(48b) Separation Of Dilute Binary Gases By Simulated-Moving Bed With Pressure-Swing Assist: The Smb/psa Process

Simulated-moving bed (SMB) processes have been studied for both gas and liquid phase separations. However, when both gas and liquid phase SMB processes are feasible for a given separation, liquid operations have been commercialized since they appear to be more economical. For cases when liquid operation is difficult, gas phase SMB processes are of interest. In the current work, we have developed a new hybrid simulated-moving bed process with pressure-swing assist (?SMB/PSA?). As developed, SMB/PSA has two advantages over the standard gas SMB process: i.) regeneration at low pressure and ii.) internal generation of desorbent (carrier gas). SMB/PSA takes advantage of gas expansion at low pressure as well as the desorbent present in the feed stream to decrease overall molar desorbent requirements. However, the addition of a pressure-swing adds complexity to a standard gas SMB process. Both 4-zone and 5-zone clean-in-place SMBs (Abel et al., J. Chromatogr. A. (2004) 1043, 201; Paredes et al., Ind. Eng. Chem. Res. (2004) 43, 6157; Xie et al., Ind. Eng. Chem. Res. (2005) 44, 9904) are easy to retrofit with a pressure-swing if only the pressure of the extract-producing zone is varied, but there are restrictions on operating parameters.

The relatively difficult gas phase separation of R- and S- enflurane enantiomers was selected as a model to determine the feasibility of the SMB/PSA process. Prior work on enantiomeric enflurane separations using a gas chromatographic simulated-moving bed (GC-SMB) is reported in the literature (Juza et al., J. Chromatogr. A. (1998) 813, 333; Biressi et al., Chem. Eng. Sci. (2000) 55, 4537; Biressi et al., J. Chromatogr. A. (2002) 957, 211). The GC-SMB studied previously was configured as a 4-zone SMB with a feed of dilute enflurane in a nitrogen carrier gas. In the current work, simulations were done with Aspen ADSIM: the pressure ratio was PH/PL = 4 bar/0.2 bar = 20, the feed temperature was 25°C, and the other operating conditions were chosen based on the literature values for the 4-zone GC-SMB and adjusted as necessary. Productivity was fixed at the same value for all of the SMB and SMB/PSA processes studied, thereby allowing for a fair comparison based on purity. Biressi et al. (J. Chromatogr. A. (2002) 957, 211) showed that the 4-zone GC-SMB with 2 beds per zone had relatively large molar desorbent requirements. For example, products with an average purity of 0.98 were obtained with a molar desorbent to feed ratio (D/F) of 54. Our simulation results for the 5-zone SMB showed some reduction in molar desorbent requirements since desorbent was added at low pressure. An average purity of 0.92 was attained at a molar desorbent to feed ratio (D/F) of ~ 8 and a volumetric purge to feed ratio (gamma) of 5.0 for the extract-producing zone. To further increase average purity, high pressure desorbent was also added and an average purity of 0.97 was attained at a molar desorbent to feed ratio (D/F) ~ 22.

The practical limitations associated with adding a pressure swing to the 4- and 5-zone SMB processes are eliminated when the SMB/PSA concept is applied to the 2-zone SMB process (Jin and Wankat, Ind. Eng. Chem. Res. (2005) 44, 1565) since both beds can be easily regenerated at low pressure; thus, the 2-zone SMB/PSA is the focus of the current work. By adding the necessary depressurization and repressurization steps, the 2-zone SMB process is converted to the 2-zone SMB/PSA for gas separations. The four-step cycle: i.) feed, ii.) depressurization/production, iii.) regeneration/production, and iv.) repressurization with carrier gas is followed by a port switch. The bed configuration in the 2-zone SMB-PSA process can consist of a single 2-zone SMB/PSA train with a tank or two 2-zone SMB/PSA interacting trains that are out-of-phase with each other. The latter configuration resembles a typical PSA process and is continuous.

Without adding any external carrier gas (D/F = 0), the decoupled 2-zone SMB/PSA with one train and a tank and one bed per zone (i.e. the tank ? (1,1) configuration) achieved an average purity of 0.84 with a recycle ratio (RR: the relative amount of material being circulated between zones I and II ) of 0.8, a volumetric purge to feed ratio (gamma: the volume ratio of purge gas to feed gas) of 2.0 per bed, and a bed purge ratio (BPR: the relative amount of total desorbent used to regenerate the raffinate-producing zone) of 0.4. Replacing the tank with a second train (i.e. the decoupled (1,1) ? (1,1) configuration) increased the average purity to 0.86 at the same values of recycle rate and bed purge ratio. This indicated that the decoupled (1,1) ? (1,1) configuration was more favorable since it eliminated mixing effects caused by the tank. Further manipulation of bed purge ratio at a recycle rate (RR) of 0.8 and a volumetric purge to feed ratio (gamma) of 2.0 per bed with no added desorbent demonstrated that either raffinate or extract products (but not both) could be produced with purity greater than 0.95. However, due to the competing effects of recycle ratio and volumetric purge to feed ratio, average purities greater than 0.86 could not be achieved by the decoupled (1,1) ? (1,1) SMB/PSA without adding external desorbent.

In an effort to further increase average purity, the recycle ratio was increased. This allowed more internally generated desorbent to be recycled and less to be used for regeneration, thereby effectively decreasing the volumetric purge to feed ratio. Consequently, external desorbent (pure nitrogen) was added; this allowed the volumetric purge to feed ratio to be manipulated independently of the recycle ratio. At a recycle ratio (RR) of 0.97 and a molar desorbent to feed ratio (D/F) of 5.5, the decoupled (1,1) ? (1,1) system produced a product with average purity of 0.92 at the optimal bed purge ratio (BPR) of 0.35 and a volumetric purge to feed ratio (gamma) of 2.0 for each bed. Subsequent increases in the molar desorbent to feed ratio (D/F) and volumetric purge to feed ratio (gamma) at a recycle rate (RR) of 0.97 and concomitant optimization of bed purge ratio (BPR) resulted in continual increases in maximum average purity up to 0.99 at D/F ~ 15.0 for the decoupled (1,1) ? (1,1) SMB/PSA.

Increasing the number of beds per zone to two caused some improvement by further lessening molar desorbent requirements. Under the same operating conditions, the (2,2) ? (2,2) configuration achieved an average purity of 0.99 at a molar desorbent to feed ratio (D/F) of 12. Other decoupled bed configurations were found to perform between the (1,1) ? (1,1) and (2,2) ? (2,2) limits, with the (2,1) ? (2,1) configuration outperforming the (1,2) ? (1,2) configuration. Bed coupling during regeneration was also investigated: at D/F values of 0.0, 5.5, and ~ 15.0, the coupled (1,1) ? (1,1) SMB/PSA achieved average purities 0.87, 0.94, and 0.99, respectively. After a comparison with the results for the decoupled (1,1) ? (1,1) SMB/PSA process reported previously, it appears that bed coupling is only marginally favorable.