Evaluating two process scale chromatography column header designs using CFD

Chromatography is an indispensable unit operation in the downstream processing of biomolecules. Scaling of chromatographic operations typically involves a significant increase in the column diameter. At this scale, the flow distribution within a packed bed could be severely affected by the distributor design in process scale columns. Different vendors offer process scale columns with varying design features. The effect of these design features on the flow distribution in packed beds and the resultant effect on column efficiency and cleanability needs to be properly understood in order to prevent unpleasant surprises on scale‐up. Computational Fluid Dynamics (CFD) provides a cost‐effective means to explore the effect of various distributor designs on process scale performance. In this work, we present a CFD tool that was developed and validated against experimental dye traces and tracer injections. Subsequently, the tool was employed to compare and contrast two commercially available header designs. © 2014 American Institute of Chemical Engineers Biotechnol. Prog., 30:837–844, 2014     

Enzymatic synthesis of farnesyl laurate in organic solvent: initial water activity, kinetics mechanism, optimization of continuous operation using packed bed reactor and mass transfer studies

The influence of water activity and water content was investigated with farnesyl laurate synthesis catalyzed by Lipozyme RM IM. Lipozyme RM IM activity depended strongly on initial water activity value. The best results were achieved for a reaction medium with an initial water activity of 0.11 since it gives the best conversion value of 96.80%. The rate constants obtained in the kinetics study using Ping-Pong-Bi-Bi and Ordered-Bi-Bi mechanisms with dead-end complex inhibition of lauric acid were compared. The corresponding parameters were found to obey the Ordered-Bi-Bi mechanism with dead-end complex inhibition of lauric acid. Kinetic parameters were calculated based on this model as follows: V _max = 5.80 mmol l⁻¹ min⁻¹ g enzyme⁻¹, K _m,A = 0.70 mmol l⁻¹ g enzyme⁻¹, K _m,B = 115.48 mmol l⁻¹ g enzyme⁻¹, K _i = 11.25 mmol l⁻¹ g enzyme⁻¹. The optimum conditions for the esterification of farnesol with lauric acid in a continuous packed bed reactor were found as the following: 18.18 cm packed bed height and 0.9 ml/min substrate flow rate. The optimum molar conversion of lauric acid to farnesyl laurate was 98.07±0.82%. The effect of mass transfer in the packed bed reactor has also been studied using two models for cases of reaction limited and mass transfer limited. A very good agreement between the mass transfer limited model and the experimental data obtained indicating that the esterification in a packed bed reactor was mass transfer limited.     

Industrial production of fructooligosaccharides by immobilized cells of <Emphasis Type="Italic">Aureobasidium pullulans</Emphasis> in a packed bed reactor

Continuous production of fructooligosaccharides (FOS) by Aureobasidium pullulans immobilized on calcium alginate beads with a packed bed was investigated at a plant scale reactor. Optimum conditions were with 770 g sucrose/l, being fed at 200 l/h at 50°C which gave a productivity of 180 g FOS/l h. Initial activity was maintained for more than 100 days. The reactor was successfully scaled up to a production scale of 1.2 m³.     

Simulation of the dynamic packing behavior of preparative chromatography columns via discrete particle modeling

Preparative packed‐bed chromatography using polymer‐based, compressible, porous resins is a powerful method for purification of macromolecular bioproducts. During operation, a complex, hysteretic, thus, history‐dependent packed bed behavior is often observed but theoretical understanding of the causes is limited. Therefore, a rigorous modeling approach of the chromatography column on the particle scale has been made which takes into account interparticle micromechanics and fluid–particle interactions for the first time. A three‐dimensional deterministic model was created by applying Computational Fluid Dynamics (CFD) coupled with the Discrete Element Method (DEM). The column packing behavior during either flow or mechanical compression was investigated in‐silico and in laboratory experiments. A pronounced axial compression–relaxation profile was identified that differed for both compression strategies. Void spaces were clearly visible in the packed bed after compression. It was assumed that the observed bed inhomogeneity was because of a force‐chain network at the particle scale. The simulation satisfactorily reproduced the measured behavior regarding packing compression as well as pressure‐flow dependency. Furthermore, the particle Young's modulus and particle–wall friction as well as interparticle friction were identified as crucial parameters affecting packing dynamics. It was concluded that compaction of the chromatographic bed is rather because of particle rearrangement than particle deformation. © 2015 American Institute of Chemical Engineers Biotechnol. Prog., 32:363–371, 2016     

