(103c) Implementation of a Methodology for Selection of an Appropriate Tracer to Measure the Residence Time Distribution (RTD) of Continuous Powder Blending Operations

Residence time distribution (RTD) is used as a tool to measure the degree of back-mixing by determining the time (mean and distribution) a particle resides in a unit operation [1]. The RTD is commonly measured experimentally by tracer pulse injection into the unit operation entrance at time t=0, and then the tracer concentration is measured, using process analytical technology (PAT), in the exit as a function of time [2]. Choosing an appropriate tracer is of great importance. It is well accepted that a tracer should be detectable, non-interactive with the system, have the same physical properties similar to those of the bulk particles, be able to mix with other system components, and not adhere to equipment surfaces [3, 4]. However, for powder systems, accomplishing all these requirements may prove challenging.

It appears that the proper selection of tracers has not received sufficient attention in recent efforts to characterize the RTD of continuous pharmaceutical powder processes. Tracers are often selected ad-hoc or based on the availability of materials with easy means of detection. Although this is an essential requirement for tracer materials (i.e., APIs often have distinct spectral peaks), it is equally important to meet other tracer properties' requirements to avoid misinterpretations of mixing behavior inside the unit. Otherwise, any models based on such results will likely fail to predict materials' evolution through the system and could have significant consequences such as in control decisions for conforming/non-conforming intermediate material and final product. Thus, an accurate measure of RTDs, a large part of which is the proper selection of tracers, is of utmost importance.

We provide a methodology to choose a tracer that can best describe the dynamic behavior of the blend in the unit. First, the material properties of the API need to be determined using a standard suite of measurements, including particle size distribution, bulk and tapped density, and flow properties. Next, the API material properties are compared to all materials in a material property database, containing over 150 commercially available materials. The comparison method is based on determining the weighted Euclidean distance in principal component space. Materials with similar properties to the API are identified and ranked for suitability to use as tracers (e.g., being detectable). Lastly, the similarity of the RTD of the API and the tracer candidate needs to be verified. To this end, under steady-state conditions, the RTD is obtained by injecting (1) a pulse of the API and (2) a pulse of the tracer.

In this study, the problem was simplified using one excipient as the base powder. However, since formulation flow properties are affected by the API concentration and the rest of the ingredients properties, the experiments were repeated for three different excipients varying in their flow behavior applicable to the continuous direct compression line. The outcome of these experiments enables the side-by-side comparison of RTDs obtained using pulses of either API or tracer for powders with a varying range of properties. Moreover, the effect of tracer amount on RTD was explored to identify an appropriate amount that is small so that it does not perturb the system yet enough to be detected.

References:

1. Engisch, W. and F. Muzzio, Using residence time distributions (RTDs) to address the traceability of raw materials in continuous pharmaceutical manufacturing. Journal of pharmaceutical innovation, 2016. 11(1): p. 64-81.
2. Levenspiel, O., Chemical reaction engineering. 1999: John Wiley & Sons.
3. Escotet-Espinoza, M.S., et al., Effect of tracer material properties on the residence time distribution (RTD) of continuous powder blending operations. Part I of II: Experimental evaluation. Powder Technology, 2019. 342: p. 744-763.
4. Danckwerts, P.V., Continuous flow systems: distribution of residence times. Chemical engineering science, 1953. 2(1): p. 1-13.