Reactive extraction of cephalosporin antibiotics in hollow fiber membrane

Non-dispersive reactive extraction of cephalosporin antibiotics has been studied using hollow fiber membrane modules. Extraction as well as stripping has been studied using a pH swing procedure. Cephalosporin was extracted from an aqueous solution of cephalosporin having a pH above the pK_a2 value to an organic phase containing Aliquat-336 as the extractant and n-heptane as the diluent. The solute was stripped from the loaded organic phase to another aqueous phase of pH maintained well below the pK_a2 value of the cephalosporin. The extraction cum stripping relies on pH dependance of the distribution coefficient of cephalosporin in aqueous phase. Reasonably high solute recovery and mass transfer rate have been achieved in the hollow fiber module. Mass transfer performance of a single module has been evaluated and experimentally observed low value of height of transfer unit (HTU) indicates good prospect of hollow fiber membrane for the extraction duty.     

Extractive separation of penicillin G by facilitated transport via carrier supported liquid membranes

The facilitated transport of penicillin G (Pen G), through a supported liquid membrane with Amberlite LA-2 dissolved in 1-decanol, supported on a microporous polypropylene membrane, were studied. The distribution coefficient was obtained from a batch extraction experiment. The effects of flow rate, carrier concentration, initial concentration of Pen G, and the pH of feed and stripping phases on the transport rate of Pen G through the supported liquid membrane were also investigated. The results are in agreement with theoretical predictions, and it is demonstrated that the transport of Pen G through the supported liquid membrane is controlled simultaneously by mass transfer across both aqueous and liquid membranes. (c) 1993 John Wiley & Sons, Inc.     

Recovery of penicillin G from fermentation broth with reactive extraction in a mixer-settler

Summary In a laboratory countercurrent mixer-settler, penicillin was recovered from its fermentation broth by extraction with Amberlite LA-2 in n-butylacetate at pH 5.0 and reextracted from the ion-pair complex containing a solvent phase with a buffer at 7.2–7.5 with an overall degree of extraction above 90 %.Symbols A amine - AHP complex - c concentration - C partition coefficent - E degree of extraction - HP penicillin acid - K_G equilibrium constant - P, P^– penicillin acid anion Indices aq aqueous phase - org organic phase - A amine - AHP complex - G overall - HP free acid - P penicillin     

Influence of pluronic F68 on oxygen mass transfer

Pluronic F68 is one of the most used shear protecting additives in cell culture cultivations. It is well known from literature that such surface‐active surfactants lower the surface tension at the gas‐liquid interface, which influences the mass transfer. In this study, the effect of Pluronic F68 on oxygen mass transfer in aqueous solutions was examined. Therefore, the gassing in/gassing out method and bubble size measurements were used. At low concentrations of 0.02 g/L, a 50% reduction on mass transfer was observed for all tested spargers and working conditions. An explanation of the observed effects by means of Higbie's penetration or Dankwerts surface renewal theory was applied. It could be demonstrated that the suppressed movement of the bubble surface layer is the main cause for the significant drop down of the k_La‐values. For Pluronic F68 concentrations above 0.1 g/L, it was observed that it comes to changes in bubble appearance and bubble size strongly dependent on the sparger type. By using the bubble size measurement data, it could be shown that only small changes in mass transfer coefficient (k_L) take place above the critical micelle concentration. Further changes on overall mass transfer at higher Pluronic F68 concentrations are mainly based on increasing of gas holdup and, more importantly, by increasing of the surface area available for mass transfer. © 2013 American Institute of Chemical Engineers Biotechnol. Prog., 29:1278–1288, 2013     

A multiphasic hollow fiber reactor for the whole-cell bioconversion of 2-methyl-1,3-propanediol to (r)-beta-hydroxyisobutyric acid

