Experimental and CFD-PBM Study of Oxygen Mass Transfer Coefficient in Different Impeller Configurations and Operational Conditions of a Two-Phase Partitioning Bioreactor

In this work, gas dispersion in a two-phase partitioning bioreactor is analyzed by calculating volumetric oxygen mass transfer coefficient which is modeled using a commercial computational fluid dynamics (CFD), code FLUENT 6.2. Dispersed oxygen bubbles dynamics is based on standard “k-ε” Reynolds-averaged Navier-Stokes (RANS) model. This paper describes a three-dimensional CFD model coupled with population balance equations (PBE) in order to get more confirming results of experimental measurements. Values of k _L a are obtained using dynamic gassing-out method. Using the CFD simulation, the volumetric mass transfer coefficient is calculated based on Higbie’s penetration theory. Characteristics of mass transfer coefficient are investigated for five configurations of impeller and three different aeration flow rates. The pitched six blade type, due to the creation of downward flow direction, leads to higher dissolved oxygen (DO) concentrations, thereby, higher values of k _L a compared with other impeller compositions. The magnitude of dissolved oxygen percentage in the aqueous phase has direct correlation with impeller speed and any increase of the aeration magnitude leads to faster saturation in shorter periods of time. Agitation speeds of 300 to 800 rpm are found to be the most effective rotational speeds for the mass transfer of oxygen in two-phase partitioning bioreactors (TPPB).     

Interpretation of static and dynamic responses of a dissolved oxygen electrode in viscous broths

The response of an amperometric oxygen electrode is studied theoretically and experimentally for the case of significant liquid film resistance at the outer side of the membrane. The behaviour of the probe in a gas stream is predicted by using a transfer function which involves an electrode model including oxygen diffusion within the electrode. The effects of the liquid film which exists at the membrane surface when dissolved oxygen concentrations are measured, are taken into account by modifying the transfer function. The new expression obtained is used to model the step response of the probe in a 2-1 stirred tank fermentor filled with water or xanthane solutions at different concentrations. First, the results are used to correct automatically the steady-state voltage readings of the probe. Secondly, the probe transfer function is used to evaluate k₁a by dynamic measurements: the response to a step change in the gas concentration is transformed via the fast Fourier transform algorithm and k₁a is identified in the Fourier domain by a Gauss-Newton algorithm. Data acquisition, Fourier transform and k₁a identification are implemented on-line on a HP-87 computer. This method of obtaining k₁a values appears to be a generalized moment method. It is shown that it is necessary to consider the liquid film dynamics around the probe in the actual fermenting conditions to evaluate k₁a successfully.     

The effect of the physical properties of the liquid phase on the gas-liquid mass transfer coefficient in two- and three-phase agitated systems

The results of studies concerning two- and three-phase systems in an agitated vessel are presented. The aim of our research was to investigate the effect of the physical properties of the liquid phase on the value of the volumetric gas-liquid mass transfer coefficient in mechanically agitated gas-liquid and gas-solid-liquid systems. Our experimental studies were conducted in a vessel with an internal diameter of 0.288 m. The flat bottom vessel, equipped with four baffles, was filled with liquid up to a height equal to the inner diameter. The liquid volume was 0.02 m³. Three high-speed impellers of a diameter equal to 0.33 of the vessel diameter were used: Rushton turbine (RT), Smith turbine (CD 6), or A 315 impeller. The measurements were carried out in coalescing and non-coalescing systems. Distilled water and aqueous solutions of an electrolyte (sodium chloride) of two different concentrations were used as the liquid phase. The gas phase was air. In the three-phase system, particles of sea sand were used as solid phase. The measurements were conducted at five different gas-flow rates and three particle loadings. Volumetric gas-liquid mass transfer coefficients were measured using the dynamic method. The presence and concentration of an electrolyte strongly affected the value of the gas-liquid mass transfer coefficient in both two- and three-phase systems. For all agitators used, significantly higher k _l a coefficient values were obtained in the 0.4 kmol m⁻³ and 0.8 kmol m⁻³ aqueous NaCl solutions compared with the data for a coalescing system (with distilled water as the liquid phase). The k _l a coefficient did not exhibit a linear relationship with the electrolyte concentration. An increase in the sodium chloride concentration from 0.4 kmol m⁻³ to 0.8 kmol m⁻³ caused a considerable decrease in the volumetric mass transfer coefficient in both the two-phase and three-phase systems. It was concluded that the mass transfer processes improved at a certain concentration of ions; however, above this concentration no further increase in k _l a could be achieved.     

