Reaction Optimisation in Process Development

One of our clients approached us with a request to optimise the reduction of an halogenated nitroheterocycle. The starting process used Ni Raney as its catalyst to yield 60% of product, along with several main impurities.

We tackled the problem by applying DoE principles in three stages. The first stage involved obtaining more information about the starting material—solubility in different solvents, stability and impurities. The goal of these experiments was mainly to find possible solvents for the reaction. We discovered that the starting material was incompatible with nucleophilic solvents, and solubility in other solvents commonly used in hydrogenation was an issue.

In the second stage we concentrated our efforts on the reaction. We first screened catalysts, since some of the impurities arose from a competitive dehalogenation during the hydrogenation of the nitro group. Fifteen catalysts from three different industrial suppliers were evaluated under standard conditions. We discovered that a commercial platinum-based catalyst allowed a 98.8% conversion in just 6 h (versus 21 h in the original process), with a very low impurity profile.

According to DoE principles, the third stage was optimisation. We designed a two-level factorial with three variables: catalyst load, temperature and pressure. A centre point was included in the design, and the nine experiments were run. The results were analyzed using Design-Ease software.

The most significant factor in the reaction was catalyst loading. Pressure, temperature and a combination of these factors were not significant, but they were included in the analysis to avoid the simplest case of just one significant factor in the model. The program found the model to be consistent.

The combination of pressure and catalyst loading was revealed as a secondary factor, but had much less influence than catalyst loading. The influence of pressure was higher when loading was low, and pressure was not relevant with high values of loading. Temperature was the third factor. With low loading, its effect was higher at high pressure, but had a small effect. Although some of these findings were obvious from the beginning of the experiment, the Design-Ease software allowed us to predict that catalyst loading could be reduced if pressure, temperature and reaction time were increased. If lower loadings were required for scale up, these options would be realistic, since the product did not decompose in the reaction conditions.