Fluoride and Arsenite Removal by Adsorption on La 2 O 3 CeO 2 /Laterite

In the present article, the adsorbent prepared from laterite with lanthanum and cerium oxides (La 2 O 3 -CeO 2 /laterite (LCL)) was e ﬃ ciently employed for the removal of arsenite and ﬂ uoride from an aqueous environment. The obtained materials were characterized by XRD, SEM, and nitrogen adsorption/desorption. The synthesized LCL exhibited a high adsorption capacity towards arsenite (As(III)) and ﬂ uoride. The adsorption of both analytes on LCL, which was well- ﬁ tted to a pseudo -second-order equation, was found to be kinetically fast in the ﬁ rst 20 minutes and reached equilibrium at around 180 minutes. Weber ’ s intraparticle di ﬀ usion model in multilinearity using the piecewise linear regression combined with Akaike ’ s criteria was addressed. The adsorption capacities of LCL calculated from Langmuir ’ s isotherm model were found to be 67.08mg · g -1 for arsenite and 58.02mg · g -1 for ﬂ uoride. Thermodynamic parameters presented an endothermic nature of arsenite adsorption but an exothermic nature for ﬂ uoride and a negative Gibbs free energy for the spontaneous process of arsenite or ﬂ uoride adsorption at the studied temperature range. The excellent adsorption performance and stability make the composite of laterite and La-Ce binary oxides an alternative e ﬃ cient and cheap adsorbent for the removal of arsenite and ﬂ uoride in an aqueous solution.


Introduction
Arsenic or fluoride-contaminated water is an environmental problem which can lead to severe human health implications and ecological concern [1,2]. These elements are increasingly introduced to water sources by natural processes or human activities [3,4]. Inorganic species of arsenic in both forms of arsenate (As(V)) and arsenite (As(III)) prevail in groundwater. However, its valence states depend on the redox envi-ronment in aqueous sources and may change from region to region, in which arsenite is considered more mobile and acutely toxic than arsenate [5]. The acceptable concentration is 10 μg·L -1 for total arsenic in drinking water [6,7]. Excess arsenic in drinking water can cause some health problems such as neurological complications, skin lesions, kidney cancers, and the retardation of mental development of children [8]. Meanwhile, fluorine is abundant in the natural environment which is released from the fluoride-containing minerals and wastewater discharge from industrial processes using fluorine compounds [9]. It can be introduced into the human body via food and water consumption [10]. It is advised that the fluoride concentration in the range 0.5-1.5 mg·L -1 is acceptable in drinking water [11]. A higher concentration will have negative effects on the metabolism of several elements such as P and Ca in the human body and can cause dental fluorosis or molting of teeth [12]. Fluoride concentration, however, has been found to exceed the permissible values in groundwater in many regions in the world, especially in developing countries.
Some studies have reported that arsenic or fluoride elimination can be achieved by various methods such as oxidation/precipitation [13,14], electrocoagulation [15], reverse osmosis and nanofiltration [16,17], ion-exchange resin [18], polymer ligand exchange [19], coagulation and microfiltration [20], dialysis [21], ion-exchange processes [18], and adsorption processes [22]. Each method has its own advantages; however, the adsorption process is still considered one of the most employed techniques due to its low cost and simple operation without the use of large amounts of chemicals or the discharge of contaminated coagulant sludge as found in other methods. Therefore, the search for new and economical adsorbents with high adsorption capacity and affinity for contaminated elements that can be effectively employed under relevant environmental conditions has recently attracted extensive attention from scientists. Natural iron-contained minerals such as diatomite, laterite, kaolinite, montmorillonite, and illite are highly effective and widely used to remove arsenic [23,24]. Among them, the laterites are promising alternative adsorbents in removing hazardous anions because of their high surface area and abundant number of adsorptive sites. Laterites are found from the weathering of parent rocks in tropical climate regions with their chemical composition varying according to the nature of parent rocks. Laterites are predominantly composed of quartz, aluminum hydroxide, kaolinite mineral, goethite, and hematite [23] and are abundantly available in several regions in the central and northern areas of Vietnam. Recently, synthesized nanosized adsorbents including La 2 O 3 , Fe 3 O 4 , CeO 2 , CuO, MnO 2 , and TiO 2 , which exhibited a potential ability to remove arsenic or fluoride from an aqueous environment, have been extensively studied [25][26][27]. Their improved adsorption could be attributed to their controllable surface properties, being suitable for arsenite and fluoride removal. Cerium and lanthanum oxides or their binary oxides have been studied for application in areas such as catalysts [28,29], electrochemistry [30], and disease therapy [31] due to their specific physical and chemical properties. For environmental applications, cerium and lanthanum oxides have demonstrated a high adsorption capacity for hazardous chemical substances, such as arsenate [32], phosphate [33], and dyes [34]. Embedding such nanoparticles on laterite is expected to produce a promising adsorbent with low cost, yet great adsorption capacity for environmental remediation applications.
In the present work, the composite of low-cost natural laterite and La-Ce bimetal oxides (LCL) is explored as a potential fluoride and arsenic adsorbent. Both equilibrium and kinetic studies are conducted to evaluate the adsorption effectiveness of LCL to remove arsenite and fluoride from water. The Langmuir and Freundlich isotherms were addressed, and the thermodynamic calculations were also performed to have more insight into the process.

