Synthesis and pseudoboehmite agglomeration

In the present study, the starting materials to obtain hydrated alumina (γ-Al₂O₃) are Al₂(SO₄)₃•18H₂O, with a purity of 95% (by weight, wt.), technical grade, Alfa-Omega Chemical S. A., HNO₃ (karal 70%), C₆H₁₄ (J. T. Baker 99.8%), NH₄OH (Karal 30%), and anhydrous NH₃ gas (Praxair, 99.98% [wt.]) as precipitant.

The preparation procedure flow chart is depicted in Figure 1. Al₂(SO₄)₃•18H₂O was dissolved into deionized water and filtered for the removal of insoluble impurities. The precursors were prepared by dropping the saturated solution of aluminum sulfate into the solution of water and ammonia gas. That was done under rigorous magnetic stirring at a temperature range of 50 °C-70 ºC. The precipitated solution was filtered, and the precipitate of pseudoboehmite was rinsed with distilled water two times and then dried at 110 °C for 12 h. The agglomeration of pseudoboehmite in solid spheres of 3 mm diameter was obtained from a colloidal solution. The colloidal solution was prepared using pseudoboehmite, nitric acid, and distilled water. This solution was then dripped into a glass column with 15 mL of C₆H₁₄ and 25 mL of NH₄OH for making the agglomeration. The phase change from pseudoboehmite agglomerates to γ-Al₂O₃ was performed by a thermal treatment at 550 °C for 6 h. The chemical reactions (1), (2), and (3) occurred during the synthesis and the gamma phase was obtained:

Precipitation step:

Calcination step:

The AA used in this study is commercial spherical with a diameter of 3 mm (Alquimia Mexicana S. de R. L.).

Source: Author’s own elaboration.

Figure 1 Synthesis and process of agglomeration of Pseudoboehmite (precursor of γ-Al₂O₃). 

Surface characterization of the adsorbents

The γ-Al₂O₃ and AA samples were dispersed into distilled water. The solutions pH were adjusted to the following ranges for analysis: between 2 mM and 12 mM when using KCl, and 10 mM when using NaOH. The zeta potential and particle size were analyzed using an electroacoustic technique with a particle size analyzer (AcoustoSizer II, ESA; Colloidal Dynamics, USA) at 25 °C ± 0.2 °C. The textural properties were characterized by N₂ adsorption (Micromeritics, ASAP 2010). The samples were vacuum degassed at 200 °C for 3 h. Nitrogen adsorption isotherms were measured at a liquid N₂ temperature (77 K) and N₂ pressures ranging from 10^-6 P/P₀ to 1.0 P/P₀. The surface area was calculated according to the Brunauer-Emmett-Teller (BET) method, and the pore size distribution was obtained according to the Barret-Joyner-Halenda (BJH) method. The crystalline properties of the samples were determined at room temperature by using X-ray powder diffraction. A Siemens D-500 diffractometer equipped with a CuKα radiation anode was used for these measurements, under the following conditions: a sweep of 5°-80º at an angle of 2θ, with a wavelength of k = 1.54 Å, a current of 30 kV, and a current of 20 mA. Solid-state nuclear magnetic resonance (NMR) measurements were made with a Bruker Avance III HD 400 spectrometer in a nominal field of 9.4 T at a Larmor frequency of 104 MHz. The ²⁷Al spectra were observed using a magic angle spinning (MAS) probe. For this process, rotors of 4 mm were used, as well as a rotation frequency of approximately 11 kHz, using as reference a solid sample of Al(NO₃)₃• 9H₂O (²⁷Al, d = 0.0 ppm). The nature of acid sites was determined by pyridine adsorbed Fourier transformed infrared (pyridine FT-IR) spectroscopy in a FT-IR Nicolet 470 instrument with a resolution of 1 cm^-1. The samples were dehydrated at 300 °C for 1 h under vacuum. This was followed by adsorption of purified pyridine at room temperature. Then, the sample was evacuated at room temperature and later at 200 °C. The pyridine-adsorbed IR spectra were recorded. The surface morphology was examined using JEOL JSM-7600-F Field-Emission scanning electron microscopy. This was equipped with an energy dispersive X-ray spectrometer (FE-SEM/EDS) to determine the chemical composition.

