Standing wave analysis of SMB chromatography: Linear systems

Optimal performance comparison of the simulated moving bed process variants based on the modulation of the length of zones and the feed concentration

The VariCol and ModiCon processes are two variants of the simulated moving bed (SMB) process, characterized by the modulation of the length of zones of the chromatographic column train and the feed concentration. These features give more flexibility than the conventional operation, leading to essential improvements in the separation and purification of mixtures. The optimal performance comparison of these two variants, the hybrid formed by their combination, and the conventional SMB process are scarce in the literature. This comparison helps discover new characteristics of each single and combined operation mode and creates guidelines to select the appropriate operation mode for possible real applications. In this work, the performance comparison of the ModiCon, VariCol, ModiCon+VariCol, and SMB processes is carried out in terms of maximal throughput for specific product purity values. Particular emphasis is placed on both the ModiCon and the hybrid ModiCon+VariCol processes characteristics. A strategy for combining and optimizing the ModiCon and the VariCol processes was determined. As a case study, the enantioseparation of guaifenesin was considered. In the ModiCon process, more than two modulation subintervals did not improve the performance in the separation. The optimal pattern, based on two subintervals, has zero feed concentration in the first subinterval and the maximal concentration in the second one. The best result for the hybrid operation (ModiCon+VariCol) was reached when the feed port moves simultaneously as the SMB process switching period. The optimal throughput of the ModiCon and the ModiCon+VariCol processes was almost doubled than that of the SMB process. These performances were based on larger zones I and II and not in zones II and III as occur with the SMB and VariCol process. The throughput in the hybrid operation increases more significantly than the ModiCon process when 5 columns were considered instead of 6. The hybrid operation could be more attractive for a system with a few numbers of columns.

Automatic control of simulated moving bed process with deep Q-network

Optimal control of a simulated moving bed (SMB) process is challenging because the system dynamics is represented as nonlinear partial differential-algebraic equations combined with discrete events. In addition, product purity constraints are active at the optimal operating condition, which implies that these constraints can be easily violated by disturbance. Recently, artificial intelligence techniques have received significant attention for their ability to address complex problems, involving a large number of state variables. In this study, a data-based deep Q-network, which is a model-free reinforcement learning method, is applied to the SMB process to train a near-optimal control policy. Using a deep Q-network, the control policy of a complex dynamic system can be trained off-line as long as a sufficient number of data is provided. These data can be efficiently generated by performing numerical simulations in parallel on multiple machines. The on-line computation of the control input using a trained Q-network is fast enough to satisfy the computational time limit for the SMB process. However, because the Q-network does not predict the future state, it is not possible to explicitly impose state constraints. Instead, the state constraints are indirectly imposed by providing a relatively large penalty (negative reward) when the constraints are violate. Furthermore, logic-based switching control is utilized to limit the ranges of the extract and raffinate purities, which helps to satisfy the state constraints and reduce the regions in the state space for reinforcement learning to explore. The simulation results demonstrate the advantages of applying deep reinforcement learning to control the SMB process.

Optimal design of a simulated-moving-bed chromatographic process for high-purity separation of acetoin from 2,3-butanediol in a continuous mode

For the high-purity production of acetoin or 2,3-butanediol (BD) from related fermentation processes, it is essential to accomplish a detailed separation between acetoin and BD in an economical mode. To address this issue, we aimed to develop a highly-efficient simulated-moving-bed (SMB) process for the continuous-mode separation of acetoin from BD with high purity and small loss. As a first step for this task, the adsorption and mass-transfer parameters of acetoin and BD on a proven adsorbent were estimated while assuming that BD isomers (meso-BD and DL-BD) would be identical in adsorption and mass-transfer behaviors. The resultant parameters from such estimation were applied to the optimal design of the acetoin-BD separation SMB. The designed SMB was then experimentally investigated, which revealed that some sign of BD isomerism occurred in the SMB column-profile data and thus had an adverse effect on the SMB separation performance. To resolve this problem, the individual parameters of BD isomers were determined on the basis of the SMB column-profile data and an inverse-method principle. The resulting parameters of BD isomers were used in the re-design of the target SMB, which was then experimentally checked for its separation performance. It was confirmed that such SMB re-designed in consideration of BD isomerism was quite effective in the continuous-mode separation of acetoin from BD with high purity (> 99.2%) and small loss (< 1.52%).