This paper describes the bioconversion of 2-methyl-1,3-propanediol to (R)-beta-hydoxyisobutyric acid (HIBA) by Acetobacter ALEI in a hollow fiber membrane bioreaction system arrangement that allows the integration of three liquid phases: the aqueous bioconversion phase, the organic phase consisting of a solution of trioctyl phosphine oxide (TOPO) in isooctane, and the third phase consisting of a basic stripping solution that allows reextraction of HIBA from the organic phase. A comparison of HIBA mass transfer experiments was carried out in the membrane reactor with two and three phases for different pH and TOPO concentrations. The use of the three-phase arrangement allows the extraction of high quantities of HIBA from the aqueous medium (higher than 85%) independently of the pH, whereas in the two-phase system the percentage of HIBA extracted from the aqueous medium was lower, 42% in the best case, and strongly influenced by the pH. The percentage of the extractive agent TOPO in the organic phase influenced on the mass transfer rate in both bi- and triphasic arrangements. By simply integrating the re-extraction phase in the system it was possible to increase the extraction yield by 2-fold, reduce the amount of TOPO by 4-fold, and operate at the more favorable pH 4. A bioconversion experiment was done in these conditions (pH = 4, TOPO = 5%) to confirm the advantages of including the third stripping solution. Fed-batch operation of the triphasic membrane reactor was maintained for more than 20 h, reaching an HIBA concentration in the stripping solution of 29 g L(-)(1).     

Equilibrium modeling of extractive enzymatic hydrolysis of penicillin G with concomitant 6-aminopenicillanic acid crystallization

In the present downstream processing of penicillin G, penicillin G is extracted from the fermentation broth with an organic solvent and purified as a potassium salt via a number of back-extraction and crystallization steps. After purification, penicillin G is hydrolyzed to 6-aminopenicillanic acid, a precursor for many semisynthetic beta-lactam antibiotics. We are studying a reduction in the number of pH shifts involved and hence a large reduction in the waste salt production. To this end, the organic penicillin G extract is directly to be added to an aqueous immobilized enzyme suspension reactor and hydrolyzed by extractive catalysis. We found that this conversion can exceed 90% because crystallization of 6-aminopenicillanic acid shifts the equilibrium to the product side. A model was developed for predicting the equilibrium conversion in batch systems containing both a water and a butyl acetate phase, with either potassium or D-p-hydroxyphenylglycine methyl ester as counter-ion of penicillin G. The model incorporates the partitioning equilibrium of the reactants, the enzymatic reaction equilibrium, and the crystallization equilibrium of 6-aminopenicillanic acid. The model predicted the equilibrium conversion of Pen G quite reasonably for different values of pH, initial penicillin G concentration and phase volume ratio. The model can be used as a tool for optimizing the enzymatic hydrolysis.     

Two-stage recovery of S-adenosylmethionine using supported liquid membranes with strip dispersion

We report the selective recovery of S-adenosylmethionine (SAM) from fermentation broths using a two-stage supported liquid membrane system with strip dispersion (SLM-SD). The system utilized two MiniModule^® hollow-fiber membrane modules as microporous supports and an organic membrane solution consisting of the extractants of sodium di-2-ethylhexyl sulfosuccinate (AOT), di-(2-ethylhexyl)phosphoric acid (DEHPA), and trioctylphosphine oxide (TOPO) in the solvent n-octanol. SAM was extracted in the first membrane module. Methionine (Met) was captured by the first stripping solution and further purified in the second membrane module. pH values in the feed phase and the first and second stripping solutions, extractant concentrations, NaCl concentration, and the SAM acceptor in the first stripping solution were optimized. Strip dispersion mixing speed, pressure differences across the membranes, and flow rates of the feed and strip dispersion phases were investigated experimentally. The optimal extractant concentrations were: AOT 2.78 wt%, DEHPA 27.0 wt%, and TOPO 1.61 wt%. The optimal pH values in the feed phase and the first and second stripping solution were 3.0, 2.5, and 1.0, respectively. SAM extraction efficiency of 98.7%, SAM recovery efficiency of 91.8% and Met removal efficiency of 85.4% were achieved within 5 h. Finally, the mass transfer analysis indicated that the mass transfer resistances from the extraction reaction and the membrane phase were predominant.     