Study on Mass Transfer of Isopropylbenzene and Oxygen in a Two-Phase Partitioning Bioreactor in the Presence of Silicone Oil

A two-phase partitioning bioreactor to treat gas effluents polluted by volatile organic compound has been developed. In this work, both the mass transfer of isopropylbenzene (IPB) and oxygen have been considered in relation to their influence on the hydrodynamics of the reactor and the type of silicone oils used as a second phase. The synergistic effect of silicone oil and stirrer speed on the global oxygen mass transfer coefficient (K _L a) and gas holdup (up to 12%) have been investigated. The addition of 10% of low viscosity silicone oil (10 cSt) in the reactor does not significantly affect the oxygen transfer rate. The very high solubility of IPB in the silicone oil leads to an enhancement of driving force term, especially for high fraction of silicone oil. However, it does not seem useful to exceed a volume fraction of 10% since K _L a _IPB decreases sharply at higher proportions of silicone oil. K _L a _IPB and K _L a _O2 evolve in the same way with the proportion of silicone oil. These results confirm the potentialities of our bioreactor to improve both the oxygen and pollutant gas transfer in the field of the treatment of gaseous pollutants, even for highly concentrated effluents.     

Non-and near-resonant (vv) energy transfer between the isotopes of N2 and of CO in liquid Ar and in the gas phase at 85 k

Rate constants for vibrational energy transfer have been measured in the system N₂(v= 1) +CO(v=0) as a function of energy mismatch by using isotopic derivatives of N₂ and of CO. These rate constants have been determined both in the gas phase and in liquid argon solution at 85 K. For non-resonant processes we find k_L=k_G for the nearest to resonant system we find k_L <k_G. This result is considered to arise because long range forces important in near-resonant energy transfer operate to a lesser extent in the liquid phase.     

Prediction of power consumption in mechanically agitated gassed reactor in viscous batch

Transport characteristics such as volumetric mass transfer coefficients, k_La, power input, P, gas hold-up, γ, and mixing time, t_m, are the key parameters in the design of mechanically agitated gasliquid contactors. For their successful design, values of the key parameters can be estimated using empirical correlations. Power input in this case is very often used as the scale of energy dissipation for other characteristics. Our goal was to propose reliable power input correlations for viscous batch processes, which are widely used in industry. The measurements were carried out in a pilot-plant vessel and also results from a laboratory vessel were used to develop the correlations. Different types of impellers and their combinations were used, including radial, axial, and combined liquid flow impellers. The power input was measured in a multiple-impeller vessel at different impeller frequencies and several gas flow rates. Correlation equations describing the behavior of particular impellers were evaluated. In addition, separate correlations for the bottom and upper sections in the multiple-impeller vessel were presented. These correlations can be used for impeller power prediction in industrial scale vessels under a wide range of operational conditions.     

Competitive facilitated transport through liquid membranes

A mathematical model is presented which solves the dimensionless, transient, non-linear partial differential equations governing the competitive facilitated transport of two gases through a liquid membrane. The model incorporates the mass transfer coefficients in the boundary conditions for the free gas concentrations. Several studies were carried out. A comparison of this model with a steady-state “equilibrium core” model was excellent. Through varying the dimensionless parameters, it was found that gas I would have a higher steady-state facilitation factor than gas 2 if k₁ >k₂ and k_-1k_-2. The boundary conditions and mass transfer coefficients were also varied to see their effects on the facilitation factors. The idea of pumping one of the gases against its concentration gradient was shown to be theoretically possible.     