Packing of large‐scale chromatography columns with irregularly shaped glass based resins using a stop‐flow method

Rigid chromatography resins, such as controlled pore glass based adsorbents, offer the advantage of high permeability and a linear pressure‐flow relationship irrespective of column diameter which improves process time and maximizes productivity. However, the rigidity and irregularly shaped nature of these resins often present challenges in achieving consistent and uniform packed beds as formation of bridges between resin particles can hinder bed consolidation. The standard flow‐pack method when applied to irregularly shaped particles does not yield well‐consolidated packed beds, resulting in formation of a head space and increased band broadening during operation. Vibration packing methods requiring the use of pneumatically driven vibrators are recommended to achieve full packed bed consolidation but limitations in manufacturing facilities and equipment may prevent the implementation of such devices. The stop‐flow packing method was developed as an improvement over the flow‐pack method to overcome these limitations and to improve bed consolidation without the use of vibrating devices. Transition analysis of large‐scale columns packed using the stop‐flow method over multiple cycles has shown a two‐ to three‐fold reduction of change in bed integrity values as compared to a flow‐packed bed demonstrating an improvement in packed bed stability in terms of the height equivalent to a theoretical plate (HETP) and peak asymmetry (A_s). © 2014 American Institute of Chemical Engineers Biotechnol. Prog., 30:1319–1325, 2014     

Roles of interstitial fraction and load conditions on the dynamic binding capacity of proteins on capillary‐channeled polymer fiber columns

Capillary‐channeled polymer (C‐CP) fibers are used as a stationary phase for ion‐exchange chromatography of proteins. Collinear packing of the fibers permits operation at high linear velocities (U_o > 100 mm s^?1) and low backpressure (<2,000 psi) on analytical‐scale columns. Rapid solvent transport is matched with very efficient solute mass transfer as fibers are virtually non‐porous with respect to the size of the target protein molecules. Lack of porosity of course limits the equilibrium binding capacity of stationary phases. Breakthrough curves and frontal analysis are used to better understand trade‐offs between the kinetic and thermodynamic properties as C‐CP fibers are applied in preparative situations. Fiber columns packed to different interstitial fraction values affect both the total fiber surface area (e.g., equilibrium binding capacity [EBC]) and the permittivity to flow and mass transport characteristics (e.g., dynamic binding capacity [DBC]). The EBC of the nylon 6 C‐CP fibers was found to be 1.30 mg g^?1, with isotherms that were best matched by a Moreau model, showing linearity up to solute concentrations of ～0.4 mg mL^?1. Isotherms generated under flow conditions were equally well approximated using Langmuir, Freundlich, and Moreau isotherm models. Fairly linear responses were seen up to the maximum load concentration of 1.2 mg mL^?1. Counterintuitively, dynamic studies revealed that conditions of high column porosity yielded a DBC that is ～70% higher than the EBC. These findings point to potential advantages in terms downstream processing applications, where protein throughput and yield are critical metrics. © 2014 American Institute of Chemical Engineers Biotechnol. Prog., 31:97–109, 2015     

Development and scale up of a capture step (expanded bed chromatography) for a fusion protein expressed intracellularly in Escherichia coli

A capture step was developed using the expanded bed adsorption technology to separate a protein of interest on a cation exchanger from a crude Escherichia coli homogenate. This method was developed in bench-top scale using a STREAMLINE 25 column (Amersham Pharmacia Biotech, Sweden) and STREAMLINE SP. The development was based on earlier experiments performed in a packed bed column (SP-Sepharose FF) to investigate the conditions for sample application, wash and elution. The packed bed method was transformed into an expanded bed method by slightly modifying the wash procedure and cleaning in place (CIP). This method was then scaled-up to pilot scale and used for production of the fusion protein according to cGMP.The yield over the step in pilot scale was 70-85% compared with only 30-50% in small scale. Pressure build-up, attachment of biomass to the adsorbent and collapses of the expanded bed were phenomena seen in small scale but not in pilot scale. The scale-up of the step significantly improved the performance of the step.     

