(565d) Development and Validation of an in-Line NIR Spectroscopic Method for Continuous Blend Potency Determination in the Feed Frame of a Tablet Press

De Leersnyder, F. - Presenter, Ghent University
Peeters, E., Laboratory of Pharmaceutical Process Analytical Technology
Hasna, D., Laboratory of Pharmaceutical Process Analytical Technology
Vanhoorne, V., Ghent University
Van Snick, B., Ghent University
Vervaet, C., Ghent University
De Beer, T., Ghent University
Hammond, S., Pfizer, Inc.
Hong, K., Pfizer, Inc.

The desire within the pharmaceutical industry to shift towards continuous processing strengthens the need to invest in process analytical technology (PAT), because continuous real-time quality monitoring and control is essential for continuous production. Due to its fast, non-invasive and non‑destructive character, near-infrared (NIR) spectroscopy is one of the most widely used PAT tools for the analysis of pharmaceutical tablets. The fact that NIR can be implemented into a process environment via fiber optics makes it an excellent in-line PAT tool.

Nowadays, NIR spectroscopy for potency analysis is one of the most widely used PAT applications for tablets [1–4]. This is often achieved by on‑line or off-line NIR examination of tablet cores, resulting in a considerable lag time between the actual tablet sampling and the subsequent quality analysis.

To avoid this time lag, in this study, a measurement set­‑up was developed and optimised for in‑line measurement of the circulating blend inside the feed frame of a tablet press (MODUL™P, GEA Process Engineering n.v., Halle, Belgium).

Secondly, an NIR calibration model was developed and validated, enabling in‑line and real‑time blend potency measurements within the optimised interfacing.

Next, the residence time distribution (RTD) inside the feed frame was studied to determine the potential lag time between the measured concentrations in the feed frame of the tablet press and the measured concentrations in the resulting tablets. Finally, influences of the process parameters on the in‑line and off‑line RTD were investigated.


During the set‑up optimisation and quantification model development study, a formulation was used consisting of sodium saccharine (JMC corporation, South Korea), microcrystalline cellulose (Avicel® PH 102, FMC biopolymer, Ireland), dibasic calcium phosphate (Emcompress™, JRS Pharma, Germany) and magnesium stearate (Mallinckrodt, Ireland)(0.5% w/w) as a lubricant for tableting.

For the RTD studies, a blend consisting of lactose (FastFlo 316, Foremost) and magnesium stearate (Mallinckrodt, Ireland)(0.5% w/w) was employed. Spiking was performed using a formulation consisting of lactose and sodium saccharine (in equal parts), and magnesium stearate (0.5% w/w) was added for lubrication.


The in‑line NIR spectra were collected in reflection mode using a diode array spectrometer: SentroPAT FO (Sentronic GmbH, Germany). A SentroProbe DL RS NIR (Sentronic GmbH, Germany) connected to the SentroPAT system, through a fiber-optic cable, was mounted inside the feed frame. Spectra were acquired at a speed of 15 spectra per second. A spectral range of 1150-2200 nm was covered with a spectral resolution of 2 nm.

The sampled tablets were inspected off-line in transmission mode using an FT-NIR analyser (Antaris II, Thermo Fisher Scientific, USA), furnished with a tablet analyser module. In transmission mode, a larger sample volume is measured, since the light penetrating through the sample is measured. This measurement mode is not suitable in‑line because of the longer integration time and the need to be able to measure through the sample, which is impossible in the current design of the feed frame.


An NIR probe was mounted on top of the feed frame in such a position that it measured the circulating blend just before compression. In order to ensure a reproducible and robust implementation of the NIR probe in the feed frame, a micrometer was installed in the tablet press, allowing a precise positioning of the probe. During tableting, NIR spectra were continuously acquired.

Initially, disturbances of the NIR signal, caused by the rotating paddle wheels of the feed frame were observed. The use of mathematical spectral filters to reduce spectral noise can be very complex and can result in deletion of important physical information about the powder blend within the feed frame [5]. Therefore small notches were created inside the paddle wheel fingers of the tablet feed frame, avoiding disturbances in the spectra and hence eliminating the need for mathematical or spectral filters.

The effect of paddle wheel type, probe height and paddle wheel speed on the stability of the NIR signal was investigated by means of a full factorial experimental design, developed in MODDE 10 (Umetrics, Umeå­, Sweden). Optimal process settings, resulting in most stable NIR spectra, were used to make NIR calibration models.


Calibration models were developed by regressing the in‑line NIR spectra against the true sodium saccharine concentration of the tablets (based on known weighed masses). In‑line spectra were acquired during tableting of five blends with different concentrations. Models were calculated for the prediction of two different API target concentrations: 5% and 20% (w/w). Several calibration models for the prediction of this API content were developed, comparing the use of single spectra or averaged spectra, as well as using partial least squares (PLS) regression versus ratio models. The best models were selected and validated by means of accuracy profiles [6–8], by processing physical mixtures with four different API concentrations (validation set).


