Roller compaction (RC) is an alternative technique to wet granulation, therefore, heat and moisture are not involved during densification of the material to improve powder flow and handling properties. In this study, binary mixtures of Acetaminophen (APAP) and Microcrystalline Cellulose (MCC, Avicel PH-200) were used as model materials to produce ribbons by roller compaction/dry granulation. APAP is an active pharmaceutical ingredient (API) with poor compressibility. This material shows elastic deformation behavior upon compression and produces weak compacts. Avicel PH-200 is a well-known excipient with great compactability. An full factorial design (2^4) was employed to evaluate the effects of roller compaction process variables on the ribbon attributes. Near-infrared (NIR) reflectance spectroscopy was used as a non-destructive method to monitor the key ribbon properties such as tensile strength, Young's modulus, relative density.

The four factors varied in the process were roller speed, raw material feeding rate, roll pressure and API load. In order to evaluate the effects of these process variables and their interactions, multiple response variables were investigated, i.e. key ribbon properties, slopes of NIR monitoring spectra and post-milling particle size distribution. The ribbon tensile strength and Young's Modulus were obtained using the three-point beam bending method. The ribbon relative density was determined using geometric and helium pycnometer true density measurements. The granules' particle size distribution of milled ribbons was obtained using the sieve analysis method.

Principal component analysis (PCA) was applied to probe the response correlation structure and create latent variables. The general linear model and the linear regression were performed to evaluate the significance of main factor effects and their interactions. The magnitude and direction of all effects were then determined and ranked. While using different response variables, the roller speed was always the most influential process variable on key ribbon properties, followed by the raw material feeding rate. From such ranking results, we were able to establish the critical process variables to control the roller compaction process. It was found that the low roll speed and high feeding rate at low roll pressure produced highest ribbon tensile strength within the design space. The significance results using key ribbon properties were consistent with the results using NIR monitoring spectra and the latent variable after PCA. The findings obtained in this study will potentially contribute to the implementation of NIR spectroscopy as a feasible online monitoring technique for the roller compaction process.

Keywords: Roller compaction; Dry granulation; Relative Density; Tensile strength; Young's Modulus; Process Variables; Process Analytical Technology; Near-infrared spectroscopy.