(418g) Developing Scale-up Approach for a Fast Reactions in Continuous Flow

Because of potential high API demand for a particular Merck product currently under development, a new chemical synthesis was developed with improved yield. One of the chemical transformations in the synthesis was developed as a continuous flow step, improving reaction conversion by over 15% compared from batch mode. The reaction exhibits very fast kinetics (in the order of ms) requiring very good mixing. Successful scaling up of such fast reactions requires two important pieces of information:

1. Understanding of how reaction performance is affected by mixing.
2. A reliable way of measuring/scaling up mixing across different scales/geometries of static mixers.

In order to gain understanding on the effect of mixing, the continuous reaction was ran at various flow rates, static mixer geometries/sizes and temperatures. The collected data for reaction conversion was analyzed statistically in order to obtain a mathematical model for its dependence on the above process parameters. The model had acceptable predictive capabilities and confirmed an expected trend: for each mixer, the conversion was increased with increasing flow rate (improved mixing) and decreasing temperature (slower kinetics, allowing for better mixing). The drawback of this approach was that all these models were static mixer specific. If a different static mixer was to be used, a new study was needed in order link the flow rates to reaction conversion. A different, more generalized method was needed to measure/characterize mixing, independent of the mixer type.

Many research attempts to characterize and scale-up continuous mixing in static mixers usually involve correlations using scale-up parameters such as Reynolds number (Re), residence time (τ) or turbulent energy rate (ε). To link all these scale up parameters to the mixing intensity, a method known as the Fourth Bourne Reaction is usually employed. Some of these parameters usually work well, however they may fail to capture any differences in the internal geometry of the static mixer. Instead of using the above scale-up parameters, the conversion from the Fourth Bourne Reaction was used directly as a surrogate process parameter. This direct method of measuring mixing is more accurate as it will capture any internal changes to the mixing pattern (such as back mixing), which may not be captured if the other scale-up approaches are used. It can be seen as a snapshot of the mixing pattern inside the static mixer.

In order to show the applicability of this method, the Fourth Bourne Reaction tests were performed on the same mixers used in the continuous reaction, within the same flow rate ranges. For each mixer type/flow rate combination, the Bourne reaction conversion (BRC) number was assigned and used as a surrogate process parameter. Then the BRC and temperature only were used in the same statistical analysis. Both BRC and temperature were found to be statistically significant and the overall model had the same predictive capabilities. The advantage of this method was the ability to substitute the mixer type/flow rate combinations with the single BRC parameter. This allowed the three model plots produced previously for each of the mixers to be reduced to only one: BRC and temperature only. This methodology also makes scale-up much more straight-forward: one need to show only that the BRC is the same across scales in order to ensure identical mixing (provided the BRC test is executed identically).