Digitalization of Polymer Production Processes with PAT

PAT Approach in Digitalization of Polymer Production Processes Polymers are high molecular weight molecules composed of repeating units bonded by covalent chemical bonds called monomers. Polymerization reactionsgenerally consist initially of monomers and the medium has low viscosity. The polymerization process begins with catalyst addition and/or heat, and viscosity increases over time. There is a direct relationship between viscosity and the molecular weight of the polymer. In other words, viscosity is an indirect parameter for tracking molecular weight. In this context, viscosity is one of the critical quality control parameters for production monitoring in polymerizations involving polymer production or processes where polymers are used at any stage. In polymerization processes, the determination of start and end points is generally carried out through "off-line" monitoring of viscosity and acid number. In the common application, this monitoring process is performed through titration on manually collected samples and viscosity measurement using Ford / DIN Cup or digital benchtop instruments. Particularly to determine the end point of the reaction and stop the polymerization at the correct point, sampling frequency increases considerably. Production monitoring becomes cumbersome, costly and prone to operator errors. An error at this point can result in irreversible consequences such as viscosity increasing to the point of gelation or even solidification, or obtaining a product with low viscosity below acceptance criteria. Beyond this, serious differences in product quality between batches and product losses due to non-conformity cause economic losses. For this reason, in fact, continuous viscosity measurement using process viscometers—a technology that is approximately 50 years old on average—at tanks or at points such as circulation / feed lines is becoming increasingly common. Additionally, with relatively newer technologies such as process spectrometers, it becomes possible to monitor acid number or color parameter in-line. Thus, with minimum sample requirement and minimum loss, the investment return is fast. The scenarios mentioned above addressing the relationship between polymerization and viscosity, which are frequently encountered, have actually caused a paradigm shift in many industries for many different critical control parameters and different products in a similar manner and with similar negative consequences. One of these is the "quality by design" approach instead of providing quality control on the finished product. "Quality by design" is a new paradigm that concerns all production including chemical manufacturing processes. In summary, this approach, which can be defined as "quality control is not performed on the final product, the product is manufactured with quality," aims to establish quality control everywhere in the process and already at the design stage. To achieve this goal, important parameters must be continuously monitored at the process stage and control scenarios must be implemented when necessary. Process analytical technology (PAT) is exactly the technology containing tools that enable this purpose. PAT bridges between traditional laboratory analyses and process automation systems. Additionally, it enables continuous and rapid process monitoring while production continues, and most importantly, rapid control interventions. PAT technology components are called PAT tools.  

2.1 Process Analytical Technology (PAT)

According to the FDA, PAT is a method of analyzing the critical quality parameters of raw materials, process inputs and outputs, and the process itself in a time-dependent manner in-line, on-line, or non-contact rather than through off-line or at-line analyses to ensure final product quality. Actually, it was not the FDA that first brought up work related to process analytics. Process analytics has actually been used in the petroleum and petrochemical industries since the 1950s. However, with the emerging needs of industries such as pharmaceuticals, food and chemicals and developments in analytical technologies, the emergence of process analytical chemistry has caused the subject to undergo a sort of reincarnation.2 Conceptually, initial work began in the early 1980s in the process analytical chemistry (PAC) concept at the University of Washington under the leadership of Bruce Kowalski. Under Kowalski's leadership, a consortium of industry representatives called the Center of Process Analytical Chemistry (CPAC) was established in 1984. This center initially focused on collaborative work with facilities that were already users of real-time analysis, and conducted studies on the analysis of existing data and "evaluation of data using the chemometrics approach." One of these facilities is Dow Chemical Company. Example applications carried out at Dow include analysis of raw materials with NIR and determination of the amount of caustic and salt at variable temperatures during the process using the chemometrics approach.3 Analyzers used in the chemical industry are generally based on spectroscopy, with examples being near infrared spectroscopy (NIR), nuclear magnetic resonance (NMR), Raman spectroscopy, and Fourier transform infrared spectroscopy (FTIR). Systems were used that not only perform efficiency and product quality optimization but also process modeling and statistical approaches. One of the commonly used statistical approaches has been the "partial least squares method (PLS)."4 Kowalski and colleagues defined the types of process analysis such as off-line, at-line, on-line, and non-invasive in their work, contributing to the terminology of concepts that are confused even today.5 Analysis types in processes can be defined as follows (Figure 1): [caption id="attachment_143774" align="aligncenter"] Figure 1. Comparison of in-line and on-line process analysis types[/caption] Off-line Analysis: This is the type in which a sample generally collected manually from the process is taken to the laboratory and analyzed. At-line Analysis: This is the analysis performed on the production site immediately near the production line in the off-line manner. On-line Analysis: This is the name given to the type of analysis in which the sample is automatically collected from the process in a bypass line with a valve or by a sample collection system and fed to the analyzer. In-line Analysis: This is the type of analysis performed within the process without the sample leaving the process. Generally, no sampling is performed; measurement is carried out through a probe or sensor immersed in the process or through a flow cell. PAT process is carried out using PAT tools by employing on-line and in-line analysis types. The purposes of using PAT tools include: shortening production cycle time, preventing product destruction due to rejection on quality grounds, providing real-time data, increasing automation, providing material and energy savings, facilitating continuous production, and so on. In fact, PAT differs somewhat from the concept previously called Process Analytical Chemistry (PAC) in past terminology. PAT is actually not just an analysis but a concept beyond analysis. Simply transferring an analysis performed off-line in the laboratory to at-line, on-line, or in-line is not necessarily PAT.7 For example, it may appear at first glance that replacing a parameter analyzed off-line in the final product in the laboratory with HPLC with only on-line NIR spectroscopy analysis is a PAT action. However, this would be incorrect. Because in this case, the process would still not be analyzed, only final product analysis would continue. According to PAT philosophy and its background; when data obtained from an on-line analyzer is examined, results that would allow conclusions regarding product quality should be obtained. This thus provides the opportunity to intervene in the process at an early stage to ensure the targeted final product quality. For example, in a chemical blending process, a new in-line NIR analyzer will begin to be used, and this analyzer will enable mixing to continue until the mixture becomes homogeneous instead of mixing for a predetermined period, and this process will be monitored in real-time by NIR. With this method, instead of a process where currently only the quality of the finished product is determined, it will be an innovation where corrective action can be undertaken when necessary. On the other hand, the purpose of this analyzer is to control and monitor the process only in such a way that it results in the correct quality. Therefore, this example meets the situation "quality cannot be tested on products; it should be built-in or by design."

