Optimization of Primary Drying by Construction of a Graphical Design Space

Posted by

SP Scientific

To view the full webinar, please visit: http://www.spscientific.com/Graphical... A design space is intended to define the boundaries of the process that produces an acceptable product. In the case of primary drying in a freeze drying operation, it is essentially definition of the time course of the variables “shelf temperature” and “chamber pressure” that allow a process to produce a product that has acceptable properties. Certainly acceptable critical quality attributes are essential, but normally one also wants to consistently avoid obvious collapse in the product, and avoid loss of chamber pressure control arising from sublimation rates in excess of equipment capability. Although “Design of Experiments (DOE)” has been suggested as a tool for design space development, our view is that DOE has no place in development of the primary drying process. Rather, construction of a design space is best accomplished by a combination of experiments and simple heat and mass transfer theory. The design space is normally presented as a simple two-dimensional plot that illustrates the boundaries of the range of acceptable operations. Such a plot is easy to understand and serves to emphasize the importance of “Quality by Design”. However, the conventional 2-D design space has two significant limitations. First, it represents behavior of the “average vial” in the freeze dryer. Clearly, it is not sufficient to guarantee quality for only the average vial. To compensate for this deficiency, a “safety margin” is normally imposed, somewhat empirically, to account for variations. Alternately, one can employ a more rigorous approach based in the statistics of variations. Secondly, the optimum process conditions are time dependent, largely due to the fact that as sublimation proceeds at constant shelf temperature and constant chamber pressure, the sublimation rate decreases and the product temperature rises. This is because of the increasing resistance of the dry layer as drying proceeds (as it’s thickness increases). Thus, to present the true optimum process (i.e., as fast as you can without loss of chamber pressure control and/or product collapse), one needs a series of pressure and shelf temperature settings, or at least several to represent “early”, “mid”, and “end” of primary drying. This is difficult to represent in one plot, but can be presented in several plots (or in an algorithm).