Equipment Requirements to Study Agitation for Gas-Liquid Fermentation

By Gregory T. Benz, Lee Enterprise Consulting

Special to The Digest

Small pilot fermenters are used with several different objectives: studying the organism, developing feeding strategies, evaluating potential yields, etc. A major purpose can be to study the mass transfer characteristics of the system, and how they relate to agitation. In the production scale, the required mass transfer rate can be achieved over a wide range of aeration and agitation conditions. However, the combined power draw of the agitator and compressor for a given mass transfer rate goes through a minimum, as illustrated in figure 1.

Figure 1 Fermenter Power Optmization

In order to develop such curves, it is necessary to develop broth-specific k_la correlations. (Reference 1) The purpose of this article is to outline the equipment needs and data to be taken to do this, as most off-the-shelf pilot fermenter systems are not well suited to study mass transfer and agitation.

Vessel Requirements

For mass transfer studies, the gas dispersion mechanism must be the same in the pilot scale as it is in the production scale. Size matters in two ways: the bubbles must be smaller than the vortex trail created by the dispersing impeller, and mass transfer directly to the liquid surface should not be more than about 10% of the total mass transfer. These two factors mean that the minimum size for piloting mass transfer is about 50 liters working volume, with a dispersing impeller diameter of at least 125 mm diameter. 100 liters or more would be even better. Small benchtop fermenters are OK for studying microorganism response to various inputs but cannot be relied upon for scalable mass transfer data.

Although most production fermenters have aspect ratios of 2-3, it is OK for the pilot scale to be a bit shorter, as long as it has at least 2 impellers so it can match full scale impeller types.

Range of Variables

Normally, mass transfer correlations are presented in the form k_la = A(P/V)^B(U_S)^C. Though this form works reasonably well, it is not perfect, and should not be used for extrapolation. What this means in practice is that the values of P/V and U_Sshould cover the same range in the pilot scale as in the production scale.

Table 1 Range of Variables: Pilot vs Production

Table 1 shows typical values of P/V, U_Sand VVM for a variety of vessel sizes at near-optimum conditions, based on a common air/ionic water correlation and an OTR of 200 mmol/l-h. The general trend is for higher P/V and lower U_Sin the pilot scale than in the production scale. However, for proper piloting of mass transfer, we need to develop a test program that brackets the anticipated full scale values in the pilot scale (Reference 2).

Table 2 gives vessel dimensions and other parameters for an array of production volumes and oxygen transfer rates.  What does this mean for modeling in the pilot scale? It would take up too much article space to illustrate the scaled-down parameters for all the production vessel sizes shown, so we will illustrate the concept by scaling down the 200 cubic meter production fermenter.

Table 3 shows the required power and airflow in several pilot fermenter volumes as a function of full scale OTR. The requirement for equal U_Sresults in a higher VVM in the small scale than in the production scale. Most off-the-shelf pilot fermenters are not set up to deliver that much air, so they should be modified to do so. The author has actually run as high as 30 VVM in a 100 l test vessel to study hydrodynamics.

Impeller System Recommendation

Though perhaps not absolutely essential, higher confidence in the derived correlations may be higher if the pilot scale uses the same general types of impellers as those that will be used in the production scale, and those should follow current best practice. Presently, generally accepted best practice in production equipment is to use a lower concave-blade radial turbine, preferably deeply concave, and one or more upper up-pumping high-solidity axial hydrofoils. These are illustrated in figures 2 and 3, respectively.

Most off-the-shelf pilot fermenters come equipped with standard Rushton impellers (Figure 4). Aside from not matching production equipment, there is a very practical reason why such impellers are not the best. When setting up an array or grid of P/V and U_Svalues to test, some combinations of low P/V and high U_Smay lead to a flooded impeller, and such test conditions must be avoided. Concave impellers have at least double the gas handling capacity as a Rushton. The best concave radial impellers have more than 5 times the gas handling capacity of a Rushton. This means that more possible variable combinations can be tested, improving the range of validity of the correlation.

Table 4 shows a typical set of impellers and shaft speeds to match the pilot conditions shown in table 3, as well as the nearest standard motor sizes and shaft speeds.

Buying Recommendations

To allow for flexibility, pilot equipment should be able to deliver more power and airflow than a scale down from the anticipated production scale would require. We recommend the capability to deliver at least double the P/V and U_Sof the full scale estimated design. Table 5 is an example of recommended impeller systems, air flows, shaft speeds and motor sizes for several pilot scales using the 200 mmol/l-h example from the previous tables.

Of course, a specific recommendation can be made for a specific set of client needs. 

Measurement/Data Needs

• Inlet airflow rate, vessel temperature and back pressure
• Ungassed liquid volume and liquid level
• Density of the liquid
• Viscosity, especially if it will vary during the test. (May have to add it as a parameter in the correlation)
• Barometric pressure at time of test
• C_satvalue at a reference gas composition and pressure at operating temperature. This must be expressed in actual concentration units, such as mg/l or mmol/l, not simply %
• Offgas analyzer measuring oxygen, nitrogen and CO₂at a minimum for aerobic fermentations
• Dissolved Oxygen (DO) at top and bottom of fermenter, or in middle if only one is available. The readings must be converted from % to concentration units, so the actual concentration at 100% must be known.
• Agitator shaft speed
• Agitator power draw, if possible. If vessel has standard baffles and the impellers are of a well-studied type, the manufacturer’s power draw calculations may be substituted in lieu of measured power. It is difficult to measure power draw of an agitator; a whole article could easily be devoted to that.

Summary

Using a pilot fermenter to study agitated mass transfer generally has requirements that are not met by off-the-shelf units. This article identifies some of the issues involved and provides guidance about how to purchase appropriate new pilot equipment or make necessary modifications to existing units to allow for such studies. The biggest modification is usually increasing the air supply.

About the Author

Gregory Benz is a member of Lee Enterprises Consulting, the world’s premier bioeconomy consulting group, with more than 100 consultants and experts worldwide who collaborate on interdisciplinary projects, including those requiring the technologies discussed in this article.  The opinions expressed herein are those of the author, and do not necessarily express the views of Lee Enterprises Consulting. Mr. Benz is also President of Benz Technology International, Inc.

References-1 optimizing fermenters 2 piloting fermenters

• “Optimize Power Consumption in Aerobic Fermenters”, G. Benz, Chemical Engineering Progress, May 2003, pp 100-103
• “Piloting Bioreactors for Agitation Scale-Up”, G. Benz, Chemical Engineering Progress, February 2008, pp32-34

Acknowledgment: Figures 2, 3 and 4 are courtesy of Chemineer, a brand of NOV. They are intended to illustrate type and are not to be taken as a product endorsement.

List of symbols

P/V                  Power/volume

U_S                   Superficial gas velocity

k_la                    Overall mass transfer coefficient

DO                  Dissolved oxygen concentration

VVM               Volume of air per Volume of liquid per Minute

A, B and C      Correlation constants

C_sat                  Saturation value of oxygen in liquid