Scale-up of microbubble dispersion generator for aerobic fermentation

A laboratory-scale microbubble dispersion (MBD) generator was shown to improve oxygen transfer to aerobic microorganisms when coupled to the conventional air-sparger. However, the process was not demonstrated on a large scale to prove its practical application. We investigated the scale-up of a spinning-disk MBD generator for the aerobic fermentation of Saccharomyces cerevisiae (baker’s yeast). A 1-L spinning-disk MBD generator was used to supply air for 1- and 50-L working volume fermentation of baker’s yeast. For the two levels investigated, the MBD generator maintained an adequate supply of surfactant-stabilized air microbubbles to the microorganisms at a relatively low agitation rate (150 rpm). There was a significant improvement in oxygen transfer to the microorganism relative to the conventional sparger. The volumetric mass transfer coefficient, k _L a, for the MBD system at 150 rpm was 765 h⁻¹ compared to 937 h⁻¹ for the conventional sparger at 500 rpm. It is plausible to surmise that fermentation using larger working volumes may further improve the k _L a values and the dissolved oxygen (DO) levels because of longer hold-up times and, consequently, improve cell growth. There was no statistically significant difference between the cell mass yield on substrate (0.43 g/g) under the MBD regime at an agitation rate of 150 rpm and that achieved for the conventional air-sparged system (0.53 g/g) at an agitation rate of 500 rpm. The total power consumption per unit volume of broth in the 50-L conventional air-sparged system was threefold that for the MBD unit for a similar product yield. Practical application of the MBD technology can be expected to reduce power consumption and therefore operating costs for aerobic fermentation.     

Influence of operating conditions and vessel size on oxygen transfer during cellulase production

The production of low-cost cellulase enzyme is a key step in the development of an enzymatic-based process for conversion of lignocellulosic biomass to ethanol. Although abundant information is available on cellulase production, little of this work has examined oxygen transfer. We investigated oxygen transfer during the growth of Trichoderma reesei, a cellulase-producing microorganism, on soluble and insoluble substrates in vessel sizes from 7 to 9000 L. Oxygen uptake rates and volumetric mass transfer coefficients (k _La) were determined using mass spectroscopy to measure off gas composition. Experimentally measured k _La values were found to compare favorably with a k _La correlation available in the literature for a non-Newtonian fermentation broth during the period of heavy cell growth.     

Temperature dependence of vibrational energy transfer in liquids: VV relaxation of CO(ν = 1) by O2 in liquid Ar from 86 to 145 K

The rate constant for the energy transfer process

has been measured for CO and O₂ dilute in liquid Ar along the coexistence curve between 86 and 145 K. The results are compared to gas phase measurements of this rate constant over the same temperature range. The ratio of the rate constants in the two phases is compared to that predicted by an application of the isolated binary collision (IBC) model for energy trasnfer, which scales the energy transfer rate constants by the rate constants for binary collisions in the two phases. The theoretical predictions of the IBC model are that the liquid state energy transfer rate constant, k^ℓ, should be 30–60% greater than the gas phase energy transfer rate constant, k^g. The measured value of k^ℓ is, within the experimental error, equal to that for k^g. Possible reasons for the discrepancy are discussed.     

Studies on thermo-acoustic parameters of poly(ethylene glycol)- 400 at different temperatures

The ultrasonic velocity and density have been measured at different temperatures between 299 and 363 K for the pure liquid sample, poly(ethylene glycol) with average molecular mass 400 g mol^?1 (PEG 400). From these, isentropic compressibility (β), intermolecular free length (L _f), acoustic impedance (Z), molar volume (V _m), Schaff’s available volume V _a(s), molar sound velocity (R _a), and molar compressibility (W) have been evaluated. The variations of these parameters with the temperature of the sample have been studied. Data so obtained are employed to compute other thermodynamic parameters. Variations in various parameters with respect to temperature are discussed in the light of the results obtained.     