ENHANCED PRODUCTION OF 6-AMINOPENICILLANIC ACID IN AQUEOUS METHYL ISOBUTYL KETONE SYSTEM WITH IMMOBILIZED PENICILLIN G ACYLASE

Enzymatic hydrolysis of penicillin G for production of 6-amino-penicillanic-acid (6-APA) was achieved by using penicillin G acylase as catalyst in an aqueous-methylisobutyl ketone (MIBK) system. The optimization was carried out and it was found that the best conversion was improved 10% more than the aqueous system, which was obtained at the conditions: initial pH 8.0, 5.0% (W/V) substrate (penicillin G), and temperature at 35°C, and the ratio of aqueous and organic phase was 3:1. The stability of the biocatalyst was studied at the operational conditions. After 5 cycles of semi-batch reactions, the residual activity of penicillin G acylase was 69.2% of the initial activity. There was no apparent loss of the yield of product. This process has a potential application in the industrial scale production of 6-APA because it simplifies the process effectively.     

Ethanol stripping by pervaporation using porous PTFE membrane

In a shell-and-tube type of module containing either porous or nonporous tubular membranes, the sweeping action of a flow inert gas in the shell side was used to strip ethanol from an aqueous ethanol solution flowing countercurrently in the tube side. A calculation of the overall mass transfer coefficient, K_G, of the membrane used was made for this system. In ethanol stripping tests using a module containing polytetrafluoreethylene (PTFE) tubular membranes, the K_G was found to be more affected by the liquid flow rate than the gas flow rate. Moreover, the gas side mass transfer coefficient, k_G, was estimated to be about 5×10⁻⁵ mol/cm²·s·atm. The liquid side mass transfer coefficient, k_L, on the other hand, was found to increase linearly with the linear velocity of the aqueous solution. Also, at an average solution temperature range of 21 to 32°C, no significant change in the K_G was observed. Comparison of the K_G of different tubular membranes revealed that the K_G of the PTFE membrane was higher than that of polypropylene or silicone membranes under the given experimental conditions.     

Reactive extraction of penicillin G in hollow-fiber and hollow-fiber fabric modules

Penicillin G (Pen G) can be rapidly extracted in hollow-fiber liquid-liquid contactors using N-lauryl-N-trialkylmethylamine (Amberlite LA-2) as the extractant. n-Butylacetate is much better than decanol as a diluent for such an extraction, although decanol can give a partition coefficient four times larger. The overall mass transfer coefficient found is a function of aqueous flow on the lumen side of the fiber, and is less dependent on shell-side flow. In backextraction, the overall mass transfer coefficient determined is only one tenth that of the forward extraction, primarily because the hydrophobic hollow fibers used have a high mass transfer resistance under these conditions. The mass transfer experiments show that hollow-fiber extraction of Pen G is competitive with centrifugal extraction. The prospects for extraction of other fermentation products with hollow fibers can be estimated based on the present study.     

A comparative study of solid and liquid non‐aqueous phases for the biodegradation of hexane in two‐phase partitioning bioreactors

