Optimization of fluoride removal from aqueous solution by Al[2]O[3] nanoparticles

2017

Optimization of fluoride removal from

aqueous solution by Al[2]O[3] nanoparticles

Hafshejani Laleh Divband

Elsevier BV

info:eu-repo/semantics/article

info:eu-repo/semantics/acceptedVersion

© Elsevier B.V

CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/

http://dx.doi.org/10.1016/j.molliq.2017.04.104

https://erepo.uef.fi/handle/123456789/4288

Downloaded from University of Eastern Finland's eRepository

Optimization of fluoride removal from aqueous solution by Al2O3 nanoparticles

Laleh Divband Hafshejani, Sareh Tangsir, Ehsan Daneshvar, Marja Maljanen, Anna Lähde, Jorma Jokiniemi, Mu. Naushad, Amit Bhatnagar

PII: S0167-7322(17)30915-7

DOI: doi:10.1016/j.molliq.2017.04.104

Reference: MOLLIQ 7260

To appear in: Journal of Molecular Liquids Received date: 1 March 2017

Revised date: 11 April 2017 Accepted date: 21 April 2017

Please cite this article as: Laleh Divband Hafshejani, Sareh Tangsir, Ehsan Daneshvar, Marja Maljanen, Anna Lähde, Jorma Jokiniemi, Mu. Naushad, Amit Bhatnagar , Optimization of fluoride removal from aqueous solution by Al2O3 nanoparticles. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Molliq(2017), doi:10.1016/j.molliq.2017.04.104

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

ACCEPTED MANUSCRIPT

Page 1 of 39

Optimization of fluoride removal from aqueous solution by Al

O

nanoparticles

Laleh Divband Hafshejani ^a,b, Sareh Tangsir ^a,b, Ehsan Daneshvar ^a, Marja Maljanen ^a, Anna Lähde ^a, Jorma Jokiniemi ^a, Mu. Naushad ^c, Amit Bhatnagar ^a,*

a Department of Environmental and Biological Sciences, University of Eastern Finland, P.O. Box 1627, FI-70211 Kuopio, Finland

b Department of Irrigation and Drainage Engineering, Faculty of Water Sciences Engineering, Shahid Chamran University of Ahvaz, Khuzestan, Iran

c Department of Chemistry, College of Science, Bld#5, King Saud University, Riyadh, Saudi Arabia

*Corresponding author: Phone: +358 503696419 email: amit.bhatnagar@uef.fi;

dr.amit10@gmail.com

ACCEPTED MANUSCRIPT

Page 2 of 39 Abstract

According to World Health Organization (WHO), fluoride has a narrow prescribed concentration level in drinking water (less than 1.5 mg L^-1) and defluoridation of water is necessary to remove elevated concentrations of fluoride from water. In the present work, aluminium acetylacetonate was used as a precursor which was dissolved in methanol to produce Al₂O₃ nanoparticles by flame spray pyrolysis (FSP) technique. The Al2O3 nanoparticles were characterized by various techniques (e.g. Brunauer-Emmet-Teller (BET), X-ray diffraction (XRD), scanning electron microscope (SEM), transmission electron microscopy (TEM) and Fourier Transform Infrared (FTIR) to get an insight of their physicochemical properties. The effects of different variable parameters such as solution pH (3-11), contact time (5–180 min), initial fluoride concentration (0-30 mg L⁻¹), solution temperature (25, 35 and 45 °C) and competing anions (chloride, nitrate, carbonate, sulfate and phosphate) were investigated to study the adsorption of fluoride from water. Among different kinetic and isotherm models studied, the pseudo-second order model best described the kinetics, while equilibrium data were well fitted by the Langmuir isotherm model.

The competing anions on defluoridation from water followed the order:

phosphate>carbonate>sulfate>nitrate>chloride. Results from this study revealed the potential utility of nano-alumina particles for defluoridation of water.

Keywords: Fluoride; Al2O3 nanoparticles; Aluminium acetylacetonate; Kinetics; Isotherms.

ACCEPTED MANUSCRIPT

Page 3 of 39 1. Introduction

Halogen elements like fluorine, chlorine, bromine, and iodine in water are one of the main environmental concerns because of their toxic nature [1]. In the Periodic Table, fluorine with atomic number nine is the most electronegative and reactive chemical elements. There is no fluorine in the nature due to its high reactivity, and it can be found as inorganic (the free anion F^- ) and organic (freons) fluorides. The abundance of inorganic fluoride is more than the organic fluoride and widely distributed in the global environment [2]. The fluorapatite (Ca₅(PO₄)₃F), fluorite (CaF2) and cryolite (Na3ALF6) are the main sources of inorganic fluoride [3]. Fluoride concentrations are generally proportional to the degree of water–rock interaction because fluoride primarily originates from the geology [4-8]. Fluoride is commonly dispersed in the geological environment [9] and largely released into the groundwater by sluggish dissolution of fluorine-containing rocks [4]. Minerals such as fluorite, biotites, topaz, and their corresponding host rocks e.g., granite, basalt, syenite, and shale release the fluoride in the groundwater [10].

Besides, effluents of various industries such as glass and ceramic production, semiconductor manufacturing, electroplating, coal fired power stations, beryllium extraction plants, brick and iron works, and aluminum smelters also contain high fluoride concentration [11].

Drinking water containing fluoride has both beneficial and detrimental effects on human health [12]. The beneficial effect on the calcification of dental enamel and maintenance of healthy bones considered in the acceptable range of fluoride concentration (1-1.5 mg L^-1) [13]. On the other hand, consumption of water containing fluoride concentration above the World Health Organization (WHO) guideline value (1.5 mg L^-1) leads to various health effects viz.

ACCEPTED MANUSCRIPT

Page 4 of 39

osteoporosis, arthritis, skeletal fluorosis and cancer [14]. Besides dental and skeletal problems, many other health problems such as disorder in calcium and phosphate metabolism in human body, dental caries and lesion of endocrine glands occur when the concentration of this anion is higher than the permitted limits [15].

Hence, the removal of fluoride from water and wastewaters is important before these are mixed up with clean natural water resources.

Conventional methods such as, nanofiltration, electrocoagulation, ion exchange, reverse osmosis, electrodialysis, precipitation and membrane separation have been applied for fluoride removal from water and wastewaters [16]. However, major disadvantages of these methods include high operational cost, energy-intensive, limited and less adaptability to a wide range of fluoride concentrations [10].