In order to determine the RTD of the feed frame, initially, the feed frame was filled with the lactose formulation. On top of this lactose blend, a pulse of 50 g of the spiking blend (saccharine/lactose) was introduced and on top of the spiking blend, 2500 g of lactose blend was added. During tableting, in‑line NIR spectra were continuously acquired and tablets were sampled for off‑line potency analysis. This experiment was repeated using different tableting parameters, according to a full factorial experimental design.

The mean RTD was calculated via the approach used by Kumar et al.[9]. The influence of the tableting parameters on the RTDs was evaluated using MODDE 10. In addition, the difference and link between in‑line measured RTDs and off‑line measured RTDs was investigated to gain a better understanding of the mechanisms inside the feed frame.


In this work, an in-line NIR method with high speed acquisition of spectral information to allow real‑time release (RTR) was developed for use in the feed frame of a tablet press.

An in-line NIR spectroscopic model was developed and validated for the monitoring of API concentration of a powder blend in the feed frame during a tableting process.

Finally, RTD of the feed frame was determined in‑line and off‑line, helping to gain a better understanding of the flow/mixing processes going on inside the feed frame.

This study showed that in-line NIR measurements in the feed frame of a high speed rotary tablet press can be applied as a method for powder composition monitoring during the tableting process. When coupled with existing tablet press control systems, this method represents, together with the knowledge of the RTD of the feed frame, a big step towards real time release testing and real time release.


[1] W. Li, A. Bashai-Woldu, J. Ballard, M. Johnson, M. Agresta, H. Rasmussen, et al., Applications of NIR in early stage formulation development. Part I. Semi-quantitative blend uniformity and content uniformity analyses by reflectance NIR without calibration models, Int. J. Pharm. 340 (2007) 97–103. doi:10.1016/j.ijpharm.2007.03.040.

[2] W. Li, L. Bagnol, M. Berman, R. a. Chiarella, M. Gerber, Applications of NIR in early stage formulation development. Part II. Content uniformity evaluation of low dose tablets by principal component analysis, Int. J. Pharm. 380 (2009) 49–54. doi:10.1016/j.ijpharm.2009.06.032.

[3] J. Gottfries, H. Depui, M. Fransson, M. Jongeneelen, M. Josefson, F.W. Langkilde, et al., Vibrational spectrometry for the assessment of active substance in metoprolol tablets: A comparison between transmission and diffuse reflectance near-infrared spectrometry, J. Pharm. Biomed. Anal. 14 (1996) 1495–1503. doi:10.1016/0731-7085(96)01800-6.

[4] E. Peeters, A.F. Tavares da Silva, M. Toiviainen, J. Van Renterghem, J. Vercruysse, M. Juuti, et al., Assessment and prediction of tablet properties using transmission and backscattering raman spectroscopy and transmission NIR spectroscopy, Asian J. Pharm. Sci. 11 (2016) 547–558. doi:10.1016/j.ajps.2016.04.004.

[5] A.J. O’Neil, R.D. Jee, A.C. Moffat, The application of multiple linear regression to the measurement of the median particle size of drugs and pharmaceutical excipients by near-infrared spectroscopy., Analyst. 123 (1998) 2297–2302. doi:10.1039/a806001k.

[6] P. Hubert, J.J. Nguyen-Huu, B. Boulanger, E. Chapuzet, P. Chiap, N. Cohen, et al., Harmonization of strategies for the validation of quantitative analytical procedures: A SFSTP proposal - Part I, J. Pharm. Biomed. Anal. 36 (2004) 579–586. doi:10.1016/j.jpba.2004.07.027.

[7] P. Hubert, J.J. Nguyen-Huu, B. Boulanger, E. Chapuzet, P. Chiap, N. Cohen, et al., Harmonization of strategies for the validation of quantitative analytical procedures A SFSTP proposal – Part II, J. Pharm. Biomed. Anal. 45 (2007) 70–81. doi:10.1016/j.jpba.2004.07.027.

[8] P. Hubert, J.J. Nguyen-Huu, B. Boulanger, E. Chapuzet, P. Chiap, N. Cohen, et al., Harmonization of strategies for the validation of quantitative analytical procedures: A SFSTP proposal–Part III, J. Pharm. Biomed. Anal. 45 (2007) 82–96. doi:10.1016/j.jpba.2007.06.032.

[9] A. Kumar, J. Vercruysse, M. Toiviainen, P.E. Panouillot, M. Juuti, V. Vanhoorne, et al., Mixing and transport during pharmaceutical twin-screw wet granulation: Experimental analysis via chemical imaging, Eur. J. Pharm. Biopharm. 87 (2014) 279–289. doi:10.1016/j.ejpb.2014.04.004.