.[31] A.K. Yadav, C.P. Kaushik, A.K. Haritash, A. Kansal, N. Rani, J. Hazard. Mater. 128
(2005) 289–293.[32] I.M.N. Vold, K.M.V. Arum, E. Guibal, O. Smidsr, Carbohydr. Polym. 54 (2003)
471–477.[33] A.A. Khan, R.P. Singh, Colloids Surf. 24 (1987) 33–42.