Among the various techniques of fluoride removal from wastewater, it is widely recognized that adsorption is one of the most appropriate and attractive for the purpose. Adsorption process is relatively simple, fast, inexpensive and appropriate for treating water over a wide pH range and low-concentration effluents [17]. Various materials including biomass, chitosan, metal oxides, activated red mud, aligned carbon nanotubes, iron and alumina-based adsorbents have been used to remove the fluoride ions from water [18]. Alumina-based adsorbents such as activated alumina, alumina impregnated carbon, dispersed alumina in charcoal and alumina–chitosan composite are known as efficient adsorbents for defluoridation of drinking water.

In the past, aluminium based micron-sized adsorbents have been used for the removal of fluoride, however, in recent years, it has been demonstrated by various researchers that metal oxide nanoparticles have shown enhanced efficiency for fluoride removal from aqueous solutions [19]. Nanomaterials as separation media for water purification are preferred over

ACCEPTED MANUSCRIPT

Page 5 of 39

conventional materials as a result of their several key physico-chemical properties such as larger surface area than bulk particles, enhanced reactivity, potential for self-assembly, high specificity, stability and conductivity [20]. Single-step flame spray pyrolysis (FSP) method has been known as best technique for producing nanoparticles such as Al₂O₃ due to their several advantages such as, cost-effectiveness and easy process-ability [21]. In addition, a broad range of precursors (e.g., metal organics, organic salts) can be used in the FSP synthesis. However, the properties of the selected precursors and solvents can have a significant effect on the properties of the produced particles [22]. For example, the solution viscosity, vapor pressure, reaction enthalpies are important parameters in the gas-to-particle conversion during the synthesis that determine the properties of the final particles [22].

In our previous study, aluminium acetylacetonate (used as a precursor for Al2O3 nanoparticles) was dissolved in “toluene” during FSP [23]. However, benzene ring in toluene can cause soot formation during the FSP synthesis [24]. This, in turn, can reduce the flame temperature and decrease the reaction rates. In addition, the soot can condensate on the particle surface altering the surface properties of the produced particles.

In the present study, the same precursor (aluminium acetylacetonate) was used, but it was dissolved in “methanol” this time to produce Al2O₃ nanoparticles by FSP method. The main aim of this study was to see the influence of dissolving medium (solvent) on the properties of Al2O3

nanoparticles by FSP method. Defluoridation experiments were performed in batch mode under different experimental conditions, such as solution pH, contact time, temperature, initial fluoride concentration and most common competing anions. The equilibrium, kinetics, isotherms and thermodynamics of defluoridation in batch system were investigated.

ACCEPTED MANUSCRIPT

Page 6 of 39 2. Materials and methods

2.1. Materials

Aluminium acetylacetonate (ReagentPlus®, 99%, Sigma-Aldrich, USA) was used as a precursor for Al₂O₃ particles. Sodium fluoride (NaF) was purchased from Sigma-ALDRICH. Stock solution of fluoride (100 mg L^-1) was prepared by dissolving required amount of NaF in 1 L of deionized water. Required concentration range was further prepared by appropriate dilution of the stock solution with deionized water. All other chemical used in this study were of analytical grade.

2.2. Synthesis of Al2O3 nanoparticles

Al2O3 nanoparticles were synthesized with flame spray pyrolysis (FSP) method. Aluminium acetylacetonate was dissolved in methanol (J.T. Baker, Baker analyzed, exceeds ACS specifications) resulting in the total concentration of 0.2 M at room temperature. The details of FSP system used in the study are described elsewhere [25]. The precursor solution feed rate was set to 5 mL min^-1. The solution was atomized with a high-pressure dispersion gas (O₂, AGA, purity ≥ 99.5%) with the flow rate of 5 L min^-1. A pre-mixed methane-oxygen flamelet with the methane (AGA, purity ≥ 99.5 %) and oxygen (AGA, purity ≥ 99.5%) flow rates of 1 L min^-1 and 2 L min^-1, respectively, were used to ignite the aerosolized precursor. In addition, a sheath air flow (filtered air) of 10 L min^-1 was used in all experiments. The produced particles were collected on the filter (Zefluor filters, 90 mm in diameter, 1.0 μm and 50/PK, Pall Corporation).

2.3. Characterization of Al2O3 nanoparticles

The crystalline phase of the produced Al2O3 nanoparticles were analyzed by XRD (Bruker D8 Discover with Cu α radiation at 40 kV and 40 mA), at 2θ between 10 and 40º (2θ resolution of 0.03º and integration time of 384 sec.). The data was analyzed with Topas 3 software including

ACCEPTED MANUSCRIPT

Page 7 of 39

Rietveld refinement that was used for the full spectrum fitting to determine the phase composition analysis. The crystalline phases were identified based on the reference data from International Center Diffraction Data (ICDD). The chemical composition of the particles was analyzed with the Fourier Transform Infrared (FTIR, ThermoNicolet 8700) spectroscopy.

Brunauer-Emmet-Teller (BET) analysis were carried out with TriStar II 3020 surface area and porosity analyzer with the nitrogen adsorption isotherms. The powder samples were dried and degassed at 120 ºC under vacuum at <1×10^-5 bar for 2 h prior the analysis. The zeta potential and hydrodynamic size of the particles was measured from the aqueous suspensions (Malvern Zeta Sizer, Nano series). Ion-exchanged water was used in the suspension. All measurements were done in triplicates and the average value was calculated based on the measurements. The shape of the particles was studied with a field-emission low-voltage electron microscope (SEM, Zeiss Sigma HDVP) operated at a 3 kV acceleration voltage. The morphology of the particles was analyzed with a transmission electron microscopy (TEM, Philips CM-200 FEG/STEM) operated at 200 kV. The particles samples were placed on the holey carbon copper grid (S147-400 Holey Carbon Film 400 Mesh Cu grid, Agar Scientific) using an ethanol suspension of the particles.

The primary particle size was analyzed from TEM images by measuring particle population of around 500 particles.

