Recent Advances in Obtaining Process Information from Near Infrared Spectra

• University of Puerto Rico
• The State University of New Jersey

Introduction

The effectiveness of near infrared (NIR) spectroscopy for monitoring and control of pharmaceutical manufacturing has been demonstrated in recent years.¹ Chemical composition and physical properties can be evaluated simultaneously by NIR spectroscopy. Due to its non-invasive nature and versatile instrumentation, NIRS is considered a Process Analytical Technology (PAT) tool allowing real time analysis that contributes to product and process understanding. Many studies have determined the chemical composition of pharmaceutical products, which showed their suitability for the purpose. This report highlights recent advancements in obtaining NIR spectral information on the physical properties of pharmaceutical products during manufacturing. Such information can be useful for monitoring and improving a pharmaceutical process.

The changes in the physical properties of materials can affect NIR spectra varying the spectral slope, intensity, position of absorbance bands, and the spectral baseline. This spectral information has been used in different studies for the construction of multivariate calibration models to determine several manufacturing parameters related to product quality. NIR spectroscopy provides information on powder density, compaction force, and crushing strength among others, contributing to an increased level of process understanding. The more recent advancements involve the study of shear stress on the NIR spectrum, the determination of powder density, and crushing strength of tablets.

Discussion

Shear stress is parallel to the surface of a material. Shear stress occurs in pharmaceutical blending as a layer of particles is forced to flow over another layer of the powder blend, and its result is a velocity profile in the different layers of powder which affects its flow properties. Manufacturing of pharmaceutical tablets implies unit operations where different forces are applied that cause deformation that affects bulk density, hardness, tablet mass and dissolution. The effect of variation in shear has been studied by using a modified Couette shear cell, where the powder is confined between two walls consisting in two concentrical cylinders.² Shear is applied when the internal cylinder rotates and the built-in array of pins from each wall interacts with the powder. The Couette cell provides an environment where uniform shear is obtained. The variation in bulk density according to shear was observed in lubricated powders. Hardness in tablets was also affected by decreasing up to 50% related to increasing in shear level.2 Another study evaluated the influence of shear strain on tablets properties compacted at different compression levels. The materials exposed to different shear strain show different tendencies to bond. Also an increase on the recovered portion of elastic energy during in-die relaxation was noted when this force also increases. Moreover, the effect of shear was lower on tablet porosity at maximum compaction and no significant effect was observed during tablet relaxation.³

Spectral variation related to shear stress in pharmaceutical blends has been recently reported by near infrared and scattering measurements. Calibration models were constructed after varying the shear level to obtain different bulk density values in blends containing 15% w/w acetaminophen, 84% silicified microcrystalline cellulose, and lubricated with 1% (w/w) magnesium stearate.⁴ The bulk density changed from 0.500 – 0.620 g/cm³ as the level of shear stress applied was varied. Reference values were obtained by weighing the powder after shear for a fixed volume, and partial least squares calibration models were constructed. Near infrared spectra showed direct relation between the baseline and the bulk density. The higher the bulk density value, the higher the baseline and slope exhibiting a more compact arrangement of the particles due to the lubrication induced by shear and reduced static as shown in Figure 1. For higher bulk density, less radiation is remitted due to favored transmittance by the reduced void space through shearing.³ This is consistent with previous studies in tablets with increasing compaction forces, where the higher compaction force, the lower the diffuse reflectance and higher the transmittance.^5,6 A control strategy by using NIR monitoring as feed forward in combination with feedback measurements was established to control tablet mass.⁷

 Figure 1. Differences in NIR spectra after subjecting to different levels of shear stress with the Couette cell.

