• A comparative study of the formulation requirements of dosator and dosing disc encapsulators, simulation of plug formation, and creation of rules for an expert system for formulation design

      Heda, Pavan Kumar; Augsburger, Larry L. (1998)
      Objectives. The objectives of this study were: (i){A0}to develop a 'semi-quantitative' definition of fluidity, lubricity and compactibility requirements for the two most commonly used types of automatic capsule filling machines - the dosator and the dosing disc type, (ii) to simulate powder plug formation and explore the low force powder compression physics of the process, (iii) to utilize artificial neural networks (ANNs) as a pattern recognition tool to forecast (a) the encapsulation performance of an automatic capsule filling machine, and (b) the impact of formulation characteristics and filling principle on product performance, and (iv) to develop new or modified decision trees for an expert system for formulation design. Methods. Laboratory research - Model formulations were developed differing either in their degree of fluidity, lubricity, or compactibility. Carr's index, ejection force, and plug strength were used for gauging the fluidity, lubricity and compactibility respectively. The model formulations were encapsulated on Hofliger-Karg (H&K) GKF- 400 and Zanasi LZ-64 encapsulators instrumented for monitoring compression and ejection forces. The formulations were evaluated for percent coefficient of fill-weight variation, ejection force, plug strength, and dissolution of model drugs, hydrochlorothiazide and ascorbic acid. A programmable tablet compaction simulator especially tooled for plug compression was used. Plugs of different heights at different punch speeds were made using a saw-tooth waveform. Non-laboratory research - A group of experts from the industry were consulted to provide a pool of expertise for identifying new and improved rules for an expert system for formulation support. A survey of the practices of formulators and industry policies concerning the development of formulations for hard gelatin capsules was conducted with a panel of experts. A literature search was also conducted to identify information on excipients specific to capsule formulation. Results and conclusions. From the laboratory research on the instrumented automatic capsule filling machines it was observed that there were differences in the formulation requirements of the dosator type machine, such as, the Zanasi LZ-64 Vs the dosing disc type machine, such as, the Hofliger Karg (H&K) GKF-400. The results show that an optimal degree of fluidity is required in a formulation for successful encapsulation on either type of filling machine. A relatively lower level of lubricant may be sufficient for encapsulation on the H&K machine as compared to the Zanasi. A higher degree of cohesiveness is needed for clean processing on the Zanasi machine as compared to the H&K. The study provided a range of formulation characteristics that would be suitable for transferring formulations between these machines. Optimal points for all the input and response variables for successful encapsulation on either type of machine were obtained using ANNs. Using the programmable tablet compaction simulator to simulate plug formation tablet compaction models such as, Heckel analysis, Shaxby-Evans's exponential relationship, Kawakita's pressure-volume relationship, and coefficient of lubrication 'R' were successfully applied to the plug formation process. It was concluded that as compressibility is required of powder formulations on automatic capsule filling machines, these tableting models could be used to characterize the physico-mechanical properties of formulations required for efficient plug formation. This research has provided a better understanding of the interplay of formulation and machine operating variables and formulation design strategies that, when incorporated into an expert system, will facilitate the determination of rational, robust formulations for capsules. The database generated in this study will be incorporated in an expert system in the form of both 'domain specific rules' and cautionary 'flags'.