Synthesis and Characterization of Boehmites Obtained from Gibbsite in Presence of Different Environments

Boehmite (γ-AlOOH), one of the two polymorphs of aluminum oxyhydroxide (the other one is diaspore, α-AlOOH), is a material useful in several commercial applications, for example, ceramics, abrasive materials, fire-retardants, adsorbents, catalysts and fillers for polymeric composites 1. It can also be used as raw material for producing transition aluminas (mainly γ -Al2O3) with a large range of applications in catalysis and adsorption, and α-Al2O3 for the production of ceramic materials (used in a broad range of applications, for instance mechanical parts, refractory materials and electric insulators) 2. To date, microand nano-boehmite crystals with different morphologies have been described 3 : crystals of one dimension larger than the others (1D crystals), crystals with two comparable dimensions, both larger than the third (2D crystals), and crystals with sizes comparable in the three dimensions (3D crystals). 1D materials includes fibers, spindles, needles, rods, wiskers and tubes 3-15. Platelets, leaves and all kind of lamellar materials can be labeled as 2D materials 3,6,16-21. 3D materials comprise crystals with near cubic or spherical morphologies 22-23 , as well as three-dimensional assemblies of 1D or 2D crystals which receive several names as “urchin-like”, “canteloup-like”, “flower-like”, “rosette-like”, “hollow microspheres” and “three-quarte-sphere-like” structures 24-32. Boehmite can be prepared by a solid state thermal transformation of gibbsite 2,33-39. Water in vapor form is released by heating gibbsite particles to 380K to around 580K 34. If favorable conditions exist for the water vapor formed in the heating process to be retained within the gibbsite particles (coarse gibbsite particles; high heating rate; previous presence of water vapour in the system), boehmite is formed. In the absence of these conditions, gibbsite could transform directly into a transition alumina (χ-alumina) 34,35,39. This mechanism, however, is still matter of discussion: for example, to Mercury et al. 36, gibbsite decomposes thermally to boehmite regardless of gibbsite particle size 36. Boehmite can also be synthesized from liquid phase by hydrothermal/solvothermal routes 9,11,27,40-52 or by sol-gel and precipitation techniques 8,53-59. Experimental conditions such as aluminum source, nature of the precipitating agent, pH, presence and nature of organic additives, temperature and time of thermal treatment (hydrothermal/solvothermal or thermal aging) may have a main influence in the resultant boehmite crystallite size and morphology 60. In most studies dealing with synthesis in liquid phase of micro or nano boehmites, aluminum sources are inorganic salts (the most common are nitrate and chloride) or organic compounds such as alkoxides. As mentioned in literature 16, 20,61, pH may significantly affect the morphology of boehmite crystals, with elongated shapes (1D morphologies) being favored in syntheses carried out in acid medium. Increasing pH (neutral or alkaline media) favors the formation of crystals with lamellar morphology (2D) or 3D near cubic or spherical morphologies. There are few studies reporting the production of boehmite from hydrothermal treatment of commercial Bayer gibbsite 46,62 . To the authors’ knowledge, there are no reports published in the scientific literature dealing with hydrothermal transformation of a Bayer commercial gibbsite in boehmite in aqueous acidic medium. In a previous study conducted in our laboratory by Rocha 63, boehmite was prepared from Synthesis and Characterization of Boehmites Obtained from Gibbsite in Presence of Different Environments


