Defense Date

11-16-2018

Graduation Date

Spring 5-10-2019

Availability

Immediate Access

Submission Type

dissertation

Degree Name

PhD

Department

Pharmaceutics

School

School of Pharmacy

Committee Chair

Carl A. Anderson

Committee Member

James K. Drennen III

Committee Member

Peter L.D. Wildfong

Committee Member

Ira S. Buckner

Committee Member

Christian Airiau

Keywords

process control, process analytical technology, PAT, NIR, Raman, fluidized bed coating, first principle control

Abstract

The conventional basic control for pharmaceutical batch processes has several drawbacks. The basic control often uses constant process settings discovered by trial and error. The rigid process operation provides limited process understanding and forgoes the opportunities of process optimization. Product quality attributes are measured by the low efficient off-line tests, therefore these cannot be used to monitor and inform the process to make appropriate adjustments. Frequent reprocessing and batch failures are possible consequences if the process is not under effective control. These issues raise serious concerns of the process capability of a pharmaceutical manufacturing process.

An alternative process control strategy is perceived as a logical way to improve the process capability. To demonstrate the strategy, a hybrid control system is proposed in this work. A challenging aqueous drug layering process, which had a batch failure rate of 30% when operated using basic control, was investigated as a model system to develop and demonstrate the hybrid control system.

The hybrid control consisted of process manipulation, monitoring and optimization. First principle control was developed to manipulate the process. It used a theory of environmental equivalency to regulate a consistent drying rate for the drug layering process. The process manipulation method successfully eliminated the batch failures previously encountered in the basic control approach. Process monitoring was achieved by building an empirical analytical model using in-line Near-Infrared spectroscopy. The model allowed real time quantitative analysis of drug layered content and was able to determine the endpoint of the process. It achieved quality assurance without relying on the end product tests. Process optimization was accomplished by discovering optimum process settings in an operation space. The operation space was constructed using edge of failure analysis on a design space. It provided setpoints with higher confidence to meet the specifications. The integration of the control elements enabled a complete hybrid control system. The results showed the process capability of the drug layering process was significantly improved by using the hybrid control. The effectiveness was substantiated by statistical evidence of the process capability indices.

Language

English

Share

COinS