A comparative study of the performance of solid and liquid non‐aqueous phases (NAPs) to enhance the mass transfer and biodegradation of hexane by Pseudomonas aeruginosa in two‐phase partitioning bioreactors (TPPBs) was undertaken. A preliminary NAP screening was thus carried out among the most common solid and liquid NAPs used in pollutant biodegradation. The polymer Kraton G1657 (solid) and the liquid silicone oils SO20 and SO200 were selected from this screening based on their biocompatibility, resistance to microbial attack, non‐volatility and high affinity for hexane (low partition coefficient: K = C_g/C_NAP, where C_g and C_NAP represent the pollutant concentration in the gas phase and NAP, respectively). Despite the three NAPs exhibited a similar affinity for hexane (K ≈ 0.0058), SO200 and SO20 showed a superior performance to Kraton G1657 in terms of hexane mass transfer and biodegradation enhancement. The enhanced performance of SO200 and SO20 could be explained by both the low interfacial area of this solid polymer (as a result of the large size of commercial beads) and by the interference of water on hexane transfer (observed in this work). When Kraton G1657 (20%) was tested in a TPPB inoculated with P. aeruginosa, steady state elimination capacities (ECs) of 5.6 ± 0.6 g m^?3 h^?1 were achieved. These values were similar to those obtained in the absence of a NAP but lower compared to the ECs recorded in the presence of 20% of SO200 (10.6 ± 0.9 g m^?3 h^?1). Finally, this study showed that the enhancement in the transfer of hexane supported by SO200 was attenuated by limitations in microbial activity, as shown by the fact that the ECs in biotic systems were far lower than the maximum hexane transfer capacity recorded under abiotic conditions. Biotechnol. Bioeng. 2010;106: 731–740. © 2010 Wiley Periodicals, Inc.     

Reactive extraction of penicillin G in a pilot plant Karr-column II. Fermentation broths

Summary Penicillin G was extracted from mycelfree fermentation broths by means of the carrier (Amberlite LA-2) in n-butylacetate at pH 5 in a 7.6 m high pilot plant Karr-column with degrees of extraction E=98–99% and penicillin enrichments up to 3. The reextraction was carried out with phosphate buffer at pH-values above 7.5 with degree of extractions E=86–88% and penicillin enrichments up to 3. The penicillin and carrier losses were negligible. The influence of the process variables on the extraction degree was investigated. The penicillin extraction of the model medium and the fermentation broths were compared. Recommendations are given for the optimal penicillin recovery with reactive extraction.Symbols a specific interfacial area with regard to the volume of the continuous phase - cA concentration of carrier - cAHP,O concentration of complex in feed - cP,cP,O concentration of penicillin acid anion in theaqueous phase, in the feed - d ₃₂ Sauter droplet diameter - E degree of extraction - f stroke frequency - V _aq throughput of the aqueous phase - V ₀ throughput of the organic phase - Z dimensionsless longitudinal coordinate of the column with regard to its active length (4m) - holdup of the organic phase     

Syngas fermentation to biofuel: Evaluation of carbon monoxide mass transfer coefficient (kLa) in different reactor configurations

Lignocellulosic biomass such as agri‐residues, agri‐processing by‐products, and energy crops do not compete with food and feed, and is considered to be the ideal renewable feedstocks for biofuel production. Gasification of biomass produces synthesis gas (syngas), a mixture primarily consisting of CO and H₂. The produced syngas can be converted to ethanol by anaerobic microbial catalysts especially acetogenic bacteria such as various clostridia species.One of the major drawbacks associated with syngas fermentation is the mass transfer limitation of these sparingly soluble gases in the aqueous phase. One way of addressing this issue is the improvement in reactor design to achieve a higher volumetric mass transfer coefficient (k_La). In this study, different reactor configurations such as a column diffuser, a 20‐μm bulb diffuser, gas sparger, gas sparger with mechanical mixing, air‐lift reactor combined with a 20‐μm bulb diffuser, air‐lift reactor combined with a single gas entry point, and a submerged composite hollow fiber membrane (CHFM) module were employed to examine the k_La values. The k_La values reported in this study ranged from 0.4 to 91.08 h^?1. The highest k_La of 91.08 h^?1 was obtained in the air‐lift reactor combined with a 20‐μm bulb diffuser, whereas the reactor with the CHFM showed the lowest k_La of 0.4 h^?1. By considering both the k_La value and the statistical significance of each configuration, the air‐lift reactor combined with a 20‐μm bulb diffuser was found to be the ideal reactor configuration for carbon monoxide mass transfer in an aqueous phase. © 2010 American Institute of Chemical Engineers Biotechnol. Prog., 2011     

