Journal of Colloid and Interface Science 298 (2006) 15 www.elsevier.

com/locate/jcis

Temperature effect on the zeta potential and uoride adsorption at the -Al2 O3/aqueous solution interface
A. Lpez Valdivieso a, , J.L. Reyes Bahena a , S. Song a , R. Herrera Urbina b
a Area de Ingeniera de Minerales, Instituto de Metalurgia, Universidad Autnoma de San Luis Potos, Av. Sierra Leona 550, Lomas 2a Seccin,

San Luis Potos, S.L.P., 78210, Mexico

b Departamento de Ingeniera Qumica y Metalurgia, Universidad de Sonora, Apartado Postal 106, Hermosillo, Sonora 83000, Mexico

Received 18 June 2005; accepted 28 November 2005 Available online 27 December 2005

Abstract The effect of temperature and pH on the zeta potential of -Al2 O3 and adsorption of uoride ions at the -Al2 O3 /aqueous solution interface has been investigated through electrophoretic mobility measurements and adsorption studies, to delineate mechanisms involved in the removal of uoride ions from water using alumina as adsorbent. When the temperature increases from 10 to 40 C, the pH of the point of zero charge (pHpzc ) shifts to smaller values, indicating proton desorption from the alumina surface. The pHpzc increases linearly with 1/T , which allowed estimation of the standard enthalpy change for the surface-deprotonation process. Fluoride ion adsorption follows a Langmuir-type adsorption isotherm and is affected by the electric charge at the -Al2 O3 /aqueous solution interface and the surface density of hydroxyl groups. Such adsorption occurs through an exchange between uoride ions and surface-hydroxyl groups and it depends on temperature, pH, and initial uoride ion concentration. At 25 and 40 C, maximum uoride adsorption density takes place between pH 5 and 6. Increasing the temperature from 25 to 40 C lowers the adsorption density of uoride. 2005 Elsevier Inc. All rights reserved.
Keywords: Fluoride removal; Alumina; Electrokinetics; Adsorption; Water treatment; Fluorosis

1. Introduction Fluoride is potentially toxic for humans. Therefore, its ingestion in food or drinking water must not exceed a narrow range of concentrations. Maximum allowed concentration of uoride in drinking water, a major source of uoride intake, depends on average daily temperature. The World Health Organization has suggested a value below 1 mg/L for areas with warm climates [1]. High concentrations of uoride occurring naturally in groundwater have caused widespread uorosis in many parts of the world. In Mexico, for example, 5 million people (about 6% of the population) are affected by uoride in drinking water [2]. Although uoride concentrations in groundwater are generally smaller than 2 mg/L, under certain geological conditions they can reach levels from 5 to 15 mg/L [3]. In addition to natural sources, industrial efuents can contribute to high lev* Corresponding author. Fax: +52 (444) 825 4326.

E-mail address: alopez@uaslp.mx (A. Lpez Valdivieso). 0021-9797/$ see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2005.11.060

els of uoride in groundwater and increase the risks for human health and the environment [4]. Adsorption from solution is widely used for uoride removal from drinking water supplies and wastewaters, with activated alumina being a common adsorbent [57]. Other materials that have also been used as adsorbents include limestone [3], ionexchange resin [8], and alum sludge [9]. The parameters that inuence the most uoride adsorption are the pH, uoride concentration, and temperature. The optimum pH for maximum uoride adsorption on -Al2 O3 occurs at about pH 5 [10], and at pH 6 for alum sludge [9]. Temperature effects on uoride adsorption on alum sludge have been reported by Sujana et al. [9], who found a decrease in adsorption with increasing temperature. Fluoride adsorption, however, has been reported to increase the dissolution rates of bayerite (-Al(OH)3 ) and boehmite ( AlOOH) at pH 5 [11] and to promote the dissolution of -Al2 O3 at pH 5.5 [12]. Fluoride-promoted dissolution of aluminum oxides could have a deleterious effect on the removal of uoride