2.4. Fluoride analysis

The concentration of fluoride in the aqueous solutions was determined by ion chromatography (Dionex, DX-120, Ion Chromatography system, USA). The mobile phase consisted of a mixture of 4.5 mM sodium carbonate (Na2CO3) and 0.8 mM sodium bicarbonate (NaHCO3) delivered at the flow rate of 1.0 mL min⁻¹. AS40 autosampler (Dionex, USA) was assembled with a 1.7 mL injection loop. A separation column, IonPac® AS 23, 4.0 mm × 250 mm (Thermo, USA), a

ACCEPTED MANUSCRIPT

Page 8 of 39

guard column, IonPac® AG 23, 4.0 mm × 50 mm (Thermo, USA), and membrane suppressor, ASRS 300, 4-mm (Thermo, USA) were used. The data acquisition was performed using a PeakNet 5.11 software (Dionex, USA).

2.5. Fluoride adsorption studies

Batch adsorption experiments for fluoride removal by Al2O3 nanoparticles were conducted at room temperature (25 ± 2 ^oC) in 50 mL capped plastic tubes containing 20 mL of fluoride solution with different initial concentrations in the range of 1–50 mg L⁻¹. A known amount of adsorbent (synthesized Al2O3 nanoparticles) was added to each tube and tubes were shaken in a shaker at room temperature with a speed of 150 rpm for a specified period of contact time to reach the equilibrium. Then samples were filtered and the residual fluoride concentration in filtrates was determined by ion chromatography. Reproducibility of the measurements was determined in duplicates and the average values are reported. Relative standard deviations of between measurements were found to be within 3.0%.

The amount of fluoride adsorbed at equilibrium time, q_e (mg g^-1), was calculated using Eq. (1).

q

=

m (1) where, C0 is the initial concentration (mg L^-1) of fluoride; Ce is the concentration (mg L^-1) of fluoride at equilibrium; V is the volume of the solution (L) and m is the mass of the adsorbent (g). The removal percentage (R%) of fluoride was calculated using Eq. (2).

R% =^(C0_C^−Ce)

0 × 100 (2) 2.5.1. Influence of solution pH, contact time, initial fluoride concentration, competing ions and temperature

The effect of solution pH on fluoride adsorption was studied by adjusting solutions pH within the range of 3-11 by using 0.1 M HCl and 0.1 M NaOH under an initial fluoride concentration of 10

ACCEPTED MANUSCRIPT

Page 9 of 39

mg L^-1, contact time of 24 h, adsorbent dosage of 0.5 g L⁻¹ and at room temperature. The effect of contact time (1 min to 24 h) was investigated with initial fluoride concentration of 10 mg L^-1, adsorbent dosage of 0.5 g L⁻¹, at optimum pH and at room temperature. The influence of competing anions (chloride, nitrate, carbonate, sulfate and phosphate) on adsorption of fluoride by Al2O3 nanoparticles was examined by adding competing anion with initial concentrations of 10, 20 and 50 mg L⁻¹ into solutions with fixed fluoride concentration (10 mg L⁻¹), adsorbent dosage of 0.5 g L⁻¹ at an optimum pH and room temperature. Adsorption experiments also were conducted at 25, 35 and 45 ºC to investigate the effect of different temperatures on the fluoride adsorption by Al₂O₃ nanoparticles.

2.6. Adsorption kinetic studies

To better understand the adsorption kinetics, the experimental data were evaluated using different adsorption kinetic models such as the pseudo-first-order [26] (Eq. (3)), pseudo-second- order [27] (Eq. (4)), Avrami [28] (Eq. (5)) and intraparticle diffusion [29] (Eq. (6)) models. The non-linear forms of these models are given below (Eqs. (3-6)).

𝑞_𝑡 = 𝑞_𝑒(1 − 𝑒^{−(𝑘1𝑡)}) (3) 𝑞_𝑡 =_1+𝑞^{𝑞𝑒2𝑘2𝑡}

2𝑘₂𝑡 (4) 𝑞_𝑡 = 𝑞_𝑒(1 − 𝑒^{(−(𝑘𝐴𝑉𝑡))𝑛𝐴𝑉}) (5) 𝑞_𝑡 = 𝐾_𝑝𝑡^1⁄2+ 𝐼 (6)

where, q_t and q_e are the amounts of fluoride ions adsorbed (mg g^-1) on the Al₂O₃ nanoparticles at time t and equilibrium, respectively; k1 is the pseudo-first-order rate constant (min^-1), k2 is the pseudo-second-order rate constant (g mg⁻¹ min⁻¹); k_AV (min^-1) and n_AV are the exponential of the Avrami model, Kp is the intra-particle diffusion rate constant (mg g⁻¹ min^1/2) and I (mg g^-1) is the intercept of intra-particle diffusion model related to the thickness of the boundary layer.

ACCEPTED MANUSCRIPT

Page 10 of 39 2.7. Adsorption isotherm studies

To better understand the adsorption isotherms, the experiment data were analyzed using different adsorption isotherm models including the Dubinin–Radushkevich [30] (Eq. (7)), Redlich- Peterson [31] (Eq. (8)), Langmuir (Eq. (9)) [32] and Freundlich [33] (Eq. (10)) models. The non- linear forms of these models are given below.

𝑞_𝑒 = 𝑞_𝑚𝑒𝑥𝑝⁡(−𝛽 [𝑅𝑇𝑙𝑛 (1 +_𝑐¹

𝑒)²]) (7) 𝑞_𝑒 = ^{𝐾𝑅𝐶𝑒}

1+𝑎_𝑅𝐶_𝑒^𝛽 (8) 𝑞_𝑒 = ^𝑏𝑞_1+𝑏𝐶^𝑚𝐶𝑒

𝑒 (9) 𝑞_𝑒 = 𝐾_𝐹𝐶_𝑒

1

𝑛 (10) where, qe (mg g⁻¹) is the equilibrium adsorption capacity of Al2O3 nanoparticles fluoride adsorption, C_e (mg L⁻¹) is the equilibrium concentration of fluoride in solution, q_m (mg g⁻¹) is the maximum adsorption capacity of Al2O3 nanoparticles for fluoride, R (8.314 kJ mol^-1) is the universal gas constant, T (K) is the absolute temperature, β is activity coefficient related to mean adsorption energy per mole, KR (L g⁻¹) and aR (L mg⁻¹) are Redlich-Peterson constant. b (L mg⁻¹) in Langmuir model is related to the energy of adsorption, KF (mg g⁻¹) (L mg⁻¹)^1/n and n (g L⁻¹) are the Freundlich equilibrium constant and exponent, respectively.

2.8. Thermodynamic studies

Thermodynamic parameters were used in order to investigate the thermodynamic feasibility and to confirm the nature of the adsorption process. The thermodynamic parameters were obtained employing the following equations [34, 35].