Continuous-mode separation of fucose and 2,3-butanediol using a three-zone simulated moving bed process and its performance improvement by using partial extract-collection, partial extract-recycle, and partial desorbent-port closing

If a multi-component monosugar mixture including fucose was used as the substrates for the Klebsiella oxytoca fermentation, it could offer the following two benefits simultaneously; (i) the removal of all monosugars other than fucose, and (ii) the acquisition of 2,3-butanediol (BD). To utilize such two benefits in favor of the economical efficiency of the fucose production process, it is essential to accomplish a high-purity separation between fucose and BD on the basis of a highly-economical mode. To address this issue, we aimed to develop a simulated moving bed (SMB) process for continuous-mode separation of fucose and BD with high purities. It was first found that an Amberchrom-CG71C resin could become a suitable adsorbent for the separation of interest. The intrinsic parameters of fucose and BD on such proven adsorbent were determined, and then applied to the optimal design of the fucose-BD separation SMB. The capability of the designed SMB in ensuring high purities and high yields was experimentally verified. Finally, we devised two potential strategies to make a further improvement in product concentrations and/or desorbent usage while keeping the purities and yields of fucose and BD almost unchanged. The first strategy was based on partial extract-collection and partial extract-discard, which was found to result in 33% higher BD product concentration. The second strategy was based on partial extract-collection, partial extract-recycle, and partial desorbent-port closing, which could lead to 25% lower desorbent usage, 33% higher BD product concentration, and 7% higher fucose product concentration.

Optimization of production rate, productivity, and product concentration for a simulated moving bed process aimed atfucose separation using standing-wave-design and genetic algorithm

The effectiveness of a simulated moving bed (SMB) technology in the continuous separation of fucose from a multi-component monosugar mixture, which stemmed from defatted microalgae, has recently been identified. To guarantee high economical efficiency of such fucose-production method, the comprehensive optimization of the relevant fucose-separation SMB process needs to be accomplished such that its production rate (Prate) and/or productivity (Prod) can be maximized while meeting the requirements on fucose product concentration (C_prod,F) and pressure drop (ΔP_SMB). To resolve this issue, the SMB optimization program based on standing-wave-design method and genetic algorithm was prepared and then applied to the fucose-separation SMB optimization. It was found that the Prate, under a given particle size, could reach its maximum when the column length was selected to create a balance between the effects of the two limiting factors related to C_prod,F and ΔP_SMB. It was also found that the Prate was governed by fucose yield, if the SMB would be in need of a relatively high C_prod,F; otherwise, the Prate was governed by feed flow rate. If the particle size of the SMB adsorbent was fixed at one of the commercially available ones, the SMB conditions leading to the highest Prate and the highest Prod coincided with each other. By contrast, if the particle size was included as one of optimization variables, the Prate and Prod represented a trade-off relationship. Finally, it was confirmed from the simultaneous optimization for Prate and Prod that the increase of particle size improved Prate at the cost of Prod, thereby causing the maximum Prod to be always attained at a smaller particle size than the maximum Prate regardless of the target C_prod,F level.

Modulator Dynamics Shape the Design Space for Stepwise-Elution Simulated Moving Bed Chromatographic Separations

For proteins and other biological macromolecules, SMB chromatography is best operated non-isocratically. However, traditional modes of non-isocratic SMB operation generate significant mobile-phase modulator dynamics. The mechanisms by which these modulator dynamics affect a separation's success, and thus frame the design space, have yet to be explained quantitatively. Here, the dynamics of the modulator (e.g., salts in ion exchange and hydrophobic interaction chromatography) are explicitly accounted for. This leads to the elucidation of two new design constraints, presented as dimensionless numbers, which quantify the effects of the modulator phenomena and thus predict the success of a non-isocratic SMB separation. Consequently, these two new design constraints re-define the SMB design space. Computational and experimental studies at the boundaries of this design space corroborate the theoretical predictions. The design of efficient and robust operating conditions through use of the new design space is also demonstrated.