Experiments
2.1. Materials. Laterite for this study was collected from the northern area of Vietnam. It was ground and washed several times by distilled water. After being dried, solids were screened and stored in a desiccator for further use. The oxide composition analyzed by the X-ray fluorescence method (XRF) is listed in Table 1.
Arsenite and fluoride stock solution (1000 mg·L -1 ) was prepared by dissolving sodium-meta-arsenite (NaAsO 2 , Merck, Germany) and sodium fluoride (NaF Merck, Germany) in distilled water. The working solution was freshly prepared from stock solution (1000 mg·L -1 ) to the desired concentration. The acid (0.01 M HNO 3 ) and base solutions (0.01 M NaOH) were employed for pH adjustments.
2.2. Synthesis of La 2 O 3 , CeO 2 , and La 2 O 3 -CeO 2 /Laterites (LCL). To synthesize La 2 O 3 and CeO 2 , gelatin (50 grams) was firstly dissolved in distilled water at 80°C before adding La(NO 3 ) 3 or Ce(NO 3 ) 4 at a La(NO 3 ) 3 /gelatin or Ce(NO 3 ) 4 /gelatin molar ratio of 1/1. The mixture was stirred properly to form a gel. The obtained gel was dried at 110°C in 12 hours and calcinated at 550°C for 4 hours to generate the corresponding oxide. An LCL sample was synthesized in a similar manner, with the molar ratio of La(NO 3 ) 3 /Ce(NO 3 ) 4 /gelatin being 1/1/1. The pH of the mixture was adjusted to 7. Laterite was added at an oxide/laterite mass ratio of 10/90 just after gel formation. This mixture was then continuously stirred to allow an even dispersion of the gel over the laterite. Then, the laterite gel was dried at 110°C in 12 hours and finally calcined at 550°C for 2 hours to obtain LCL.