Batch adsorption experiments

F⁻ adsorption kinetics experiments were conducted by mixing suspensions containing 10 mg/L, 30 mg/L, 50 mg/L, 100 mg/L, and 150 mg/L of fluoride ion in deionized water with 0.3 grams of adsorbent material to the levels of pH of 5, 7, and 9, which was adjusted with hydrochloric acid. Samples were left for five days at different temperatures (25 °C and 35 °C); this was because the groundwater temperature varies from 25 °C to 50 ºC. The highest temperatures are located near faults and fractures. High temperatures match with the highest groundwater concentrations of As and F^- in Guanajuato, Mexico (^{Morales, Villanueva-Estrada, Rodríguez & Armienta, 2015}). The samples in solution were analyzed immediately at the desired mixing time. The residual F− concentration in solution samples was measured using a fluoride selective electrode (Thermo Scientific Orion 4-Star Plus). That was in accordance with the standard method (^{Wang et al. 2014}). After five days, the solution was filtered, and the final concentrations of fluoride ions were determined. It was considered that equilibrium was reached when the fluoride concentration obtained was the same, or that there was a difference of 5% over the earlier recorded period.

The adsorption isotherms were obtained by mixing solutions containing 7.5 g/L of γ-Al₂O₃ or AA, and 0 to 20 mg/L of F⁻ for 2 h. The suspension also contained 1:1 volume of total ionic strength adjustment buffer (TISAB II) solution to control the suspension of pH at 7.0 ± 0.2 and the ionic strength at 0.04 M., a liter of the TISAB II solution contained, 1% acetic acid (C₂H₄O₂), 8% sodium acetate (C₂H₃NaO₂), 6% sodium chloride (NaCl), <1% Trans-1,2-Diamino cyclohexane tetra acetic acid (CDTA), and >84% deionized water (H₂O). After mixing, the solutions samples were analyzed to obtain the equilibrium F− concentration.

Powder X-ray diffraction (XRD)

The crystalline properties of synthesized and agglomerate alumina (γ-Al₂O₃) and AA are shown by the XRD pattern of high angle (10°-80° 2θ) in Figure 2. The material of γ-Al₂O₃ exhibits some broad crystalline peaks around 35°-40° and especially at 45° and 67º 2θ. This is due to the reflections (of X-rays) from the 311, 400, and 440 crystal planes, in relation with the cubic γ-Al₂O₃, according to JCPDS Card No. 10-0425 (^{Renuka, Shijina & Praveen, 2012}). This revealed that the γ-Al₂O₃ presented a low degree of crystallization. However, the XRD pattern for the AA demonstrated a significant number of peaks, which suggests that more than one phase is present. On the other hand, other peaks that do not fit within the phase γ-Al₂O₃ can be indexed as pseudoboehmite (γ-AlOOH) phase (JCPDS Card No. 21-1307) (^{Barzegar & Ebadzadeh, 2011}). Therefore, the AA material may be composed of γ-Al₂O₃ and pseudoboehmite such as in a main phase, as shown in Figure 2 (gray). Due to this, the AA material showed a greater degree of crystallinity compared with γ-Al₂O₃.

Source: Author’s own elaboration

Figure 2 XRD patterns of γ-Al₂O₃ (gray) and AA (black), JCPDS cards: Pseudoboehmite (21-1307), and γ-Al₂O₃ (10-0425). 

Adsorption/desorption of N₂

Figure 3 illustrates the N₂ adsorption-desorption isotherms and the Barrett-Joyner-Halenda (BJH) pore size distribution curve of γ-Al₂O₃ and AA. According to the classification made by IUPAC (^{Sing et al., 1985}), the N₂ adsorption-desorption isotherms of γ-Al₂O₃ and AA are found to be of a type IV. Two well distinguished regions in the adsorption isotherm are evident: (i) monolayer-multilayer adsorption and (ii) capillary condensation. The first increase in adsorption at relative pressure P/P0 < 0.2 is due to a multilayer adsorption on the surface, while the second increase at P/P0 = 0.42-0.97 arises from capillary condensation in the mesopores with nitrogen multilayers adsorbed on the inner surface (^{Gregg, Sing & Salzberg, 1967}). These materials show irregular shape isotherms with the hysteresis loop type E, which indicate that the pores in the material have an inkwell-type shape. As seen in Figure 3b, a relatively narrow pore size distribution is detected, and the average BJH pore diameter, calculated by using the adsorption branch data of the N₂ adsorption-desorption isotherms, is about 5.8 nm; this measure belongs to the γ-Al₂O₃, and 3.5 nm is the measure for the AA. These values are congruent due to the higher crystallinity that AA presents, as shown in Figure 2. Table 1 shows the results of the surface area BET, the pore size distribution, as well as the average pore diameter of γ-Al₂O₃ and AA.