Introduction
Boehmite (γ-AlOOH), one of the two polymorphs of aluminum oxyhydroxide (the other one is diaspore, α-AlOOH), is a material useful in several commercial applications, for example, ceramics, abrasive materials, fire-retardants, adsorbents, catalysts and fillers for polymeric composites 1 . It can also be used as raw material for producing transition aluminas (mainly γ -Al 2 O 3 ) with a large range of applications in catalysis and adsorption, and α-Al 2 O 3 for the production of ceramic materials (used in a broad range of applications, for instance mechanical parts, refractory materials and electric insulators) 2 .
Boehmite can be prepared by a solid state thermal transformation of gibbsite 2, [33][34][35][36][37][38][39] . Water in vapor form is released by heating gibbsite particles to 380K to around 580K 34 . If favorable conditions exist for the water vapor formed in the heating process to be retained within the gibbsite particles (coarse gibbsite particles; high heating rate; previous presence of water vapour in the system), boehmite is formed. In the absence of these conditions, gibbsite could transform directly into a transition alumina (χ-alumina) 34,35,39 . This mechanism, however, is still matter of discussion: for example, to Mercury et al. 36 , gibbsite decomposes thermally to boehmite regardless of gibbsite particle size 36 .
In most studies dealing with synthesis in liquid phase of micro or nano boehmites, aluminum sources are inorganic salts (the most common are nitrate and chloride) or organic compounds such as alkoxides. As mentioned in literature 16,20,61 , pH may significantly affect the morphology of boehmite crystals, with elongated shapes (1D morphologies) being favored in syntheses carried out in acid medium. Increasing pH (neutral or alkaline media) favors the formation of crystals with lamellar morphology (2D) or 3D near cubic or spherical morphologies.
There are few studies reporting the production of boehmite from hydrothermal treatment of commercial Bayer gibbsite 46,62 . To the authors' knowledge, there are no reports published in the scientific literature dealing with hydrothermal transformation of a Bayer commercial gibbsite in boehmite in aqueous acidic medium. In a previous study conducted in our laboratory by Rocha 63 , boehmite was prepared from a commercial Bayer gibbsite by an hydrothermal treatment carried out at 160 o C in distilled water and in aqueous acidic solutions (using acetic or nitric acid), with reaction times ranging between 24h and 168 h. At that temperature, full conversion of gibbsite into boehmite by hydrothermal treatment conducted in the presence of acetic acid was not observed. In the case of hydrothermal treatments in a medium containing nitric acid it was not observed by XRD the presence of residual gibbsite after 120h treatment indicating that the conversion was complete. SEM images obtained in all syntheses (with both neutral and acidic media) indicated that the gibbsite undergoes a process of dissolution -the phase transition of gibbsite into boehmite under hydrothermal conditions seems to occur by a dissolution-precipitation mechanism, as suggested by Egorova et al 62 . Diamond-like tricky near "3D" (in both pure water and nitric acid solution media) and lamellar "2D" (in acetic acid solution medium) morphologies of boehmite crystals were obtained. In the lamellar "2D" crystals, the thickness was in the nanoscale. These morphologies were distinct from that of the original gibbsite particles (which also shows a "3D" morphology).
It has been suggested in Rocha's mentioned work 63 that the presence of acetate ion in the aqueous reaction medium would be essential for obtaining "2D" morphology. In the literature, the only paper found indicating an effect of acetic acid as a directing morphology agent is the work of Zhan et al. 64 . However, what has been studied in this work was not an hydrothermal conversion of gibbsite to boehmite but an hydrothermal aging process of a commercial nano-boehmite.
This paper deepens Rocha's study, assessing how acidic pH and the presence of acetate ion in the reactional medium affect the characteristics of produced boehmites crystals.

Materials
Commercial gibbsite (Hydrogard GP -ALCOA) has the following chemical composition: 65.2% Al 2 O 3 ; 0.017% SiO 2 ; 0.013% Fe 2 O 3 ; 0.34% Na 2 O; 34.2% LOI (1000 o C). Gibbsite powder consists of platelet aggregates having a density of 2.42 g/cm 3 . All the other reagents (potassium acetate and glacial acetic acid supplied by Brazilian companies; 65% nitric acid from Sigma Aldrich) had commercial grade and were used as received.

Methods
All materials were produced by hydrothermal processes (autogenous pressure at 160 o C) differenced by the number and type of reagents added to the starting gibbsite, as listed below: • process 1: in pure deionized water medium (molar ratio: Processes 3 and 4 were carried out to deepen the observation made in previous work done in our laboratory by Rocha 63 , according to which the acetate ion present in the reaction medium would be essential to obtain particle morphology with at least one nanometric dimension (morphologies "2D" or "1D").
All preparations were performed in intervals of 72h or 168h, in 350mL maximum capacity autoclaves made in steel with an inner PTFE cladding (manufactured in our laboratory). The properties of PTFE used for the construction of our equipment limited its operating temperature -we observed a deformation of the cladding above 160 o C leading to leaks that prevented the establishment of hydrothermal conditions in the synthesis process.
Cooling after hydrothermal treatment occurred naturally to room temperature. The suspensions containing potassium acetate were centrifuged to remove the KAc from the solution prior to air drying due to hygroscopicity of KAc. This procedure was repeated three times, replacing the supernatant solution of each centrifugation by deionized water The identification of the samples prepared in this study, as well as the initial pH of the original suspensions of the hydrothermal processes, are listed in Table 1.

