Page 100 - Adsorption by Powders and Porous Solids
P. 100

CHAPTER 3. METHODOLOGY OF ADSORPTION                            85
    described just  above, but it may allow the balance to be used over a smaller range,
    where a higher sensitivity is available.
      A pertinent question is: as the'volume of the adsorbed phase increases, do we have to
    &e  into account the corresponding increase of buoyancy? (e.g. the buoyancy doubles
    after saturation of an adsorbent with 50% porosity). The answer is no, provided we want
    U, asess the surface excess mass mu. As illustrated in Figure 3.22, because of the buoy-
    mq effect, we do not measure the total mass of the adsorbed layer (shaded + hatched
    m) but simply a surface excess mass (hatched area only). Thus, adsorption gravime-
       and the Gibbs representation are highly compatible (Findenegg, 1997).
      Finally, it may be worth considering the effect of buoyancy correction on the mag-
    nitude of the error in the determination of the BET(NJ  surface area. This is indicated
    in Figure 3.23 for surface areas ranging from 1 to 1000 mZ g-'  and adsorbent densi-
    ties in the range 1 to 8. The buoyancy is calculated by taking into account the density
    of  nitrogen vapour at a temperature of 77 K and a pressure of  100 mbar, which is
    assumed enough to complete the monolayer. For a sample of 1 mZ g-', with a density
    of  2 (for instance, a finely ground quartz), the error amounts to 80%! At the other
    extreme, for a sample of 1000 m2 g-',  with a density of  1.5 (for instance, an active
    carbon), the error will be as small as 0.1%.

    3.4.5.  Adsorption equilibrium
    By convention, 'adsorption isotherms' are generally assumed to correspond to a ther-
    modynamic equilibrium: if  this is not true, use of  the term  'adsorption isotherm'
    is questionable. The confiiation of  adsorption equilibrium is the~efore of  crucial
    significance.


              % Error





















    Figure 3.23.  Gas  adsorption  gravimetry:  percentage  error in  surface area as a  result of  neglecting
    totally the buoyancy effect on the adsorbenr (assumptions: nitrogen BET, monolayer at 100 rnbar, bulk
    density of adsorbent from 1 to 8).
   95   96   97   98   99   100   101   102   103   104   105