Nien-Hwa Linda Wang

A versatile model for simulation of reaction and nonequilibrium dynamics in multicomponent fixed-bed adsorption processes

Abstract Mixtures of large biochemicals can exhibit complex chromatographic behavior. In addition to mass transfer mechanisms, the presence of nonequilibrium ad

Immobilized Metal Ion Affinity Chromatography (IMAC) Chemistry and Bioseparation Applications

Abstract This review discusses the principles of immobilized metal ion affinity chromatography (IMAC) and its applications to protein separations. IMAC functions by binding the accessible electron-donating pendant groups of a protein - such as histidine, cysteine, and tryptophan - to a metal ion which is held by a chelating group covalently attached on a stationary support. A common chelating group is iminodiacetate. The ions commonly used are of borderline or soft metals, such as Cu2+, Ni2+, Co2+, and Zn2+. Protein retention in IMAC depends on the number and type of pendant groups which can interact with the metal. The interaction is affected by a variety of independent variables such as pH, temperature, solvent type, salt type, salt concentration, nature of immobilized metal and chelate, ligand density, and protein size. Proteins are usually eluted by a decreasing pH gradient or by an increasing gradient of a competitive agent, such as imidazole, in a buffer. There are still several unresolved issues in...

Simulated Moving Bed Equipment Designs

Abstract Simulated Moving Bed or SMB chromatography is a continuous adsorption technique that increases throughput, purity and yield relative to batch chromatography. SMB utilizes a series of columns with periodically moving inlet and outlet ports. This technique has been applied for the production of petrochemicals and sugars, and to a limited extent, pharmaceuticals. The recent surge in SMB research has led to various novel SMB schemes or configurations. This review presents a survey and comparison of SMB schemes and equipment, specifically SMB valve designs, which are needed to implement the various schemes. Illustrated examples include SMBs with three, four, five, eight, nine, or twelve zones, Varicol scheme, Japan Organo process, fast startup and shutdown, and online decoupled regeneration. The SMB valve designs reviewed are classified into central and distributed valves designs, which include two‐way‐valve designs and rotary‐valve designs. We also present a novel SMB design, the Versatile SMB or V‐SMB, which can implement all the novel schemes by using a one‐rotary‐valve‐per‐column design. Select‐Trapping valves are used to interrupt the stream between the columns and to act as a junction that allows streams to be partially or totally added or removed, or to flow through to the next column. The streams can also be redirected to another SMB zone. Compared to other SMB valve designs, the V‐SMB: (1) can be configured for open‐loop, multi‐zone and zone bypass operation; (2) does not require a carousel for column rotation; (3) does not contribute to product contamination; (4) allows independent port switchings, which is required for certain schemes; (5) allows additional columns to be added easily; and, (6) requires only constant speed pumps. The V‐SMB has successfully purified enantiomers, sugars, organic acids, an antibiotic, and biosynthetic human insulin.

Computer simulations of the dynamics of multicomponent ion exchange and adsorption in fixed beds. Gradient-directed moving finite element method

Abstract In most fixed-bed separation processes, interference (competition of solutes for sorbent sites) coupled with mass transfer effects results in complex concentration waves. A general rate equation model was developed to simulate the column dynamics of multicomponent adsorption and ion exchange. This model takes into account axial dispersion, film and intraparticle diffusion, size exclusion and interference effects for systems with nonlinear isotherms or variable separation factors. The model equations were solved with a new method of orthogonal collocation on gradient-directed moving finite elements. Both the boundaries and the number of finite elements were adjusted to focus the collocation points in the region of steep concentration gradients. The computation time for simulating a step change was reduced to about 20% of that of orthogonal collocation on fixed finite elements. Column dynamics of step input and elution processes were simulated with the rate equation model. The predicted effluent histories are shown to be in close agreement with the data on protein elution. This model is important for studying interference phenomena coupled with mass transfer effects, especially for systems with a large number of components, small feed pulses, small interparticle diffusivities and large sorbent particles.