Process Viscometers

Polymers whose production and use are widely carried out in industry include: adhesives such as phenolic resins, polyurethane types, paper chemicals such as wet-strength chemicals, starch-based polymers such as sizing used in the textile industry and adhesives used in the corrugated board industry, alkyd resins, paints and paint raw materials, PE, PP, and others. In these processes, viscosity is one of the most important critical quality control parameters, and in-line measurement is becoming widespread. Process viscometer types are classified in a manner similar to laboratory equipment, such as rotational type, vibrating type, capillary type, and others. However, since they will operate continuously under harsh conditions most of the time in the process, the initial condition is the ability to work under harsh and variable process conditions such as high pressure and high temperature, resistance to wear and corrosion, and minimal influence from bubbles and solid particles are important factors in the selection of the model and principle of the viscometer to be used. When it comes to viscosity, another characteristic that should be considered is the rheology of the fluid to be worked with. Because polymers are generally non-Newtonian and commonly pseudoplastic in character. This has two important implications: First, the viscosity values measured in the process will not be the same order as those measured in the laboratory. Second, with incorrect viscometer use, the shear effects in the process—for example, when stirrer speed / pump flow rate changes, viscosity will change. [caption id="attachment_143778" align="aligncenter"] Figure 2. Example of a process viscometer and installation types[/caption] For this reason, when working with a non-Newtonian fluid, it is recommended to use viscometers referred to as high-shear type that generate their own shear rate at high levels. For example, a vibrating-type high-shear viscometer can provide shear above 200 s-1 vibration frequency. This minimizes the effect of process effects on the measured viscosity. Also, as is known, when the shear rate applied to pseudoplastic fluids is increased significantly, the fluid exhibits behavior mimicking a Newtonian fluid, and if the shear rate is increased further, viscosity does not increase. For this reason, high-shear-rate viscometers provide more stable measurements in these fluids and most importantly, higher reproducible measurements. Process viscometers are used by being mounted at suitable points on the tank body or pipe lines using various mounting flanges. A cable connects the sensor to the electronic converter components. These converters generally serve as both a viscosity indicator and a transmitter that enables signal transmission. An example of the installation of a process viscometer is shown representatively in the figure.  