Enzymatic conversion of sunflower oil to biodiesel in a solvent-free system: Process optimization and the immobilized system stability

The feasibility of using the commercial immobilized lipase from Candida antarctica (Novozyme 435) to synthesize biodiesel from sunflower oil in a solvent-free system has been proved. Using methanol as an acyl acceptor and the response surface methodology as an optimization technique, the optimal conditions for the transesterification has been found to be: 45 ^oC, 3% of enzyme based on oil weight, 3:1 methanol to oil molar ratio and with no added water in the system. Under these conditions, >99% of oil conversion to fatty acid methyl ester (FAME) has been achieved after 50 h of reaction, but the activity of the immobilized lipase decreased markedly over the course of repeated runs. In order to improve the enzyme stability, several alternative acyl acceptors have been tested for biodiesel production under solvent-free conditions. The use of methyl acetate seems to be of great interest, resulting in high FAME yield (95.65%) and increasing the half-life of the immobilized lipase by about 20.1 times as compared to methanol. The reaction has also been verified in the industrially feasible reaction system including both a batch stirred tank reactor and a packed bed reactor. Although satisfactory performance in the batch stirred tank reactor has been achieved, the kinetics in a packed bed reactor system seems to have a slightly better profile (93.6 ± 3.75% FAME yield after 8–10 h), corresponding to the volumetric productivity of 48.5 g/(dm³ h). The packed bed reactor has operated for up to 72 h with almost no loss in productivity, implying that the proposed process and the immobilized system could provide a promising solution for the biodiesel synthesis at the industrial scale.     

Performance optimization of continuous countercurrent tangential chromatography for antibody capture

Recent studies have demonstrated that continuous countercurrent tangential chromatography (CCTC) can effectively purify monoclonal antibodies from clarified cell culture fluid. CCTC has the potential to overcome many of the limitations of conventional packed bed protein A chromatography. This paper explores the optimization of CCTC in terms of product yield, impurity removal, overall productivity, and buffer usage. Modeling was based on data from bench‐scale process development and CCTC experiments for protein A capture of two clarified Chinese Hamster Ovary cell culture feedstocks containing monoclonal antibodies provided by industrial partners. The impact of resin binding capacity and kinetics, as well as staging strategy and buffer recycling, was assessed. It was found that optimal staging in the binding step provides better yield and increases overall system productivity by 8–16%. Utilization of higher number of stages in the wash and elution steps can lead to significant decreases in buffer usage (～40% reduction) as well as increased removal of impurities (～2 log greater removal). Further reductions in buffer usage can be obtained by recycling of buffer in the wash and regeneration steps (～35%). Preliminary results with smaller particle size resins show that the productivity of the CCTC system can be increased by 2.5‐fold up to 190 g of mAb/L of resin/hr due to the reduction in mass transfer limitations in the binding step. These results provide a solid framework for designing and optimizing CCTC technology for capture applications. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:430–439, 2016     

Pilot scale recovery of monoclonal antibodies by expanded bed ion exchange adsorption

The aim of the investigations was to estimate the scale up properties of an efficient chromatographic first capture step for the recovery of murine IgG₁ from undiluted and unclarified hybridoma cell culture broth using an ion exchange matrix in expanded bed mode. The tested new sulfopropyl-based ion exchange matrix (Streamline^TM SP XL, Amersham Pharmacia Biotech) stands out due to its enhanced capacity compared to its precursor (Streamline^TM SP). Defining the working pH in preliminary electrophoretic analyses (titration curve, SDS-PAGE) and small-scaled chromatographic binding studies showed, that the optimal value for the IgG purification was pH 4.6, where a co-chromatography of the medium supplement albumin (500 mg l^-1, pI = 4.8) could not be avoided. Further scouting experiments dealt with the dynamic capacity of the matrix, which was evaluated by frontal adsorption analysis. In packed bed mode no break-through of the target protein was achieved even after 6.5 mg IgG per ml matrix were applied. These results could not be reproduced in expanded bed mode with cell-free supernatant, where the dynamic capacity was found to be only 1.5 mg IgG/ml SP XL. Processing cell-containing broth resulted in an additional decrease of the value down to 0.5 mg ml^-1, presumably caused by the remarkable biomass adsorption to the matrix. The search for the reasons led to the examination of the hydrodynamic conditions. Buffer experiments with a tracer substance (acetone) pointed out, that the flow in expanded bed was significantly more influenced by back-mixing effects and channel formations than in packed bed. These effects could be compensated with an enhanced viscosity of the liquid phase, which was achieved by the addition of glucose. As a result of the improved hydrodynamic conditions in the expanded bed, the dynamic capacity could be increased from 0.5 to more than 4.5 mg IgG/ml matrix for the processing of cell culture broth with 400 mM glucose. Finally, the scale up from a Streamline^TM 25 to a Streamline^TM 200 column was performed under conditions, which proved to be optimal: 100 L of unclarified hybridoma broth were concentrated with a binding rate of 95% in less than 3.5 hours. Loading the column no break-through of the target protein was achieved. However, the eluate still contained debris and cells, which points out the major disadvantage of the method: the biomass attachment to the matrix.     