Shear stress also affects the NIR spectra of tablets. The spectral differences were then related to the dissolution profile of acetaminophen tablets. A novel PLS-2 NIR method was developed for the simultaneous quantification of shear level and dissolution profile. A comparison analysis for testing dissolution was performed between the present NIR method and the reference USP apparatus 2 showing its suitability for the purpose.⁸ This non-destructive testing method based on the NIR spectrum reduced the analysis time from 120 min to just 2 min. The study indicated a possible approach for real time release testing where a NIR spectrum could be used to identify the formulation, determine the drug concentration, evaluate with the compaction pressure used was adequate, and then predict the dissolution profile. This study was followed by another study based on a broad experimental design where the effect of varying the blender speed (150 rpm, 200, and 250 rpm), feed frame speed (20 rpm, 25, and 30 rpm), and also the compaction force (8 kN, 16 kN, and 24 kN). The calibration model was calculated by principal component regression (PCR) means, in which correlations between the spectra obtained for tablets from the experimental design and dissolution profession profile were established. The results show the successful determination of dissolution profile of tablets obtained by continuous direct compression.⁹

The effect of shear stress on tablets was further studied with angleresolved elastic scattering measurements. The experiment used a laser focused on the sample and the scattered light was collected by an elliptical mirror and transferred to a charge-coupled device for detection. The shear stress de-agglomerates the particles, and reduces the average size of the particles. These studies indicated that compaction realigns the particles in the formulation and reduces the distance between the particles. The scattering decreased by increasing the shear level. The angle-resolved elastic scattering showed that shear stress reduced the particle size and provided better mixing which reduced scattering while increasing the isotropy of the scattering. These measurements confirm the impact of shear in tablets and provide a better explanation for the changes observed in the NIR spectra.¹⁰

The recent studies also include contributions to improve roller compaction processes. During roller compaction baseline shifts and changes in the slope of the NIR spectra are observed due to the increasing solid fraction (SF).¹¹ The SF was calculated as the ribbon bulk density divided by material bulk density. The NIR spectral changes are attributed to a reduction of the NIR radiation that reaches the detector. This study developed calibration models for SF in a placebo formulation using partial least squares (PLS) and spectral slopes algorithms.¹¹ The performance of the models were evaluated in compacts and ribbons. The effect of un-modeled chemical variability was also assessed by predicting validation samples containing drug concentration -varied from 3–17% w/w- which was not present in the calibration set. The calibration model developed with the spectral slope show a better performance in terms of robustness, and better handling of the un-modeled chemical data.¹¹ Another study evaluates the influence of wavelength range of NIR instruments and algorithms on the performance of NIR models; These aimed to predict crushing strength on tablets. The obtained results confirm that the predictive ability of the models is influenced to a large extent to several factors involved in spectral characteristics and extraction of the information.¹²

In this report several studies which extract information from NIR spectra are summarized, showing the capability of NIR calibration models to assess pharmaceutical quality parameters. However, the successful extraction of such data requires taking into consideration the construction of the multivariate models, and the selection of appropriate chemometric tools. The good predictive ability of an NIR model depends on several experimental and instrumental factors. Obtaining a representative calibration set that contains all chemical and physical variability of the samples to be predict is mandatory.¹³ The chemical composition of the pharmaceutical mixture and its NIR fingerprint is also necessary to select the best instrumental set up.^11,12 The quality of the spectral data and the selection of a proper algorithm that better relate the characteristic bands with the property of interest influence to a large extend the performance of the calculated models.^11,12 NIR spectra can provide valuable information on both the chemical composition and physical properties of pharmaceutical products which is useful to improve manufacturing.