Source: Author’s own elaboration.

Figure 3 Graphs a) N₂ adsorption-desorption isotherm and b) pore size distribution curve of the γ-Al₂O₃ and AA. 

Morphology and surface composition of the adsorbents

Figure 4 shows the FE-SEM images of γ-Al₂O₃ and AA. The powder from the γ-Al₂O₃ obtained through the process of hydrolysis-precipitation and agglomerates consists of nanofibers, which are about 100 nm in length. That tends to form amorphous agglomerates due to their high surface energy. Therefore, high porosity is generated. In comparison with the morphology of the commercial AA powder (Figure 4b), the powder exhibits a higher degree of agglomeration in the surface and, consequently, it has a lot of small pores.

Source: Author’s own elaboration.

Figure 4 FE-SEM/EDS images of a) γ-Al₂O₃ and b) AA. 

The FE-SEM images of γ-Al₂O₃ and AA (Figure. 4) show the spots in places where the electron beam determined the elemental EDS (energy dispersive X-ray spectroscopy) through the analysis. The EDS is employed to identify the elemental compositions from both alumina materials. This technique can be applied to investigate the properties of a broad range of chemicals. Elemental concentrations of aluminum and oxygen are investigated using EDS at one point from both samples. The weight concentrations of both samples are compared in Table 2. The analysis confirms the high purity of the alumina samples. The values obtained in the elemental chemical composition of aluminum and oxygen from both samples are congruent with values reported in the literature related to pure alumina (^{Ilyas, Pendyala, Narahari & Susin, 2017}).

²⁷ Al MAS NMR

Figure 5 displays the ²⁷Al MAS NMR spectra that belong to the γ-Al₂O₃ and AA. The first one is Figure 5a, which exhibits characteristic profiles of γ-Al₂O₃, such as signals between 0 ppm and 5 ppm from the octahedral aluminum Al^VI sites. Moreover, the presence of signals in the region that are around 34 ppm is evident. This can be associated to aluminum in pentacoordinate sites Al^V, and signals between 58 ppm and 62 ppm from the sites of aluminum in tetrahedral geometry Al^IV. The second material (Figure 5b) exhibits signals between 0 ppm and 5 ppm from the octahedral aluminum Al^VI sites, and it signals between 58 ppm and 62 ppm from the aluminum in tetrahedral geometry Al^IV sites (^{Hill, Bastow, Celotto & Hill, 2007}; ^{Li, Liu & Liu, 2013}). For the sample AA, no resonance related to Al^V was found. The integration reveals values between 15% and 20% from the tetrahedral sites, which are in accordance with the γ-Al₂O₃ model (^{Chagas et al., 2014}). The γ-Al₂O₃, as seen in Figure 5a, exhibited aluminum pentacoordinate sites Al^V which, according to the literature, arise from the structure defects from γ-Al₂O₃. Also, a high Al^V amount implies low crystallinity (^{Tsyganenko & Mardilovich, 1996}). The ratio of tetrahedral to octahedral Al is about 1:3, a characteristic from the γ-Al₂O₃ phase (^{Sifontes et al., 2014}).

Source: Author’s own elaboration.

Figure 5 ²⁷Al NMR spectra of γ-Al₂O₃ and AA. 

