Real-time monitoring and fault detection of pulsed-spray fluid-bed granulation using near-infrared spectroscopy and multivariate process trajectories
Graphical abstract
Introduction
Fluid bed granulation (FBG) is a typical wet granulation technique that plays important roles in the pharmaceutical industry, fine chemicals industry, and food science. FBG accomplishes mixing, spraying, and drying steps with the same equipment to simplify the granulation process and improve efficiency. Wet granulation is a complex evolutionary process involving multiple simultaneous activities including wetting and nucleation, consolidation and growth, and breakage and attrition (Iveson, Litster, Hapgood, & Ennis, 2001). In FBG, the coalescence and breakage of particles are the main reasons for the change of particle size (Terrazas-Velarde, Peglow, & Tsotsas, 2011a). A fine powder and mist contact and then coalesce into a larger particle via the formation of liquid bridges between particles (Terrazas-Velarde, Peglow, & Tsotsas, 2011b). However, if the moisture content of a particle is excessive, the result could be a poor fluidization level or even bed collapse (Tian, Wei, Zhao, Li, & Qu, 2018).
Pulsed spray has drawn increasing attention as an effective process control tool to achieve a consistent and desirable product quality in FBG (Liu, Wang, Schlindwein, & Li, 2013; Schaafsma, Kossen, Mos, Blauw, & Hoffmann, 1999). The mode of binder addition in pulsed-spray fluid-bed granulation (PSFBG) is interruption of the liquid feed according to a predetermined sequence, allowing the intermittent drying and rewetting of the granules. Compared with the conventional continuous binder spray addition method, the pulsed spray method has the ability to compensate for excessive moisture levels and control particle size (Schaafsma et al., 1999). Pulsed spray improves control of the moisture content of granules to lower the risk of overwetting and bed collapse.
Previous PSFBG studies have focused on the design of experiments approach to investigate the effect of pulsed spray parameters on granulation behavior (Ehlers et al., 2009; Liu et al., 2013; Närvänen, Lipsanen, Antikainen, Räikkönen, & Yliruusi, 2008), population balance modeling (Liu, Yoon, & Li, 2016) and computational fluid dynamics simulations (Liu & Li, 2014) to simulate the granulation process. However, consideration of the pulsed spray parameters at the same time as other critical process parameters makes PSFBG simulation very complex. Experimental work has been carried out in our laboratory to improve understanding of the PSFBG process, including using the definitive screening design method to investigate the effects of pulsed spray parameters along with other operating factors on granule quality (Zhao, Li, Qu, Tian, & Wei, 2019) and employing near-infrared (NIR) spectroscopy combined with multivariate data analysis to monitor the granule moisture content, particle size, and granule yield (Tian et al., 2018).
The main objective of granulation is to produce granules with low batch-to-batch quality variability to meet rigorous specifications of the final product (FDA, 2004; Haas, Ierapetritou, & Singh, 2017). To obtain a consistent final product with predefined characteristics, the granulation process needs to be monitored and controlled. However, very few attempts have been made to develop quality control strategies for the PSFBG process. Therefore, further studies of this type are urgently required.
Controlling the PSFBG process is important to improve batch-to-batch reproducibility. The United States Food and Drug Administration (FDA) established the Process Analytical Technology (PAT) guidelines to monitor process performance and ensure product quality (Haas et al., 2017, Igne et al., 2014). The multivariate statistical process control (MSPC) strategy is a commonly used quality control method that has the ability to prevent process failure and improve product quality by identifying drifts in a process before they become critical. The effective diagnostics of MSPC can identify precisely when and where faults occur to lower warranty claim frequency and the risk of recall (Kona et al., 2013). MSPC enables on-line or in-line quality control to minimize the need for expensive and time-consuming off-line testing. MSPC has been successfully applied to granulation processes. For example, microwave resonance technology combined with multivariate data analysis was used as a novel PAT tool to elucidate the processes involved in FBG (Lourenco, Herdling, Reich, Menezes, & Lochmann, 2011). In other work, in-line spatial filter velocimetry and product temperature measurements were used to implement the batch statistical process control of FBG (Burggraeve et al., 2011) and an MSPC strategy based on in-line NIR spectroscopy was developed and applied to FBG (Kona et al., 2013, Rantanen et al., 2001).
NIR spectroscopy is an attractive PAT tool and has been widely used in many fields (Dixit et al., 2017; Nee et al., 2018Nee, Bryan, Levitskaia, Kuo, & Nilsson, 2018; Singh et al., 2016). NIR spectroscopy is a simple, rapid, low-cost, and nondestructive method that requires minimal sample pretreatment and is environmentally friendly because it does not consume solvents or generate chemical waste. Moreover, NIR spectra can provide both physical and chemical information about samples. Some applications have already been developed to monitor FBG using in-line NIR spectroscopy as a PAT tool (Heigl, Koller, Glasser, Muzzio, & Khinast, 2013; Lee et al., 2011; Peters et al., 2018).