Speedy standing wave design and simulated moving bed splitting strategies for the separation of ternary mixtures with linear isotherms

Simulated Moving Bed (SMB) has advantages over batch chromatography in terms of productivity and solvent efficiency. However, SMB applications in large scale production are still limited because of the many design parameters that must be specified and the multiple splitting strategies that can be implemented. To overcome these barriers, this study extends the Speedy Standing Wave Design (SSWD) method of Weeden and Wang for binary linear systems to ternary linear adsorption systems. The dimensionless operating parameters, sorbent productivity, and solvent efficiency can be quickly calculated without process simulations. SSWD also gives an overview of the productivity and solvent efficiency as a function of two key dimensionless groups. This overview can be used for optimization of separation costs and for comparison of splitting strategies. The SSWD method was verified using rate model simulations for the separation of three amino acids. The simulated yields agree with the SSWD target yields within 1% for all components. The example was also used to illustrate the key design rules for ternary separations. High productivity and solvent efficiency can be achieved with a large difference in the retention factors of the heavy key and light key, which are the components that define the split of the feed between extract and raffinate products. For ternary ideal systems, solvent efficiency is inversely proportional to the largest difference in retention factors. For this reason, minimizing the overall range of retention factors can significantly improve the solvent efficiency and product concentration without sacrificing productivity. If more than one SMB is needed, the easiest split should be done first for higher productivity, solvent efficiency, and product concentration. In the example case study, both the productivity and solvent efficiency were about an order of magnitude higher when the easiest split was done in the first ring. The SSWD method can be used to design a wide array of multi-component separations with high yield, productivity, and solvent efficiency.

Standing wave design and optimization of a simulated moving bed chromatography for separation of xylobiose and xylose under the constraints on product concentration and pressure drop

The feasibility of a simulated moving bed (SMB) technology for the continuous separation of high-purity xylobiose (X2) from the output of a β-xylosidase X1→X2 reaction has recently been confirmed. To ensure high economical efficiency of the X2 production method based on the use of xylose (X1) as a starting material, it is essential to accomplish the comprehensive optimization of the X2-separation SMB process in such a way that its X2 productivity can be maximized while maintaining the X2 product concentration from the SMB as high as possible in consideration of a subsequent lyophilization step. To address this issue, a suitable SMB optimization tool for the aforementioned task was prepared based on standing wave design theory. The prepared tool was then used to optimize the SMB operation parameters, column configuration, total column number, adsorbent particle size, and X2 yield while meeting the constraints on X2 purity, X2 product concentration, and pressure drop. The results showed that the use of a larger particle size caused the productivity to be limited by the constraint on X2 product concentration, and a maximum productivity was attained by choosing the particle size such that the effect of the X2-concentration limiting factor could be balanced with that of pressure-drop limiting factor. If the target level of X2 product concentration was elevated, higher productivity could be achieved by decreasing particle size, raising the level of X2 yield, and increasing the column number in the zones containing the front and rear of X2 solute band.

Design of simulated moving bed for separation of fumaric acid with a little fronting phenomenon

The production of fumaric acid through a biotechnological pathway has grown in importance because of its potential value in related industries. This has sparked an interest in developing an economically-efficient process for separation of fumaric acid (product of interest) from acetic acid (by-product). This study aimed to develop a simulated moving bed (SMB) chromatographic process for such separation in a systematic way. As a first step for this work, commercially available adsorbents were screened for their applicability to the considered separation, which revealed that an Amberchrom-CG71C resin had a sufficient potential to become an adsorbent of the targeted SMB. Using this adsorbent, the intrinsic parameters of fumaric and acetic acids were determined and then applied to optimizing the SMB process under consideration. The optimized SMB process was tested experimentally, from which the yield of fumaric-acid product was found to become lower than expected in the design. An investigation about the reason for such problem revealed that it was attributed to a fronting phenomenon occurring in the solute band of fumaric acid. To resolve this issue, the extent of the fronting was evaluated quantitatively using an experimental axial dispersion coefficient for fumaric acid, which was then considered in the design of the SMB of interest. The SMB experimental results showed that the SMB design based on the consideration of the fumaric-acid fronting could guarantee the attainment of both high purity (>99%) and high yield (>99%) for fumaric-acid product under the desorbent consumption of 2.6 and the throughput of 0.36L/L/h.