ΔG⁰ = −𝑅𝑇𝑙𝑛𝐾 (11)

ACCEPTED MANUSCRIPT

Page 11 of 39

𝐾 = 𝑏 × 55.5 (12)

𝑙𝑛^𝐾_𝐾¹

2= ^∆𝐻_𝑅⁰(_𝑇¹

1−_𝑇¹

2) (13)

∆𝐺⁰ = ∆𝐻⁰− 𝑇∆𝑆⁰ (14) where, ΔGºis the standard free energy change (kJ mol⁻¹), R is the universal gas constant (8.314 J K^-1 mol^-1), T is the absolute temperature (in Kelvin), K is the equilibrium constant and b is related to the Langmuir constant. The value 55.5 corresponds to the molar concentration of the solvent (water, in this study), ΔH^º is standard enthalpy change (kJ mol⁻¹) and ΔSº is standard entropy change (J mol⁻¹ K⁻¹).

2.9. Error analysis

In this study, coefficient of determination (R²) and residual root mean square error (RMSE) were used to evaluate the fit goodness of applied models. These parameters were used by other research also for the selection of best model [23, 35, 36]. Two parameters are represented below:

1. The coefficient of determination (R²) 𝑅² = ⁡ (∑^𝑛_𝑖=1(𝑂_𝑖 − 𝑂_𝑎𝑣𝑒). (𝑃_𝑖 − 𝑃_𝑎𝑣𝑒))²

∑^𝑛_𝑖=1(𝑂_𝑖− 𝑂_𝑎𝑣𝑒)². ∑^𝑛_𝑖=1(𝑃_𝑖− 𝑃_𝑎𝑣𝑒)² (15)

2. Residual root mean square error (RMSE)

𝑅𝑀𝑆𝐸 = √ 1

𝑛 − 2∑(𝑂_𝑖− 𝑃_𝑖)²

𝑛 𝑖=1

(16)

where:

n: number of observations

Oi: i^th value of the observed measurement P_i: i^th value of the predicted measurement

ACCEPTED MANUSCRIPT

Page 12 of 39 Oave: mean of the observed values

P_ave: mean of the predicted values

3. Results and discussion

3.1. Characterization of Al2O3 nanoparticles

Table 1 shows the properties of Al2O3 nanoparticles produced from methanol solution with the FSP method. Contrary to the expectations, the solvents (i.e. methanol vs. toluene) had only minor effect on the particle properties that produced by FSP method. The produced particles consisted mainly -Al2O₃ phase identified according to the ICDD crystallographic database (PDF4+ file 04-007-2615). The crystalline size of the particles was below 10 nm. However, the hydrodynamic size of the particles as well as the zeta potential values were slightly smaller compared to the previous studies. The formation of Al2O3 particles from toluene precursor solution during the flame spray pyrolysis is described elsewhere [25].

Table 1.

Figure 1 shows the SEM and TEM images of the produced particles. The primary particles were spherical and loosely agglomerated forming a porous network of the particles. The primary particle size was around 9 nm (±4 nm) according to the TEM images which corresponds well with the crystalline size of the particles. However, the powder consisted also a small number of large particles with size above 100 nm. The hydrodynamic size of the particles when suspended in water was 168 nm. The zeta-potential of the suspensions was 58.4 mV indicating the stability of the suspension. The nitrogen adsorption-desorption isotherms are presented in Figure 2. The

ACCEPTED MANUSCRIPT

Page 13 of 39

adsorption isotherm can be classified as Type IV according to IUPAC classification indicating mesoporosity of the powder. The surface area of the powder was calculated as 213 m² g^-1 with the average pore size around 93 nm.

Figure 1.

Figure 3 shows the FTIR spectra of Al2O3 nanoparticles produced with the FSP. The FTIR spectra below 1000 cm^-1 shows typical peaks of alumina, which correspond to the stretching vibrations of Al-O bond [37, 38]. The bands at 3458 cm^-1 and 1628 cm^-1 arise from the surface bound water and can be assigned to the stretching and bending vibration of O-H, respectively.

The peaks observed at 1400 cm^-1 and 1348 cm⁻¹ have been previously explained as the combination of the neutral O2 species adsorbed on alumina surface and Al=O bond (1345 cm⁻¹) [39, 40]. However, the peaks can also be caused by the C-C deformation which could be due to the residual carbon absorbed on the surface of the alumina powder even though the sooting tendency of methanol is much lower compared to the toluene [24].

Figure 2.

Figure 3.

3.2. Effect of solution pH on fluoride adsorption byAl2O3 nanoparticles

The effect of solution pH (initial and final pH) on adsorption capacity of Al2O3 nanoparticles towards fluoride removal was studied and results are shown in Figure 4. The maximum adsorption of fluoride by Al₂O₃ nanoparticles (9.16 mg g^-1) occurred at final pH 5.4. At acidic pH, Al2O3 nanoparticles have positive charge due to the presence of more H⁺ ions, which

ACCEPTED MANUSCRIPT

Page 14 of 39

increased the electrostatic attraction between Al2O3 nanoparticles surface and negatively charged fluoride ions. Also, adsorption depends on the surface chemical characteristics of the adsorbent especially the surface charge. At pH of lower than pHpzc, surface ofAl2O3 nanoparticles have positive charge (due to protonation reactions), thus, ability ofAl₂O₃ nanoparticles increased for adsorption of fluoride ions. The further increase in solution pH (> 5.4), the adsorption of fluoride decreased. At basic pH, the OH^- ions and fluoride ions compete for the same active sites;

therefore, active sites for fluoride ion are decreased. Similar results have been reported by other researchers where, adsorption capacity of fluoride was decreased in the basic pH range [16, 41].

Figure 4.

3.3. Effect of contact time on fluoride adsorption by Al₂O₃ nanoparticles

The effect of contact time was examined on fluoride adsorption by Al₂O₃ nanoparticles and the results are presented in Figure 5. It can be seen that the adsorption capacity of Al2O3

nanoparticles for removing fluoride was increased as the contact time increased and equilibrium was achieved within 60 min. In the beginning stages of contact time, fluoride adsorption rate on Al2O3 nanoparticles was high as a large number of adsorption sites were readily available. When the available free sites were filled up by the fluoride ions gradually, adsorption process became slow [34, 42]. Similar results have been reported by other researchers where adsorption of fluoride increased with increasing contact time and after reaching to the equilibrium, adsorption remained constant [35, 41]. Therefore, 60 min was chosen for the equilibrium studies. A comparison of the results between this study and other studies suggest that the process of

ACCEPTED MANUSCRIPT

Page 15 of 39

fluoride adsorption by Al2O3 nanoparticles in this study is fairly faster than by other adsorbents [35, 43, 44].