Separation of lactic acid from acetic acid using a four-zone SMB.

A simulated moving bed (SMB) process has been developed to separate l‐(+)‐lactic acid from acetic acid, a major impurity in the fermentation broth of Lactobacillus rhamnosus. Poly(4‐vinylpyridine) resin (PVP) was selected as the adsorbent. Adsorption isotherms and mass transfer parameters of the organic acids were estimated from single‐column frontal tests. Experimental results show that the Langmuir isotherms obtained from the frontal tests can be used in the design of an SMB process to achieve 99.9% purity and over 93% yield of lactic acid. The column profiles and effluent histories, however, deviate from rate model predictions based on the Langmuir isotherms. They agree more closely with the predictions based on a modified Langmuir isotherm for lactic acid. The standing wave design method for systems with modified Langmuir isotherms is developed in this study. Rate model simulations show that the process based on the modified design method can achieve high purity (>99.9%) and high yield (>99.9%). For this nonlinear system, accurate isotherm model and model parameters are needed in the design, and the zone flow rates must be closely monitored and controlled in order to ensure high purity and high yield in the SMB process.

Standing wave design and experimental validation of a tandem simulated moving bed process for insulin purification

A tandem simulated moving bed (SMB) process for insulin purification has been proposed and validated experimentally. The mixture to be separated consists of insulin, high molecular weight proteins, and zinc chloride. A systematic approach based on the standing wave design, rate model simulations, and experiments was used to develop this multicomponent separation process. The standing wave design was applied to specify the SMB operating conditions of a lab‐scale unit with 10 columns. The design was validated with rate model simulations prior to experiments. The experimental results show 99.9% purity and 99% yield, which closely agree with the model predictions and the standing wave design targets. The agreement proves that the standing wave design can ensure high purity and high yield for the tandem SMB process. Compared to a conventional batch SEC process, the tandem SMB has 10% higher yield, 400% higher throughput, and 72% lower eluant consumption. In contrast, a design that ignores the effects of mass transfer and nonideal flow cannot meet the purity requirement and gives less than 96% yield.

Adsorption dynamics of single and binary surfactants at the air/water interface

Abstract Models of diffusion-controlled adsorption of mixed surfactants at the air/water interface reveal a rich variety of interesting behavior. The models predict the following. A maximum in the dynamic adsorption concentration for the less surface-active surfactant can occur if this component has a higher initial bulk concentration. The maximum is even more pronounced for small diffusion lengths, at which the dynamic subsurface concentration of the fast-adsorbing component can temporarily exceed that of the bulk. If an impurity is more surface-active than the main component, then small amounts of the impurity can drastically alter the adsorption behavior of the latter, as shown in some examples. These phenomena can be partly explained in terms of the timescales of adsorption for single surfactants described by the Henrys law or Langmuir isotherms, and partly by the competition of the components for surface sites. Since dynamic surface concentration data are not yet available for binary surfactants, the model predictions are used to calculate dynamic surface tensions and compare them with dynamic tension data. The results have implications for the dynamic and equilibrium selectivities of foam fractionation processes and for the dynamics of free-surface flows.

Experimental probing and modeling of key sorbent–solute interactions of norephedrine enantiomers with polysaccharide-based chiral stationary phases