Thermal Analysis (Gravimetric and Differential -TGA and DTA)
Samples prepared in the same way as the samples used for XRD were analysed in a Shimadzu DTG-60H equipment -Simultaneous TG-DTA (NUPEGEL-USP), operated in heating and cooling cycle from room temperature to 1100°C with heating rate of 10°C min -1 and N 2 atmosphere.

Scanning Electron Microscopy (SEM)
SEM images were obtained on two equipments: a JEOL model JSN 5600 LV, operating at 15 kV and 20 kV (NUPEGEL-USP), and a FEI model Inspect F50 operating at 5kV (PMT-EPUSP). In the case of images obtained on JEOL equipment, powder samples have been previously dispersed in deionized water using an ultrasonic device, dried directly over metal supports and coated with gold. For the SEM images took in the FEI equipment, dry powders were deposited on metallic sample holders covered with a double-sided carbon tape, receiving then a gold conducting coating.
Boehmite was present in all powders produced. A 72h hydrothermal reaction at 160 o C was insufficient for complete conversion of gibbsite to boehmite in most cases -the exception was the material produced in aqueous solution of potassium acetate (sample f). There is no clear explanation for this exception with the available data in this moment. Literature reports that the boehmite rate formation in hydrothermal processes increase over pH 5.0 due to the enhancement of dissolution-reprecipitation 20 . This could explain the complete transformation of gibbsite in boehmite observed in sample f, as the initial pH of this preparation was 8.6 (see Table 1; reaction time = 72h, acetate ion present in the hydrothermal medium). However, if the complete transformation was due only to the alkaline pH of the starting suspension, the same should be observed for sample b, which had an initial pH of 9.3, and this did not occur: in this sample (reaction time = 72h; pure deionized water medium) the total transformation of gibbsite in boehmite was not observed.
A reaction time of 168h at 160 o C was sufficient for complete conversion of gibbsite to boehmite for most cases. By the XRD data, the exception was the material produced in an aqueous solution of potassium acetate acidified to pH 2.0 with nitric acid (sample i): a small peak corresponding to (002) gibbsite reflection was observed in the diffractogram of this sample. The interpretation of this exception is not yet clear. In the literature, Phambu 65 mentions that the acetate anion can adsorb on the surface of gibbsite forming a complex linking two Al atoms, called a "bridging complex". It is reasonable to assume that this absorption could occur in our case and could reduce the rate of formation of the boehmite by slowing the gibbsite dissolution process. Therefore, the hydrothermal reaction time of 168h at 160 o C would be insufficient to promote the full conversion of gibbsite. A similar effect of reducing boehmite formation rate caused by the presence of tartrate anion in the hydrothermal medium is mentioned by Pardo et al. 20 . However, if the presence of the acetate anion in the reaction medium was the only explanation for slowing down the boehmite formation, the same behavior would be observed also in the materials produced by the processes 2 and 3. Nevertheless, in samples f-g (process 3) it was exactly the opposite: the complete transformation of gibbsite to boehmite has been observed after a 72h hydrothermal reaction. In these samples, however, the initial pH was alkaline, which as mentioned above tends to accelerate the gibbsite dissolution rate.
Sharper peaks (with smaller peak width at half height) were observed in the XRD patterns of materials produced from gibbsite suspensions with initial alkaline pH (processes 1 and 3, respectively b-c and f-g samples) compared to those observed in materials produced with initial acidic pH condition (processes 2 and 4, respectively d-e and h-i samples). This observation indicates the presence of larger crystallites in those materials as might be expected due to the enhancement of dissolution-reprecipitation in alkaline pH.
The lack of a simple and consistent explanation for some of our experimental results indicates that the studied system is complex, requiring more detailed studies to understand the mechanism of boehmite formation by hydrothermal treatment of gibbsite in aqueous medium in the presence of acetate anion.