2.3.
Apparatus. X-ray diffraction (XRD) was measured on a Bruker D8 ADVANCE with CuKα, λ = 0:15406 nm. The Xray fluorescence spectrometer (XRF) used was a ZSX Primus IV (Rigaku, Japan). Thermal analysis (TG-DTA) was recorded (1) Adsorption Kinetics. The adsorption capacity (q t ) and equilibrium adsorption capacity (q e , mg·g -1 ) were calculated as equation (1) and equation (2), respectively: where V is the solution volume (L); m is the weight of the adsorbent (g); C o , C t , and C e (mg·L -1 ) are the concentrations of the adsorbate at the initial, certain, and equilibrium times.
The kinetics data were analyzed using the pseudo-firstorder kinetic and pseudo-second-order kinetic models. The pseudo-first-order kinetic and the pseudo-second-order kinetic equations in the nonlinear form [35,36] are written as equation (3) and equation (4), respectively: where k 1 (min -1 ) and k 2 (g·mg -1 ·min -1 ) are the rate constants for the pseudo-first-order kinetic and the pseudo-secondorder kinetic models, respectively. The values of k 1 , k 2 , and q e were obtained by the nonlinear regression method with the Solver function in Microsoft Excel.
For the study of kinetics diffusion, the intraparticle diffusion model proposed by Weber and Morris [37] is used (equation (5)) where k i (mg·g -1 ·min -0.5 ) is the intraparticle diffusion rate constant, x i (mg·g -1 ) is a constant related to the thickness of the boundary layer. The values of k i and x i were obtained from the slope of the linear plot of q t versus t 1/2 . If intraparticle diffusion is the rate limiting step, then a linear plot of q t vs. t 0:5 will give an intercept equal to zero. Otherwise, the intraparticle diffusion is not the only rate-limiting step. The plot of Weber's model often consists of multilinearity in nature. A statistical approach of the piecewise linear regression combined with Akaike's criteria proposed by Malash and El-Khaiary [38] for the analysis of kinetics data was described in S1.
(2) Thermodynamics Adsorption. Thermodynamic studies were performed in batch condition. LCL (2.5 g) was added to 500 mL of the arsenite solutions (5 and 10 mg·L -1 ) and fluoride solutions (5 and 10 mg·L -1 ) in a 1-litre beaker attached with a mechanical stirrer at room temperature. Four mL of the solution was withdrawn at different times. The concentration of arsenite or fluoride in the supernatant was determined by the corresponding analysis method.
The temperature effect on arsenite or fluoride adsorption was investigated using 0.5 g of LCL in a 100 mL flask containing arsenite solution (10 mg·L -1 ) or fluoride solution (10 mg·L -1 ). The flasks were magnetically stirred for 6 hours at a fixed temperature to obtain the adsorption/desorption equilibrium. The equilibrium concentration of arsenite or fluoride in the supernatant was measured. The distribution coefficient K d is described as follows [39]: The change of standard Gibbs free energy of adsorption (ΔG°) is expressed by where ΔG°, ΔH°, and ΔS°are the change of standard Gibbs free energy, enthalpy, and entropy. The relationship betweenK d andΔH°andΔS°can be demonstrated as follows: (3) Equilibrium Study. The equilibrium study was performed at ambient temperature. Briefly, 0.5 g of adsorbent (LL, CL, and LCL) was introduced into eleven 200 mL flasks filled with 100 mL of different arsenite or fluoride concentrations (i.e., 1, 5, 10, 20, 30, 40, 50, 100, 120, 150, and 200 mg·L -1 ). The flasks were stirred by a shaker for 12 hours at room temperature. Afterwards, the liquid was obtained by centrifugation.
Two isotherm models of Freundlich and Langmuir were employed to analyze the equilibrium data. The Freundlich isotherm model is described as follows [40]:

Journal of Nanomaterials
where K F is the Freundlich constant, n is the empirical parameter. The Langmuir isotherm model is expressed as follows [41]: where q m is the maximum monolayer capacity amount (mg·g -1 ), K L is the Langmuir equilibrium constant (L·mg -1 ). The model parameters were determined using nonlinear regression.

Desorption of Adsorbents.
The saturated NaCl solution was used to recover LCL nanoparticles. A saturated NaCl solution was allowed to run through a column containing the used LCL nanoparticles several times to completely eliminate the adsorbates (fluoride or arsenite). The recovered LCL nanoparticles were separated from the column and cleaned in both acid and base solutions and subsequently dried at 110°C for 12 hours before reuse. The morphology of the obtained samples was performed by SEM observation (Figure 3). Both La 2 O 3 and CeO 2 materials exhibit large agglomerates of fine particles of around 50-200 nm. It is clear that laterite exhibits a highly porous structure with a heterogeneous texture. The LCL composite presents the embroiling of fine oxide particles in the laterite matrix.