However, an MSPC strategy using in-line NIR spectroscopy for the PSFBG process has not been reported. Moreover, the NIR spectra acquired from different granulation batches are unsynchronized in the time series. Batches of variable length can be attributed to different process times caused by the different levels of binder addition during the spraying stage. The duration of a granulation process is rarely fixed as a criterion; the end point of a batch process is usually based on quality requirements rather than on time (Wan, Marjanovic, & Lennox, 2014). Nevertheless, most multivariate methods applied to batches depend on the basic assumption that all batch trajectories are properly synchronized (Gonzalez-Martinez, Ferrer, & Westerhuis, 2011), because unsynchronized batches destroy the linear structure of datasets and lead to biased results (Wan et al., 2014). Therefore, appropriate data synchronization is a crucial preprocessing step for the successful use of these multivariate methods.
Different alignment algorithms for batch synchronization have been proposed to address datasets with varying batch-to-batch processes (Gonzalez-Martinez et al., 2011; Liu et al., 2017; Ramaker, van Sprang, Westerhuis, & Smilde, 2003). Correlation optimized warping (COW) is a typical alignment algorithm that was originally designed to warp single- and multiple-wavelength chromatographic profiles (Nielsen, Carstensen, & Smedsgaard, 1998). COW has recently received extensive attention to deal with the problem of uneven-length batch monitoring (Kassidas, MacGregor, & Taylor, 1998). COW corrects a sample vector toward a reference one by translating, expanding, and contracting localized segments within the trajectories to achieve a minimum dissimilarity value between the two trajectories without losing important information (Zhao, Mo, Gao, Lu, & Yuan, 2011). The COW method has been used with MSPC as a preprocessing tool to align uneven trajectories based on Raman spectroscopy datasets (Liu et al., 2017). The signal is aligned in a piecewise manner by the COW method, and the optimal alignment is determined based on the correlation between the aligned signal and reference fragments (Tomasi, van den Berg, & Andersson, 2004). The segment strategy correlation is used instead of distance as a criterion to improve the performance of the COW method.
The overall objective of this study is to use synchronized NIR spectra to develop an MSPC strategy for use as a quality control method in PSFBG. NIR spectroscopy is used as a PAT tool to monitor the PSFBG process in real time. The COW method is used to synchronize the time-varying process batch trajectories according to the granulation phase before MSPC development. Three types of control charts—principal component (PC) score charts, Hotelling's T2 control charts, and distance to model X (DModX) control charts—are developed based on the synchronized NIR spectra obtained from the granulation batches. Then, the NIR spectra for new batches are used to verify the feasibility of the developed models. Normal operating conditions (NOC) batches and abnormal operating conditions (AOC) batches are used to estimate the capability of the developed models to diagnose faults. The developed PAT tool can be used to establish better process control and lower test cost.
Section snippets
Materials
The formulation blends consisted of unmodified maize starch (Dongyuan Pharm Co. Ltd., Liaoning, China) and lactose α-monohydrate (Hengxin Chemical Reagent Co. Ltd., Shanghai, China). An aqueous solution of polyvinylpyrrolidone (PVP K-30, ISP Technologies, Inc., USA) was used as the binder for granulation.
Experimental setup
A schematic representation of the PSFBG experimental setup is provided in Fig. 1. Granulation experiments were performed in a bench-scale fluid bed granulator in top-spray mode (FBL10, Xiaolun
Alignment principles identified
FBG batch evolution is a typical batch process composed of three stages from a mechanistic point of view: mixing, spraying, and drying (Fransson & Folestad, 2006). These three process stages were chosen as the principles to align the process data that were most relevant for each type of process dataset.
COW application
B4 with a medium duration was chosen as a reference for the COW synchronization method (Liu et al., 2017). First, the linear interpolation method was used to generate a segment of equal length to
Conclusions
The feasibility the quality control of PSFBG via multivariate process trajectories was analyzed. NIR spectroscopy was chosen as an in-line PAT tool to monitor the PSFBG process. The COW algorithm was used to synchronize the time-varying process batch trajectories in accordance with the granulation phase. The COW algorithm was found to be a practical tool to handle time-varying process data. MPCA was used to develop different types of MSPC models, including PC score charts, Hotelling's T2
Conflict of interest
None declared.
Acknowledgments
This work was supported by the National Science and Technology Major Project (grant number 2018ZX09201011-002). The authors gratefully thank Mr. Haifan Han, Thermo Fisher Scientific Inc., for constructive discussions and technical advice about the NIR equipment. Special thanks to Naomi Twery for editing this manuscript.
References (48)
- et al.
Bed collapse behavior of pulsed fluidized beds of nano-powder
Advanced Powder Technology
(2014) - et al.
Real time in-line monitoring of large scale Bacillus fermentations with near-infrared spectroscopy
Journal of Biotechnology
(2014) - et al.
