(577e) Speedy Standing Wave Design, Optimization, and Scaling Rules of Simulated Moving Bed Systems with Linear Isotherms

SMBs consume orders of magnitude less solvent, require an order of magnitude less sorbent, take up less space, and require less manpower than batch chromatography operations. These advantages make SMBs economical for large-scale separations. SMBs, however, have not been widely used for large-scale production. One key barrier is the complexity of SMB design and optimization. A four-zone SMB for a binary, linear adsorption system has 23 variables, which include nine material properties and 14 design parameters (two yields, five operating parameters, and seven equipment parameters). Trial and error design of the 14 parameters is not feasible. Furthermore, scaling rules of SMB systems have not been well established. The most commonly used method for scale-up is to design and test the SMB at lab scale and then simply increase the column diameter to meet the production requirements, while keeping the same column length and the fluid zone velocities for the production scale SMB. This method, however, does not guarantee an optimal large-scale process.

In a previous study, we developed for size exclusion SMB (SEC-SMB) systems the analytical expressions of solvent consumption and sorbent productivity as a function of key dimensionless groups. The resulting Speedy Standing Wave Design (SSWD) equations could be used to quickly optimize SEC-SMB design parameters to achieve maximum productivity, minimum solvent consumption, or minimum cost if cost functions are known.

In this study, the SSWD is extended to systems with linear adsorption isotherms, which can result from various separation mechanisms, such as size-exclusion, ion-exclusion, or complexation. Moreover, in deciding whether simulated moving bed (SMB) is technically or economically feasible for a separation, many nonlinear systems are first approximated as linear systems for preliminary studies. Furthermore, general scaling rules based on the key dimensionless groups in the SSWD equations are also developed. The SSWD equations are simplified for diffusion-controlled systems. The separation of acetic acid from glucose in biomass hydrolysate is used as an example to illustrate the effects of various material properties and key dimensionless groups on sorbent productivity and solvent consumption. The SSWD equations are also used in a search routine, along with cost functions, to determine the SMB design with the lowest separation cost for this example.

The SSWD equations also allow estimation of the maximum SMB productivity based on intrinsic material properties and specified yields. For an isocratic system, the desorbent flowrate is always over two times the feed flowrate for the maximum productivity design. The optimal column configuration for either maximum productivity or minimum solvent consumption is controlled by the ratio of the effective diffusivities of the two components and the selectivity. Optimal large-scale SMB process can be found using the SSWD method if the intrinsic parameters are known. For lab-scale testing of a scaled-down process, one can use smaller particles to speed up the testing process, if all other material properties are the same. By keeping the values of key dimensionless groups constant, the lab-scale SMB experiments can be performed four times faster using particles with ½ the diameter, and the lab-scale SMB will have the same dimensionless solvent consumption and productivity as the plant-scale SMB.