Process for the preparation of high surface area alpha alumina and the use thereof

Abstract

The present invention refers to a process for the preparation of a high surface area nanoparticulate alpha alumina.

Claims

1. A method comprising: (a) preparing a high-surface area nanoparticulate alpha alumina, having a BET surface area of at least 130 m.sup.2/g and having a particle size in the range of 1 to 50 nm measured by TEM, by a process comprising subjecting a precursor consisting of (A) -AlOOH.Math.xH.sub.2O with x in the range of 0.1x0.5 and optionally (B) a metal and/or a metal compound to a milling process in a ball mill with a milling jar and balls in a weight ratio of balls to said -AlOOH.Math.xH.sub.2O of 1 to 200 in a temperature range below the conversion temperature of nanocrystalline -AlOOH.Math.xH.sub.2O to -Al.sub.2O.sub.3 to yield said high-surface area nanoparticulate alpha alumina as a milling product; and (b) using the high-surface area nanoparticulate alpha alumina prepared as catalyst or catalyst support for metal catalysts or in ceramics applications.

Description

(1) The invention is further illustrated by the following Figures. In the Figures, it is shown:

(2) FIG. 1: PXRD patterns of (a) boehmite of d.sup.Scherrer=17 nm (denoted as -AlOOH-17), (b) -AlOOH-17 after 3 h ball milling, (c) -AlOOH-17 after 12 h ball milling, and (d) after post-calcination of 3 h ball milled -AlOOH-17 at 550 C. in static air for 10 h. The crystalline phases identified in the samples are indicated on the corresponding diffractograms (B=boehmite, -AlOOH, ICDD PDF 21-1307; D=diaspore, -AlOOH, ICDD PDF 05-0355; =corundum, -Al.sub.2O.sub.3, ICDD PDF 46-1212). Milling conditions: Vibration mill with WC jar and WC balls, milling time=3-12 h, bpr=40.5, f.sub.mill=25 Hz.

(3) FIG. 2: (a) TEM, (b) particle size distribution, and (c) HRTEM of nanocrystalline alpha alumina obtained by 3 h ball milling of -AlOOH-17 followed by post-calcination at 550 C. in static air for 10 h.

(4) FIG. 3: PXRD patterns of the boehmite sample described in the caption of FIG. 1(a) before and (b) after ball milling under the same conditions as in the caption of FIG. 1, except at lower bpr of 27. The crystalline phase identification is as provided in the caption of FIG. 1.

(5) FIG. 4: PXRD patterns of the boehmite sample described in the caption of FIG. 1(a) before and (b) after ball milling under the same conditions as in the caption of FIG. 1, except at lower bpr of 27.6 and using steel milling balls. The crystalline phase identification is as provided in the caption of FIG. 1, except unmarked reflection due to iron-containing phases.

(6) FIG. 5: PXRD patterns of the boehmite sample described in the caption of FIG. 1(a) before and (b) after ball milling in planetary mill. The crystalline phases identified in the samples are indicated on the corresponding diffractograms (B=boehmite, -AlOOH, ICDD PDF 21-1307; =corundum, -Al.sub.2O.sub.3, ICDD PDF 46-1212; Fe=metallic Fe, ICDD PDF 06-0696; *=Fe.sub.3O.sub.4, magnetite, ICDD PDF 26-1136). Milling conditions: Planetary mill with steel jar and steel balls, milling time=3 h, bpr=41.4, rpm=650.

(7) FIG. 6: PXRD patterns of (a) boehmite of d.sup.Scherrer=20 nm (denoted as -AlOOH-20), (b) -AlOOH-20 after 3 h ball milling under the same conditions as in the caption of FIG. 1, and (c) 3 h ball milled -AlOOH-20 after post-calcination at 550 C. in static air for 10 h. The crystalline phase identification is as provided in the caption of FIG. 1.

(8) FIG. 7: PXRD patterns of (a) boehmite of d.sup.Scherrer=52 nm (denoted as -AlOOH-52), (b) -AlOOH-52 after 3 h ball milling under the same conditions as in the caption of FIG. 1, and (c) 3 h ball milled -AlOOH-52 after post-calcination at 550 C. in static air for 10 h. The crystalline phase identification is as provided in the caption of FIG. 1.

