School of Pharmacy, University of Maryland, 20 N. Pine Street, Baltimore, MD 21201, USA.
Int J Pharm. 2013 Aug 16;452(1-2):63-72. doi: 10.1016/j.ijpharm.2013.04.039. Epub 2013 Apr 22.
Fluid bed is an important unit operation in pharmaceutical industry for granulation and drying. To improve our understanding of fluid bed granulation, in-line near infrared spectroscopy (NIRS) and novel environmental temperature and RH data logger called a PyroButton(®) were used in conjunction with partial least square (PLS) and principal component analysis (PCA) to develop multivariate statistical process control charts (MSPC). These control charts were constructed using real-time moisture, temperature and humidity data obtained from batch experiments. To demonstrate their application, statistical control charts such as Scores, Distance to model (DModX), and Hotelling's T(2) were used to monitor the batch evolution process during the granulation and subsequent drying phase; moisture levels were predicted using a validated PLS model. Two data loggers were placed one near the bottom of the granulator bowl plenum where air enters the granulator and another inside the granulator in contact with the product in the fluid bed helped to monitor the humidity and temperature levels during the granulation and drying phase. The control charts were used for real time fault analysis, and were tested on normal batches and on three batches which deviated from normal processing conditions. This study demonstrated the use of NIRS and the use of humidity and temperature data loggers in conjunction with multivariate batch modeling as an effective tool in process understanding and fault determining method to effective process control in fluid bed granulation.
流化床是制药工业中用于制粒和干燥的重要单元操作。为了提高我们对流化床制粒的理解，在线近红外光谱（NIRS）和新型环境温度和 RH 数据记录仪 PyroButton(®) 与偏最小二乘法（PLS）和主成分分析（PCA）一起使用，开发了多元统计过程控制图（MSPC）。这些控制图是使用从批实验中获得的实时水分、温度和湿度数据构建的。为了展示它们的应用，使用统计控制图（如 Scores、Distance to model (DModX) 和 Hotelling's T(2)）来监测制粒和随后干燥阶段的批处理演变过程；使用经过验证的 PLS 模型预测水分水平。两个数据记录仪一个放置在进入制粒机的空气进入制粒机的碗室通风区的底部附近，另一个放置在与流化床中的产品接触的制粒机内部，以帮助监测制粒和干燥阶段的湿度和温度水平。控制图用于实时故障分析，并在正常批次和三个偏离正常加工条件的批次上进行了测试。本研究证明了 NIRS 的使用以及湿度和温度数据记录仪与多元批处理建模相结合，是流化床制粒中理解过程和确定故障的有效工具，也是有效过程控制的方法。