Amides as a model system of low molar mass algal organic matter. Influence on the adsorption of p-nitrophenol on activated carbon

In this study, the adsorption equilibrium and diffusivity parameters of p-nitrophenol were estimated for water containing different concentrations of secondary amides. Commercial powdered activated carbon was used as an adsorbent. The external mass transfer coefficient (k_f), the surface diffusion coefficient (D_s) and the standard free Gibbs energy were calculated for p-nitrophenol in the presence of different secondary amide concentrations. The analysis established that there are correlations between structural parameters of amides, on the one hand, and diffusion and thermodynamic parameters for p-nitrophenol adsorption process, on the other. It was noticed that voluminous hydrophobic amides decreased the adsorption capacity of p-nitrophenol on activated carbon. On the basis of the results obtained for external mass transfer coefficients, it is assumed that amides cause the reduction of adsorption capacity of p-nitrophenol onto activated carbon by concentrating at the solid/liquid interface.     

A faradaic impedance study of the reduction of oxygen from aqueous alkaline solution at the dropping mercury electrode

The reduction of oxygen to hydrogen peroxide at a dropping mercury electrode in an aqueous solution of 1 M KNO₃+0.04 M KOH (pH=12.35) has been studied by means of impedance measurements as a function of frequency and d.c. potential. The reaction appears to be nearly reversible in the dc sense, but quasi-reversible in the ac sense. The impedance data obey the Randles' equivalent circuit with the following apparent values for the kinetic parameters: standard heterogeneous rate constant k_sh^a=0.035 cm s^?1 and cathodic transfer coefficient α^a_c=0.22. The results are interpreted in terms of a two-step charge transfer mechanism with the step O₂+eO₂^? being rate-determining.     

Numerical study of nitrogen desorption by rapid oxygen purge for a medical oxygen concentrator

Efficient desorption of selectively adsorbed N₂ from air in a packed column of LiX zeolite by rapidly purging the adsorbent with an O₂ enriched gas is an important element of a rapid cyclic pressure swing adsorption (RPSA) process used in the design of many medical oxygen concentrators (MOC). The amount of O₂ purge gas used in the desorption process is a sensitive variable in determining the overall separation performance of a MOC unit. Various resistances like (a) adsorption kinetics, (b) column pressure drop, (c) non-isothermal column operation, (d) gas phase mass and thermal axial dispersions, and (e) gas-solid heat transfer kinetics determine the amount of purge gas required for efficient desorption of N₂. The impacts of these variables on the purge efficiency were numerically simulated using a detailed mathematical model of non-isothermal, non-isobaric, and non-equilibrium desorption process in an adiabatic column. The purge gas quantity required for a specific desorption duty (fraction of total N₂ removed from a column) is minimum when the process is carried out under ideal, hypothetical conditions (isothermal, isobaric, and governed by local thermodynamic equilibrium). All above-listed non-idealities (a?Ce) can increase the purge gas quantity, thereby, lowering the efficiency of the desorption process compared to the ideal case. Items (a?Cc) are primarily responsible for inefficient desorption by purge, while gas phase mass and thermal axial dispersions do not affect the purge efficiency under the conditions of operation used in this study. Smaller adsorbent particles can be used to reduce the negative effects of adsorption kinetics, especially for a fast desorption process, but increased column pressure drop adds to purge inefficiency. A?particle size range of ??300?C500???m is found to require a?minimum purge gas amount for a given desorption duty. The purge gas requirement can be further reduced by employing a pancake column design (length to diameter ratio, L/D<0.2) which lowers the column pressure drop, but hydrodynamic inefficiencies (gas mal-distribution, particle agglomeration) may be introduced. Lower L/D also leads to a smaller fraction of the column volume that is free of N₂ at the purge inlet end, which is required for maintaining product gas purity. The simulated gas and solid temperature profiles inside the column at the end of the rapid desorption process show that a finite gas-solid heat transfer coefficient affects these profiles only in the purge gas entrance region of the column. The profiles in the balance of the column are nearly invariant to the values of that coefficient. Consequently, the gas-solid heat transfer resistance has a minimum influence on the overall integrated N₂ desorption efficiency by O₂ purge for the present application.     