(9) FIG. 8: PXRD patterns of boehmite of d.sup.Scherrer=4 nm (denoted as -AlOOH-4) (a) before and (b) after 3 h ball milling under the same conditions as in the caption of FIG. 1. The crystalline phase identification is as provided in the caption of FIG. 1.

(10) FIG. 9: Thermogravimetric analysis of boehmite compounds described in Examples 1 through 8. (a) -AlOOH-4, (b) -AlOOH-4 after drying at 140 C. in static air for 12 h (denoted as -AlOOH-4-dried), and (c) -AlOOH-17, -AlOOH-20, and -AlOOH-52.

(11) FIG. 10: PXRD patterns of (a) -AlOOH-4, (b) -AlOOH-4 after drying at 140 C. in static air for 12 h (denoted as -AlOOH-4-dried), (c) -AlOOH-4-dried after 3 h ball milling under the same conditions as in the caption of FIG. 1, and (d) 3 h ball milled -AlOOH-4-dried after post-calcination at 550 C. in static air for 10 h. The crystalline phase identification is as provided in the caption of FIG. 1.

(12) FIG. 11: PXRD pattern of -Al.sub.2O.sub.3 supported cobalt oxide catalyst obtained by ball milling -AlOOH-17 and Co.sub.3O.sub.4 under the same conditions as in the caption of FIG. 1. The crystalline phases identification is as provided in the caption of FIG. 1, except =Co.sub.3O.sub.4, ICDD PDF 42-1467; *=WC, ICDD PDF 51-0939.

(13) FIG. 12: PXRD patterns of (a) -Al.sub.2O.sub.3 derived by calcination of -AlOOH-17 at 550 C. in static air for 5 h (denoted as -Al.sub.2O.sub.3-17) and (b) -Al.sub.2O.sub.3-17 after ball milling under the same conditions as in the caption of FIG. 1. The crystalline phases identified in the samples are indicated on the corresponding diffractograms (=-Al.sub.2O.sub.3, ICDD PDF 50-0741; =corundum, -Al.sub.2O.sub.3, ICDD PDF 46-1212). This figure is of a comparative example.

(14) FIG. 13: PXRD patterns of (a) -Al.sub.2O.sub.3 derived by calcination of -AlOOH-4 at 550 C. in static air for 5 h (denoted as -Al.sub.2O.sub.3-4) and (b) -Al.sub.2O.sub.3-4 after ball milling under the same conditions as in the caption of FIG. 1. The crystalline phases identified in the samples are indicated on the corresponding diffractograms (=-Al.sub.2O.sub.3, ICDD PDF 50-0741; =corundum, -Al.sub.2O.sub.3, ICDD PDF 46-1212). This figure is of a comparative example.

METHODS

Ball Milling

(15) The ball milling experiments were performed using a Retsch Mixer Mill MM400, which involves a horizontal vibration motion at a set frequency (herein referred as vibration mill), having two milling jars (25 cm.sup.3) made of either WC or stainless steel. The material of milling balls used is also either WC (ball diameter, d.sub.ball=1.2 cm) or stainless steel (d.sub.ball=1.5 cm). The ball milling experiments were also performed using a Fritsch Pulverisette, which involves a planetary motion to achieve a high-energy input (herein referred as planetary mill). It is equipped with two milling jars (45 cm.sup.3) made of stainless steel. The material of milling media was also stainless steel. In a typical experiment, the milling jar is filled with a precursor powder(s) and milling balls at a desired bpr (defined in Equation 4) in an open environment and placed in the vibration or planetary mill after properly closing, followed by setting the milling frequency (f.sub.mill) for vibration mill or revolutions per minute (rpm) for planetary mill and duration and starting the milling experiment.

(16) $\begin{array}{cc}\mathrm{bpr}=\frac{\mathrm{total}\mathrm{weight}\mathrm{of}\mathrm{all}\mathrm{balls}\mathrm{used}}{\mathrm{weight}\mathrm{of}\mathrm{powder}}& \left(\mathrm{Eq}.4\right)\end{array}$
where: bpr is ball to powder ratio.

