Practical Limitations of Tableting Indices

Abstract

The purpose of this study was to utilize tableting indices to distinguish between materials with varying degrees of compactibility by establishing a quantitative relationship between indices and compactibility. Compactibility in this study is restricted to tablet strength and friability alone. Nine mixtures with varying degrees of compactibility were tableted and the tensile strength and friability of the tablets were determined. The tableting indices of these mixtures were determined using an Instron® Universal testing machine. An artificial neural network program was used to establish a quantitative relationship between indices and tablet strength and friability. Six new powders were used to validate the models describing the relationship between indices and tablet strength and friability. These powders were compressed into tablets and their strength and friability were determined. Their indices were also determined. The established models were used to predict tablet strength and friability from index values. The predicted values were compared with the experimentally determined values. There was little correlation between the predicted and experimentally determined values for tablet strength and friability. It was also found that materials or mixtures having almost similar indices had remarkably different compactibilites. It was concluded that models created to predict compactibility using one set of materials may not be able to successfully predict the compactibility of a new material. This calls into question the practicality of indices.

Keywords:

• Artificial neural network
• Compactibility
• Compressibility
• Indentation hardness
• Tableting indices
• Tensile strength