Mixing in liquid-impelled loop reactors

The liquid-impelled loop reactor is a new column-type bioreactor. The design of this device is based on the principle of the air-lift loop reactor. In the external-loop configuration used in this work, descending perfluorochemical drops bring about circulation of the continuous aqueous phase. Mixing of this continuous phase is characterized per section of the rector. Axial-dispersion coefficients for the tube with two-phase flow are determined and correlated with the energy dissipation in the tube. Comparisons with similar systems such as bubble columns and air-lift loop reactors are made. Overall mixing parameters are derived and used for calculation of the number of circulatins needed to achieve a certain degree of mixing. The hydrodynamic model from previous work is tested for the reactor configurations of this work. It can be useful to calculate circulation times.     

Oxygen transfer in liquid-impelled loop reactors using perfluorocarbon liquids

Oxygen transfer in the liquid-impelled loop reactor is described for a setup in which the perfluorochemical FC40 is aerated externally. Two sizes of reactors are investigated. The mass-transfer coefficient k appears to be lower with a factor of about 0.6 compared to gas liquid systems. the specific exchange area in the present experimental setup is found to be favorable when compared with gas liquid bioreactors at the same superficial dispersed-phase velocities. However, slow coalescence of the dispersed-phase drops in the phase separation section limits the dispersed-phase flow rate seriously. In Case this become crucial from the point of view of oxygen supply, special measures need to be found or alternatives such as combined sparging of air and solvent. (c) 1992 John Wiley & Sons, Inc.     

A simplified model of hydrodynamics in airlift loop reactors

As a function of the gas throughput the following parameters were measured in an external loop reactor with a riser diameter of 0.6 m and a gassed liquid height of 8.6 m: integral and local values of gas hold-up; liquid velocities; mixing times and axial dispersion coefficients of the liquid phase. The height of the reactor could be altered by reconstruction. Measurements were also carried out with lower heights than 8.6 m. Besides pure water, aqueous solutions of coalescing, non-coalescing and viscosity-increasing substances were used as model systems. With the results a general relationship between superficial gas velocity, gas hold-up and liquid velocity was established. This hydrodynamic model uses the relative velocity between gas and liquid phase as the fundamental parameter. The generally valid model consists of one term for the homogeneous and of two additional terms for the heterogeneous flow regime.     

Substrate effects in tower loop reactors

Summary Hansenula polymorpha was cultivated in a bubble column loop bioreactor employing ethanol and/or glucose as substrates. By varying the substrate concentration, the cultivations were carried out in non-limited, substrate limited and oxygen transfer limited growth ranges. The influence of the transitions from one range to another on reactor performance (OTR,k _L a, a) and cell productivity () were investigated. When employing ethanol as a substrate, the concentration considerably influences the fluid dynamics, mass transfer phenomena and cell productivity. When employing glucose as a substrate, glucose repression occurs. At high glucose concentrations no transition into the oxygen transfer limited growth is possible. The ethanol produced during the glucose repression influences the fluid dynamics, mass transfer phenomena and productivity. With decreasing glucose concentration the glucose repression can be gradually eliminated.     

Model for the production of L-tryptophan from L-serine and indole by immobilized cells in a three-phase liquid-impelled loop reactor