WC Removal from the Ball Milled Sample

(17) The appearance of the sample after ball milling in vibration mill was light gray to gray due to the abrasion of WC vial and balls during milling. The tungsten content in the sample, as determined by EDX analysis, after 3 h ball milling was less than 7 wt.-%. As WC has a strong absorption in PXRD, it complicates the identification of other phases present in the sample. Thus, it was removed from the sample, for the accurate characterization purpose, by adapting a literature method (Archer et al., Journal of Analytical Atomic Spectrometry 2003, 18, 1493-1496). In a typical WC removal method adapted in the present invention, one gram ball milled sample was dispersed in 30 cm.sup.3 solution of 5% HNO.sub.3 in 30% H.sub.2O.sub.2 in a round-bottom (RB) flask. The latter was attached to the condenser, slowly heated to 80 C., and stirred at this temperature for 1-2 minutes. The heating bath is removed and the sample was allowed to stand at room temperature for 30 minutes. The white (whitish) solid was separated by centrifugation and dried at 70 C. for 12 h. This treatment did not cause any change in the crystallinity of the sample or on the textural properties as confirmed by the powder X-ray diffraction and N.sub.2 physisorption analyses of the sample before and after the treatment.

N.SUB.2 .Physisorption

(18) N.sub.2 physisorption at 196 C. was carried out using a Micromeritics 3Flex Surface Characterization Analyzer. The samples were evacuated at 140 C. for 12 h prior to the measurement. The specific surface area (S.sub.BET) was calculated from the adsorption data in the relative pressure range of 0.05 to 0.3 using the Brunauer-Emmett-Teller (BET) method (Brunauer et al., Journal of the American Chemical Society 1938, 60, 309-319).

Thermogravimetric Analysis

(19) Thermogravimetric analysis (TGA) was performed in a NETZSCH STA 449 F3 Jupiter analyzer connected to a NETZSCH QMS 403 D Aolos quadrupole mass spectrometer gas analysis system. Analyses were performed in Ar (40 cm.sup.3 STP per min). About 15 mg of sample was placed in the sintered alumina crucible and the temperature was raised from 40 C. to 900 C. at 10 C./min. Throughout the temperature ramp, atomic mass unit (AMU) 18 (H.sub.2O) was monitored; in addition, full scans over the whole mass range were performed at different temperatures.

Energy-Dispersive X-Ray Spectroscopy

(20) Energy-dispersive X-ray spectroscopy (EDX) analysis was performed on a Hitachi S-3500N instrument. The microscope was equipped with a Si(Li) Pentafet Plus detector from Texas Instruments.

Particle Size Measurement

(21) Transmission Electron Microscopy

(22) Transmission electron microscopy (TEM) and high resolution (HR) TEM images were recorded using a Hitachi HF-2000 microscope with a cold field-emission cathode at maximum acceleration voltage of 200 kV. Samples were prepared by sprinkling dry specimen on the TEM grid consisting of a lacy carbon film supported by a copper grid. Particle size distribution was determined by measuring the diameters of more than 200 particles from several TEM images of the same sample. The number average particle size (d50) was calculated by adding the measured diameters of all particles together and dividing by the number of particles measured. The d50 was used as Sauter mean diameter (SMD, d.sub.32) to estimate the surface area based on particle size using the Equation 5 (Sauter, VDI-Forschungsheft Nr. 279 (1926) and Nr. 312 (1928) ISSN 0042-174X; Wang and Fan, Ch. 2, pp 42-76 in Woodhead Publishing Series in Energy (2013) ISBN 9780857095411).

(23) $\begin{array}{cc}\mathrm{Nanoparticle}\mathrm{surface}\mathrm{area}=\frac{6}{d_{32}.Math.}& \left(\mathrm{Eq}.5\right)\end{array}$

(24) where: d.sub.32 is the Sauter mean diameter, which is defined as an average of particle size, and is the density of the powder.

(25) Powder X-Ray Diffraction

(26) Powder X-ray diffraction (PXRD) was measured using a Stoe STADI P diffractometer operating in reflection mode with Cu K radiation. Data were recorded in the 10-70 2 range with an angular step size of 0.04. The average particle size was also determined from diffractograms (denoted by d.sup.Scherrer) by applying the Scherrer equation (Equation 6) to the three most intense reflections and taking the average of the obtained values. Thus, for boehmite starting materials, reflections centered at 2 14.4, 28.2, and 38.3; and for -Al.sub.2O.sub.3 samples, reflections centered at 2 35.2, 43.3, and 57.4 were used for the Scherrer determination.