Improvement of the performances of a tandem simulated moving bed chromatography by controlling the yield level of a key product of the first simulated moving bed unit

One of the trustworthy processes for ternary separation is a tandem simulated moving bed (SMB) process, which consists of two subordinate four-zone SMB units (Ring I and Ring II). To take full advantage of a tandem SMB as a means of recovering all three products with high purities and high economical efficiency, it is important to understand how the separation condition in Ring II is affected by that in Ring I, and further to reflect such point in the stage of designing a tandem SMB. In regard to such issue, it was clarified in this study that the Ring I factors affecting the Ring II condition could be represented by the yield level of a key product of Ring I (Y_key^RingI). As the Y_key^RingI level became higher, the amount of the Ring I key-product that was reloaded into Ring II was reduced, which affected favorably the Ring II separation condition. On the other hand, the higher Y_key^RingI level caused a larger dilution for the stream from Ring I to Ring II, which affected adversely the Ring II separation condition. As a result, a minimum in the desorbent usage of a tandem SMB occurred at the Y_key^RingI level where the two aforementioned factors could be balanced with each other. If such an optimal Y_key^RingI level was adopted, the desorbent usage could be reduced by up to 25%. It was also found that as the throughput of a tandem SMB became higher, the factor related to the migration of the Ring I key-product into Ring II was more influential in the performances of a tandem SMB than the factor related to the dilution of the stream from Ring I to Ring II.

Highly efficient recovery of xylobiose from xylooligosaccharides using a simulated moving bed method

Xylobiose (X2), which is currently available from xylooligosaccharides (XOS), has been reported to have outstanding prebiotic function and to be highly suitable for application in food industries. This has sparked an interest in the economical production of X2 of high purity (> 99%) in food and prebiotic industries. To address such issue, we developed a highly-efficient chromatographic method for the recovery of X2 from XOS with high purity and high recovery. As a first step for this work, an eligible adsorbent for a large-scale separation between X2 and other XOS components was selected. For the selected adsorbent, a single-column experiment was carried out to determine the intrinsic parameters of all the XOS components, which were then used in the optimal design of the continuous X2-recovery process based on a simulated moving bed (SMB) chromatographic method. Finally, the performance of the designed X2-recovery SMB process was verified by the relevant SMB experiments, which confirmed that the developed process in this study could recover X2 from XOS with the purity of 99.5% and the recovery of 92.3% on a continuous-separation mode. The results of this study will be useful in enabling the economical production of high-purity X2 on a large scale.

Enantioseparation of racemic aminoglutethimide using asynchronous simulated moving bed chromatography

The separation of aminoglutethimide enantiomers by the continuous multicolumn chromatographic processes were investigated experimentally and theoretically, where the columns were packed with cellulose tris 3,5-dimethylphenyl-carbamate stationary phase (brand name Chiralcel OD) and mobile phase was a mixture of n-hexane and ethanol with monoethanolamine additive. The continuous enantioseparation processes included a synchronous shifting process (SMB) and an asynchronous shifting process (VARICOL), which allowed reducing the column number (here from six-column SMB to five-column VARICOL process). Transport-dispersive model with the consideration of both intraparticle mass transfer resistance and axial dispersion was adopted to design and optimize the operation conditions for the separation of aminoglutethimide enantiomers by SMB process and VARICOL process. According to the optimized operation conditions, experiments were carried out on VARICOL-Micro unit using five-column VARICOL process with 1/1.5/1.5/1 configuration and six-column SMB process with 1/2/2/1 configuration. Products of R-aminoglutethimide (R-AG) enantiomer and S-aminoglutethimide (S-AG) enantiomer with more than 99.0% purity were obtained continuously from extract stream and raffinate stream, respectively. Furthermore, the experiemntal data obtained from five-column VARICOL process were compared with that from six-column SMB process, the feasibility and efficiency for the separation of guaifenesin enantiomers by VARICOL processes were evaluated.

Effect of adsorbent particle size on the relative merits of a non-triangular and a triangular separation region in the optimal design of a three-zone simulated moving bed chromatography for binary separation with linear isotherms

The design approaches for a three-zone simulated moving bed (SMB) chromatography with linear isotherms can be classified into two categories, depending on whether the SMB design is based on a classical region (i.e., triangular region of the triangle theory) in the first quadrant (m2, m3) plane or on a non-triangular separation region in the third quadrant (m2, m3) plane. The SMBs based on the classical and the non-triangular design approaches, which are named here as (m(+))_SMB and (m(-))_SMB respectively, are compared in this study using the Pareto solutions from the simultaneous optimization of throughput and desorbent usage under the constraints on product purities and pressure drop. The results showed that the (m(-))_SMB approach led to significantly lower desorbent usage than the (m(+))_SMB approach, which was due to the fact that the flow-rate-ratios from the (m(-))_SMB approach are extremely lower than those from the (m(+))_SMB approach. This factor also enables the (m(-))_SMB to have a significantly lower pressure drop, thereby making its throughput less restricted by a pressure-drop constraint. Due to such advantage of the (m(-))_SMB, it can make a further substantial improvement in throughput by modulating its adsorbent particle size properly. This issue was investigated using a model separation system containing succinic acid and acetic acid. It was confirmed that if the adsorbent particle size corresponding to the boundary between a mass-transfer limiting region and a pressure-drop limiting region is adopted, the (m(-))_SMB can lead to 82% higher throughput and 73% lower desorbent usage than the (m(+))_SMB.