Production of L-tryptophan from L-serine and indole catalyzed by Escherichia coli, immobilized in k-carrageenan gel beads, is technically feasible in the liquidimpelled loop reactor (LLR), using an organic solvent, e.g. n-dodecane.With L-serine in large excess intrinsic reaction kinetics is approximately first order with respect to indole, with a reaction constant of 8.5×10^–5 m³ kg _dw ^–1 s^–1.The overall process kinetics is jointly controlled by intrinsic kinetics and by intraparticle mass transfer resistance, which can be quantified using an effectiveness factor.Mass transfer of indole from the organic to the aqueous phase and from the aqueous to the gel phase are relatively fast and thus have negligible influence in the overall process kinetics, under the operational conditions tested. However, they may become important if the process is intensified by increasing the cell concentration in the gel and/or the gel hold-up in the reactor.A simple model which includes indole mass balances over the aqueous and organic phases, mass transfer and reaction kinetics, with parameters experimentally determined in independent experiments, was successful in simulating L-tryptophan production in the LLR.List of Symbols a, b, c coefficients of the equilibrium curve for indole between organic and aqueous phases - A, B, C, D, E, F auxiliary variables used in liquid-liquid mass transfer studies - a _x specific interfacial area referred to the volume of the aqueous phase (m^–1) - A _x interfacial area (m²) - a _Y specific interfacial area referred to the volume of the organic phase (m^–1) - A _Y interfacial area (m²) - C _b substrate concentration in the bulk of the aqueous phase (kg m^–3) - C _e substrate concentration in exit stream (kg m^–3) - C _E biocatalyst concentration referred to the aqueous phase (kg m^–3) - C _{E s} biocatalyst concentration referred to the volume of gel (kg m^–3) - C _s substrate concentration at the gel surface (kgm^–3) - d, e, f coefficients of the equilibrium curve for indole between aqueous and organic phases - dp particle diameter (m) - K ₂ kinetic constant (s^–1) - K ₁ kinetic constant K₂/K_M (kg^–1 m³ s^–1) - K _M Michaälis-Menten constant (kgm^–3) - K _X mass transfer coefficient referred to the aqueous phase (ms^–1) - K _Xa_X volumetric mass transfer coefficient based on the volume of the aqueous phase (s^–1) - k _Y mass transfer coefficient referred to the organic phase (ms^–1) - K _Ya_Y volumetric mass transfer coefficient based on the volume of the organic phase (s^–1) - N _X mass flux of indole from organic to aqueous Phase (kg m^–2s^–1) - N _Y mass flux of indole from aqueous to organic phase (kg m^–2s^–1) - Q _e volumetric flow rate in exit stream (m³s^–1) - Q _f volumetric flow rate in feed stream (m³s^–1) - _obs observed reaction rate (kg s^–1 m^–3) - intrinsic reaction rate (kg s^–1 m^–3) - Re Reynolds number - Sc Schmidt number - Sh Sherwood number - t time (s) - u superficial velocity (m s^–1) - V _max maximum reaction rate (kg s^–1m^–3) - V _S volume of the support (m³) - V _X volume of aqueous phase (m³) - V _Y volume of the organic phase (m³) - X indole concentration in the aqueous phase (kgm^–3) - Y indole concentration in the organic phase (kg m^–3 Greek Letters overall effectiveness factor - _e external effectiveness factor - _i internal effectiveness factor - Thiele module A fellowship awarded to one of us (D.M.R.)by INICT is gratefuly acknowledged.     

Macromixing characteristics of gas-liquid jet loop reactors

Measurements of liquid macromixing characteristics are reported for a half industrial scaled jet loop reactor operating with air-water mixtures. Based on a model of loop reactors with sections of different mixing behavior the single circulation dispersion coefficient can be split into its components caused by the riser and the downcomer. The dispersion coefficient of the riser is about 100 times greater than that of the downcomer. The addition of gas involves greater dispersions coefficients. The comparison of the mixing times of the JLR with those of stirred vessels leads to the conclusion that the JLR is equivalent or even superior to stirred vessels.     

Hydrodynamic effects on cells in agitated tissue culture reactors

Tissue cells are known to be sensitive to mechanical stresses imposed on them by agitation in bioreactors. The amount of agitation provided in a microcarrier or suspension bioreactor should be only enough to provide an effective homogeneity. Three distinct flow regions can be identified in the reactor: bulk turbulent flow, bulk laminar flow, and boundary-layer flows. Possible mechanisms of cell damage are examined by analyzing the motion of microcarriers or free cells relative to the surrounding fluid, to each other, and to moving or stationary solid surfaces. The primary mechanisms of cell damage appear to result from (a) direct interaction between microcarriers and turbulent eddies, (b) collisions between microcarriers in turbulent flow, and (c) collisions against the impeller or other stationary surfaces. If the smallest eddies of turbulent flow are of the same size as the microcarrier beads, they may cause high shear stresses on the cells. Eddies the size of the average interbead spacing may cause bead-bead collisions which damage cells. The severity of the collisions increases when the eddies are also of the same size as the beads. Bead size and the interbead distance are virtually equal in typical microcarrier suspensions. Impeller collisions occur when the beads cannot avoid the impeller leading edge as it advances through the liquid. The implications of the results of this analysis on the design and operation of tissue culture bioreactors are also discussed.     