A. Lpez Valdivieso et al. / Journal of Colloid and Interface Science 298 (2006) 15

from solution. Because of the solubility of the adsorbent increases with temperature, higher temperatures are expected to inuence uoride adsorption onto alumina. Temperature also affects the surface charge on -alumina and shifts the pHpzc to lower values as the temperature increases [13]. In addition, small changes in temperature have a strong inuence on the stability of surface complexes. This work reports the results of an investigation on the adsorption of uoride onto -Al2 O3 and the zeta potential at the -Al2 O3 /aqueous uoride interface at 25 and 40 C as a function of pH. In addition, the zeta potential of -alumina was determined at 10, 20, 30, and 40 C in 0.01 mol/L NaNO3 as a function of pH, and adsorption isotherms at 25 and 40 C were determined for uoride ions at pH 5 and 9. 2. Experimental materials and methods The -Al2 O3 powder for adsorption and electrokinetic experiments was obtained from Taimei Chemicals Co. Ltd and used as received. This alumina has a purity of 99.99%, a specic surface area of 17.4 m2 /g as determined by the BET method using nitrogen as the adsorbent, and a mean particle size of 150 nm. All chemicals used in this study were analytical grade reagents. Sodium nitrate was used as a supporting electrolyte at a concentration of 0.01 mol/L to maintain constant the ionic strength and so the electrical double layer thickness in electrokinetic studies. Dilute solutions of NaOH and HNO3 were used for pH adjustments. Sodium uoride was used as uoride source. All solutions and suspensions were prepared with deionized water (resistivity 17.4 M /cm at 25 C) obtained with a Barnstead E-pure apparatus. Alumina suspensions were prepared in polypropylene bottles. Suspensions for electrophoresis measurements were prepared as follows: a measured amount (50 mg) of -Al2 O3 was added to 100 ml of 0.01 mol/L aqueous sodium nitrate, and the suspension was aged for 24 h at its natural pH and 20 C. The same aging procedure was used for the electrokinetic studies in the presence of uoride ions. After the aging period, the suspension was transferred to a water-jacketed stirred vessel with ports for a thermometer, a pH electrode, a Teon stirrer, and a temperaturepH sensor. The temperature of the suspension was adjusted to the appropriate value (10, 20, 30, or 40 C) by pumping water from a temperature-controlled water bath through the jacket of the vessel. Once the desired temperature was reached, the pH was adjusted to the desired value. The suspension was then conditioned for 30 min, after which the pH was recorded and reported as nal. After conditioning, the suspension was quickly transferred to the electrophoresis cell of a unit manufactured by Zeta Meter, Inc., New York. This company supplied the reservoir into which the electrophoresis cell was placed for mobility measurements at the various temperatures considered in this study. Water at a constant and the desired temperature was pumped through the reservoir. This water pumping was stop before carrying out the mobility measurements because it induced particle movement in the electrophoresis cell with no applied voltage to the system. At least 10 different particles were tracked to deter-

mine the average electrophoretic mobility, which was converted to zeta potential using the Henry equation [14], = 6EM/f (a), (1)

where is the zeta potential, the water viscosity at the test temperature, EM the electrophoretic mobility, the water dielectric constant at the test temperature, the electrical double layer thickness, and a the particle radius, which was taken as 75 nm. The f (a) value was determined at the 0.01 mol/L ionic strength used in this study as described by Hunter [14]. Adsorption experiments were conducted as follows: 500 mg alumina was added to 100 ml aqueous solution of 0.01 mol/L NaNO3 containing 3 or 10 mg/L total uoride. These concentrations of uoride in drinking water cause uorosis in teeth and bones [24]. The alumina suspension was aged for 24 h at 25 or 40 C. After aging, the pH was adjusted to the desired value and the suspension was conditioned further for 30 min. The pH of this suspension was then measured and reported as nal. An aliquot was taken and centrifuged at 5000 rpm for 30 min. The solution recovered after centrifugation was used for determining the concentration of total uoride remaining in the solution. The total uoride concentration in the solution was measured with an ion-selective electrode connected to an Orion Specic Ion Meter Model SA-720 and coupled to a standard calomel electrode as reference. The range of uoride concentrations that can be analyzed by this technique varied from 0.1 to 1000 mg/L. For avoiding interference with the electrode performance, an ionic strength xer and buffer solution [TISAB: 1 M NaCl + 1 M acetic acid + 1 M CDTA (1,2-cyclohexylenediamine tetraacetic acid)], set to a pH in the range 5.05.3 with 6 N NaOH, was added during actual measurements [15]. A given volume of TISAB solution was added to a volume of uoride-containing supernatant and allowed to stand for 2 h at 25 C before the electromotive force of the electrode was read. 3. Results and discussion 3.1. Effect of temperature on pHpzc The zeta potential of -Al2 O3 in 0.01 M aqueous NaNO3 solutions at 10, 20, 30, and 40 C was determined as a function of pH, and the results are presented in Fig. 1. This gure shows a series of curves that shift to the left as the temperature increases. The pH of zeta potential reversal (pHpzc ) decreases from 9.6 to 8.1 when the temperature increases from 10 to 40 C. A similar temperature effect on the electrokinetic behavior of Al2 O3 has been reported by Fuerstenau and Raghavan [16]. pHpzc values of Al2 O3 from potentiometric titration data at different temperatures also decrease with increasing temperature [13,17,18]. These results indicate that increasing temperature favors proton desorption or hydroxide adsorption taking into account the following surface equilibria: AlOH+ (surf) = AlOH(surf) + H+ , 2 AlOH(surf) + OH = AlO (surf) + H2 O. (2) (3)

