Pharmaceutical Technology - May 2018

Pharmaceutical Technology eBook - Biologics and Sterile Drug Manufacturing

Issue link: https://www.e-digitaleditions.com/i/988595

Contents of this Issue

Navigation

Page 20 of 61

Pharmaceutical Technology BIOLOGICS AND STERILE DRUG MANUFACTURING 2018 21 FIGURES ARE COURTESY OF THE AUTHORS. The system-dependent coefficients are deter- mined by fitting Equation 2 to experimental results for a given system type (e.g., bioreactor, fermenter, aerator, submerged air turbine). All three coeffi- cients are positive, and α and β range between 0 and 1. A final useful term for bioreactors is the ves- sel volumes per minute (VVM), which is defined as the gas flow rate divided by the liquid volume. Study setup and procedures To test the performance of different impeller config- urations in a bioreactor, an experimental setup was prepared with the parameters listed in Table I. The ex- perimental setup was comparable to a 100-L bioreac- tor equipped with a dish bottom that complies with American Society of Mechanical Engineer codes, four standard baffles, and a five-inch ring sparge. The impeller configurations were comprised of single, dual, and triple SPX Flow Lightnin A320 (down-pumping) and A340 (up-pumping) wide- blade hydrofoils. Photographs of the A320 and A340 are shown in Figure 1. Seven test cases, listed in Table II, were designed to establish a base case (Case 1), investigate the effect of agitator power input and gas f low rate (Cases 2 and 3), determine the impact of removing the agitator and using the sparged gas as a bubble column (Case 4), and compare different impeller configurations (Cases 5–7). All test cases, except for Cases 2 and 4, were operated at a power input of 0.4 HP/kgal. All test cases, except for Case 3, were operated at a VVM of 0.1 1/min. The experiments used water with sparged air, and test cases were compared qualitatively and quantitatively. Blend tests were conducted using the iodine-thiosulfate method (1), and blend time was measured by an observer with a stopwatch. Three trials per case were performed. The set of three blend times was averaged, and a 95% confi- dence interval was calculated. Mass transfer tests were conducted according to the Excess Sulfite Batch method (2) and k L a was measured. Three trials per case were performed. The set of three k L a measurements was standardized to 20 °C, aver- aged, and a 95% confidence interval was calculated. In addition to the experiments, transient 3D computational f luid dynamics (CFD) simulations were generated using MStar Simulation's DMT Software. The simulations assumed a single-phase system composed of water and that the flow pattern was not affected by the gas. Based on experimental observation, the impellers adequately dispersed the gas with minimal f low pattern disruption, so the assumption was considered to be acceptable. Experimental blend time and mass transfer results The visual blend test results were compiled and are displayed in Figure 2. The dual A320 down-pumping base case exhib- ited the largest variation amongst blend-time mea- Table I: Experimental setup Parameter Value Tank diameter (T) [in] 17.5 Liquid level (Z) [in] 21.0 Z/T 1.2 Liquid volume (V) [gal] 20.7 Figure 1: SPX Flow Lightnin A320 impeller (left) and A340 impeller (right).

Articles in this issue

Archives of this issue

view archives of Pharmaceutical Technology - May 2018 - Pharmaceutical Technology eBook - Biologics and Sterile Drug Manufacturing