Customer

• Company B were interested in developing a continuous reactor system for production of an important quaternary ammonium monomer:

Project Aims

• To determine suitability for different reactor modes (continuous and batch).
• To obtain calorimetry data + other information for use during batch scale-up

Approach

• A combination of reaction calorimetry and process simulation was used
• A chemical model for the system was constructed from the process chemistry network
• Reaction kinetics were determined and added to the model in order to construct detailed process models
• Process models were then used to simulate large scale behaviour in plug-flow (PFR), continuous stirred-tank (CSTR), and batch reactors.

Results

• Very fast initial reaction phase (Qr rose rapidly on addition of chloride, fell rapidly after first addition)
• Slower second stage (Qr does not fall between successive additions indicating chloride accumulation)
• Long (>10h) stir-out period was required to achieve complete conversion
• Stir-out period was found to be affected by pH and agitation
  • Reaction stalled if pH was kept constant at 9.5
  • Reaction stalled if stirring rate dropped below 450 r.p.m.
• Agitation was an important variable due to the phenomenon of two discrete mixing zones, attributed to NaCl precipitation and accumulation.
• Kinetic model constructed from Qr data predicted temperature rise of 25 → 172°C in < 1 second!

Process Modelling Conclusions

• A PFR would require addition of reagents along length of reactor to regulate exotherm and control pH. This was deemed too complex to develop and difficult to operate at production scale for this process
• The long stir-out period would be unsuitable for CSTR operation. A ‘train’ of reactors would be needed to provide the residence time required to achieve high conversion (each with its own temperature and pH control). This was also felt to be overly complex to develop

Outputs

• Process models suggested the process was unsuitable for continuous operation modes
• Using date already obtained, a batch model was constructed to simulate scale-up in 2,000 gallon Pfaudler vessel
• Chloride accumulation predicted in the batch model matched that observed in calorimeter runs, thus validating the results from simulation experiments
• Thermal safety data (heats of reaction) for the process was also provided

For more information on batch-to-continous consulting, click here