(66i) Palladium Catalyst Removal From Reaction Mixtures by Fixed Bed Adsorption

The removal of palladium catalyst from a Heck coupling reaction mixture was investigated using an external fixed bed adsorption column containing Quadrapure TU resin. The objective was to develop mathematical models that accurately predict Pd breakthrough for the Heck coupling-Quadrapure TU system. Local equilibrium, mass transfer limited, and reaction limited models were investigated. The local equilibrium model predicts breakthrough assuming the resin is fully utilized. However, experimental breakthrough occurs before the bed is completely saturated. As a result, more complex models were developed that account for mass transfer resistance and slow surface reactions that reduce adsorption efficiency. The mass transfer models require a mass transfer coefficient; therefore, a mass transfer coefficient was estimated using the linear mass transfer model to replicate experimental data from Spring 2008 Column Run 7 (CR7-S08). The resulting mass transfer coefficient was 7.82 x 10-3 cm/hr. Mass transfer resistance did not accurately represent the Heck coupling-Quadrapure TU system due to inconsistencies between predicted and experimental breakthrough times. Therefore, two additional models were investigated that assume the surface reaction between the Pd containing species and the adsorbent is the rate limiting step. The first model represents the decay of the resin adsorbent as it becomes saturated throughout the packed bed, in which the reaction rate is time dependent. This model is characterized by two parameters, the Damköhler Number and Activity Number, which may be calculated using two different methods. In the first method, Activity Number is dependent on the Damköhler Number. The parameters are calculated simultaneously in the second method. The resin decay model was fit to CR1-F08 and CR2-F08. For both column runs, Method 2 predicted Damköhler Numbers greater than Method 1. Activity Numbers determined using Method 1 were greater than the values obtained using Method 2 for both runs. Reasons for inconsistencies between methods have not been determined. The second reaction limited model is a reaction-deactivation model, in which the rate is dependent on the concentration of Pd containing species. However, a solution to this model has not yet been developed. Currently, surface reaction models assume a first order reaction between Pd and Quadrapure TU. However, a reaction rate order of 2.5 was estimated using both Integral and Differential methods. While the surface reaction may be the limiting step in the Heck coupling-Quadrapure TU system, a solution to the model is needed before the accuracy of the reaction-deactivation model can be determined.