Balancing the Ethanol Formation in Continuous Bioreactors with Ethanol Stripping

Ethanol microbially produced in continuously operated aerated bioreactors is partly discharged with the liquid flow and is partly stripped off with the gas phase. The calculation of ethanol formation rates – as required, for example, in data evaluation by Metabolic Flux Analysis – solely based on the liquid‐borne ethanol, while neglecting the ethanol stripping, inevitably results in defective data interpretation. The proportion of stripped ethanol can be measured or, alternatively, calculated. The developed structured model describes the ethanol stripping as a two‐step process while differentiating between the phase transition from the liquid into the gas and the discharge of the evaporated ethanol with the off‐gas. As shown by model analysis as well as stripping experiments, the stripping rate is mainly determined by the ethanol discharge rather than by the phase transition. Consequently, the ethanol‐stripping rate depends on the specific gas flow rate and the partition coefficient for ethanol (at 30 °C, K_L/G = 3125 L/L), but not on the mass transfer coefficient. As shown by model simulations, the lower the dilution rate and the larger the gas flow rate of an aerated chemostat with a microbial ethanol formation, the higher the proportion of stripped ethanol. Under practically relevant conditions, more than 30 % of the produced ethanol may be stripped off. Exhaust‐gas coolers, actually used to reduce water losses by evaporation, do not prevent but slightly affect ethanol stripping.     

Development of a novel bioreactor system for treatment of gaseous benzene

A novel, continuous bioreactor system combining a bubble column (absorption section) and a two-phase bioreactor (degradation section) has been designed to treat a gas stream containing benzene. The bubble column contained hexadecane as an absorbent for benzene, and was systemically chosen considering physical, biological, environmental, operational, and economic factors. This solvent has infinite solubility for benzene and very low volatility. After absorbing benzene in the bubble column, the hexadecane served as the organic phase of the two-phase partitioning bioreactor, transferring benzene into the aqueous phase where it was degraded by Alcaligenes xylosoxidans Y234. The hexadecane was then continuously recirculated back to the absorber section for the removal of additional benzene. All mass transfer and biodegradation characteristics in this system were investigated prior to operation of the integrated unit, and these included: the mass transfer rate of benzene in the absorption column; the mass transfer rate of benzene from the organic phase into the aqueous phase in the two-phase bioreactor; the stripping rate of benzene out of the two-phase bioreactor, etc. All of these parameters were incorporated into model equations, which were used to investigate the effects of operating conditions on the performance of the system. Finally, two experiments were conducted to show the feasibility of this system. Based on an aqueous bioreactor volume of 1 L, when the inlet gas flow and gaseous benzene concentration were 120 L/h and 4.2 mg/L, respectively, the benzene removal efficiency was 75% at steady state. This process is believed to be very practical for the treatment of high concentrations of gaseous pollutants, and represents an alternative to the use of biofilters.     

Predispersed solvent extraction of succinic acid aqueous solution by colloidal liquid aphrons in column

A study of the PDSE (predispersed solvent extraction) for succinic acid by colloidal liquid aphrons was conducted. The organic phase contaning TOA (tri-n-octylamine) and 1-octanol permits a selective extraction of succinic acid from its aqueous solution. There was no difference of the extractability of PDSE and that of conventional mixer-settler type extraction. Taking into account the no mechanical mixing in PDSE, it was concluded that the PDSE process is more adaptive than the conventional mixer-settler type extraction process. From mass transfer analysis at the various concentration of TOA in counter-current continuous operation, the concentration of TOA had no influence on the mass transfer coefficient. The loading values in continuous PDSE were almost same as those in batch operation.     