(27) $\begin{array}{cc}{d}^{\mathrm{Scherrer}}=\frac{K.Math.}{.Math.\cos }& \left(\mathrm{Eq}.6\right)\end{array}$

(28) where: d.sup.Scherrer is the average size of the crystalline domains, K is a dimensionless shape factor (0.9), is the X-ray wavelength used to irradiate the sample (i.e. in this case of Cu X-rays, it was 1.5406 Angstroms), is the Bragg angle determined from the respective reflection, and is the line broadening at half the maximum intensity (FWHM), determined according to the Equation 7:

(29) $\begin{array}{cc}={\left({}_{\mathrm{sample}}^2-{}_{\mathrm{instrument}}^2\right)}^{0.5}& \left(\mathrm{Eq}.7\right)\end{array}$

(30) where: .sub.sample is the FWHM of the corresponding reflection used to determine the average crystallite size and .sub.instrument is the FWHM of the due to instrument determined using NIST Si standard

(31) The invention is further illustrated by the following Examples.

EXAMPLES

Example 1

(32) Boehmite of d.sup.Scherrer=17 nm (denoted as -AlOOH-17) was obtained from Sasol. One gram of -AlOOH-17 was charged in the WC milling vial together with WC milling balls to achieve a bpr of 40.5. The milling vial was placed in the vibration mill and the ball milling was performed in the closed environment for 3-12 h. After the experiment the powder was removed from the vial and subjected for WC removal, followed by characterization by PXRD, N.sub.2 physisorption, and TEM/HRTEM analysis.

(33) The PXRD analysis (FIG. 1) evidenced the full transformation of the boehmite phase in the precursor (a) after 3 h ball milling (b). Based on PXRD phase analysis, the ball milled sample was composed of more than 70% -Al.sub.2O.sub.3 and the remaining being -AlOOH (diaspore) as can be seen from the relative intensity of the corresponding reflections. The S.sub.BET of the starting boehmite was 89 m.sup.2/g and after 3 h ball milling the S.sub.BET of 125 m.sup.2/g was obtained. The PXRD analysis of -AlOOH-17 sample after 12 h ball milling showed the presence of solely -Al.sub.2O.sub.3 (FIG. 1c). The S.sub.BET of this sample was determined to be 116 m.sup.2/g.

(34) A part of the 3 h ball milled sample after WC removal was subjected to calcination at 550 C. (2 C./min) in static air for 10 h. The calcination fully transformed the remaining diaspore to -Al.sub.2O.sub.3, producing the material with sole desired -Al.sub.2O.sub.3 phase (FIG. 1d). The S.sub.BET of the calcined sample was 120 m.sup.2/g. TEM image of this sample showed rounded particles in the range of 4-32 nm with TEM-based number average particle size, d50, of 13 nm (std. dev.=6 nm) (FIG. 2a, b). Using this d50 in Equation 5, the nanoparticle surface area was calculated to be 118 m.sup.2/g (based on density of alpha alumina of 3.9 g/cm.sup.3). This matches very well with the BET surface area of 120 m.sup.2/g reported above for this sample. Furthermore, application of Scherrer equation to the diffractogram of this sample, according to the procedure described in methods, provided d.sup.Scherrer=18 nm. The high resolution TEM analysis evidenced a crystalline nature of the sample also at the surface (FIG. 2c). Determination of d spacing at different areas provided an average value of 2.079 , which is similar to 2.085 for 113 plane of -Al.sub.2O.sub.3.

Example 2

(35) The ball milling experiment was performed under the same conditions and using the precursor as described in Example 1, except at lower bpr (reduced number of WC balls).

(36) The PXRD analysis evidenced the full transformation of the boehmite phase in the precursor after 3 h ball milling (FIG. 3). Based on PXRD phase analysis, the ball milled sample was composed of -Al.sub.2O.sub.3 as the predominant phase with minor -AlOOH, like in Example 1. The surface area of the ball milled powder was 110 m.sup.2/g.

Example 3

(37) The ball milling experiment was performed under the same conditions and using the precursor as described in Example 1, except at lower bpr and using steel balls.

(38) The PXRD analysis evidenced the full transformation of the boehmite phase in the precursor after 3 h ball milling (FIG. 4). Based on PXRD phase analysis, the ball milled sample was composed of -Al.sub.2O.sub.3 as the predominant phase with minor -AlOOH, like in Example 2. The sample was additional composed of iron-based phases, which originated from the abrasion of steel balls during milling. The content of iron as determined by EDX analysis was 1.5 wt.-%. The surface area of the ball milled powder was 108 m.sup.2/g.

