Page 43 - Adsorbents fundamentals and applications
P. 43

28   SORBENT SELECTION: CRITERIA

                     However, PSA is also used for drying of air and industrial gases. More appli-
                     cations continue to be discovered for PSA, including air prepurification (i.e.,
                     purification of air prior to cryogenic distillation by removal of carbon dioxide,
                     water, and hydrocarbons; see Rege et al., 2000). For air prepurification, PSA has
                     been used only after ca. 1990. TSA and PSA are both currently being used.
                       Besides TSA and PSA, inert purge cycles have also been used for desorbing
                     weakly adsorbed compounds. This cycle is similar to the temperature swing cycle
                     except that preheating of the purge gas is not required. Usually a fraction of the
                     light product (raffinate) is used as the inert purge gas. The inert purge cycles are
                     similar to TSA and will not be discussed separately.
                       Simulated moving bed (SMB) or the UOP Sorbex processes are used only for
                     liquid-phase separations. No gas phase separation has been performed commer-
                     cially by SMB due to the axial dispersion problem in the fixed beds.


                     3.2.1. Temperature Swing Adsorption
                     In this process cycle, the bed is regenerated by raising the temperature. The most
                     convenient way for raising the temperature is by purging the bed with a preheated
                     gas. This is the oldest and most completely developed adsorption cycle. Because
                     heating is a slow and often a rate-limiting step, the length of each cycle usually
                     ranges from several hours to over a day. In order to make the time length of the
                     adsorption step comparable with that of regeneration, the cycle is used only for
                     purification purposes.
                       Dual-bed systems are most commonly used for TSA. A number of the dual
                     bed arrangements can be made in which no external source of regenerant (e.g.,
                     air, steam, or other available gases) is used. A three-bed design has also been
                     used for systems with a long length of unused bed (LUB, which is approxi-
                     mately one-half the mass transfer zone or the span of the concentration front).
                     In the three-bed system, a guard bed is located between the adsorber and that
                     being regenerated (Chi and Cummings, 1978). When the concentration of the
                     adsorber effluent reaches nearly that of the feed stream, the beds are switched.
                     The loaded primary adsorber goes to regeneration, the former guard bed assumes
                     the primary position, and the freshly activated bed becomes the guard. By this
                     rotation, the LUB section is always contained in the guard bed, and the primary
                     adsorber is always fully loaded to its capacity when regeneration begins. Thus
                     both a high-purity product and an economic regeneration are achieved. The cost
                     of the additional bed, however, may outweigh its benefits for many common
                     purification processes.
                       The extensive literature on fixed-bed adsorber modeling and analysis has been
                     discussed elsewhere (Wankat, 1986; Yang, 1987; Suzuki, 1990; Tien, 1994; Bas-
                     madjian, 1997; Crittenden and Thomas, 1998). The theoretical analyses primarily
                     resort to equilibrium theory, that is, mass and heat transfer rates are assumed to
                     be instantaneous. Because the adsorption and desorption steps in TSA are oper-
                     ated slowly, each spanning for hours, the equilibrium theories are indeed good.
                     Often, quantitative agreements are obtained between theory and experiment (see
   38   39   40   41   42   43   44   45   46   47   48