FT-IR pyridine adsorption

The spectra of the pyridine adsorption-desorption were obtained for the γ-Al₂O₃ and AA at a temperature of 25 ºC, as shown in Figure 6. This indicates that only Lewis acid sites were detected, but no Brönsted acid sites were observed from the γ-Al₂O₃. It was possible to observe the IR bands at 1593 cm^-1, 1575 cm^-1, and 1445 cm^-1, which is specific from the vibration mode of pyridine when interacting with medium or weak Lewis acid sites (^{Liu, 2008}; ^{Rana, Huidobro, Ancheyta & Gómez, 2005}). Besides the already mentioned bands, other band was observed at 1615 cm^-1, associated at pyridine adsorbed on Lewis acid sites. On the other hand, for the adsorbent AA, a band at 1489 cm^-1 associated with pyridine adsorbed on both Lewis and Brönsted acid sites was observed (Li et al., 2013; ^{Xue, Li, Wang, Liu & Xu, 2010}). The results may be relative to the unsaturated alumina atoms (Al^IV and Al^V) and the specific surface alumina areas. It is agreed that the weak Lewis acidity mainly arises from pentacoordinate aluminum atoms while the medium strong Lewis acidity mainly relates to aluminum atoms in tetrahedral geometry (Li et al., 2013). This is corroborated in Figure 5. These features influence directly the surface acidity distribution of γ-Al₂O₃ and AA. As demonstrated in Figure 6, the material AA presents a higher percentage of Brönsted acidity and, according to reported studies, this benefits the fluorine adsorption reaction. That is due to the proton donation.

Source: Author’s own elaboration.

Figure 6 FT-IR spectra of the pyridine adsorption of the γ-Al₂O₃ and AA. 

Zeta potential and particle size

The plot of pH versus zeta potential seen in Figure 7 shows the zeta potential measurement of the γ-Al₂O₃ and AA. Specifically, the adsorption capacity of a material is determined by its chemical composition and distribution in a particle surface charge. Zeta potential (PZ) is widely used for electrical charge magnitude quantification in a double layer (^{Wang & Keller, 2009}). The isoelectric point of AA and γ-Al₂O₃ is found at pH 8.6 and 7.72, respectively. This is consistent with the reported data in the literature. This indicates that the synthesized and agglomerate alumina surface is slightly basic. This result corroborates that the basic sites concentration is slightly higher for the γ-Al₂O₃ than AA, as determined by FT-IR pyridine adsorption (Figure 6). The reported values show that the potential zeta for AA or γ-Al₂O₃ are around pHPZC 8.94 (^{Cheng et al., 2014}). In other researches, different pHPZC values were reported: 8.3 (^{Bowen, Carry, Luxembourg & Hofmann, 2005}), 8.5 (^{Zamorategui, Ramírez, Martínez & Serafín, 2016}), 8.0 for a commercial AA (^{Ku & Chiou, 2002}), and (^{Choi & Chen, 1979}) reported that the pHPZC, which belongs to the adsorbents AA, ranged from 6.2 to 8.9. Therefore, the PZ pertaining to synthesized and agglomerate alumina and commercial AA, which was employed in this work, was within the range of values reported in the literature. These potential changes are related to the flocculation phenomena and the neutralized surface charge. The differences in the values of IEP are attributed to the acidity from both materials. This acidity depends on the coordination of aluminum atoms, as observed in Figure 5; this is due to change in surface charge. The AA presented a higher acidity due to the presence of Lewis and Brönsted acid sites, in comparison with the γ-Al₂O₃. This can be observed in the FT-IR spectra of the pyridine adsorption. Figure 7b shows the size of particles that belongs to the γ-Al₂O₃ (0.27 µm) and AA (0.5 µm). The particle size of γ-Al₂O₃ is smaller than that of the AA. This is due to a combination of phases, such as gamma-alumina and pseudoboehmite as the main phase. Their entirety was not affected by dehydration and agglomeration of crystals.

Source: Author’s own elaboration.

Figure 7 Graphs of a) potential zeta, b) particle size of the γ-Al₂O₃, and AA. 