Thermal Analysis (Gravimetric -TGA and Differential -DTA )
DTA curves of the hydrothermal materials and of the original gibbsite are shown in Figure 2. DTA curve of the original gibbsite is consistent with what is presented in the literature 34,66-67 an intense endotherm around 300°C-320 o C and a second endotherm, less intense, above 500°C. The first endotherm probably corresponds mainly to the thermal transformation of gibbsite to boehmite, as the gibbsite crystals are coarse and so favorable conditions to gibbsite → boehmite transformation exist, as previously mentionned. However, the conditions under which the thermal analysis was performed do not allow to exclude the possibility of formation of χ-alumina 34,35,39 . The second endotherm corresponds to the dehydroxylation of boehmite thermally formed to transition alumina (γ-alumina) 34,67 .
DTA results corresponding to different materials prepared are consistent with XRD results. In most materials produced with a reaction time of 72 h (samples b, d and h) the conversion of gibbsite to boehmite was not complete, as evidenced by the presence of an endotherm around 300 o C ( Figure 3 a-b). In the materials where high gibbsite to boehmite conversion was observed by XRD (samples c, e, f, g and i)  Table 1.
almost only the second endotherm (that corresponds to the dehydroxylation of boehmite) was observed. It is interesting to note that in the case of the sample i where the presence of gibbsite is still observed by XRD it was not observed corresponding endotherm around 300 o C. In contrast, for the sample e, where the presence of residual gibbsite was not observed by XRD, a small endotherm was observed with maximum around 280 o C, which may indicate the presence of residual gibbsite. These two observations indicate that the experimental techniques has limitations, and it is always important for the characterization of a sample a joint analysis of the results of more than one technique. TGA data of the original gibbsite shows a weight loss of 35.2%, close to the theoretical value. The difference between the theoretical and experimental values can be attributed to moisture present in the original gibbsite.
Estimates of the relative amounts of gibbsite and boehmite, calculated using as a basis the theoretical mass  Table 1. (All images from JEOL equipment, NUPEGEL-USP) loss of these two phases (34.6% for gibbsite and 15.0% for boehmite), assuming that only these two phases are present in all materials produced, are shown in Table 2.
In materials where intense reflections of gibbsite were observed by XRD (samples b, e and h), the weight loss is greater below 400 o C, which is consistent with the existence of significant amounts of residual gibbsite (not hydrothermally transformed) in these samples. Also in these same materials (samples b, e and h), the first endotherm (which corresponds to the thermal decomposition of gibbsite) was always be observed at temperatures around 15 o C lower than the temperature of the same endotherm in the original gibbsite. As discussed below (in the presentation and discussion of the SEM results), the gibbsite undergoes a dissolution during the hydrothermal process, which could increase the reactivity of the residual gibbsite particles, resulting in a decrease in it thermal transformation temperature.
In two materials (samples e and i), where the existence of residual gibbsite was not clearly observed by XRD and DTA, the partial weight loss values below 400 o C could indicate its presence.
In the remaining materials (samples c, f and g) the partial weight loss values below 400 o C are quite insignificant. The partial mass weight values above 400 o C are very close to the theoretical value of dehydroxylation of boehmite. These results confirm the results of XRD and DTA: in these materials the transformation of gibbsite to boehmite was almost complete.
Regarding the temperature of the second endotherm (related to dehydroxylation of boehmite), the produced materials can be divided in two groups ( Table 2). The first group, comprising the samples b, c, f and g have the temperature of this endotherm around 540 o C, close to the temperature observed in the original gibbsite (520 o C, sample a; decomposition of boehmite formed by thermal transformation of gibbsite). The second groupsamples d, e, h and i -show endotherm temperatures around 490 o C. As will be seen below (SEM images), this difference may be related to the morphology of boehmite crystals: in this second group of samples, boehmite crystals have elongated ("near 1D") shape. Thermal transformation of boehmite into γ-alumina could be favored by this morphology. Figure 3 shows a typical image of the particles found in the original Bayer commercial gibbsite. This material has particles that have an hexagonal and/or pseudohexagonal shape often with some irregular contours, indicating that the material could have undergone a grinding in their fabrication process ( Figure 3A).