Process Spectrometers

Spectroscopy is the expression of the relationship between electromagnetic waves and the analyte as a function of wavelength. It can be fundamentally classified as atomic and molecular spectroscopy. However, inline process spectroscopy is generally concerned primarily with molecular spectroscopy that defines the relationship of electromagnetic waves with molecular bonds. When matter interacts with light, three different effects occur: absorption, emission, or scattering. These interactions can be detected at different wavelengths using different spectroscopic methods. Among these methods, near infrared (NIR), ultraviolet-visible light (UV-Vis) spectrometer are examples of the most commonly used in-line process spectrometers in chemical processes and especially in polymer production processes. Each spectroscopic technique is associated with the analysis of different molecules. For example, while UV-VIS spectroscopy uses ultraviolet and visible light in the 200-780 nm wavelength range, NIR spectroscopy is a method operating with NIR light and in the 800-2500 nm wavelength range. Process spectrometers fundamentally operate based on Beer-Lambert's law. This law expresses the relationship between the concentration of a light-absorbing molecule in a liquid, the optical path between light and detector, and the absorption of light. Thanks to this relationship, the concentration of the analyte can be determined by measuring the light absorption. In-line process spectrometers are generally produced and used as a set containing lamp and detector modules in the form of a sensor body that is the same diameter as a pipe or in the form of a probe with a rod geometry. [caption id="attachment_143781" align="aligncenter"] Equation 1. Beer-Lambert Law[/caption] [caption id="attachment_143782" align="aligncenter"] Figure 3. In-line Spectrometer Example[/caption] Since the optical path between the light source lamp module and the detector module is fixed and produced to suit the measurement range, according to Beer-Lambert's law, for the same analyte, the light absorption is in a linear relationship directly with the concentration of the analyte in the liquid. The relevant concentration can be monitored in an average range of 0-5 as concentration unit (CU) or absorption unit (AU). This raw data in the visible region can be converted to color unit systems such as Lovibond and Saybolt, or mathematically converted to turbidity units such as ftu and ppm in the NIR region. Process spectrometers can be used together with process viscometers for monitoring acid number and viscosity in polymerization reactions. For example, let us consider a polymerization in which acid number decreases while viscosity increases. With the process viscometer mounted on the reactor, viscosity increase can be monitored and acid number can be monitored with the process spectrometer. Because, while little change is expected in the C-H region peaks at 1200 and 1700 nm, a strong peak is expected to be observed in the -OH region at 1410 nm. This indicates high hydroxyl number and low acid number. By monitoring the peak around 1410 nm, the change in acid number can be tracked, and chemometrically this value can also be expressed numerically for acid number. In polymer production processes and facilities, PAT use is not limited to the polymerization stage but also provides significant benefits at different stages of the process. For example, in polyurethane and related raw material production processes, leakage or escape of analytes with toxic properties such as DNT or TDA to condensate or heat exchanger systems can also be detected with PAT tools. These aromatic derivative analytes containing toluene can be measured in-line by UV absorption. Beyond this, in separation / filtration operations, at unit outlets and inlets, concentration/color or turbidity can be measured using NIR and UV / VIS spectroscopy to perform separation control. In summary, since the polymer industry serves many industries from food to packaging, it grows in parallel with the increasing consumption rate worldwide. For this reason, the need to use limited raw materials efficiently arises. This can only be achieved through increased efficiency. In this regard, digitalization configured to reduce human errors and losses will increase the quality and efficiency of processes and consequently products. Investments in these areas and modernization to be done regarding PAT will be investments that provide rapid returns.     References 1. FDA. Guidance for Industry PAT: A Framework for Innovative Pharmaceutical Development, Manufacuring, and Quality Assurance. FDA official document http:// www.fda.gov/cvm/guidance/published.html (2004) doi:http://www.fda.gov/CDER/guidance/6419fnl.pdf. 2. McLennan, F. & Kowalski, B. R. Process Analytical Chemistry. (Springer Netherlands, 1995). 3. Pell, R. J., Beth, M., Beebe, K. R. & Koch, M. V. Process analytical chemistry and chemometrics , Bruce Kowalski ' s legacy at The Dow Chemical Company. 321–331 (2014) doi:10.1002/cem.2535. 4. Rathore, A. S., Bhambure, R. & Ghare, V. Process analytical technology (PAT) for biopharmaceutical products. Anal. Bioanal. Chem. 398, 137–154 (2010). 5. Callis, J. B., Illman, D. L. & Kowalski, B. R. Process Analytical. 59, (1987). 6. Mandenius, C. F. & Gustavsson, R. Mini-review: Soft sensors as means for PAT in the manufacture of bio-therapeutics. J. Chem. Technol. Biotechnol. 90, 215–227 (2015). 7. Workman, J., Koch, M. & Veitkamp, D. J. Process analytical chemistry. Anal. Chem. 75, 2859–2876 (2003). 8. Hinz, D. C. Process analytical technologies in the pharmaceutical industry: The FDA's PAT initiative. Anal. Bioanal. Chem. 384, 1036–1042 (2006). 9. Read, E. K. et al. Process analytical technology (PAT) for biopharmaceutical products: Part I. Concepts and applications. Biotechnology and Bioengineering vol. 105 276–284 (2010). 10. MIVI, the versatile direct insertion viscometer and density-meter - Sofraser. https://www.sofraser.com/products/mivi-viscometer/. 11. Proses Viskozimetresi - Proses Analitik Teknolojiler. http://prosesviskozimetresi.com/. 12. Claßen, J., Aupert, F., Reardon, K. F., Solle, D. & Scheper, T. Spectroscopic sensors for in-line bioprocess monitoring in research and pharmaceutical industrial application. Anal. Bioanal. Chem. 409, 651–666 (2017). 13. Bakeev, K. A. Process Analytical Technology: Spectroscopic Tools and Implementation Strategies for the Chemical and Pharmaceutical Industries: Second Edition. Process Analytical Technology: Spectroscopic Tools and Implementation Strategies for the Chemical and Pharmaceutical Industries: Second Edition (John Wiley & Sons, Ltd, 2010). doi:10.1002/9780470689592. 14. Chemical Process Control Solutions. https://www.optek.com/en/chemical/applications.asp. 15. Bruder, S., Reifenrath, M., Thomik, T., Boles, E. & Herzog, K. Parallelised online biomass monitoring in shake flasks enables efficient strain and carbon source dependent growth characterisation of Saccharomyces cerevisiae. Microb. Cell Fact. 15, 127 (2016).   Dr. Hulki Özel Sales and Application Manager Pikolab Engineering