Figure 5.

3.4. Effect of initial fluoride concentration by fluoride adsorption byAl2O3 nanoparticles Initial fluoride concentration is one of effective parameters for adsorption capacity. Therefore, fluoride adsorption was studied by taking different initial fluoride concentrations (0-30 mg L⁻¹) and keeping all other parameters constant (Al2O3 nanoparticles dosage: 0.5 g L^-1, temperature: 25

°C, shaking speed: 150 rpm, contact time: 60 min and natural pH). The results are presented in Figure 6. According to Fig. 6, fluoride adsorption (%) reduced with increasing the initial fluoride concentration. Considerable change was not observed in adsorption capacity for the values of initial fluoride concentration over 15 mg L^-1. It might be due to the saturation of active sites of the adsorbent surfaces at higher fluoride concentrations (>15 mg L^-1) [45]. These results are similar to the results reported by Tangsir et al. [35].

Figure 6.

3.5. Effect of temperature on fluoride adsorption byAl₂O₃ nanoparticles

The temperature can be effective on the improvement of anions’ adsorption by nanoparticles. In this study, change of fluoride adsorption by Al2O3 nanoparticles was investigated at 35 and 45

°C besides room temperature (25 °C). The results are shown in Figure 7. Increasing of temperature from 25 to 45 °C resulted in the increase of fluoride adsorption from 9.8 to 13.8 mg

ACCEPTED MANUSCRIPT

Page 16 of 39

g^-1. These results confirm that the adsorption process is endothermic in nature. It is suggested that the increases in adsorption capacity by increasing temperature might be attributed to increased interaction between fluoride ions and Al2O3 nanoparticles [35].

Figure 7.

3.6. Effect of co-existing (competing) anions on fluoride adsorption byAl₂O₃ nanoparticles Effect of some commonly found anions present in water (Cl^-, NO₃^-, SO₄^2-, CO₃^2- and PO₄^3-) was examined on fluoride adsorption. The initial concentrations of these ions was studied as 10, 20 and 50 mg L^-1 and fluoride initial concentration was kept constant (10 mg L^-1) in all the tests.

The results are presented in Figure 8. It is clear that presence of all ions influenced the fluoride adsorption but with different impact. The most negative effect was observed by phosphate followed by carbonate, sulfate and nitrate. Phosphate, carbonate and sulfate ions are the inner- spherically sorbing anions similar to the fluoride, and their presence might have caused the decrease in fluoride adsorption due to the competition for same sites [46]. However, presence of chloride (outer-spherically sorbing anion) did not have considerably effect on fluoride adsorption. These results are similar to the results reported in previous study [35]. On the other hand, decrease in fluoride adsorption in presence of carbonate was likely due to the significant increase of pH of the solution. It was also confirmed from our experiments on the effect of pH that the fluoride removal decreases in highly alkaline pH. Similar results were also reported previously [19].

Figure 8.

ACCEPTED MANUSCRIPT

Page 17 of 39 3.7. Adsorption kinetic studies

The results of adsorption kinetic models are shown in Table 2 and Figure 9 (a–b). As can be seen from Table 2, fluoride adsorption by Al2O3 nanoparticles follows the pseudo-second-order kinetic model better as compared to the other studied models with the higher value of R² and lower value of RMSE (R² = 0.98 and RMSE= 0.57). Pseudo-second-order model fitting has been found applicable for fluoride adsorption onto various adsorbents by other researchers also [34, 43]. Furthermore, theoretical value (9.70 mg g^-1), calculated by pseudo-second-order rate model, was also close to the experimental value (9.80 mg g^-1). Theoretically, the fluoride adsorption on to solid particles can be described by the three steps: (i)transport of fluoride ions to the external surface of the Al2O3 nanoparticles, (ii) fluoride adsorption on to particle surfaces, (iii) adsorbed fluoride ions are transferred to the internal surfaces for porous materials [47].

As can be seen from Fig. 9(b), the intercept of the curve does not pass through the origin which indicates that intra-particle diffusion was not the only rate-controlling step in adsorption kinetics and combination of surface reaction and diffusion boundary layer and intra-particle diffusion may also be controlling steps in the fluoride adsorption process [48].

Table 2 Figure 9.

3.8. Adsorption isotherm studies

The results of adsorption isotherm models are shown in Table 3 and Figure 10. As can be seen from Table 3, the Langmuir model with the higher value of R² and lower value of RMSE (R² = 1 and RMSE= 0.46) presented a better fit for the fluoride adsorption by Al2O3 nanoparticles as

ACCEPTED MANUSCRIPT

Page 18 of 39

compared to other isotherm models. Therefore, a good fit of experimental data with the Langmuir model would indicate monolayer adsorption on homogenous surface of Al₂O₃ nanoparticles. Also, it shows that the all sites on surface of Al2O3 nanoparticles have equal energy for fluoride adsorption. These results are similar to the results reported by other researchers where Langmuir isotherm model has shown good agreement with the experimental data for adsorption of fluoride ions [34, 41]. In this study, the maximum capacity of Al2O3

nanoparticles for fluoride adsorption was obtained as 13.70 mg g⁻¹.

The value of mean adsorption energy, E, can be calculated from Dubinin- Radushkevich parameter as follows [23, 49].

E =

√2β (18) This parameter is useful for predicting the type of adsorption. When the value of mean adsorption energy is from 1 to 8 kJ mol⁻¹, then the adsorption is physical in nature and if it is from 8 to 16 kJ mol⁻¹, then the adsorption is chemical in nature. In this study, the E value obtained is 0.92 kJ mol⁻¹, indicating the adsorption of fluoride onto the Al2O3 nanoparticles is physical in nature.

A dimensionless constant separation factor was used to estimate whether adsorption process is favorable or unfavorable. It is calculated from the Langmuir isotherm model parameter by the following equation:

𝑅_𝐿 = 1

1 + 𝑏𝑐₀ (19)

where, b (L mg⁻¹) is the Langmuir constant and C0 (mg L⁻¹) is the initial fluoride concentration^. According to Table 3, the R_L values for different concentrations are between 0 and 1, which shows favorable adsorption of fluoride onto the Al2O3 nanoparticles. The n value in Freundlich

ACCEPTED MANUSCRIPT

Page 19 of 39

model is 3.17 (between 1 and 10) which indicates favorable adsorption of fluoride by Al2O3

nanoparticles.