Hydrodynamic and mass transfer studies in an external-loop air-lift bioreactor for immobilized animal cell culture

Air-lift bioreactors containing suspended or immobilized animal cells have been used for the production of a variety of high-value biologicals. In the bioprocessing industry, there is a need to study and quantify the relationships between bioreactor-system properties such as mixing, flow, mass transfer, and cell processes. In the present study, the performance of a 1-L external-loop air-lift bioreactor was investigated by studying gas-liquid oxygen transfer, mixing time, liquid velocity and gas hold-up at various aeration rates. These studies were performed over a range (0-25%) of loadings of small (500-800 μm) calcium alginate beads to investigate the effect of using various concentrations of cell immobilization matrices on the physical properties of the system. At an aeration rate of 0.5 vvm, the mixing time was decreased by 50%, from 75 s at 0% bead loading to 38 s at 10% bead loading. A minimum liquid velocity of 10 cm/s was required to keep the alginate beads in suspension. As bead loading increased, flow within the reactor went from turbulent conditions to the transition zone. At all bead loadings tested, the gas hold-up increased by only 2% with an increase in aeration rate from 0.1 to 1.0 vvm, regardless of whether the total reactor volume (i.e., liquid and beads) or the liquid volume was used in calculating the hold-up. A mathematical correlation was developed for expressing the dependence of the volumetric mass-transfer coefficient, k₁a, on aeration rate (vvm) and microbead loading. With this equation it was possible to predict, within 20%, the k₁a knowing the gas-flow rate and the volume percentage of microbeads present in the bioreactor. A theoretical study was also performed to calculate the oxygen transfer from the bulk liquid to the center of microcapsules containing animal cells using experimental k₁a data. The results suggest that whereas there is no oxygen limitation at 10 to 15% microcapsule loading, there is a potential mass-transfer problem at 25% loading if the bioreactor is operated at an aeration rate of less than 1.06 vvm.     

Simultaneous high-throughput determination of interaction kinetics for drugs and cyclodextrins by high performance affinity chromatography with mass spectrometry detection

The individual determination of the apparent dissociation rate constant (k_d,app) using high performance affinity chromatography (HPAC) is a tedious process requiring numerous separate tests and massive data fitting, unable to provide the apparent association rate constant (k_a) and equilibrium binding constant (K_a). In this study, a HPAC with mass spectrometry detection (HPAC-MS/MS) was employed to determine the drug-cyclodextrin (CD) interaction kinetics with low sample loading quantity (<10 ng per injection for single compound) and high-throughput yield as twenty drugs determined in one injection. The k_d,app measured by HPAC-MS/MS approach were 0.89 ± 0.07, 4.34 ± 0.01, 1.48 ± 0.01 and 7.77 ± 0.04 s⁻¹ for ketoprofen, trimethoprim, indapamide and acetaminophen, with k_d,app for acetaminophen consistent with that from the HPAC method with UV detector in our previous studies. For twenty drugs with diverse structures and chemical properties, good correlationship was found between k_d,app measured by single compound analysis method and high-throughput HPAC-MS/MS approach, with the correlation coefficient of 0.987 and the significance F less than 0.001. Comprehensive quantification of k_a,app, k_d,app and K_a values was further performed based on the measurement of k_d,app by peak profiling method and K_a by the peak fitting method. And the investigation of the drug-CD interaction kinetics under different conditions indicated that the column temperature and mobile phase composition significantly affected the determination of k_a,app, k_d,app and K_a while also dependent on the acidity and basicity of drugs. In summary, the high-throughput HPAC-MS/MS approach has been demonstrated high efficiency in determination of the drug-CD primary interaction kinetic parameter, especially, k_d,app, being proven as a novel tool in screening the right CD for the solubilization of the right drug.     