Results and Discussion
The specific surface areas of the obtained materials were calculated by the Brunauer-Emmett-Teller (BET) model using adsorption data with relative pressure in the range of 0.05-0.35 (Figure 4). It was found that the specific surface areas of lanthanum oxide, cerium oxide, laterite, and LCL are 47.3, 48.3, 4.0, and 14.4 m 2 ·g -1 , respectively. Laterite exhibits a significantly low surface area, but the combination of laterite with lanthanum and cerium oxides greatly enhances the specific surface area which can lead to an improvement in its adsorption capacity.
The element composition of the LCL sample was analyzed by EDX spectroscopy. The main elements include Al, Si, Fe, La, and Ce in laterite and added oxides. It is notable that the La/Ce molar ratio in the product was close to the initial mixture. The distribution of elements in the obtained composite was studied by EDX mapping ( Figure 5) showing that elements of Ce and La were highly dispersed in the laterite matrix. Some other elements (S, Mi, K, Ca, Ti, and C) of   Journal of Nanomaterials around 5.8 atomic % as impurities in laterites were found in EDX mapping (data not shown).  Figure 6). No significant change in arsenite adsorption capacity was observed in the pH range of 2 to 9.2. This is consistent with the existence of the solely neutral molecule H3AsO3 in aqueous media at pH < 9:2 [42]. Hence, the electrostatic interaction could play a negligible role in the arsenite removal due to uncharged arsenite species at this pH range. The pH PZC of LCL was estimated by the pH drift method [43] (Figure 7). It was notable that the pH PZC    (Figure 7). The shift of pH PZC could be due to the result from the negatively charged inner-sphere complexes between arsenite and the adsorbent [44,45], and effective removal was achieved. Similarly, the fluoride adsorption capacity varies slightly from 1.88 at pH 2 to 1.1 mg·L -1 at pH 7 and increases to 1.88 mg·g -1 at pH 9.2 ( Figure 6). The pH PZC of LCL does not vary much from 6.6 (in an electrolyte solution) to~6.5 (with 2 mg·L -1 F) (Figure 7). The anionic displacement and adsorption via coulombic forces on the surface of metal oxide is widely used to explain the result in the fluoride adsorption of porous materials [40,41]. However, the electrostatic interaction does not play an important role in this case, and the capillary attraction could explain for the fluoride adsorption of LCL due to its porous structure with a heterogeneous texture.

Adsorption Kinetics.
The kinetic experiments were conducted at two concentrations of arsenite or fluoride (Figures 8(a) and 8(b)).
More than 90% of arsenite (C o = 5 mg · L −1 ) and fluoride (C o = 10 mg · L −1 ) were removed within the first 120 min. The adsorption occurred in three distinct stages for both the arsenite and fluoride: (1) a rapid removal in the first 20 min (>50%), (2) relatively slower removal after 20 min until 120 min, and (3) little removal beyond 120 min. The fitting of experimental data to the pseudo-first-order and pseudosecond-order reaction models shows that the pseudo-second-order model provides the best fit (R 2 > 0:990) for both arsenite and fluoride ( Table 2). The plot of Weber's model in multilinearity was performed. It was found that the three-segment linear model gives the best goodness of fit with the lowest AIC (Table 3 and Figures 8(c) and 8(d)). As seen from Table 3, the intercept of the first segment with 95% confidence interval in the plot varies from a negative value to a positive value (except for 10 mg·L -1 of arsenite) ( Table 4) indicating that the intercept passes the origin or the intraparticle diffusion is the rate-limiting process. In the second and third stages, all values of the intercept varying from a positive value to a positive value are significantly different from zero    Journal of Nanomaterials implying that the line does not pass the origin [38]. These results reveal that the intraparticle diffusion is not the only rate-limiting process.