Modelling of multi-phase flow in jet loop reactors

A model for the steady state flow in jet loop reactors with two-phase flow has been developed. The model can be used to predict the critical effect in a reversed flow jet loop fermenter. The calculated critical effect according to the model corresponds well to experimental data.     

Cultivation of Hansenula polymorpha in tower loop reactors

Summary Hansenula polymorpha was cultured for extended periods in an air lift tower loop reactor (15 cm in diameter with a 275 cm bubbling layer height) with ethanol and/or glucose as the substrates. At constant operation conditions variations of the following parameters were measured: the consumption of the substrate and oxygen, the production of CO₂ and biomass, the longitudinal concentration profile of dissolved oxygen, the oxygen and substrate yield coefficients, the respiratory quotient and the specific interfacial area and volumetric mass transfer coefficient. The influence of the microorganisms on the oxygen transfer rate is discussed especially in the case of glucose repression.     

Acceleration of mass transfer in methane-producing loop reactors

Gas bubbles entrapped in methanogenic granules subjected to hydrostatic pressure oscillations during recirculation in loop reactors will induce intraparticle liquid flows, and thereby enhance mass transfer in excess of diffusion. This breathing particle concept was clearly demonstrated in a well defined inorganic model system. The experimental results could be described satisfactory with a structured mathematical model, while a 30% improvement is predicted for methanogenic loop reactors as compared to constant pressure systems. It is concluded that acceleration of mass transfer in gas-producing systems offers challenging perspectives for both heterogeneous catalysis and biological fermentations.     

Comparison of single- and three-stage tower loop reactors

Summary Hansenula polymorpha was cultured for long periods in 254 cm high single and three-stage countercurrent tower loop reactors 20 cm in diameter using ethanol as a substrate in the absence and presence of antifoam agents (Desmophen 3600 and/or soy oil). In the absence of antifoam agents in the three-stage column, much higher volumetric mass transfer coefficients were attained than in the corresponding single-stage column. The cell productivity in the former, however, was only slightly higher than in the single-stage column due to considerable enrichment of the cells in the foam and nonuniform cell concentration distribution in the three-stage column. In the presence of antifoam agents the three-stage column has a higher cell productivity, OTR, k_L a and a lower specific energy requirement with regard to the absorbed oxygen and/or produced cell mass than the single stage column. The reactor performance is especially high if the bubbling layer height is reduced to 20 cm. Soy oil has considerably less foam eliminating property than Desmophen. Since the soy oil is metabolized by the yeast, large amounts are needed to operate these reactors.     

Characterisation of liquid and gas flow in jet loop reactors

The fluid dynamics in jet loop fermenters has been subject to experimental and theoretical investigation. It is demonstrated that by determination of Euler number, Bodenstein number, and residence time distribution for the gas phase it is possible to perform a reliable characterization of the fermenters. It can be shown that the investigated jet loop fermenters with internal loop closely resemble ideally mixed tanks.     

Heterogeneous one-dimensional model for fixed bed enzyme reactors

A one-dimensional model, taking into account the diffusion of substrate between the liquid phase and the solid support, has been used to describe fixed bed enzyme reactors. The equations were solved numerically, and the values of the different parameters were calculated by a nonlinear regression method. The model was applied to different systems. The results are presented and discussed.     

Hydrodynamic properties of a double-helical model for DNA.

The translational and rotational diffusion coefficients of very short DNA fragments have been calculated using a double-helical bead model in which each nucleotide is represented by one bead. The radius of the helix is regarded as an adjustable parameter. The translational coefficient and the perpendicular rotation coefficient agree very well with experimental values for oligonuclotides with 8, 12, and 20 base pairs, for a single value of the helical radius of about 10 A. We have also calculated a nuclear magnetic resonance relaxation time in which the coefficient for rotation about the main axis is involved. As found previously with cylindrical models, the results deviate from experimental values, indicating that the internal motion of the bases has a remarkable amplitude. An attempt to quantify the extent of internal motions is presented.     