The key interactions of a chiral solute, norephedrine or 2-amino-1-phenyl-1-propanol (PPA), with three commercially important polysaccharide-based chiral stationary phases, amylose Tris(3,5-dimethylphenylcarbamate) (ADMPC), amylose Tris((S)-alpha-methylbenzylcarbamate) (ASMBC) and cellulose Tris(3,5-dimethylphenylcarbamate) (CDMPC) are studied in detail using different experimental techniques and molecular simulations. The HPLC retention factors of the enantiomers of PPA in n-hexane/2-propanol (90/10, v/v) at 298 K vary significantly with these sorbents. The enantioselectivities of -PPA versus +PPA are 2.4, 1.0, and 0.8 (reversal in the elution order), respectively. The observed changes in the wavenumbers and the intensities of the amide bands of these polymers in the attenuated total reflection-infrared spectroscopy (ATR-IR) spectra upon absorption of each enantiomer are different. The IR wavenumbers, and the H-bonding interaction energies of the polymer side chains with each enantiomer (polymer-solute) in four different binding configurations are estimated and ranked using the density functional theory (the DFT/B3LYP/6-311+g(d,p) level of theory). X-ray diffraction (XRD) results show that the polymer crystallinity increases significantly upon absorption of each enantiomer. The helical pitches and the inter-rod packing for these polymers are inferred from the XRD results and incorporated into the molecular dynamics (MD) simulations. The elution orders predicted for the enantiomers of PPA using MD simulations of the polymer-PPA binary systems are consistent with the chromatography results. The enantioselectivity observed in ADMPC is hypothesized to be due to having three simultaneous interactions (two H-bond and one pi-pi) of the polymer with -PPA versus two interactions (one H-bond and one pi-pi) with +PPA.

Optimization of throughput and desorbent consumption in simulated moving-bed chromatography for paclitaxel purification.

In simulated moving-bed (SMB) applications, throughput and desorbent consumption are two key factors that control process cost. For a given adsorbent volume and product purity requirements, throughput and desorbent consumption depend on desorbent composition, column configuration, column length to diameter ratio, and adsorbent particle size. In this study, these design parameters are systematically examined for paclitaxel purification. The results show that if adsorbent particle size, column dimensions and column configuration are fixed, the higher the product purity required, the lower the throughput. If product purity and yield are fixed, the larger the solute migration speed ratio, the higher the throughput, and the lower the desorbent consumption. If total bed volume and product purities are fixed, the longer the separation zones, the higher the throughput, but the higher the desorbent flow-rate. An intermediate configuration gives the minimum desorbent consumption. If there are no limits on pressure drop or zone flow-rate, the larger the column length to diameter ratio, the smaller the adsorbent particle size, the higher the throughput, and the lower the desorbent consumption. If the maximum zone flow-rate is controlled by the pressure drop limit and not by the standing waves requirement, the longer the columns, the lower the zone flow-rates and the lower the throughput. For 150 microns adsorbent particles and a maximum zone flow-rate of 300 ml/min, a design with optimal throughput and desorbent consumption is found for paclitaxel purification.

Effect of the solute molecular structure on its enantioresolution on cellulose tris(3,5-dimethylphenylcarbamate)

The effects of the molecular structures for 13 structurally similar chiral solutes on their HPLC retention and enantioresolutions on a commercially important polysaccharide-based chiral stationary phase, cellulose tris(3,5-dimethylphenylcarbamate) (CDMPC) are studied. Among these 13 solutes, only methyl ephedrine (MEph) shows significant enantioresolution. The retention factors of these chiral solutes vary significantly from 0.7 to 3.2 in n-hexane/2-propanol (90/10, v/v) at 298 K. The retention factors of some simpler non-chiral solutes having similar but fewer functional groups than their chiral counterparts are also studied under the same conditions and are compared to those of the chiral solutes. The H-bonding interactions between the functional groups of the solute and the C=O and NH functional groups of the polymer are probed with attenuated total reflection-infrared spectroscopy (ATR-IR) for the polymer, for binary sorbent-solute systems. The CDMPC IR amide band wavenumbers change significantly, indicating H-bonding interactions of the polymer C=O and NH groups with the solutes. The elution orders predicted for the enantiomers of these chiral solutes using molecular dynamics (MD) simulations of the polymer-solute binary systems are consistent with the HPLC results. The CDMPC cavity nano-structure and the potential interactions with chiral solutes are proposed based on HPLC data, IR data, and the simulations. The results are consistent with the three-point attachment model and support the hypothesis that significant enantioresolution requires at least three different synergistic interactions which can be a combination of steric hindrance, H-bonding, or pi-pi interactions.