GENERAL RESEARCH Optimum Design of Multiple-Impeller Self-Inducing System Swapnil S. Patil and Jyeshtharaj B. Joshi* Institute of Chemical Technology, University of Mumbai, Matunga, Mumbai 400 019, India An attempt has been made to eliminate the unstable behavior of self-inducing reactors. For this purpose, a multiple-impeller system was developed. Experiments were performed in 1-m-i.d. tank, with impeller parameters varied over a wide range. Recommendations are made for optimum reactor design. 1. Introduction In the category of dead-end gas-liquid systems (reac- tions with pure gas), self-inducing impellers find a prominent place because of their high efficiency in gas induction per unit power consumption. 1,2 These impel- lers, similar to other contactor systems (surface aera- tors, sparged loop reactors, etc.), show a region of instability over a certain range of impeller speeds. Patil and Joshi 3 discussed this subject for a single-impeller system. Recently, it has been observed that the unstable region can be conveniently eliminated using a dual- impeller system, which is the subject of this paper. A schematic representation of a self-inducing impeller is shown in Figure 1. To begin, we briefly describe stable and unstable operations. When the impeller is not rotating, the height of the liquid in the standpipe is equal to the height of the liquid in the reactor (Figure 2A). As the impeller speed is increased, the liquid level in the standpipe drops progressively (Figure 2B and 2C). This happens as a result of a decrease in pressure in the vicinity of the inducing impeller. At a certain critical speed, the liquid level reaches fairly close to the impeller (Figure 2D). At this condition, the gas is entrained as in surface aeration. In Figure 3, the critical impeller speed is indicated by point A. The speed at which the gas induction process begins is called the critical impeller speed (N C ) for the onset of induction. With increasing impeller speed (N > N C ), the impeller becomes increas- ingly exposed to the gas-liquid interface by increasing the size of the vortex (Figure 2E). The larger size of the gas-liquid interface allows the system to entrap more bubbles, and the gas hold-up in the inducing zone increases. If the liquid circulation rate also increases, the rate of bubble carriage increases, resulting in an increase in the rate of gas induction (Figure 3, line AB). However, at a certain impeller speed, the gas induction suddenly drops (point B). The impeller speed at which this drop is observed is called the first transition speed (N t1 ). Line BC in Figure 3 shows the resulting drop in the rate of gas induction (ΔQ Gt1 ). In the case illustrated in Figure 3, it can be seen that the value of ΔQ Gt1 is 19.5% of Q G at point B. As N is increased beyond N t1 , the rate of gas induction (Q G ) increases (line CD), but the increase in Q G is much slower. The gas induction rate in state 1 (line AB) is higher than that in state 2 (line CD). Further, the fluctuations in the rate of gas induction are relatively less in state 1 than in state 2. This is illustrated in Figure 3 by the error bars. The physical significance of the transition (line BC) can be understood from Figure 2. Any increase in the rate of gas induction along line AB is due to the increase in the vortex size as shown in Figure 2D and 2E. However, at the critical speed (N t1 ), the vortex size becomes so large that the impeller is flooded with the gas phase (Figure 2F). This results in a drop in the liquid circula- tion rate and a drop in the rate of gas induction. However, the hydrostatic head of liquid does not allow the impeller to emerge completely into the gas phase. Under these conditions, the liquid level shows very wide fluctuations due to the opposing actions of the impeller motion and the static liquid head. As a result, the rate of induction also shows large fluctuations. One more important observation is the existence of hysteresis behavior between the two states 1 and 2. For instance, a decrease in speed along the line DC (state 2) results in a decrease in Q G , even when N is below N t1 . At point E, the rate of gas induction increases steeply. The speed corresponding to this point is called the second transi- tion speed (N t2 ). When the impeller speed is less than N t2 , again the induction behavior coincides with state 1. Depending on the impeller diameter and location (with respect to the stator), the second transition speed can be either less than or equal to the first transition speed (N t2 e N t1 ). Thus, the state 2 operation (line CD) is unsuitable because (i) the rate of induction is approximately 20% lower than that in state 1 and (ii) the fluctuating behavior in state 2 means greater fluctuations in the power drawn. Therefore, the system becomes subject to mechanical problems. The foregoing discussion attests to the importance of eliminating the unstable behavior of inducing impellers. However, this problem has not been addressed in the published literature. Therefore, it was thought desirable to undertake a systematic work in this direction. A precalibrated turbine-type anemom- eter was used for the measurement of gas induction. The power consumption was calculated by measuring * To whom correspondence should be addressed. E-mail: jbj@udct.ernet.in. Tel.: 91 022 2414 56 16. Fax: 91 022 2414 56 14. 1261 Ind. Eng. Chem. Res. 2003, 42, 1261-1265 10.1021/ie020220g CCC: $25.00 © 2003 American Chemical Society Published on Web 02/12/2003