A. Lpez Valdivieso et al. / Journal of Colloid and Interface Science 298 (2006) 15

Fig. 1. The zeta potential of -Al2 O3 as a function of pH at various temperatures using 0.01 M NaNO3 as the supporting electrolyte.

Fig. 2. The zeta potential of -Al2 O3 as a function of pH at 25 C in the absence and presence of 3 and 10 mg/L initial concentration of uoride ions using 0.01 M NaNO3 as the supporting electrolyte.

Halter [13] correlated the shift in the pHpzc with the water dissociation constant as a function of temperature. For a given pH value, an increase in temperature decreases the zeta potential, also indicating proton desorption from the surface. Another aspect that should be considered in the interpretation of zeta potential results is the effect of temperature and pH on the solubility of -Al2 O3 . Using the appropriate thermodynamic data, Reyes Bahena et al. [10] have calculated that the minimum solubility of this alumina occurs at about pH 7 at 25 C. It is well known that metal oxide solubility increases with temperature. Berube and de Bruyn [19] have derived a thermodynamic relation for the double layer on oxides surfaces in which the pHpzc depends on temperature as follows, 1 4.6R pKw pHpzc = 2 H S, T (4)

Fig. 3. The zeta potential of -Al2 O3 as a function of pH at 40 C in the absence and presence of 3 and 10 mg/L initial concentration of uoride ions using 0.01 M NaNO3 as the supporting electrolyte.

where H and S are the standard heat and the standard ionic entropy of transferring the potential-determining ions H+ and OH from the bulk solution to the interfacial region at the pzc, pKw is the negative log of the dissociation constant for water, and pHpzc is the point of zero charge of the oxide. They have shown that a plot of [ 1 pKw pHpzc ] against reciprocal temper2 ature was linear for TiO2 . From the slope and intercept of the line H and S values were estimated. The results of this work for -Al2 O3 agree with the thermodynamic relation derived by Berube and de Bruyn [19]. A value of 46.4 kJ/mol and 80 J/mol K was obtained for H and S , respectively. Tombcz et al. [20] have reported a H 0 of 68 kJ/mol for -alumina, Halter [13] 18.36 kJ/mol for -alumina, and Kallay et al. [21] 50 kJ/mol for -Al2 O3 . 3.2. Effect of pH, temperature, and F added on zeta potential The zeta potential of -Al2 O3 in the absence and presence of 3 and 10 mg/L F was determined at 25 and 40 C as a function of pH. Fig. 2 shows that at 25 C uoride ions affect the zeta potential, which is shifted more to the left than at 40 C (see Fig. 3). The adsorption of uoride ions has been proposed to be specic through exchange of surface OH groups [10,11].

Then the decrease in positive zeta potentials can be accounted for by the following surface-exchange reactions: AlOH+ (surf) + F = AlF(surf) + H2 O, 2 AlOH(surf) + F = AlF(surf) + OH . F (5) (6)

These surface-exchange reactions suggest that all charges are located at the surface [22], which is in agreement with the zeta potentials measured at the -Al2 O3 /aqueous uoride interface in this investigation. At 40 C, no signicant changes in the zeta potential were found when 3 mg/L F were added owing to increased proton desorption making the surface of alumina more negatively charged and enhancing the electrostatic repulsion of F . A dose of 10 mg/L F , however, shifts the pHiep toward a smaller value and makes the zeta potentials at all pH values investigated more negative. This electrokinetic behavior indicates that F chemisorbs onto -Al2 O3 as adsorption takes place even though the surface is electrically negative. 3.3. Effect of pH and temperature on uoride adsorption The adsorption density of F on -Al2 O3 at 25 and 40 C with aqueous solutions containing 3 and 10 mg/L initial u-

A. Lpez Valdivieso et al. / Journal of Colloid and Interface Science 298 (2006) 15

Fig. 4. Adsorption of uoride on -Al2 O3 as a function of pH at 25 and 40 C, 3 and 10 mg/L initial concentration of uoride ions.