Biosynthesis of γ-aminobutyric acid (GABA) using immobilized whole cells of <Emphasis Type="Italic">Lactobacillus brevis</Emphasis>

On an industrial scale, the production of γ-aminobutyric acid (GABA) from the cheaper sodium L-glutamate (L-MSG) is a valuable process. By entrapping Lactobacillus brevis cells with higher glutamate decarboxylase (GAD) activity into Ca-alginate gel beads, the biotransformation conditions of L-MSG to GABA were optimized with the immobilized cells. The cells obtained from a 60-h culture broth showed the highest biotransformation efficiency from L-MSG to GABA. The optimal cell density in gel beads, reaction pH and temperature were 11.2 g dry cell weight (DCW) l⁻¹, 4.4 and 40°C respectively. The thermal stability of immobilized cells was significantly higher than free cells. Under the optimized reaction conditions, the yield of GABA reached above 90% during the initial five batches and the yield still remained 56% in the tenth batch. Continuous production of GABA was realized with a higher yield by incorporating cell re-cultivation using the packed bed reactor.     

Highly enantioselective esterification of racemic ibuprofen in a packed bed reactor using immobilised Rhizomucor miehei lipase

A systematic study of the enantioselective resolution of ibuprofen by commercial Rhizomucor miehei lipase (Lipozyme(R) IM20) has been carried out using isooctane as solvent and butanol as esterificating agent. The main variables controlling the process (temperature, ibuprofen concentration, ratio butanol:ibuprofen) have been studied using an orthogonal full factorial experimental design, in which the selected objective function was enantioselectivity. This strategy has resulted in a polynomial function that describes the process. By optimizing this function, optimal conditions for carrying out the esterification of racemic ibuprofen have been determined. Under these conditions, enantiomeric excess and total conversion values were 93.8% and 49.9%, respectively, and the enantioselectivity was 113 after 112 h of reaction. These conditions have been considered in the design of a continuous reactor to scale up the process. The esterification of ibuprofen was properly described by pseudo first-order kinetics. Thus, a packed bed reactor operating as a plug-flow reactor (PFR) is the most appropriate in terms of minimizing the residence time compared with a continuous stirred tank reactor (CSTR) to achieve the same final conversion. This reactor shows a similar behavior in terms of enantioselectivity, enantiomeric excess, and conversion when compared with batch reactors. A residence-time distribution (RTD) shows that the flow model is essentially a plug flow with a slight nonsymmetrical axial dispersion (Peclet number = 43), which was also corroborated by the model of CSTR in series. The stability of the system (up to 100 h) and the possibility of reutilization of the enzyme (up to four times) lead to consider this reactor as a suitable configuration for scale up of the process.     

Understanding and modeling retention of mammalian cells in fluidized bed centrifuges

Within the last decade, fully disposable centrifuge technologies, fluidized‐bed centrifuges (FBC), have been introduced to the biologics industry. The FBC has found a niche in cell therapy where it is used to collect, concentrate, and then wash mammalian cell product while continuously discarding centrate. The goal of this research was to determine optimum FBC conditions for recovery of live cells, and to develop a mathematical model that can assist with process scaleup. Cell losses can occur during bed formation via flow channels within the bed. Experimental results with the kSep400 centrifuge indicate that, for a given volume processed: the bed height (a bed compactness indicator) is affected by RPM and flowrate, and dead cells are selectively removed during operation. To explain these results, two modeling approaches were used: (i) equating the centrifugal and inertial forces on the cells (i.e., a force balance model or FBM) and (ii) a two‐phase computational fluid dynamics (CFD) model to predict liquid flow patterns and cell retention in the bowl. Both models predicted bed height vs. time reasonably well, though the CFD model proved more accurate. The flow patterns predicted by CFD indicate a Coriolis‐driven flow that enhances uniformity of cells in the bed and may lead to cell losses in the outflow over time. The CFD‐predicted loss of viable cells and selective removal of the dead cells generally agreed with experimental trends, but did over‐predict dead cell loss by up to 3‐fold for some of the conditions. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:1520–1530, 2016     

A practical strategy for using miniature chromatography columns in a standardized high‐throughput workflow for purification development of monoclonal antibodies