Example 4

(39) The ball milling experiment was performed using the precursor as described in Example 1 in planetary ball mill using milling jar and balls made of steel.

(40) The PXRD analysis evidenced the full transformation of the boehmite phase in the precursor after 3 h ball milling to -Al.sub.2O.sub.3 (FIG. 5). No -AlOOH formation found in this case. The iron content in the sample was 22 wt.-% as determined by EDX. The S.sub.BET after correction for iron content was 105 m.sup.2/g and d.sup.Scherrer=18 nm.

Example 5

(41) Boehmite of d.sup.Scherrer=20 nm (denoted as -AlOOH-20) was prepared by a hydrothermal treatment of aluminum hydroxide (Al(OH).sub.3, Sigma-Aldrich) adapting a literature method (Santos et al., Materials Research 2009, 12, 437-445, Filho et al., Materials Research 2016, 19, 659-668). In a typical synthesis, aluminum hydroxide was dispersed in deionized water in a molar Al:H.sub.2O ratio of 50 in a Teflon-lined autoclave. The hydrothermal nucleation of boehmite was achieved by heating the above-prepared autoclave to 200 C. and holding at this temperature for 72 h under autogenous pressure. The white solid was recovered by centrifugation, followed by drying in flowing air at 70 C. for 12 h. The formation of phase pure boehmite was confirmed by PXRD (FIG. 6a), with d.sup.Scherrer=20 nm.

(42) The above sample was ball milled under the same conditions as in Example 1.

(43) The PXRD analysis evidenced the full transformation of the boehmite phase in this precursor after 3 h ball milling (FIG. 6a, b). Based on PXRD phase analysis, the ball milled sample was composed of more than 70% -Al.sub.2O.sub.3 and the remaining being -AlOOH as can be seen from the relative intensity of the corresponding reflections. The S.sub.BET of the starting boehmite was 64 m.sup.2/g and after ball milling the S.sub.BET of 115 m.sup.2/g was obtained.

(44) Like in Example 1, a part of the ball milled sample after WC removal was subjected to calcination at 550 C. (2 C./min) in static air for 10 h. The calcination fully transformed the remaining diaspore to -Al.sub.2O.sub.3, producing the material with sole desired -Al.sub.2O.sub.3 phase (FIG. 6c). The S.sub.BET of the calcined sample was 136 m.sup.2/g and d.sup.Scherrer=18 nm.

Example 6

(45) Boehmite of d.sup.Scherrer=52 nm (denoted as -AlOOH-52) was prepared by a hydrothermal treatment of aluminum hydroxide hydrate (Al(OH).sub.3.Math.xH.sub.2O, Sigma-Aldrich) using the equivalent procedure as used in Example 5. The formation of phase-pure boehmite was confirmed by PXRD (FIG. 7a), with d.sup.Scherrer=52.

(46) The above sample was ball milled under the same conditions as in Example 1.

(47) The PXRD analysis evidenced the full transformation of the boehmite phase in this precursor after 3 h ball milling (FIG. 7a, b). Based on PXRD phase analysis, the ball milled sample was composed of more than 70% -Al.sub.2O.sub.3 and the remaining being -AlOOH as can be seen from the relative intensity of the corresponding reflections. The S.sub.BET of the starting boehmite was 10 m.sup.2/g and after ball milling the S.sub.BET of 103 m.sup.2/g was obtained.

(48) Like in Example 1, a part of the ball milled sample after WC removal was subjected to calcination at 550 C. (2 C./min) in static air for 10 h. The calcination fully transformed the remaining diaspore to -Al.sub.2O.sub.3, producing the material with sole desired -Al.sub.2O.sub.3 phase (FIG. 7c). The S.sub.BET of the calcined sample was 130 m.sup.2/g and d.sup.Scherrer=19 nm.

Example 7

(49) Boehmite of d.sup.Scherrer=4 nm (denoted as -AlOOH-4) was obtained from Sasol. This sample was ball milled under the same conditions as detailed in Example 1. After the experiment the powder was removed from the vial and subjected for WC removal, followed by characterization by PXRD.

(50) The PXRD analysis (FIG. 8) evidenced that the boehmite is not fully transformed after 3 h ball milling. The ball milled sample was composed of three phases, unconverted boehmite precursor, intermediate diaspore phase, and -Al.sub.2O.sub.3.