References

1. Singh, R., Sahay, A., Karry, K. M., Muzzio, F., Ierapetritou, M., Ramachandran, R. Implementation of an advanced hybrid MPC-PID control system using PAT tools into a direct compaction continuous pharmaceutical tablet manufacturing pilot plant. International Journal of Pharmaceutics 2014, 473(1–2), 38–54.
2. Mehrotra, A., Llusa, M., Faqih, A., Levin, M., Muzzio, F. J.. Influence of shear intensity and total shear on properties of blends and tablets of lactose and cellulose lubricated with magnesium stearate. International Journal of Pharmaceutics 2007, 336(2), 284–291.
3. Pawar, P., Joo, H., Callegari, G., Drazer, G., Cuitino, A. M., Muzzio, F. J. The effect of mechanical strain on properties of lubricated tablets compacted at different pressures. Powder Technology 2016, 301, 657–664.
4. Román-Ospino, A. D., Singh, R., Ierapetritou, M., Ramachandran, R., Méndez, R., Ortega- Zuñiga, C., Romañach, R. J. Near infrared spectroscopic calibration models for real time monitoring of powder density. International Journal of Pharmaceutics 2016, 512(1), 61–74.
5. M. Otsuka, H. Tanabe, K. Osaki, K. Otsuka, Y. Ozaki. Chemoinformetrical evaluation of dissolution property of indomethacin tablets by near-infrared spectroscopy. Journal of Pharmaceutical Sciences 2007, pp. 788–801.
6. Blanco, M., Alcalá, M. (2006). Content uniformity and tablet hardness testing of intact pharmaceutical tablets by near infrared spectroscopy: A contribution to process analytical technologies. Analytica Chimica Acta, 557(1–2), 353–359.
7. R., Singh, Román-Ospino, A. D., Romañach, R. J., Ierapetritou, M., Ramachandran. Real time monitoring of powder blend bulk density for coupled feed-forward/feed-back control of a continuous direct compaction tablet manufacturing process. International Journal of Pharmaceutics 2015, 495(1), 612–625.
8. Hernandez, E., Pawar, P., Keyvan, G., Wang, Y., Velez, N., Callegari, G., Romañach, R. J. Prediction of dissolution profiles by non-destructive near infrared spectroscopy in tablets subjected to different levels of strain. Journal of Pharmaceutical and Biomedical Analysis 2016, 117, 568–576.
9. Pawar, P., Wang, Y., Keyvan, G., Callegari, G., Cuitino, A., Muzzio, F. Enabling real time release testing by NIR prediction of dissolution of tablets made by continuous direct compression (CDC). International Journal of Pharmaceutics 2016, 512(1), 96–107.
10. Hernández, E., Pawar, P., Rodriguez, S., Lysenko, S., Muzzio, F. J., Romañach, R. J. Effect of Shear Applied during a Pharmaceutical Process on Near Infrared Spectra. Applied Spectroscopy 2016, 70(3), 455–466.
11. Talwar, S., Nunes, C., Stevens, T., Nesarikar, V., Timmins, P., Anderson, C. A., Drennen Iii, J. K. (n.d.). Understanding the Impact of Chemical Variability and Calibration Algorithms on Prediction of Solid Fraction of Roller Compacted Ribbons Using Near-Infrared (NIR) Spectroscopy 2016. https://doi.org/10.1177/0003702816671960
12. Talwar, S., Igne, B., Anderson, C., Drennen III, J. Evaluation of the effect of near infrared spectrometer wavelength range and calibration algorithms on prediction of crushing strength of pharmaceutical tablets. Journal of Near Infrared Spectroscopy 2016, 24(5), 413.
13. Càrdenas, V., Blanco, M., Alcalà, M. Strategies for Selecting the Calibration Set in Pharmaceutical Near Infrared Spectroscopy Analysis. A Comparative Study. Journal of Pharmaceutical Innovation 2014, 9(4), 272–281.

Author Biographies

Vanessa Cárdenas holds a PhD title in chemistry from the Universitat Autónoma de Barcelona (UAB). Her research has been focused on the development of NIR calibration models for pharmaceutical quality control and enhancement of modelling strategies. As a postdoctoral researcher at the University of Puerto Rico- Mayaguez, she is experienced in process analytical technologies and chemometrics for improvement of pharmaceutical manufacturing.

Andres Román Ospino, PhD is currently a post-doctoral research associate in the Chemical and Biochemical Engineering at Rutgers. Andres obtained his Ph.D. in Applied Chemistry at UPR- Mayagüez in 2016.

Rodolfo J. Romañach, PhD is Professor of Chemistry and Site Leader for the Engineering Research Center for Structured Organic Particulate Systems at Mayagüez. His research involves near infrared and Raman spectroscopy and multivariate methods for continuous improvement in manufacturing.