Page 324 - Adsorption by Powders and Porous Solids
P. 324

CHAPTER 10. ADSORPTION BY METAL OXIDES                        313
     between  two  adjacent A  (or  B) layers  is 0.28 nrn whereas the A-B  distance is
     0.20 nm.
      Bayerite does not occur in nature, but it can be made in a number of different ways
     (e.g. by the hydrolysis of an aluminium alkoxide). The OH layers in bayerite appear
     to be stacked in the order ABABAB.. . . Within the double layer the A-B  distance is
    0.21 nrn and between the double layers the A-B  distance is 0.26 xun. The density of
     bayerite is correspondingly a little higher than that of gibbsite.
      Although deposits of nordstrandite have been found, this modification is not easy
     to prepare in a relatively pure form. For this reason, the exact structure is still under
     discussion. However, the layer stacking is likely to be made up of a combination of
     both bayerite and gibbsite.
      In its most common industrial form, gibbsite is a sandy material, of c. 50-100  pm
     grain size. Each grain is itself a dense agglomerate of smaller hexagonal crystals, typ-
     ically 5-15  pn in size, and the BET-nitrogen surface area is usually not more than
     0.2 mZ g-'.  Other fonns of gibbsite have also been subjected to physisorption studies.
     These include  loose, thin  hexagonal crystals with  BET-nitrogen areas  of  5  and
     15 m2 g-'  corresponding to  mean  crystal  sizes  of  1  and  0.2 pn, respectively
     (Rouquerol et  al., 1975), and  also porous  aggregates. For  example, Ramsay and
     Avery (1979) found that a batch of very pure gibbsite powder gave a Type TV nitro-
     gen isotherm with an H1 hysteresis loop at high p/pO. It appeared that the gibbsite
     was mesoporous  and  possibly  also macroporous, the effective pore  width  being
     mainly > 20 nm. The BET-nitrogen area was 41 m2 g-l, which was consistent with
     the mean thickness of c. 25 nrn of the thin hexagonal platelets, as determined by elec-
     tron  microscopy and X-ray  line broadening.  This rather ill-defined porosity  was
     ascribed to the space between the gibbsite crystallites and was found to persist after
     heat treatment at 40°C. A batch of gibbsite of lower BET-nitrogen area (5.6 m2 g-')
     was used by Stacey (1987). In this case, the polycrystalline grains of mean diameter
     75 pn were composed of 0.3 pn platy crystallites.
      A sample of microcrystalline nordstrandite was found to be somewhat mesoporous
     by Aldcroft and Bye (1967). The nitrogen hysteresis loop was Type H3, which indi-
     cated the existence of slit-shaped pores between the crystallites. The BET-nitrogen
     area of 34 m2  g-l appeared to represent the external area of the crystallites.
       A  relatively low-area  (8 m2 g-')  aged  sample of  bayente (Bye and  Robinson,
     1964) which gave a reversible Type I1 nitrogen isotherm, was shown to be essentially
     non-porous (Payne and Sing, 1969). Electron microscopy revealed that this sample
     was composed of discrete conical crystals.


     10.3.3.  Aluminium oxide-hydroxides
     There are two well-known oxide-hydroxides (Al00H) with  closely related struc-
     tures: diaspore and boehmite. Diaspore occurs in some types of clay and bauxite. It
     has been produced by the hydrothermal treatment of corundum, a-A120,. Whereas
     boehmite is characterized by cubic close-packing of the anions, diaspore has a hexag-
     onal close-packed structure. This difference probably accounts for the direct thermal
     transformation of diaspore to corundum at relatively low temperatures (450-600°C).
   319   320   321   322   323   324   325   326   327   328   329