Optimization of Lactic Acid Production by Pellet-Form Rhizopus oryzae in 3-L Airlift Bioreactor Using Response Surface Methodology

The influence of two key environmental factors, pH and oxygen transfer coefficient (k _La), was evaluated on the lactic acid production as the main answer and, on the size of cell pellets of the fungal strain Rhizopus oryzae KPS106, as second dependant answer by response surface methodology using a central composite design. The results of the analysis of variance and modeling demonstrated that pH and k _La had a significant effect on lactic acid production by this strain. However, no interaction was observed between these two experimental factors. pH and k _La had no significant influence on the pellet size. Optimal pH and k _La of the fermentation medium for lactic acid production from response surface analysis was 5.85 and of 3.6 h^?1, respectively. The predicted and experimental lactic acid maximal values were 75.4 and 72.0 g/l, respectively, with pellets of an average of 2.54?±?0.41 mm. Five repeated batches in series were conducted with a mean lactic acid production of 77.54 g/l. The productivity was increased from 0.75 in the first batch to 0.99 g/l h in the last fifth batch.     

New algorithms for the computation of column dynamics of multicomponent liquid phase adsorption

New and efficient numerical algorithms were developed for simulating column dynamics of multicomponent liquid phase adsorption. Simple and realistic models are used for the simulation. Langmuir form of isotherm and linear driving force rate expressions are employed in the model equations. Algorithms were formulated for three different rate control mechanisms, namely, film diffusion control, particle diffusion control and combined film and particle diffusion control. The algorithms derived are explicit with the exception of the requirement of solving a nonlinear equation in one single variable which is the concentration of a reference species. Thus the tedious iterative calculation procedure for solving simultaneous nonlinear equations in a multicomponent fixed bed system is avoided. Example calculations indicated very good numerical accuracy as verified from an independent check by means of an overall mass balance.Nomenclature C Concentration of speciesi in the solution. mol/l - h _i Dimensionless length variable for speciesi - H _i Dimensionless total length parameter for speciesi. - k _li Liquid phase mass transfer coefficient for speciesi, l/sec - K _li Overall liquid phase mass transfer coefficient for speciesi. l/sec - k _si Solid phase mass transfer coefficient for speciesi, l/sec - K _si Overall solid phase mass transfer coefficient for speciesi. l/sec - L Total bed height, m - Q _i Concentration of speciesi in the adsorbent. mol/g solid - t Absolute time, second - V Linear superficial flow rate, m/see - X _i Normalized liquid phase i species concentration - Y _i Normalized liquid phasei species concentration - Z Axial distance of the bed, m     

Ammonia absorption into water in a packed tower II: Measurement of mass transfer coefficients

A simple method is outlined for the experimental determination of mass transfer coefficients in gas absorption. The modification of existing theory shows that the mass transfer coefficient should be constant within a tower but hydrodynamic conditions limit this to a region of the tower below five to ten tower diameters where the water radial distribution is constant. The theory is extended to meet all possible operational variables but in most cases the result demands computer techniques to achieve solutions. Details are given of experimental procedures which are necessary to enable meaningful data to be obtained. One important point in experiment design is to keep the gas rate constant and to vary the liquid rate. In this way, plots of ln y versus the tower height Z give a straight line whose slope is approximately proportional to K_Ga. Using the technique, laboratory data have been obtained which parallel the operation of a number of industrial installations.     

Oxygen Transfer in Solid-State Cultivation Under Controlled Moisture Conditions

The aim of this work was to study oxygen transfer as a function of the initial moisture content in solid-state cultivation under controlled moisture conditions. The use of controlled moisture conditions prevents drastic changes in the medium during cultivation, allowing the use of a pseudo-steady-state model to estimate the overall oxygen mass transfer coefficient (K _L a) in the biofilm around the solid particles. Drechslera (Helminthosporium) monoceras, an aerobic mold that produces allergenic proteins, was cultured on wheat bran in a packed bed column bioreactor. The bed height (30 mm) and air flow rate (0.4 L/min) were selected to implement moisture control. The results show that there is an optimal moisture content (35 %) at which a lower biofilm thickness and packing of the bed improves K _L a. However, a higher biomass growth was obtained at 45 % moisture. The different patterns of biomass growth demonstrate the importance of the balance between aerial and film growth in solid-state cultivation. These results contribute to the understanding of oxygen transfer in solid fermentation, optimization of processes, and production of allergen extracts from D. (Helminthosporium) monoceras biomass.