Interpretation of gas residence time distributions in large airlift tower loop reactors

Gas-residence time distribution (RTD) response curves measured in a 23 m high pilot plant airlift tower loop reactor, which consisted of a riser, a special downcomer construction and at the top of the riser a large head. The measurements were evaluated by means of a deterministic dispersion model, which yielded the following particular parameters for the riser, downcomer and the head: Gas-Bo numbers, gas-mean residence times, gas holdups, liquid velocities, gas and liquid circulation times as well as a fraction of the large and small bubbles in a model medium (water) and during cultivation of baker's yeast.List of Symbols A cross section - Bo Bodenstein number - Bo _d (= l _d w _G,d /D _d ) - Bo _h (= l _h w _G,h /D _h ) - Bo _r (= l _r w _G,r /D _r ) - D longitudinal dispersion coefficient - E gas holdup - E(t) RTD-density function - L, l length parameter - q fraction of the gas throughput which is not recirculated (approximately equal to fraction of the large bubbles) - r fraction of the throughput which is recirculated (approximately equal to the fraction of the small bubbles) - t _c circulation time - t _cL liquid circulation time - t _c,L ^* liquid circulation time calculated from the measured w _Ld in the downcomer - V ^h hydrodynamical calculated gas-liquid volume - V _d ^h (=V _d+0.75/2 V _k ) - V _k ^h =(0.25V _k ) - V _r ^h = (V _r+0.75/2 V _k ) - V _L liquid volume - V _G dispersed gas volume - V _G ^* gas throughput at the gas distributor (given in m³/h) under standard conditions, 1 bar and 25°C) - V _G,d ^* gas throughput in downcomer (=V _G ^* ) - V _G,h ^* gas throughput in head (=V _G ^* ) - V _G,r ^* gas throughput in riser (V _G ^* (1+) - w _g gas velocity - w _G,rel relative gas velocity with respect to the liquid velocity w _L - w _G,d gas velocity in the downcomer (=w _G,rel –w _Ld ) - w _G,h gas velocity in the head (=w _G,rel ) (since wLh = o) - w _G,r gas velocity in the riser (=w _G,rel +w _Lr ) - w _L liquid velocity - w _L,d liquid velocity in the downcomer measured with mass flow meter - w _sg ·w _SL superficial gas and liquid velocities - first moment of the response curve - mean residence time Indices d downcomer - G gas phase - h head - L liquid phase - r riser - h hydrodynamic (upper position) Dedicated to the 65th birthday of Proffessor Fritz Wagner.The authors gratefully acknowledge the financial support by the Krupp Industrietechnik, Grevenbroich and the support of Pleser Co, Darmstadt. H. M. Rüffer thanks the Verband der Chemischen Industrie for a Fond der Chemie scholarship, and W. Liwei thanks the government of Lower Saxony for a graduate scholarship.     

Effect of scale-up on average shear rates for aerated non-Newtonian liquids in external loop airlift reactors

Average shear rates have been estimated experimentally in a 700-dm3 external loop airlift reactor. Aqueous pseudoplastic carboxymethylcellulose and xanthan gum solutions were used to simulate non-Newtonian behavior of biological media. Average shear rates of non-Newtonian solutions were found by analogy with Newtonian glycerol solutions using downcomer liquid velocity as the measurable parameter. Due to the complexity of local shear rate measurement, an average shear rate was assumed to exist and is proportional to superficial gas velocity. The data from this work and those in the literature were used in producing a new correlation for estimating average shear rates as a function of superficial gas velocity, geometry, and dispersion height. Wall shear rates were found to be significant. The ratio of wall shear rates to bulk shear rates were varied from 5% to 40%. Furthermore, it has been found that shear rates generated in airlift loop reactors are lower than those generated in bubble columns. Copyright 1999 John Wiley & Sons, Inc.