Micro process analytical technologies combine process analytics with micro reaction technology. In-line and on-line reaction monitoring require suitable analytical instrumentation to successfully meet the challenges of the reduced dimensions of the micro process.
Process analytical technologies (PAT) have become part of the industrial manufacturing process. It is only for the development or validation of the analytical method in form of a feasibility study that the dimensions of such large-scale processes are reduced to laboratory scales.
PAT is based upon spectroscopic and chromatographic as well as integral methods . The conditions of the production process often demand for greater robustness, stability and performance of the analytical instruments, because of the close proximity to the manufacturing line. Process monitoring and control require prompt or real-time data recording, processing and feeding the data back to the process control unit. These constraints necessitate in-line, on-line or at least at-line analytical methods .
Micro processes are conducted in very small-scale reactors and mixing devices in a Lego-like manner, cf. fig. 1.
Compared to their large-scale counterparts, the micro devices provide a highly efficient heat transfer. Typical yields may range from milligrams to a few grams depending if the reaction is conducted in batch or flow mode [3-5].
Micro processes with respect to scale, volumetric flow and yield demand for micro analytics if implemented in-line or on-line. Instruments installed at-line only require a sample cell of suitable size and corresponding sensitivity. For in-line and on-line monitoring of micro processes with spectroscopic methods, the probes or sample cells are either installed within the reaction vessel or via a by-pass analogously to large-scale facilities. Alternatively, the reaction proceeds within the sample cell of a spectrometer. Miniaturized analytical devices are preferred for the first variant, whereas standard laboratory instruments may be used for the second. So-called bench-top instruments are hence of special interest to micro process analytical technology.
Instruments of reduced dimensions may have the size of a microwave oven [6-9].
The analytical reaction monitoring of a Knoevenagel condensation will be described in the following sections. The reaction proceeds within a micro reactor assembly and is monitored with in-line near-infrared (NIR) and on-line NMR spectroscopy, cf. fig. 1.
Micro Reaction Assembly, Micro Process Analytics Set-up and Experimental Procedures
Micro reaction components by Ehrfeld Mikroreaktionstechnik BTS are assembled according to fig. 1. The educts, 30 mL of 2-methylpropanal and 29 mL of malonic acid diethylester, are mixed together in a vessel equipped with a stirrer and a transflectance NIR immersion probe Falcata by Hellma Analytics. As catalysts 1 mL of piperidine and 3 mL of acetic acid were added to the solution. The vessel was connected to a pump, Smartline 100 from Knauer, such that the solution flowed into the micro reaction system with a rate of 10 mL/ min. The residence time of a solution fraction in the reactor amounted to 3 min by a temperature of 60 °C. After the reaction time, the solution, then containing product, was re-circulated into the storage vessel.
The micro reaction assembly was controlled through the software LabVision 2.10 from HiTec Zang. The overall process was followed by in-line NIR spectroscopy. For the spectral recording, an Antaris II NIR-spectrometer from Thermo Scientific was used. During ten hours, samples were measured with increasing time intervals. At each time point, 32 spectra ranging from 4,000-10,000 cm-1 were accumulated. The spectra were processed using the Result Software Suite from Thermo Scientific. On-line reaction monitoring was achieved by using a picoSpin 45 1H-NMR spectrometer from Thermo Scientific operating at 45 MHz proton Larmor frequency and T=42 °C. The sample cell possessed a total volume of 200 µL and an active volume of 20 nL. It was found suitable for flow injection. Due to its effective sensitivity and its electronic lock system, the NMR spectrometer could be used for observing the neat liquids without addition of deuterated solvents. The reaction mixture was thus transferred into the instrument via a by-pass of 35 cm total length. Setting the valve, the flow was stopped such that 16 spectra could be accumulated per sample. Spectral processing was achieved via MestReNova 9.0.1 from Mestrelab Research on a laptop computer. The following sample was introduced, transferring the previous one back to vessel 1, cf. bottom scheme in fig. 1.
As a reference, the reaction was conducted in a three neck round bottom flask equipped with magnetic stirrer, heating bath, reflux condenser and Dean-Stark apparatus. The educts, 55 mL 2-methylpropanal, 76 mL malonic acid diethylester were dissolved in chloroform and heated to reflux in the presence of 5 mL acetic acid and 3 mL piperidine. During four hours, samples were taken with increasing time intervals and cooled until at-line 1H-NMR spectroscopic investigation. The corresponding NIR spectra were recorded in-line.
Results and Discussion
Through in-line NIR and on-line NMR spectroscopy, the Knoevenagel condensation of 2-methylpropanal and malonic acid diethylester could be monitored in the micro reaction assembly. The changing composition could be clearly recognized in the NMR spectra by the aldehyde resonance at 9.5 ppm, the methylene resonance of the malonic acid diethylester at 3.3 ppm and the double bond resonance of the product at 6.6 ppm. Increase and decrease of reaction components could be identified from the bands at 5200 cm-1 and 4150 cm-1 within the NIR spectra, cf. fig. 2. The process NIR probe and the low-field NMR spectrometer proved hence suitable for the monitoring of reactions in micro reactors with respect to sensitivity, resolution, robustness and comparable dimensions.
A detailed analysis of the spectroscopic data showed that the reaction within the three neck round bottom flask proceeded faster approximately by a factor of 10 and reached a higher product-educt ratio due to the higher reaction temperature by about 28 °C and the continuous elimination of the water occurring along with the condensation reaction. Fig. 3 presents the corresponding concentrations as functions of time (c-t diagrams) as extracted from the 1H-NMR spectra. Increase and decrease were plotted as percent of the starting concentrations, the latters normalized to 100 and 0. The c-t diagram of the micro reaction displayed a modulation of the product and educt curves which did not occur in the curves of the reaction in the three neck round bottom flask. The modulations could be interpreted in terms of the products remaining in the reaction mixture and being circulated. Yet, periodical changes in the instrument environment could also cause such modulations. A model neglecting the modulations could be computed for the c-t curves of the educts and products. As to the reaction order, no simple scheme could be devised likely due to the piperidine catalysis. The model should however be sufficient for envisaged process control as long as it proves robust and reproducible.
Analytical bench-top instruments could be combined successfully with micro reaction technology. A Knoevenagel condensation was monitored through an on-line low-field NMR and an in-line NIR spectrometer, the latter equipped with a transflectance immersion probe. The specificity of the NMR signals allowed for a simple, univariate and thus direct analysis of the product and educt concentrations within the reaction mixture. On the basis of the kinetic data, a model shall be derived that may allow for a real-time feedback-control of the micro process.
Prof. Dr. Martin Jäger
Niederrhein University of Applied Sciences
Department of Chemistry and ILOC