MTCiting: Continuous Flow Synthesis in Pharmaceutical Manufacturing

The adoption of true, online HPLC is impeded by the real-world requirements of implementation. Notably, samples for HPLC analysis must be removed from the reaction mixture, quenched and diluted before initiating analysis. There are numerous physical and engineering issues associated with these steps that cause online application to be problematic. In this work, the authors report and demonstrate that DirectInject-LC technology mitigates these issues. In a first-time demonstration, a photocatalytic reaction that is performed both in batch and continuous flow is analyzed in real-time by online HPLC PAT.

A photoredox C−N coupling reaction between 4-bromobenzotrifluoride and pyrrolidine, catalyzed by Ru(bpy)₃(PF₆)₂ and NiBr₂·3H₂O was selected as a model reaction and performed initially in an EasyMax™ batch reactor with irradiation through the reactor window. Both EasySampler™ and ReactIR™ probes were inserted into the reaction vessel in order to collect both HPLC and FTIR data for tracking starting material and product. DirectInject-LC provided a liquid chromatography data point every 7.5 minutes; the IR data point was accumulated every 30 seconds and strong correlation between the data from both techniques was observed. Next, the EasySampler probe was adapted for flow measurements using a custom designed flow cell arrangement. The reaction was performed using a Vapourtec UV-150 plug flow photoreactor using a 405 nm LED array. The reaction was first performed using different residence times from 1-60 minutes holding the temperature constant at 35 °C. 

It was determined that residence times greater than 30 minutes yielded conversion greater than 95%. The authors commented that due to the online HPLC there was no need to stop and resume the flow, which permitted real-time monitoring for each tested condition and minimized resource-intensive use of both time and materials. Varying reaction temperature was also evaluated and using a residence time of 5 minutes at a temperature of 60 °C provided an accelerated reaction rate while maintaining excellent conversion. 

The authors commented that “The DILC™ system has demonstrated significant effectiveness in flow reactions by facilitating reaction optimization, enhancing sustainability, and minimizing human intervention. Furthermore, the DILC™ has proven to be an excellent online tool for calibrating spectroscopic data. This innovative methodology not only streamlines analytical workflows but also lays the groundwork for future advancements in automated in-process control (IPC) technologies.”

Qin, Y., Ayesa, U., Lévesque, F., & Ji, Y. (2025). Novel Application of Direct Inject Liquid Chromatography (DILC™) as Real-time Process Analytical Technology for Flow Reactions. ChemRxiv. https://doi.org/10.26434/chemrxiv-2025-c3680

In the second step of a new continuous process for rocuronium bromide (Roc-Br), the development and optimization of the N-alkylation of AMPO, continuous reaction and crystallization are presented. The authors comment that the N-alkylation of the acetylated intermediate (AMPO) using Allyl-Br to produce Roc-Br proved challenging due to slow kinetics, high reagent requirements, and extensive downstream processing. Multiple unit procedures were necessary to successfully perform this process in batch, resulting in long cycle times and increased solvent waste. Key to the establishment of a continuous manufacturing process to improve the production required thorough understanding and refining of reaction conditions, reactor design, and crystallization strategy. They state that through kinetic modeling and real-time monitoring using ReactIR and conductivity measurements, they successfully accomplished optimization of the Step 2 N-alkylation leading to reduced reaction time and reagent usage. Through implementation of a five-stage CSTR cascade, greater than 98% rocuronium bromide purity was achieved with steady-state performance and minimal residual starting material. Solubility and solvent screening studies resulted in choosing methyl propyl ketone (MPK) as an effective antisolvent, enabling high-purity rocuronium bromide recovery with yields up to 94%. The process was successfully scaled to a 500 milliliter CSTR, and inline particle analysis confirmed robust crystal growth dynamics.

To establish a baseline for the N-alkylation reaction, ReactIR was used for real-time monitoring along with Reaction Lab™ software. A kinetic model was developed that predicted reaction profiles and identified critical parameters supporting design of experiments (DOE) with the goal of optimizing the AMPO conversion to Roc-Br. An initial goal was to investigate reaction rate based on published information. To accomplish this, ReactIR measurements tracked the decrease of the Allyl-Br peak at 932 cm^–1 and the shift in carbonyl peaks from 1730 cm^–1 in AMPO to 1755 cm^–1 as Roc-Br is produced. This showed that after approximately 6 hours, only ∼50% conversion was achieved, prompting the use of temperature ramps, which proved to have no significant effect on the rate or extent of conversion. To address this and improve efficiency, the reaction concentration was significantly increased by reducing the solvent volume. This resulted in the Roc-Br synthesis to be complete in 6 hours, as compared to 120 hours needed for using the solvent volume reported in the literature. Further experimental work at the higher reaction concentration showed that temperature is key in driving reaction rate, and that 2–3 equiv of Allyl-Br are sufficient to achieve high conversion within a practical time frame.

Because the 2 N-alkylation reaction has relatively long reaction times, it was evident that a plug flow reactor would not be applicable for implementation in a continuous process. For this reason, the five-stage CSTR cascade was implemented in order to accommodate the extended reaction time. The final stage of the cascade was equipped with a ReactIR probe system to monitor the conversion. In the work to optimize crystallization, Roc-Br solubility was examined in various solvents and it was confirmed that Roc-Br is highly soluble in dichloromethane (DCM) but exhibits poor solubility in diethyl ether (Et₂O) and methyl tert-butyl ether (MTBE). Both Et₂O and MTBE have been reported in the literature as effective antisolvents for crystallizing Roc-Br. Based on this insight, further solubility studies were conducted to investigate the potential application of DCM as the primary solvent and Et₂O or MTBE as antisolvents for Roc-Br crystallization. These antisolvents proved unsatisfactory and therefore a more in-depth solvent screening study was conducted to identify more effective alternatives for Roc-Br isolation. Ketone-based antisolvents were selected for further evaluation and methyl propyl ketone (MPK) was chosen, producing crystalline Roc-Br with improved processability, drying characteristics and the desired crystalline form.

Gamekkanda, J. C., Fejedelem, Z., Neshchadin, A., Zhu, G., Aloglu, A. K., Barral, R. C., Hearl, M., Orfanos, D., Ingraham, H. M., Keenan, S. G., Liu, F., Bharate, P., Ramnath, A., Raul, U., Singh, B., Su, Q., Yu, F., Li, Y., Born, S. C., . . . Mascia, S. (2025). Continuous manufacturing of rocuronium bromide. Part 2: Step 2 process development. Organic Process Research & Development. https://doi.org/10.1021/acs.oprd.5c00206