Table 3 Figure 10.

3.9. Thermodynamic parameters

The thermodynamic parameters were calculated and the value of standard free energy change was found negative (−33.52 kJ mol⁻¹), which revealed that the adsorption of fluoride onto the Al2O3 nanoparticles is spontaneous [34]. The positive value of standard enthalpy change (23.01 kJ mol⁻¹) reflected that adsorption process of fluoride by Al₂O₃ nanoparticles is endothermic reaction. The positive value of standard entropy change (189.62 J mol⁻¹ K⁻¹) shows the affinity of the Al2O3 nanoparticles for fluoride ions.

4. Conclusions

Precursor, solvent and technique that are used for preparation of metal nanoparticles have the most important roles on their sorption capacity. Previously we used toluene as solvent of aluminium acetylacetonate. Here, we used methanol by the same technique to know more about the effect of solvents on properties of synthesized nanoparticles adsorbents. In comparison to our previous study, there is not significant difference between sorption capacities of Al2O3

nanoparticles due to using different solvents. The FTIR spectra below 1000 cm^-1 (648, 600 and 563 cm^-1) shows typical peaks of alumina which correspond to the stretching vibrations of Al-O bond. The fluoride adsorption by Al2O3 nanoparticles was higher at pH below point of zero charge of the material due to the electrostatic attraction between Al₂O₃ nanoparticles surface and

ACCEPTED MANUSCRIPT

Page 20 of 39

negatively charged fluoride ions. The rate of adsorption was fairly rapid and maximum fluoride uptake was attained within 60 min. The pseudo-second-order model describes the defluoridation kinetic well. The equilibrium adsorption data fitted well with Langmuir isotherm model with maximum sorption capacity of 13.7 mg g^-1. It was also observed that the presence of phosphate ions as competing anions has deleterious effect on removal of fluoride, while chloride anions show negligible effect. In general, it can be concluded that Al2O3 nanoparticles cab be successfully used for the defluoridation of water.

Acknowledgments

Authors (LD and ST) are grateful to Ministry of Science, Research and Technology, Iran for providing scholarships to conduct their sabbatical (research) in Department of Environmental Science, University of Eastern Finland. Authors are also thankful to Marie Pétel and Gérald Neyrinck for their help in lab experiments.

References

[1] F.L. Theiss, S.J. Couperthwaite, G.A. Ayoko, R.L. Frost, A review of the removal of anions and oxyanions of the halogen elements from aqueous solution by layered double hydroxides, J.

Colloid Interface Sci. 417 (2014) 356-368.

[2] N.N. Greenwood, A. Earnshaw, Chemistry of the Elements, Pergamon Press, Oxford, United Kingdom, 1984.

[3] J.A. Camargo, Fluoride toxicity to aquatic organisms: a review, Chemosphere 50 (2003) 251- 264.

[4] D. Banks, C. Reimann, O. Røyset, H. Skarphagen, O.M. Sæther, Natural concentrations of major and trace elements in some Norwegian bedrock groundwaters, Appl. Geochem. 10 (1995) 1-1616.

ACCEPTED MANUSCRIPT

Page 21 of 39

[5] J. Dowgiałło, Thermal water prospecting results at Jelenia Góra-Cieplice (Sudetes, Poland) versus geothermometric forecasts, Environ. Geol. 39 (2000) 433-436.

[6] B. Frengstad, D. Banks, U. Siewers, The chemistry of Norwegian groundwaters: IV. The pH- dependence of element concentrations in crystalline bedrock groundwaters, Sci. Total Environ.

277 (2001) 101-117.

[7] S. Ayoob, A.K. Gupta, Fluoride in drinking water: a review on the status and stress effects, Crit. Rev. Environ. Sci. Technol. 36 (2006) 433-487.

[8] G. Chae, S. Yun, B. Mayer, K. Kim, S. Kim, J. Kwon, K. Kim, Y. Koh, Fluorine

geochemistry in bedrock groundwater of South Korea, Sci. Total Environ. 385 (2007) 272-283.

[9] I. Abe, S. Iwasaki, T. Tokimoto, N. Kawasaki, T. Nakamura, S. Tanada, Adsorption of fluoride ions onto carbonaceous materials, J. Colloid Interface Sci. 275 (2004) 35-39.

[10] A. Bhatnagar, E. Kumar, M. Sillanpää, Fluoride removal from water by adsorption—a review, Chem. Eng. J. 171 (2011) 811-840.

[11] F. Shen, X. Chen, P. Gao, G. Chen, Electrochemical removal of fluoride ions from industrial wastewater, Chem. Eng. Sci. 58 (2003) 987-993.

[12] S. Jagtap, M.K. Yenkie, N. Labhsetwar, S. Rayalu, Fluoride in drinking water and defluoridation of water, Chem. Rev. 112 (2012) 2454-2466.

[13] V. Marinho, Fluoride gel inhibits caries in children who have low caries-risk but this may not be clinically relevant, Evid.-Bbased Dent. 5 (2004) 95-95.

[14] A. Prüss-Üstün, R. Bos, F. Gore, J. Bartram, Safer water, better health: costs, benefits and sustainability of interventions to protect and promote health. World Health Organization, 2008.

[15] Y. Çengeloğlu, E. Kır, M. Ersöz, Removal of fluoride from aqueous solution by using red mud, Sep. Purif. Technol. 28 (2002) 81-86.

[16] Z. Jin, Y. Jia, K. Zhang, L. Kong, B. Sun, W. Shen, F. Meng, J. Liu, Effective removal of fluoride by porous MgO nanoplates and its adsorption mechanism, J. Alloys Compd. 675 (2016) 292-300.

[17] S.M. Maliyekkal, S. Shukla, L. Philip, I.M. Nambi, Enhanced fluoride removal from

drinking water by magnesia-amended activated alumina granules, Chem. Eng. J. 140 (2008) 183- 192.

[18] L.M. Camacho, A. Torres, D. Saha, S. Deng, Adsorption equilibrium and kinetics of fluoride on sol–gel-derived activated alumina adsorbents, J. Colloid Interface Sci. 349 (2010) 307-313.