Fig. 5. Adsorption isotherms of uoride ions on -Al2 O3 as a function of equilibrium uoride ion concentration at 25 and 40 C at pH 5.0 and 9.0.

oride is presented in Fig. 4 as a function of pH. At the two temperatures, uoride adsorption depends on pH and shows maxima between pH 5 and 6 in agreement with the work of Clifford et al. [23] and Churchill [24]. A decrease in uoride adsorption below pH 5 can be associated with uoride-promoted dissolution of alumina [11,12]. Reyes Bahena et al. [10] have constructed the solubilitypH diagram for -Al2 O3 and reported that the solubility of alumina increases signicantly as the pH decreases below about pH 5. On the other hand, uoride has been found to enhance the solubility of aluminum hydro(oxides) at low pH [11,12], which is due to the formation of aluminumuoro complexes species in solution. Below pH 5, the aluminumuoro complexes AlF+ , AlF2+ , AlF3 (aq), and 2 AlF are predominant in solution, as shown by the aluminum 4 uoride species distributionpH diagram reported by Reyes Bahena et al. [10]. Above pH 6, uoride adsorption density falls sharply. Under these conditions, the exchange reactions (5) and (6) between surface hydroxide groups and the adsorbing uoride ion are less favorable. This is because the concentration of aqueous hydroxide ions in solution rises, which leads to a decrease of the surface density of AlOH+ sites and an increase of that of AlO 2 sites. In addition, increasing the pH favors the electrostatic repulsion between the negatively charged surface of alumina and the anionic uoride. At the optimum pH for uoride removal from solution, the adsorption density decreases as the temperature increases from 25 to 40 C. This is typical of an exothermic adsorption process. The same temperature effect has been reported when alum sludge is used as adsorbent for uoride [9]. In addition, since temperature affects the surface charge, it also affects the electrostatic contribution to the free energy of adsorption. The adsorption density of uoride on -Al2 O3 at 25 and 40 C was determined at pH 5 and 9 and is plotted as a function of equilibrium uoride concentration in Fig. 5. It was found that the experimental data ts well the Langmuir equation, supporting the mechanism of ion exchange between surface OH groups and uoride ions. As the equilibrium concentration of

uoride in solution increases, the adsorption density increases and tends to reach a plateau. The calculated values of the adsorption density at the plateaus have been found to be 8.26 and 6.62 mol F /m2 at pH 5 for 25 and 40 C, respectively. At pH 9, these values are 2.13 and 1.91 mol F /m2 for 25 and 40 C, respectively. The OH surface density on alumina has been reported to be about 8 OH/nm2 or 1.28 C/m2 in charge units [25], equivalent to 13 mol OH/m2 . Comparing this value with those of uoride adsorption density, it can be noted that not all the surface OH groups are exchanged by uoride at pH 5 and only a low percentage has been replaced at pH 9. It is then evident that full surface coverage is not reached because of the electrostatic contribution to the adsorption process. At pH 5 and 25 C, the surface charge density of alumina is on the order of 0.06 C/m2 for an indifferent electrolyte concentration of 0.01 mol/L [25,26]. The uoride adsorption density of 8.26 mol/m2 at pH 5 corresponds to 0.8 C/m2 in charge units, which is an order of magnitude greater than the surface charge density. Accordingly, uoride ions not only adsorbed on positively charged AlOH+ sites but also on the electrically un2 charged AlOH sites, as shown by Eqs. (5) and (6). At pH 9, about the pHpzc , the surface density of the AlOH+ sites is small. 2 This means that low adsorption of uoride ions is needed to negatively charge the alumina surface, which leads to electrostatic repulsion of the uoride anions. 4. Conclusions The surface charge at the -Al2 O3 /aqueous solution interface is affected by the temperature so that the point of zero charge decreases as the temperature increases. The pzc temperature dependence allowed the standard enthalpy of surface charge formation to be determined. As the temperature and pH increase, the deprotonation of the alumina surface increases, as well as the stability of the surface OH groups, leading to a decrease in uoride adsorption. The interaction of uoride with alumina is the result of a chemical afnity for the surface AlOH+ and AlOH sites and electrostatic attraction. Fluoride is 2 more efciently removed from solution between pH 5 and 6.