The emergence of monoclonal antibody (mAb) therapies has created a need for faster and more efficient bioprocess development strategies in order to meet timeline and material demands. In this work, a high‐throughput process development (HTPD) strategy implementing several high‐throughput chromatography purification techniques is described. Namely, batch incubations are used to scout feasible operating conditions, miniature columns are then used to determine separation of impurities, and, finally, a limited number of lab scale columns are tested to confirm the conditions identified using high‐throughput techniques and to provide a path toward large scale processing. This multistep approach builds upon previous HTPD work by combining, in a unique sequential fashion, the flexibility and throughput of batch incubations with the increased separation characteristics for the packed bed format of miniature columns. Additionally, in order to assess the applicability of using miniature columns in this workflow, transport considerations were compared with traditional lab scale columns, and performances were mapped for the two techniques. The high‐throughput strategy was utilized to determine optimal operating conditions with two different types of resins for a difficult separation of a mAb monomer from aggregates. Other more detailed prediction models are cited, but the intent of this work was to use high‐throughput strategies as a general guide for scaling and assessing operating space rather than as a precise model to exactly predict performance. © 2014 American Institute of Chemical Engineers Biotechnol. Prog., 30:626–635, 2014     

Control of 2-chlorophenol vapour emissions by a trickling biofilter

This research work investigates the biodegradation of 2-chlorophenol vapours in a trickling biofilter packed with a ceramic material, and seeded with a pure strain of Pseudomonas pickettii. The process was tested at laboratory scale over 260 days of operation under varying loading conditions. More than 98% degradation efficiencies were achieved for loading rates up to 82.5gm(-3)h(-1). Process analysis, performed using data on 2-chlorophenol concentration profiles along the biofilter bed, shows that best biofilter performance (i.e. maximum degradation capacity and efficiency) can be obtained for a narrow range of operating conditions, which can be ensured by proper sizing of biofilter diameter and height.     

Expanded and packed bed albumin adsorption on fluoride modified zirconia

The expanded bed characteristics of 75-103microm fluoride-modified zirconia (FmZr) particles synthesized by a fed batch oil emulsion process were investigated. These particles are distinguished from commercially available expanded-bed adsorbents by virtue of their high density (2.8 g/cc) and the mixed mode protein retention mechanism which allows for the retention of both cationic and anionic proteins. The linear velocity versus bed porosity data agree with the Richardson-Zaki relationship with the terminal velocity in infinite medium of 2858.4 cm/h and a bed expansion index of 5.1. Residence time distribution (RTD) studies and bovine serum albumin (BSA) adsorption studies were performed as a function of the height of the settled bed to the column diameter (H:D) ratio and degree of bed expansion with superficial velocities of 440 to 870 cm/h. The settled bed, a 2x expanded bed, and a 3x expanded bed were studied for the H:D ratios of 1:1, 2:1, and 3:1. The dynamic binding capacity (DBC) at 5% breakthrough was low (2-8 mg BSA/mL settled bed) and was independent of the H:D ratio or the degree of bed expansion. The saturation DBC was 32.3 +/- 7.0 mg BSA/mL settled bed. The adsorption-desorption kinetics and intraparticle diffusion for protein adsorption on FmZr (38-75 micrometer) were investigated by studying the packed bed RTD and BSA adsorption as a function of temperature and flow rate. The data show that the adsorption-desorption kinetics along with intraparticle diffusion significantly influence protein adsorption on FmZr. Low residence times ( approximately 0.8 min) of BSA result in a DBC at 5% breakthrough which is 3.5-fold lower compared to that at 6-fold higher protein residence time. At low linear velocity (45 cm/h) the breakthrough curve is nearly symmetrical and becomes asymmetrical and more dispersed at higher linear velocity (270 cm/h) due to the influence of slow adsorption-desorption kinetics and intraparticle diffusion.Copyright 1998 John Wiley & Sons, Inc. Bioeng 60: 333-340, 1998.     

Lactate from cultures of Lactobacillus casei recovered in a fluidized bed column using ion exchange resin

Summary Lactic acid produced by continuous culture of L.casei in an upflow packed bed reactor, was recovered with Amberlite IRA 400 in a fluidized bed column. Bed expansions of 1.25 and 2.25 were applied. Reutilization did not alter the capability of net recovery of 0.048 ± 0.01 g lactic acid/g resin. When 2200 cm/h of ascensional velocity was used, (bed expansion of 2.25), the resin adsorbed 39.3% of the initial lactic acid and 63.5% was eluted. This resin supported the highest exchange capacity of 0.126 g lactic acid/g resin. Applying high flow rates, the process has potential industrial applications due to the short time employed.     