Scanning Electron Microscopy (SEM)
The other three images shown in Figure 3 refers to particles observed in materials produced by hydrothermal treatments during 48h and 72h at different initial conditions, as follows: Figure 3B: sample d; Figure 3C: sample h; Figure 3D: same initial conditions of the sample b, but with synthesis time of only 48h. The identification of the synthesis conditions of these samples are listed in Table 1. In these images we can see that the particles of the original gibbsite undergo a dissolution process, as evidenced by the presence of "holes" in gibbsite particles (no matter the composition of the initial solution). Gibbsite particles presenting these "dissolution figures" were observed for all the materials produced by hydrothermal synthesis for at least 48 h treatments.
Gibbsite particles undergoing dissolution process with neoformed boehmite crystals on its surface have been observed. Those particles coexist in the same samples with boehmite crystal agglomerates that seems to be the result of the complete transformation of other gibbsite particles ( Figure 3B). In other words, in some gibbsite particles mainly dissolution occurs, while in others it may happen simultaneously dissolution of gibbsite and boehmite recrystallization on the surface of dissolving gibbsite particles. With the data we have so far we cannot explain these observations. Boehmite is formed by a recrystallization process, regardless of their morphology ("3D" particles for sample b, Figure 3D; particles with elongated shape (near "1D") for samples d and h, Figures 3B and 3C). Recrystallization can occur either on the surface of gibbsite particles that are dissolving, or in the form of single crystals separated from the gibbsite particles. This was observed in all the preparations, regardless of the initial condition. So, the transformation of gibbsite into boehmite under hydrothermal conditions by a dissolution-precipitation mechanism, as suggested by Egorova 62 , seems to apply.
The morphology of the formed boehmite crystals is dependent on the initial conditions of hydrothermal synthesis (Figure 4). Two typical morphologies were observed according to the pH of the initial suspension of the hydrothermal process. Hydrothermal processes from suspensions of initial alkaline pH (processes 1 and 3) resulted in boehmite particles with morphologies "3D" (Figure 4A, cubes, process 1; samples b and c) or "2D" (Figure 4C, thick plates, process 3, samples f and g), while processes where the initial suspension had an acidic pH (processes 2 and 4) resulted in elongated shape (near "1D") boehmite particles ( Figure 4B, process 2, samples d and e; Figure 4D, process 4, samples h and i). This observation is consistent with that reported by Pardo et al. 10,20,47 . In these works, in which boehmite is produced from xerogels treated by hydrothermal processes, elongated particles (elongated shape attributed to the increase in the proportion of (100) and/or (001) lateral faces 55 ) are obtained in processes conducted in pH 5.0 aqueous medium, while processes carried out in aqueous medium at alkaline pH (pH = 10.0) lead to "platelet-like" particles.
The hypothesis that presence of acetate ion in the reaction medium was sufficient to produce boehmite crystals with morphologies with one or two dimensions in nanometer range (thin "2D" or "1D" particles) suggested by Rocha 63 was not confirmed since the hydrothermal process 3 (aqueous solution of potassium acetate at its natural alkaline pH) resulted in thick lamellar boehmite particles ( Figure 4C). However, apparently the presence of acetate ion has led to a decrease in thickness of the boehmite crystals formed comparing to the thickness of the crystals prepared in deionized water medium, which has a "3D" morphology ( Figure 4A).
The presence of acetate ion in an acidic reactional medium (processes 2 and 4) leads to the formation of elongated particles with nanometer scale thickness ( Figures 4B and 4D). This morphology is achieved by the combined effect of acidic pH of the starting suspension and the presence of acetate anion in the reactional medium.
Additional studies realized recently in our laboratory show that only acidic pH would not be sufficient to produce elongated crystals, as can be seen in Figure 5, which shows an agglomerate of boehmite crystals produced by a hydrothermal process performed in a medium of pure deionized water acidified with nitric acid (molar ratios: 1 mol Al : 50 mols H 2 O; pH = 2.0, adjusted with HNO 3 ). Boehmite crystals have the form of thin plates with nanometric thickness, not the form of diamond-like tricky "near 3D" crystals as observed in the previous work of Rocha 63 . To date we have no adequate explanation to this discrepancy.  Table 1. (Images A and B from FEI equipment, PMT-USP; Images C and D from JEOL equipment, NUPEGEL-USP))

Conclusions
In this article a procedure to obtain boehmite crystals with different shapes by a hydrothermal process conducted at 160 o C using a Bayer commercial gibbsite as raw material is reported. The morphology of the boehmite crystals is affected by the factors studied in this work, namely pH and presence of the acetate anion in the reaction medium.

References
The absence of the acetate anion in the reaction medium at initial alkaline pH favors the formation of crystals with "near 3D" dimensions. The presence of this anion in an initial alkaline medium leads to the formation of thick plates (near "2D" crystals). The presence of acetate anion in the reaction medium with an initial acid pH leads to boehmite crystals with elongated shape (near "1D" dimension crystals).