Thermodynamic Parameters.
A thermodynamic study was performed in the temperature range of 283-313 K ( Figure 9). By increasing temperature from 283 to 313 K, arsenite removal efficiency increases but that for fluoride decreases. At 283 K, arsenite and fluoride adsorption capacities were 8.3 and 10.9 mg·g -1 , respectively, and by increasing solution temperature to 313 K, adsorption capacity increases to 13.38 mg·g -1 for arsenite but decreases to 10.2 mg·g -1 for fluoride. The positive values of ΔH°and ΔS°indicate the endothermic nature and high randomness at the surface of solid/solution phases in the arsenite adsorption system ( Table 5). The increase in the randomness may be related to the fact that the displacement of adsorbed water molecules or hydroxyl ion by the arsenite species enhances translational entropy which creates the prevalence of randomness in the system [46]. The low exothermic enthalpy value and negative   9 Journal of Nanomaterials entropy change in fluoride adsorption imply the possible domination of physical forces. However, the negative Gibbs free energy indicates the spontaneous process of arsenite or fluoride adsorption at the studied temperature range.

Adsorption Equilibrium.
Langmuir and Freundlich isotherm models were employed to estimate arsenite and fluoride sorption behavior by LCL. Both arsenite and fluoride adsorption data were found to be fitted better for the Langmuir model (R 2 = 0:99 and 0.99 for arsenite and fluoride, respectively) than the Freundlich model (R 2 = 0:96 and 0.98 for arsenite and fluoride, respectively) ( Figure 10).
Based on the Langmuir isotherm, the maximum monolayer capacity amount of LCL was found to be 67.08 mg·g -1 for arsenite and 58.02 mg·g -1 for fluoride (Table 6). It was notable that laterite has low arsenite or fluoride adsorption. However, the combination of laterite with oxides significantly increases the adsorption capacity. Although the arsenite and fluoride adsorption capacity of LCL is lower than the lanthanum oxide and cerium oxides, they are compatible  compared to other materials so far reported as shown in Table 7. Since the laterite natural mineral is inexpensive, the high adsorption capacity makes it a great potential adsorbent for the removal of arsenite and fluoride from aqueous sources.
3.2.5. Competing Anions. The arsenite or fluoride removal performance in the presence of Fe(III), Mn(II), Cl-, and SO 4 2was conducted with the initial concentrations of arsenite and fluoride of 5 mg·L -1 and 10 mg·L -1 , respectively (Table 8). It was found that in the presence of 2-fold excess of iron (III), the adsorption capacity for arsenite improved from 3.14 to 3.38 mg·g -1 and that for fluoride almost doubled from 8.84 to 16.2 mg·g -1 with 10-fold excess of iron (III). Mn(II) also slightly enhanced the arsenite and fluoride removal ability of LCL. The anions such as Cland SO 4 2did not show any observable effect on the adsorption of fluoride; however, that of arsenate was reduced by half in the presence of 20-fold excess of Cl -, and by one-third in the presence of 40-fold excess of SO 4 2-.
3.2.6. Reusability. The reusability of LCL is one of the important criteria for its practical applications. The arsenite or fluoride amount adsorbed on LCL was desorbed with a NaClsaturated solution. Before reuse, the recycled LCL was dried at 110°C for 12 h. After eight cycles, the adsorption capacity declines by approximately 8% for arsenite and 10% for fluoride ( Figure 11) indicating that LCL was stable after several adsorption cycles.

Conclusions
Laterite has been effectively used as a matrix material carrying La-Ce binary oxides to form efficient adsorbents. The La 2 O 3 -CeO 2 /laterite exhibits high fluoride and arsenite removal efficiency in a wide pH range. The combination of laterite with lanthanum and cerium oxides enhances significantly the arsenite and fluoride adsorption capacity compared with laterite. The arsenite and fluoride adsorption capacity of the La 2 O 3 -CeO 2 /laterite are 10.5 and 13.6 times higher than those of laterite, respectively. It is a promising adsorbent due to the high recyclability and arsenite and fluoride adsorption capacity. The easy availability and abundance of laterite in various areas in Vietnam make LCL an inexpensive, feasible, and applicable material in practice.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.