ACCEPTED MANUSCRIPT

Page 22 of 39

[19] E. Kumar, A. Bhatnagar, U. Kumar, M. Sillanpää, Defluoridation from aqueous solutions by nano-alumina: characterization and sorption studies, J. Hazard. Mater. 186 (2011) 1042-1049.

[20] K. Hristovski, A. Baumgardner, P. Westerhoff, Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns: from nanopowders to aggregated nanoparticle media, J.

Hazard. Mater. 147 (2007) 265-274.

[21] S. Pisduangdaw, J. Panpranot, C. Methastidsook, C. Chaisuk, K. Faungnawakij, P.

Praserthdam, O. Mekasuwandumrong, Characteristics and catalytic properties of Pt–Sn/Al2O3

nanoparticles synthesized by one-step flame spray pyrolysis in the dehydrogenation of propane, Appl. Catal. A Gen. 370 (2009) 1-6.

[22] C.D. Rosebrock, T. Wriedt, L. Mädler, K. Wegner, The role of microexplosions in flame spray synthesis for homogeneous nanopowders from low‐cost metal precursors, AIChE J. 62 (2016) 381-391.

[23] L.D. Hafshejani, A. Hooshmand, A.A. Naseri, A.S. Mohammadi, F. Abbasi, A. Bhatnagar, Removal of nitrate from aqueous solution by modified sugarcane bagasse biochar, Ecol. Eng. 95 (2016) 101-111.

[24] C.D. Rosebrock, N. Riefler, T. Wriedt, L. Mädler, S.D. Tse, Disruptive burning of

precursor/solvent droplets in flame‐spray synthesis of nanoparticles, AIChE J. 59 (2013) 4553- 4566.

[25] L.D. Hafshejani, S. Tangsir, H. Koponen, J. Riikonen, T. Karhunen, U. Tapper, V. Lehto, H. Moazed, A.A. Naseri, A. Hooshmand, Synthesis and characterization of Al2O3 nanoparticles by flame spray pyrolysis (FSP)—Role of Fe ions in the precursor, Powder Technol. 298 (2016) 42-49.

[26] S. Lagergren, Zur Theorie der Sogenannten Adsorption Gelöster Stoffe, Kungliga Svenska Vetenskapsakademiens, Handlingar 24 (1898) 1-39.

[27] Y. Ho, G. McKay, Pseudo-second order model for sorption processes, Process Biochem. 34 (1999) 451-465.

[28] M. Avrami, Kinetics of phase change. I General theory, J. Chem. Phys. 7 (1939) 1103-1112.

[29] W.J. Weber, J.C. Morris, Kinetics of adsorption on carbon from solution, J. Sanit. Eng. Div.

89 (1963) 31-60.

[30] B. McEnaney, Estimation of the dimensions of micropores in active carbons using the Dubinin-Radushkevich equation, Carbon 25 (1987) 69-75.

[31] O. Redlich, D.L. Peterson, A useful adsorption isotherm, J. Phys. Chem. 63 (1959) 1024- 1024.

ACCEPTED MANUSCRIPT

Page 23 of 39

[32] I. Langmuir, The constitution and fundamental properties of solids and liquids. PART I.

SOLIDS., J. Am. Chem. Soc. 38 (1916) 2221-2295.

[33] H.M.F. Freundlich, Über die Adsorption in Lösungen, Zeitschrift für Physikalische Chemie 57A (1906) 385 - 470.

[34] Y. Zhang, Y. Jia, Fluoride adsorption onto amorphous aluminum hydroxide: Roles of the surface acetate anions, J. Colloid Interface Sci. 483 (2016) 295-306.

[35] S. Tangsir, L.D. Hafshejani, A. Lähde, M. Maljanen, A. Hooshmand, A.A. Naseri, H.

Moazed, J. Jokiniemi, A. Bhatnagar, Water defluoridation using Al2O3 nanoparticles synthesized by flame spray pyrolysis (FSP) method, Chem. Eng. J. 288 (2016) 198-206.

[36] L.D. Hafshejani, S.B. Nasab, R.M. Gholami, M. Moradzadeh, Z. Izadpanah, S.B.

Hafshejani, A. Bhatnagar, Removal of zinc and lead from aqueous solution by nanostructured cedar leaf ash as biosorbent, J. Mol. Liq. 211 (2015) 448-456.

[37] B.P. Dhonge, T. Mathews, S.T. Sundari, C. Thinaharan, M. Kamruddin, S. Dash, A. Tyagi, Spray pyrolytic deposition of transparent aluminum oxide (Al2O3) films, Appl. Surf. Sci. 258 (2011) 1091-1096.

[38] P. Manivasakan, V. Rajendran, P.R. Rauta, B.B. Sahu, B.K. Panda, Effect of mineral acids on the production of alumina nanopowder from raw bauxite, Powder Technol. 211 (2011) 77-84.

[39] A. Raveh, Z. Tsameret, E. Grossman, Surface characterization of thin layers of aluminium oxide, Surf. Coat. Technol. 88 (1997) 103-111.

[40] J.M. Roscoe, J.P. Abbatt, Diffuse reflectance FTIR study of the interaction of alumina surfaces with ozone and water vapor, J. Phys. Chem. A 109 (2005) 9028-9034.

[41] Q. Zhou, X. Lin, B. Li, X. Luo, Fluoride adsorption from aqueous solution by aluminum alginate particles prepared via electrostatic spinning device, Chem. Eng. J. 256 (2014) 306-315.

[42] S. Budyanto, Y. Kuo, J. Liu, Adsorption and precipitation of fluoride on calcite nanoparticles: A spectroscopic study, Sep. Purif. Technol. 150 (2015) 325-331.

[43] Y. Zhang, X. Lin, Q. Zhou, X. Luo, Fluoride adsorption from aqueous solution by magnetic core-shell Fe₃O₄@ alginate-La particles fabricated via electro-coextrusion, Appl. Surf. Sci. 389 (2016) 34-45.

[44] Y. Tang, X. Guan, T. Su, N. Gao, J. Wang, Fluoride adsorption onto activated alumina:

Modeling the effects of pH and some competing ions, Colloids Surf. Physicochem. Eng. Aspects 337 (2009) 33-38.

ACCEPTED MANUSCRIPT

Page 24 of 39

[45] D. Dayananda, V.R. Sarva, S.V. Prasad, J. Arunachalam, N.N. Ghosh, Preparation of CaO loaded mesoporous Al2O3: Efficient adsorbent for fluoride removal from water, Chem. Eng. J.