A. Lpez Valdivieso et al. / Journal of Colloid and Interface Science 298 (2006) 15

Acknowledgments The authors thank Aurora Robledo Cabrera for her assistance in the preparation of the manuscript. Juan Luis Reyes Bahena thanks the Consejo Nacional de Ciencia y Tecnologia (CONACyT) for scholarship Grant 66764. This work has been carried out by the partial nancial support of SEMARNATCONACyT through Contract 2002-C01-0362. References
[1] World Health Organization, Environmental Health Criteria 36, International Programme on Chemical Safety, Fluorine and Fluorides, Geneva, 1984. [2] F. Diaz Barriga, R. Leyva Ramos, J. Quistian, J.B. Loyola Rodrguez, A. Pozos, M. Grimaldo, Fluoride 30 (4) (1997) 219222. [3] E.J. Reardon, Y. Wang, Environ. Sci. Technol. 34 (2000) 32473253. [4] Y.X. Chen, M.Q. Lin, Z.L. He, Y.D. Diao, C. Chen, D. Min, Y.Q. Liu, M.H. Yu, Fluoride 30 (1996) 712. [5] R. Leyva Ramos, A. Jurez Martnez, Avances Ing. Qum. (1991) 107 111. [6] Y.C. Wu, A. Nitya, J. Environ. Eng. Div. 105 (1979) 357367. [7] J.J. Schoeman, G.R. Botha, Water S A 11 (1985) 2532. [8] M. Haron, W.W. Yunus, J. Environ. Sci. Health Part A Toxic/Hazard. Subst. Environ. Eng. 36 (2001). [9] M.G. Sujana, R.S. Thakur, S.B. Rao, J. Colloid Interface Sci. 206 (1) (1998) 94101.

[10] J.L. Reyes Bahena, A. Robledo Cabrera, A. Lpez Valdivieso, R. Herrera Urbina, Sep. Sci. Technol. 37 (2002) 19731987. [11] J.P. Nordin, D.J. Sullivan, B.L. Phillips, W.H. Casey, Geochim. Cosmochim. Acta 63 (21) (1999) 35133524. [12] S.M. Kraemer, V.Q. Chiu, J.G. Hering, Environ. Sci. Technol. 32 (1998) 28762882. [13] W.E. Halter, Geochim. Cosmochim. Acta 63 (19/20) (1999) 30773085. [14] R.J. Hunter, Zeta Potential in Colloid Science, Academic Press, New York, 1981, p. 386. [15] Anon, Standard Methods for the Examination of Water and Wastewater, Fluoride, American Public Health Association, 1989, pp. 4.844.93. [16] D.W. Fuerstenau, S. Raghavan, in: M.C. Fuerstenau (Ed.), Flotation A.M. Gaudin Memorial, vol. 1, SMEAIME, New York, 1976, pp. 2165. [17] P.H. Tewari, A.W. McLean, J. Colloid Interface Sci. 40 (1972) 267272. [18] S. Mustafa, B. Dilara, Z. Neelofer, A. Naeem, S. Tasleem, J. Colloid Interface Sci. 204 (1998) 284293. [19] Y.G. Berube, P.L. de Bruyn, J. Colloid Interface Sci. 27 (1968) 305318. [20] E. Tombcz, M. Szekeres, E. Klumpp, Langmuir 17 (2001) 14201425. [21] N. Kallay, S. Zalac, T. Preocanin, W. Arudzinski, in: A.T. Hubbard (Ed.), Encyclopedia of Surface and Colloid Science, Dekker, New York, 2002, pp. 21292139. [22] T. Hiemstra, W.H. van Riemsdijk, J. Colloid Interface Sci. 225 (2000) 94 104. [23] D. Clifford, J. Matson, R. Kennedy, Ind. Water Eng. 15 (1978) 612. [24] H.V. Churchill, Draft Analytical ReportNew Orleans Area Water Supply StudyUSEPA 906/10-74-002 Lower Mississippi River facility, Slidell; La. Surveillance and Analysis Div. Region 6: Dallas, TX, 1974. [25] Z. Todorovic, S.K. Milojic, J. Serb. Chem. Soc. 69 (2004) 10631072. [26] M. Kosmulski, in: A.T. Hubbard (Ed.), Encyclopedia of Surface and Colloid Science, Dekker, New York, 2002, pp. 16271636.