(646g) Industrial Scale Downstream Process Development for Low Concentration Amino Acid Product Using PAT

This work utilizes PAT to develop a downstream purification process on an industrial scale for a low yielding synthetic amino acid from a dilute chemical reaction product. Cooling crystallization optimized by use of inline IR measurement was used to remove 90% of the highest impurity paving the way for efficient use of continuous chromatography to remove lower concentration impurities. The steps leading up to the final design of the Downstream Process (DSP) process are outlined in this work.

The food grade product, an amino acid (AA) is used globally for animal feed as well as human consumption. Production is targeted to be on several MT/year scale. The process of synthesis of the AA involves several reactions involving organic and inorganic compounds.

The resulting reaction product is dilute, over 50% water and contains a higher molar ratio of impurities in the form of organic and inorganics compounds than the product AA. The salts in the reaction product are metal sulfate and sulfite, both at concentrations higher than the AA. Ion exclusion chromatography pulse test results were promising with AA eluting after all impurities. However, scale up to continuous chromatographic separation using Simulated Moving Bed (SMB) chromatography failed due to pressure build up in the columns with the sulfates precipitating out and high ionic load causing osmotic pressure to build up, shattering the resins.

To allow the use of SMB chromatography, precipitation of sulfates prior to chromatography was imperative. Various methods of sulfate salt crystallization including evaporative, cooling and antisolvent crystallization were studied. Evaporative crystallization lead to formation of anhydrous sulfates while cooling crystallization forms decahydrate sulftates, removing considerable amounts of solvent in the process. In line ReactIR® was used to determine the end point for cooling crystallization when the sulfate concentration in the ML plateaued. Given the translucent nature of decahydrate crystals Inline particle track FBRM® was not capable of accurate crystal count measurement. Nucleation and growth kinetics were estimated by off line microscopy as well as real time solution concentration measurement from ReactIR. Use of PAT allowed for faster optimization of the salt precipitation step with 90% of sulfates removed prior to SMB. Salt precipitation lead to some product loss which required an additional cake wash step to recover AA and recycle back into the process with the feed. Permeability of decahydrates crystals were considerably better than anhydrous crystals allowing for better wash and product loss reduction in salt precipitation step.

A second round of crystallization process is required for purification of the AA in the SMB extract from 95% purity to commercial grade product. Inline particle track FBRM(R) was utilized for kinetic measurements in that process.