Example 8

(51) The boehmite precursor in Example 7 (i.e. -AlOOH-4) was subjected to drying at 140 C. in static air for 12 h. The obtained sample was denoted as -AlOOH-4-dried. This drying step enabled to reduce the additional water content in -AlOOH-4 as this might have effect on the efficiency of milling as observed in Example 7. The TGA analysis of the dried sample evidenced a reduced loss of additional water (7 wt.-%, see shaded area in FIG. 9b) compared to 13 wt.-% in the undried sample (FIG. 9a). The content of additional water in -AlOOH-17, -AlOOH-20, and -AlOOH-52 was up to 3 wt.-% (FIG. 9c). The drying procedure did not change the crystallinity of the sample as evidenced by its PXRD (FIG. 10, see a, b).

(52) The dried sample was ball milled under the same conditions as detailed in Example 1. After the experiment, the powder was removed from the vial and subjected for WC removal, followed by characterization by PXRD.

(53) The PXRD analysis evidenced the full transformation of -AlOOH-4-dried after 3 h ball milling (FIG. 10c). Based on PXRD phase analysis, the ball milled dried sample was composed of more than 70% -Al.sub.2O.sub.3 and the remaining was -AlOOH as can be seen from the relative intensity of the corresponding reflections. The S.sub.BET of the starting boehmite was 365 m.sup.2/g and after ball milling the S.sub.BET of 132 m.sup.2/g was obtained.

(54) Like in Example 1, a part of the ball milled sample after WC removal was subjected to calcination at 550 C. (2 C./min) in static air for 10 h. The calcination fully transformed the remaining diaspore to -Al.sub.2O.sub.3, producing the material with sole desired -Al.sub.2O.sub.3 phase (FIG. 10d). The S.sub.BET of the calcined sample was 140 m.sup.2/g and d.sup.Scherrer=19 nm.

Example 9

(55) The ball milling experiment was performed under the same conditions and using the precursor boehmite as described in Example 1, except Co.sub.3O.sub.4 (Aldrich) was added to the milling jar together with -AlOOH-17 in an amount 15 wt.-% of Co calculated to the amount of alumina.

(56) The PXRD analysis (FIG. 11) evidenced the presence of alpha Al.sub.2O.sub.3 and Co.sub.3O.sub.4 phases. WC was also additionally present in this sample as this sample was not subjected to WC removal. The S.sub.BET of the sample was 90 m.sup.2/g and d.sup.Scherrer of Co.sub.3O.sub.4 obtained by the application of Scherrer equation to 311 crystal plane (2=36.8) was 10 nm. The content of Co was 15 wt.-% according to EDX analysis.

Comparative Example A

(57) For comparison purpose boehmite precursor in Example 1 (i.e. -AlOOH-17) was calcined at 550 C. (1 C./min) in static air for 5 h. This pretreatment transformed -AlOOH-17 to -Al.sub.2O.sub.3 (see line (a) in FIG. 12). The latter sample is denoted as -Al.sub.2O.sub.3-17 and has the S.sub.BET of 109 m.sup.2/g. This sample was used as the precursor for comparative ball milling experiment under the same conditions as detailed in Example 1.

(58) The PXRD analysis (see line (b) in FIG. 12) evidenced the full transformation of -Al.sub.2O.sub.3-17 to -Al.sub.2O.sub.3 after 2 h ball milling. The S.sub.BET of the obtained sample was 64 m.sup.2/g.

Comparative Example B

(59) For yet another comparison purpose boehmite precursor in Example 9 (i.e. -AlOOH-4) was calcined at 550 C. (1 C./min) in static air for 5 h. This pretreatment transformed -AlOOH-4 to -Al.sub.2O.sub.3 (see line (a) in FIG. 13). The latter sample is denoted as -Al.sub.2O.sub.3-4 and has the S.sub.BET of 258 m.sup.2/g. This sample was used as the precursor for comparative ball milling experiment under the same conditions as detailed in Example 1.

(60) The PXRD analysis (see line (b) in FIG. 13) evidenced the full transformation of -Al.sub.2O.sub.3-4 to -Al.sub.2O.sub.3 after 2 h ball milling. The S.sub.BET of the obtained sample was 86 m.sup.2/g.