Design and Optimization of the Single-Stage Continuous Mixed Suspension-Mixed Product Removal Crystallization of 2-Chloro-N-(4-methylphenyl)propenamide.

A continuously operated single-stage mixed suspension-mixed product removal (MSMPR) crystallizer was developed for the continuous cooling crystallization of 2-chloro-N-(4-methylphenyl)propanamide (CNMP) in toluene from 25 to 0 °C. The conversion of the previous batch to a continuous process was key to developing a methodology linking the synthesis and purification unit operations of CNMP and gave further insight in the development of continuous process trains for active pharmaceutical ingredient materials. By monitoring how parameters such as cooling and agitation rates influence particle size and the yield, two batch start-up strategies were compared. The second part of the study focused on developing and optimizing the continuous cooling crystallization of CNMP in the MSMPR crystallizer in relation to the yield by determining the effects of varying the residence time and the agitation rates. During the MSMPR operation, the plot of the focused beam reflectance measurement total counts versus time oscillates and reaches an unusual state of control. Despite the oscillations, the dissolved concentration was constant. The yield and production rate from the system were constant after two residence times, as supported by FTIR data. The overall productivity was higher at shorter residence times (τ), and a productivity of 69.51 g/h for τ = 20 min was achieved for the isolation of CNMP.

Binary Solvent Swap Processing in a Bubble Column in Batch and Continuous Modes.

A lab-scale bubble column was investigated as an alternative means to achieve a low-temperature binary solvent swap of solutions containing pharmaceutical materials at atmospheric pressure, for batch and continuous configurations. The rate of solvent evaporation was predicted by first-principles vapor-liquid equilibrium (VLE) thermodynamic modeling and compared to experimentally achieved results. For batch configurations, evaporation rates of up to 5 g/min were achieved at gas flow rates up to 2.5 L/min (0.21 m/s superficial velocity) and temperatures up to 50 °C. This achieved 99 mol % purity of the desired solvent within three "put and take" evaporations from a 50:50 starting mixture. The evaporation rate profiles for the duration of the experiments were calculated, and the changing concentration profile was predicted within satisfactory error margins of <5%. Continuous process modeling explored a multistage equilibrium configuration and could predict the approach to attaining steady-state operation for various operating conditions. All rates of evaporation and resulting changes in solution concentration were measured, and direct comparison of model predictions fell within instrumentation error margins, as previously. This underlined the capability of the model to provide accurate representations of predicted evaporation rates and binary solution concentration changes during operation.

Dissolution Kinetics of a BCS Class II Active Pharmaceutical Ingredient: Diffusion-Based Model Validation and Prediction.

In this work, a diffusion-theory-based model has been devised to simulate dissolution kinetics of a poorly water-soluble drug, ibuprofen. The model was developed from the Noyes-Whitney equation in which the dissolution rate term is a function of the remaining particulate surface area and the concentration gradient across the boundary layer. Other dissolution parameters include initial particle size, diffusion coefficient, material density, and diffusion boundary layer thickness. It is useful for predicting nonsink circumstances under which pure API polydisperse powders are suspended in a well-mixing tank. The model was used to compare the accuracy of simulations using spherical (single dimensional characteristic length) and cylindrical particle (multidimensional characteristic lengths) geometries, with and without size-dependent diffusion layer thickness. Experimental data was fitted to the model to obtain the diffusion layer thickness as well as used for model validation and prediction. The CSDs of postdissolution were also predicted with this model, demonstrating good agreement between theory and experiment.