248 (2014) 430-439.

[46] S. Swain, S. Mishra, T. Patnaik, R. Patel, U. Jha, R. Dey, Fluoride removal performance of a new hybrid sorbent of Zr (IV)–ethylenediamine, Chem. Eng. J. 184 (2012) 72-81.

[47] M. Mohapatra, S. Anand, B. Mishra, D.E. Giles, P. Singh, Review of fluoride removal from drinking water, J. Environ. Manage. 91 (2009) 67-77.

[48] W. Ma, T. Lv, X. Song, Z. Cheng, S. Duan, G. Xin, F. Liu, D. Pan, Characteristics of selective fluoride adsorption by biocarbon-Mg/Al layered double hydroxides composites from protein solutions: Kinetics and equilibrium isotherms study, J. Hazard. Mater. 268 (2014) 166- 176.

[49] A. Bhatnagar, E. Kumar, M. Sillanpää, Nitrate removal from water by nano-alumina:

Characterization and sorption studies, Chem. Eng. J. 163 (2010) 317-323.

ACCEPTED MANUSCRIPT

Page 25 of 39

(A)

(B)

Figure 1. SEM (A) and TEM (B) images of Al2O3 nanoparticles.

ACCEPTED MANUSCRIPT

Page 26 of 39

Al

O

– Adsorption Al

O

– Desorption

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Relative pressure (p/p

)

0 100 200 300 400 500 600 700 800 900 1,000

Q u a n ti ty A d so rb ed (c m

/g ST P )

Figure 2. Nitrogen adsorption/desorption isotherms of Al₂O₃ nanoparticles.

ACCEPTED MANUSCRIPT

Page 27 of 39

Figure 3. FTIR spectra of Al₂O₃ nanoparticles.

ACCEPTED MANUSCRIPT

Page 28 of 39

Figure 4. Effect of solution pH on fluoride adsorption efficiency by Al₂O₃ nanoparticles (initial fluoride concentration: 10 mg L⁻¹, contact time: 24 h, temperature: 25 °C, adsorbent dose: 0.5 g L⁻¹ and initial pH = 3-11).

ACCEPTED MANUSCRIPT

Page 29 of 39

Figure 5. Effect of contact time on fluoride adsorption by Al2O3 nanoparticles (initial fluoride concentration: 10 mg L⁻¹, contact time: 5–180 min, temperature: 25 °C, adsorbent dose: 0.5 g L⁻¹ and natural pH).

ACCEPTED MANUSCRIPT

Page 30 of 39

Figure 6. Effect of initial fluoride concentration on fluoride adsorption by Al2O3 nanoparticles (initial fluoride concentration: 0-30 mg L⁻¹, contact time: 60 min, temperature: 25 °C, adsorbent dose: 0.5 g L⁻¹ and natural pH).

ACCEPTED MANUSCRIPT

Page 31 of 39

Figure 7. Effect of temperature on fluoride adsorption by Al₂O₃ nanoparticles (initial fluoride concentration: 10 mg L⁻¹, contact time: 60 min, temperature: 25, 35 and 45 °C, adsorbent dose:

0.5 g L⁻¹ and natural pH).

ACCEPTED MANUSCRIPT

Page 32 of 39

Figure 8. Effect of competing anions on fluoride adsorption by Al₂O₃ nanoparticles (initial concentration: 10, 20 and 50 mg L⁻¹, contact time: 60 min, temperature: 25 °C, adsorbent dose:

0.5 g L⁻¹ and natural pH).

ACCEPTED MANUSCRIPT

Page 33 of 39

Figure 9. (a) Kinetic modeling and, (b) Intra-particle diffusion model for fluoride adsorption by Al2O3 nanoparticles.

(a)

(b)

ACCEPTED MANUSCRIPT

Page 34 of 39

Figure 10. Isotherm modeling for fluoride adsorption by Al2O3 nanoparticles.

ACCEPTED MANUSCRIPT

Page 35 of 39 Table 1. Properties of the Al2O3 nanoparticles.

Primary Particle size [nm]

Crystalli ne size [nm]

Surface area [m²/g]

Hydrodynamic size [nm]

zeta-potential [mV]

Al2O3 9.8 ± 4.3 <10 nm 213 168 58.4

ACCEPTED MANUSCRIPT

Page 36 of 39

Table 2. Kinetic parameters of fluoride adsorption by Al2O3 nanoparticles.

Pseudo-first-order

qe(exp) (mg g^-1) k1 (min^-1) qe(cal) (mg g^-1) RMSE R²

9.12 0.53 9.80 1.01 0.95

Pseudo-second-order

qe(exp) (mg g^-1) k2 (g mg^-1 min^-1) qe(cal) (mg g^-1) RMSE R²

9.70 0.07 9.80 0.57 0.98

Intra-particle diffusion

I (mg g^-1) kP (mg g^-1 min^-0.5) - RMSE R²

4.76 0.55 - 1.95 0.77

Avrami

qe(exp) (mg g^-1) KAV(min^-1) nav qe(cal) (mg g^-1) RMSE R²

9.12 0.15 3.56 9.80 1.01 0.95

ACCEPTED MANUSCRIPT

Page 37 of 39

Table 3. Adsorption isotherm parameters of fluoride adsorption on Al2O3 nanoparticles.

Freundlich

KF (mg g⁻¹)(L mg⁻¹)^1/n n (g L⁻¹) - RMSE R²

5.12 3.17 - 0.88 0.98

Langmuir

b (L mg⁻¹) qm (mg g⁻¹) RL RMSE R²

0.53 13.70 0.06-1 0.46 1

Redlich–

Peterson

b (L mg^-1) a (L mg^-1) n RMSE R²

0.86 9.26 0.92 0.51 0.99

Dubinin- Radushkevich

β qm (mg g⁻¹) nav

E(kJ mol^-1) RMSE R²

0.59 13.09 0.92 0.65 0.99

ACCEPTED MANUSCRIPT

Page 38 of 39 Graphical abstract

ACCEPTED MANUSCRIPT

Page 39 of 39 Research Highlights

 Defluoridation of aqueous solutions was studied by Al₂O₃ nanoparticles.

 Maximum adsorption capacity for fluoride was 13.7 mg g^-1 by Al₂O₃ nanoparticles.

 FTIR results confirm the complexation of Al₂O₃ nanoparticles with fluoride ions.