Date of Award

Spring 2008

Document Type


Degree Name

Master of Science in Pharmaceutical Engineering - (M.S.)


Chemical Engineering

First Advisor

Piero M. Armenante

Second Advisor

S. Mitra

Third Advisor

Robert Benedict Barat


Stirred tank reactors are commonly used in the pharmaceutical industry for synthesis of Active Pharmaceutical Ingredients (API's) and their intermediates. Typically, these vessels are glass-lined and are provided with a single retreat-blade glass-lined impeller and a single baffle. Despite their ubiquitous utilization in the pharmaceutical industry for at least the past 40 years, the mixing characteristics of these systems have not been studied to any great extent, making it difficult to predict mixing performance in any given operation.

In this work, the velocity distribution inside the typical glass-lined vessel/impeller system was experimentally quantified using Laser Doppler Velocimetry (LDV), which is a non-intrusive experimental method used to determine the local velocity distribution (including its fluctuating component) in a fluid placed inside any transparent piece of equipment. Two different reactor configurations were investigated, i.e., a flat-bottom tank and a hemispherical-bottom tank. In each case, two baffling configurations were studied, i.e., a partially baffled tank with a single beaver-tail baffle (the most common baffled configuration used in the pharmaceutical industry), and an unbaffled system. The three velocity components (tangential, axial, and radial) at 13 radial locations on 7 horizontal planes in case of flat-bottom and 5 Horizontal planes in case of hemispherical-bottom tank in the two baffling configurations mentioned above were experimentally determined by LDV.

In the unbaffled flat-bottom reactor case, the tangential component of the velocity appears to dominate over the other velocity components at nearly every location, with tangential velocity typically on the order of 40% to 50% of the impeller tip speed. The radial and axial velocities, especially in the region just below the impeller, were found to be very small, with magnitudes typically smaller than 15% for the axial component and 5% to 10% for the radial component. In general, the presence of a hemispherical bottom did not alter significantly the magnitude of the velocity components except in the lower portion of the tank, where the hemispherical bottom generated a stronger axial and radial recirculation pattern. The velocity distribution in the single-baffle case was found to be only partially different from the unbaffled case, and primarily in the upper portion of the tank, where the baffle is. The velocity distribution in the lower portion of these vessels was not significantly affected by the presence of the baffle.

In conclusion, the dominance of the tangential velocity and the small value of the radial and especially axial velocity in all the system investigated here indicate a poor vertical recirculation of the fluid inside the tank and poor mixing performance of these types of reactors.