Adsorption isotherms of fluoride

Adsorption depends on ions in fluid (adsorbate), which diffuses through the solid surface (adsorbent). In this surface, they bond or are held there by weak intermolecular forces (^{Li et al., 2011}). Adsorption studies pointed out the most important characteristics which determined adsorbent suitability for practical applications: the adsorption capacity, selectivity for fluoride ions, regenerability, compatibility, particle and pore size, and low cost. It is important to consider that fluoride removal efficiency always depends on a raw water quality profile, i.e., initial fluoride concentration, pH, temperature, contact time, and adsorbent dosage (^{Bhatnagar et al., 2011}; ^{Mohapatra, Anand, Mishra, Giles & Singh, 2009}; ^{Rao, 2003}; ^{Tomar & Kumar, 2013}). The Langmuir adsorption model is the most common model used to quantify the amount of adsorbate on an adsorbent as a function of partial pressure or concentration at a given temperature. Adsorption isotherms for the γ-Al₂O₃ and AA are shown in Figure 8, and these go in accordance with the Langmuir equation which is represented as:

In this equation, q _e (mg • g-1) is the solution that was adsorbed by the surface, ^b is the constant equilibrium of adsorption, and q _m is the capacity of adsorption in a saturated single layer. Finally, C _e (mg • l-1) is the solution in equilibrium state (^{Rahimi & Vadi, 2014}).

Source: Author’s own elaboration.

Figure 8 Adsorption isotherms of fluoride of γ-Al₂O₃ and AA, at a) 25 ºC and b) 35 ºC, at pH different. 

Figure 8 illustrated the adsorption isotherms of γ- Al₂O₃ and AA. Figure 8a shows the adsorption isotherms of F− in the γ-Al₂O₃ and AA at 25 ºC, at pH 5, 7, and 9. The results indicate a higher capacity of adsorption from γ-Al₂O₃ than AA at pH 5, approximately 15 mg of F-/g and 11 mg of F-/g, respectively. Figure 8b shows the adsorption isotherms of F− from the γ-Al₂O₃ and AA at 35 ºC, at pH 5, 7, and 9. These results indicate a slightly higher capacity of adsorption for γ-Al₂O₃ than for AA at pH 5. That is approximately 12 mg in F-/g and 11 mg in F-/g, respectively. The best-fit parameters are listed in Table 3 and 4. The adsorption capacity results for γ-Al₂O₃ at pH 7 are approximately 10 mg of F-/g and 9 mg of F-/g for AA at 25 ºC. Moreover, at 35 ºC, the adsorption capacity of 8 mg in F-/g and 9 mg in F-/g was revealed. The qmax value for AA was 16.7 mg/g. The γ-Al₂O₃ with a q_max (27.5 mg/g) had a much higher level than the AA at 25 ºC. The qmax value for AA was 15.3 mg/g. The γ-Al₂O₃ presented a q_max (23.3 mg/g), which had a much higher level than the AA, at 35 ºC and at pH 5. In this study, solutions of 50 ppm of fluoride ion were used.

The results indicated that the commercial agglomerate AA presented a higher crystallinity than the γ-Al₂O₃, synthesized and agglomerate. This is attributed to what is shown in a diffractogram from a hybrid material that belongs to a γ-Al₂O₃ and a pseudoboehmite as a main phase (> %). On the other hand, the adsorption of fluoride ions occurs on the surface of the adsorbent material. Adsorption process is controlled by the diffusion of the fluoride ion in the liquid, and the fluoride ion must diffuse into the adsorbent agglomerate. This is the reason why the amount of fluoride ion adsorbed is less for the AA than for the γ-Al₂O₃, since the phenomenon is influenced by the size of the pores through which the fluoride ion is transported. The fluoride was adsorbed in the active sites of the activated alumina that are formed in the aluminum atoms of the alumina framework. No adsorption studies at a lower pH were explored due to the loss of active sites. The reason for this is the dissolution of aluminum present in the alumina.

The capacity of the alumina materials for adsorbing the fluoride was considerably dependent of the solution pH. The effect of the solution pH in the adsorption capacity can be ascribed to the electrostatic or coulombic interactions between the alumina surface and the fluoride ions in the aqueous solution. The γ-Al₂O₃ and AA surface was charged positively at a pH level lower than the PZC of 7.72 and 8.6, respectively. Hence, the fluoride ion was attracted to both materials’ surface. On the contrary, the surface was negatively charged at a pH level higher than 7.72 and 8.6. Hence, the fluoride ion was repelled from the surface.