Batch statistical process control of a fluid bed granulation process using in-line spatial filter velocimetry and product temperature measurements
European Journal of Pharmaceutical Sciences
(2011) - et al.
Granule size control and targeting in pulsed spray fluid bed granulation
International Journal of Pharmaceutics
(2009) - et al.
Process control and end-point determination of a fluid bed granulation by application of near infra-red spectroscopy
International Journal of Pharmaceutics
(1997) - et al.
Real-time alignment of batch process data using COW for on-line process monitoring
Chemometrics & Intelligent Laboratory Systems
(2006) - et al.
Real-time synchronization of batch trajectories for on-line multivariate statistical process control using dynamic time warping
Chemometrics and Intelligent Laboratory Systems
(2011) - et al.
Quantitative on-line vs. off-line NIR analysis of fluidized bed drying with consideration of the spectral background
European Journal of Pharmaceutics and Biopharmaceutics
(2013) - et al.
Modeling strategies for pharmaceutical blend monitoring and end-point determination by near-infrared spectroscopy
International Journal of Pharmaceutics
(2014) - et al.
Nucleation, growth and breakage phenomena in agitated wet granulation processes: A review
Powder Technology
(2001)
Application of in-line near infrared spectroscopy and multivariate batch modeling for process monitoring in fluid bed granulation
International Journal of Pharmaceutics
In line NIR quantification of film thickness on pharmaceutical pellets during a fluid bed coating process
International Journal of Pharmaceutics
A feasibility research on the monitoring of traditional Chinese medicine production process using NIR-based multivariate process trajectories
Sensors and Actuators B-Chemical
Novel description of a design space for fluidised bed granulation
International Journal of Pharmaceutics
Two-compartmental population balance modeling of a pulsed spray fluidized bed granulation based on computational fluid dynamics (CFD) analysis
International Journal of Pharmaceutics
Using the Box–Behnken experimental design to optimise operating parameters in pulsed spray fluidised bed granulation
International Journal of Pharmaceutics
Multivariate statistical process control (MSPC) using Raman spectroscopy for in-line culture cell monitoring considering time-varying batches synchronized with correlation optimized warping (COW)
Analytica Chimica Acta
Combining microwave resonance technology to multivariate data analysis as a novel PAT tool to improve process understanding in fluid bed granulation
European Journal of Pharmaceutics and Biopharmaceutics
Reduction of energy consumption in batch fluidized bed layering granulation processes by temporal separation
Chemical Engineering Research & Design
Controlling granule size by granulation liquid feed pulsing
International Journal of Pharmaceutics
Combinations of NIR, Raman spectroscopy and physicochemical measurements for improved monitoring of solvent extraction processes using hierarchical multivariate analysis models
Analytica Chimica Acta
Aligning of single and multiple wavelength chromatographic profiles for chemometric data analysis using correlation optimised warping
Journal of Chromatography A
Development, validation and transfer of a near infrared method to determine in-line the end point of a fluidised drying process for commercial production batches of an approved oral solid dose pharmaceutical product
Journal of Pharmaceutical and Biomedical Analysis
Real-time process monitoring in a semi-continuous fluid-bed dryer – Microwave resonance technology versus near-infrared spectroscopy
International Journal of Pharmaceutics
Cited by (10)
Control of segregation by non-uniform aeration in a fluidized bed spray granulator
2024, Powder TechnologyMultivariate process analytical technology tools for fluidized bed granulation and drying analysis: A review
2024, Journal of Drug Delivery Science and TechnologyLiquid atomization into gas–solid fluidized beds—A review spanning the micro- to macro-scale
2024, Powder TechnologyParticle circulation and coating in a Wurster fluidized bed under different geometries
2024, Powder TechnologyAdvances in numerical simulation of unit operations for tablet preparation
2023, International Journal of PharmaceuticsInvestigation on the granulation behavior of iron ore fine in a horizontal high-shear granulator
2022, ParticuologyCitation Excerpt :Most of the natural iron ore needs to be agglomerated into sinter before it can be added into the blast furnace for reduction in the blast furnace ironmaking process. As the first step in the iron ore sintering process, granulation is a size enlargement step where fine particles are adhered onto coarser nuclei particles (Fernández-González, Ruiz-Bustinza, Mochón, González-Gasca, & Verdeja, 2016; Realpe & Velázquez, 2008; Zhou & Zhou, 2021) and the granules are formed under inter-particle forces and mechanical forces act by granulating device (Huang, Lv, Bai, Qiu, & Lu, 2014; Zhao, Li, Qu, Tian, & Wei, 2020). The objective of the granulation is to improve material properties such as increasing particle size and strength, stabilizing the chemical composition which plays a vital role in the iron ore sintering process, because the appropriate sized granules help to increase the sintering speed and the quality of the sinter (Gan et al., 2014; Tang, Xue, Yang, Zhang, & Huang, 2018; Wu, Zhu, Bei, Zhang, & Zhai, 2015).