Recovery of L-lysine from dilute water solutions by liquid pertraction

Kinetics of liquid membrane (Pertraction) recovery of L-lysine from dilute aqueous solutions is studied in a tow-compartment glass cell. A 5% (vol) solution of the cation exchange carrier di(2-ethylhexyl)phosphoric acid in n-decane was used as intermediate, membrane liquid. The third stripping phase was 1/v hydrochloric acid. The reaction mechanism and stoichiometry were defined, and on the basis of the proposed mathematical model of the process and the experimental data obtained, the mass transfer coefficients were evaluated. It was found that overall transfer rate is controlled by the eddy diffusion of transported species in the donor and membrane liquids. The results proved the feasibility of the pertraction process for recovery and concentration of L-lysine from its dilute aqueous solutions.     

Effect of organic solvents on oxygen mass transfer in multiphase systems: Application to bioreactors in environmental protection

The absorption of oxygen in aqueous–organic solvent emulsions was studied in a laboratory-scale bubble reactor at a constant gas flow rate. The organic and the gas phases were dispersed in the continuous aqueous phase. Volumetric mass transfer coefficients (k_La) of oxygen between air and water were measured experimentally using a dynamic method. It was assumed that the gas phase contacts preferentially the water phase. It was found that addition of silicone oils hinders oxygen mass transfer compared to air–water systems whereas the addition of decane, hexadecane and perfluorocarbon PFC40 has no significant influence. By and large, the results show that, for experimental conditions (organic liquid hold-up ≤10% and solubility ratio ≤10), the k_La values of oxygen determined in binary air–water systems can be used for multiphase (gas–liquid–liquid) reactor design with applications in environmental protection (water and air treatment processes).     

Systematic Approach to Solvent Selection for Biphasic Systems with a Combination of COSMO‐RS and a Dynamic Modeling Tool

Biphasic aqueous‐organic systems are important reaction systems for catalytic processes. This is especially true for biocatalysis where the range of accessible products can be significantly extended. In such systems, the aqueous phase is the reactive phase in which the biocatalyst is dissolved and the organic phase is nonreactive and acts as substrate reservoir and as in situ product extraction solvent. Here, the choice of the nonreactive phase is highly important for the overall performance of the system. In this contribution, a systematic approach to solvent selection for biphasic aqueous‐organic systems is presented with respect to partition coefficients. The model reaction is the stereoselective carbon‐carbon coupling of two 3,5‐dimethoxy‐benzaldehyde molecules to (R)‐3,3',5,5'‐tetramethoxy‐benzoin catalyzed by benzaldehyde lyase (EC 4.1.2.38) from Pseudomonas fluorescens. A systematic approach to solvent selection consisting of two steps is proposed: Firstly, the conductor‐like screening model for real solvents (COSMO‐RS) is used to facilitate a fast solvent screening. Since this is an ab initio approach it allows a pre‐screening without laborious experimental input. The proposed ranking of solvents, based on the ratio of partition coefficients at infinite dilution, is a sound basis for the successive steps. Secondly, a dynamic model is fitted to experimental data in order to obtain detailed and reliable results for mass transfer and partition coefficients. Therefore, the method makes efficient use of the experimental data and substantiates quantitative results with guided experiments.     

Physical and reactive extraction equilibria of penicillin G in a hydrogen-bond acceptor solvent system

Physical and reactive extraction equilibria of penicillin G were investigated experimentally and theoretically in the existence of n-butyl acetate as a hydrogen-bond acceptor solvent. Physical extraction equilibrium experiments were carried out varying the pH of aqueous phase and overall penicillin concentration. We compared the experimental data with the calculated results from four physical extraction equilibrium models suggested here and obtained the most reasonable model. Also, penicillin G was reactively extracted using Amberlite LA-2 in n-butyl acetate. The experimental variables were pH of the aqueous phase, overall amine concentration, and overall penicillin concentration. A combined equilibrium model including our physical extraction equilibrium expression and the reactive extraction equilibrium expression suggested by Reschke and Schügerl was used so as to analyze the current reactive extraction equilibrium system. The calculated results from the reactive extraction equilibrium model were in good agreement with the experimental data.