Scale-up run

A procedure for preparation of sodium hyponitrite involving reduction of nitrogen oxide by the ketyl, followed by extraction into water [1], had been operated routinely on the small scale. A 41-fold scaled up run exploded and ignited after 200 ml of water had been added as part of the work-up. This was attributed to the presence of an unusually large proportion of coagulated unreacted sodium in the ketyl. Small-scale operation with precautions is urged [2],... 

TIP Consistent sample preparation can eliminate questions about the validity of assay results and may eliminate the need for duplicate assays. This can save valuable time during scale-up runs. 

TIP Coordinate IPC and final product assays with analysts prior to scale-up run. 

The focus of the scale-up run determines how the operations are approached during the run. If the focus is to corroborate process details worked out in lab, then data should be collected, and the process may be optimized during the run. 

Good teamwork is essential for efficient operations on scale. The team may be composed of chemists, engineers, operators, analysts, and others. It is important to appreciate the experience of other team members, especially the operators, who are most familiar with the scale-up equipment. Roles of team members must be clarified for instance, direct, hands-on involvement with operations may vary depending on whether the operations must be conducted by union members. In order to make efficient use of equipment and people, goals, critical timing, and personnel involvement must be organized before the scale-up run. 

TIP Optimizing a process during a scale-up run is always tempting and occasionally may be necessary. Considerable resources are at stake with a scale-up run, however. When in doubt, perform the operations you know will work. Examine any proposed processing changes later in the laboratory. 

Process considerations to prepare for and execute a scale-up run can be broken down into seven areas safety, documentation, personnel, equipment ability, analytical support, chemicals, and operations. The following checklists will help prepare for a scale-up run and provide guidelines for implementing the run in the pilot plant or the kilo laboratory. Section IV.B reviews the salient points of the checklists, and IV.C provides special reminders about kilo lab procedures. 

When processing problems occur on scale, they are due to the actions required to carry out the processing, to the chemical nature of the reaction components used, or to a combination of both. When difficulties arise, those responsible for the operations usually suspect the process inputs, and those who developed the process in the laboratory tend to suspect the physical manipulations of the processing. A more objective analysis often reveals that if a successful use-test was run on all components of the reaction prior to the scale-up run, difficulties on scale are probably due to the processing. Thus the processing must be optimized to achieve the desired results. The role of chemical inputs in processing difficulties on scale can be determined by laboratory use-tests.The rationale is shown in Figure 15.1. 

To identify the cause of processing difficulties by use-testing, the most direct approach is to use batches of chemicals that have given suitable results in laboratory runs and sequentially substitute a single input from each of the components used in the scale-up run. If suitable results are obtained with this run, the substituted component is acceptable if processing is not satisfactory, and no significant process deviations occurred, the substituted component was responsible for the undesired result. 

Using this method, the preparation of 1 could be accomplished starting from a protected (R)-4-fluorophenylalanine 2 (Fig. 3). Upon deprotection and diazotiza-tion, the desired product could be obtained in a modest yield of 50% for two steps via the isolated intermediate 3. However, the starting amino acid 2 is fairly expensive and the optical purity of the product 1 varied from a low 78% to a high 97% in scaled-up runs, indicating that the process is not robust. Both problems render this method impractical for production at a large scale. 

When scaling up run times on a preparative column from small-scale analytic testing, a simple procedure is followed ... 

---