Monitoring Crystallisation Processes in Pharmaceuticals Manufacturing Using FT-NIR Spectrometry

Crystallization is a significant step in the manufacture of pharmaceutical ingredients. The FDA is now encouraging the use of modern process analytical technologies (PATs) in pharmaceutical production and quality control. Accordingly, the industry is moving away from empirical to science-based standards for manufacturing control. As a results, new technologies are entering the manufacturing sector of the pharmaceutical industry.

Process Analytical Technologies are used for the analysis and control of manufacturing processes to assure an acceptable quality at the completion of the process. This is based on measuring critical quality parameters and performance attributes of raw and in-process materials and processes. The crystallization process is one of these. It requires the mixture and conditions to be tightly controlled in order to precipitate the correct crystalline form. An example of this, was at one of the world’s leading pharmaceutical companies, which had the requirement to tightly control the water content in an IPA/drug mixture in a glass-lined reactor such that the water content was just above 2.5%.

Experimental

Bruker Optics supplied a MATRIX-F fiber optic-based process Fourier transform NIR spectrometer and two 100-m lengths of low OH fused silica fiber optic cable along with a custom designed transmission probe contained in a 2-m hastelloy C22 DIP Pipe. The DIP pipe was needed for the glass-line reactor because no side ports were available for a standard probe. The DIP pipe was designed to enter the reactor through a flange at the top. The design and realization of this DIP pipe was done by a team of engineers using a sophisticated Computer Aided Design software to optimize the product. The initial calibration development was done in a pilot plant and then transferred to the reactor once the MATRIX-F had been installed. Spectra (32 scans,  of roughly 15 s, at a resolution of 8 cm-1) were measured constantly over several batches. The spectra collected throughout the process, with different water contents are shown in Figure 1.

FT-NIR absorption spectra of the pilot plant mixture collected throughout the course of the reaction. Significant differences can be clearly observed. They are correlated to changes in the water content.

Figure 1. FT-NIR absorption spectra of the pilot plant mixture collected throughout the course of the reaction. Significant differences can be clearly observed. They are correlated to changes in the water content.

Time stamped samples were collected and analyzed off-line before the water content was correlated to the appropriate time stamped in-line NIR spectrum. The resulting calibration model for water content is shown in Figure 2 where a prediction accuracy of 0.02% is achieved over a water content range of 1.9-3.7%.

Cross validation results of a PLS based model for prediction of the water content. The data shows a very high correlation coefficient (99.76) and a low error (0.02 absolute).

Figure 2. Cross validation results of a PLS based model for prediction of the water content. The data shows a very high correlation coefficient (99.76) and a low error (0.02 absolute).

Quantitative Analysis

Near-infrared spectra result from combination and overtone bands of C-H, N-H, O-H, etc. vibrations. Since most reaction mixtures contain some organic components with these bonds, they are ideal for near-infrared analysis. The OPUS/QUANT quantitative analysis software package uses partial lease squares (PLS) to develop quantitative models. Typically, the development of a model requires measuring samples that contain a range of concentrations of the components of interest. The unique Quant self-optimisation routine is then applied to develop the calibration model. In this example, NIR spectra of the reaction mixture were constantly collected in a pilot plant and correlated through their time stamp to samples pulled from the reactor and then analyzed off-line.

Measurement Options

Bruker Optics offers a wide range of instrumentation to meet every specific needs. For process applications the MATRIX-F is recommended due to its multiplexing capability, ruggedness and easy serviceability. A wide range of process measurement accessories are available for in-line and on-line measurements of liquids, solids and slurries.

The MATRIX-F can also be used in an at-line application by adding a simple fiber coupled vial holder and using disposable glass vials. This can be an alternative solution for developing calibration models before going on-line. Near-infrared spectra can be collected directly from liquid streams using in-line flow cells or transmission probes with no sample preparation. The use of fiber optics makes it possible to locate the instrument in a distant control room or in an enclosure, in a hazardous location close to the measurement site. High quality spectra can generally be collected in less than a minute and the quantitative analysis of multiple components easily performed. All of these factors make NIR spectroscopy a quick, reliable in-line tool for monitoring processes.

MATRIX-F individually enclosed modules designed to easily fit into standard 19-inch racks and enclosures. The spectrometer for this experiment was housed in an environmental enclosure 100 m from the sample point along with a touch-screen computer. The results are transmitted to the Process Control system via a 4-20 mA interface and the water content is controlled by a closed loop based on the NIR result. Bruker Optics is constantly improving its products and reserves the right to change specifications without notice.

Figure 3. MATRIX-F individually enclosed modules designed to easily fit into standard 19-inch racks and enclosures. The spectrometer for this experiment was housed in an environmental enclosure 100 m from the sample point along with a touch-screen computer. The results are transmitted to the Process Control system via a 4-20 mA interface and the water content is controlled by a closed loop based on the NIR result. Bruker Optics is constantly improving its products and reserves the right to change specifications without notice.

Implementation

In a process environment the MATRIX-F can be used, along with the process software ADIO, to measure and analyze the sample and send the results to a DCS through a variety of I/O options such as 4-20 mA, Modbus, Profibus, Industrial Ethernet, etc.

This information has been sourced, reviewed and adapted from materials provided by Bruker Optics.

For more information on this source, please visit Bruker Optics.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Bruker Optics. (2023, April 10). Monitoring Crystallisation Processes in Pharmaceuticals Manufacturing Using FT-NIR Spectrometry. AZoM. Retrieved on April 18, 2024 from https://www.azom.com/article.aspx?ArticleID=5983.

  • MLA

    Bruker Optics. "Monitoring Crystallisation Processes in Pharmaceuticals Manufacturing Using FT-NIR Spectrometry". AZoM. 18 April 2024. <https://www.azom.com/article.aspx?ArticleID=5983>.

  • Chicago

    Bruker Optics. "Monitoring Crystallisation Processes in Pharmaceuticals Manufacturing Using FT-NIR Spectrometry". AZoM. https://www.azom.com/article.aspx?ArticleID=5983. (accessed April 18, 2024).

  • Harvard

    Bruker Optics. 2023. Monitoring Crystallisation Processes in Pharmaceuticals Manufacturing Using FT-NIR Spectrometry. AZoM, viewed 18 April 2024, https://www.azom.com/article.aspx?ArticleID=5983.

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Your comment type
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.