Computational fluid dynamics simulation improves the design and characterization of a plug-flow-type scale-down reactor for microbial cultivation processes

The scale-up of bioprocesses remains one of the major obstacles in the biotechnology industry. Scale-down bioreactors have been identified as valuable tools to investigate the heterogeneities observed in large-scale tanks at the laboratory scale. Additionally, computational fluid dynamics (CFD) simulations can be used to gain information about fluid flow in tanks used for production. Here, we present the rational design and comprehensive characterization of a scale-down setup, in which a flexible and modular plug-flow reactor was connected to a stirred-tank bioreactor. With the help of CFD using the realizable k-ε model, the mixing time difference between a 20 and 4000 L bioreactor was evaluated and used as scale-down criterion. CFD simulations using a shear stress transport (SST) k-ω turbulence model were used to characterize the plug-flow reactor in more detail, and the model was verified using experiments. Additionally, the model was used to simulate conditions where experiments technically could not be performed due to sensor limitations. Nevertheless, verification is difficult in this case as well. This was the first time a scale-down setup was tested on high-cell-density Escherichia coli cultivations to produce industrially relevant antigen-binding fragments (Fab). Biomass yield was reduced by 11% and specific product yield was reduced by 20% during the scale-down cultivations. Additionally, the intracellular Fab fraction was increased by using the setup. The flexibility of the introduced scale-down setup in combination with CFD simulations makes it a valuable tool for investigating scale effects at the laboratory scale. More information about the large scale is still necessary to further refine the setup and to speed up bioprocess scale-up in the future.     

Continuous biosynthesis of biodiesel from waste cooking palm oil in a packed bed reactor: optimization using response surface methodology (RSM) and mass transfer studies

This study aimed to develop an optimal continuous procedure of lipase-catalyzes transesterification of waste cooking palm oil in a packed bed reactor to investigate the possibility of large scale production further. Response surface methodology (RSM) based on central composite rotatable design (CCRD) was used to optimize the two important reaction variables packed bed height (cm) and substrate flow rate(ml/min) for the transesterification of waste cooking palm oil in a continuous packed bed reactor. The optimum condition for the transesterification of waste cooking palm oil was as follows: 10.53 cm packed bed height and 0.57 ml/min substrate flow rate. The optimum predicted fatty acid methyl ester (FAME) yield was 80.3% and the actual value was 79%. The above results shows that the RSM study based on CCRD is adaptable for FAME yield studied for the current transesterification system. The effect of mass transfer in the packed bed reactor has also been studied. Models for FAME yield have been developed for cases of reaction control and mass transfer control. The results showed very good agreement compatibility between mass transfer model and the experimental results obtained from immobilized lipase packed bed reactor operation, showing that in this case the FAME yield was mass transfer controlled.     

Isolation of a recombinant formate dehydrogenase by pseudo-affinity expanded bed adsorption.

Formate dehydrogenase (FDH) is an enzyme of industrial interest, which is recombinantly expressed as an intracellular protein in Escherichia coli. In order to establish an efficient and reliable purification protocol, an expanded bed adsorption (EBA) process was developed, starting from the crude bacterial homogenate. EBA process design was performed with the goal of finding operating conditions which, on one hand, allow efficient adsorption of the target protein and which, on the other hand, support the formation of a perfectly classified fluidised bed (expanded bed) in the crude feed solution. A pseudo-affinity ligand (Procion Red HE3B) was used to bind the FDH with high selectivity and reasonable capacity (maximum equilibrium capacity of 30 U/ml). Additionally, a simplified modelling approach, involving small packed beds for generation of process parameters, was employed for defining the operating conditions during sample application. In combination with extended elution studies, a process was set up, which could be scaled up to 7.5 l of adsorbent volume yielding a total amount of 100,000 U of 94% pure FDH per run. On this scale, 19 l of a benzonase-treated E. coli homogenate of 15% wet-weight (pH 7.5, 9 mS/cm conductivity) were loaded to the pseudo-affinity adsorbent (0.25 m sed. bed height, 5 x 10(-4) m/s fluid velocity). After a series of two wash steps, a particle-free eluate pool was obtained with 85% yield of FDH. This excellently demonstrates the suitability of expanded bed adsorption for efficient isolation of proteins by combining solid-liquid separation with adsorptive purification in a single unit operation.