Continuous recovery of valine in a model mixture of amino acids and salt from Corynebacterium bacteria fermentation using a simulated moving bed chromatography

The economical efficiency of valine production in related industries is largely affected by the performance of a valine separation process, in which valine is to be separated from leucine, alanine, and ammonium sulfate. Such separation is currently handled by a batch-mode hybrid process based on ion-exchange and crystallization schemes. To make a substantial improvement in the economical efficiency of an industrial valine production, such a batch-mode process based on two different separation schemes needs to be converted into a continuous-mode separation process based on a single separation scheme. To address this issue, a simulated moving bed (SMB) technology was applied in this study to the development of a continuous-mode valine-separation chromatographic process with uniformity in adsorbent and liquid phases. It was first found that a Chromalite-PCG600C resin could be eligible for the adsorbent of such process, particularly in an industrial scale. The intrinsic parameters of each component on the Chromalite-PCG600C adsorbent were determined and then utilized in selecting a proper set of configurations for SMB units, columns, and ports, under which the SMB operating parameters were optimized with a genetic algorithm. Finally, the optimized SMB based on the selected configurations was tested experimentally, which confirmed its effectiveness in continuous separation of valine from leucine, alanine, ammonium sulfate with high purity, high yield, high throughput, and high valine product concentration. It is thus expected that the developed SMB process in this study will be able to serve as one of the trustworthy ways of improving the economical efficiency of an industrial valine production process.

Simulated moving bed separation of agarose-hydrolyzate components for biofuel production from marine biomass

The economically-efficient separation of galactose, levulinic acid (LA), and 5-hydroxymethylfurfural (5-HMF) in acid hydrolyzate of agarose has been a key issue in the area of biofuel production from marine biomass. To address this issue, an optimal simulated moving bed (SMB) process for continuous separation of the three agarose-hydrolyzate components with high purities, high yields, and high throughput was developed in this study. As a first step for this task, the adsorption isotherm and mass-transfer parameters of each component on the qualified adsorbent were determined through a series of multiple frontal experiments. The determined parameters were then used in optimizing the SMB process for the considered separation. Finally, the optimized SMB process was tested experimentally using a self-assembled SMB unit with four zones. The SMB experimental results and the relevant computer simulations verified that the developed process in this study was quite successful in the economically-efficient separation of galactose, LA, and 5-HMF in a continuous mode with high purities and high yields. It is thus expected that the developed SMB process in this study will be able to serve as one of the trustworthy ways of improving the economic feasibility of biofuel production from marine biomass.

Optimal design and experimental validation of a simulated moving bed chromatography for continuous recovery of formic acid in a model mixture of three organic acids from Actinobacillus bacteria fermentation

The economically-efficient separation of formic acid from acetic acid and succinic acid has been a key issue in the production of formic acid with the Actinobacillus bacteria fermentation. To address this issue, an optimal three-zone simulated moving bed (SMB) chromatography for continuous separation of formic acid from acetic acid and succinic acid was developed in this study. As a first step for this task, the adsorption isotherm and mass-transfer parameters of each organic acid on the qualified adsorbent (Amberchrom-CG300C) were determined through a series of multiple frontal experiments. The determined parameters were then used in optimizing the SMB process for the considered separation. During such optimization, the additional investigation for selecting a proper SMB port configuration, which could be more advantageous for attaining better process performances, was carried out between two possible configurations. It was found that if the properly selected port configuration was adopted in the SMB of interest, the throughout and the formic-acid product concentration could be increased by 82% and 181% respectively. Finally, the optimized SMB process based on the properly selected port configuration was tested experimentally using a self-assembled SMB unit with three zones. The SMB experimental results and the relevant computer simulation verified that the developed process in this study was successful in continuous recovery of formic acid from a ternary organic-acid mixture of interest with high throughput, high purity, high yield, and high product concentration.