(83p) Visualisation of Packed Bed Behaviour Using 3D Confocal Microscopy and Microfluidics

A microfluidic ion exchange chromatography column was constructed using a conventional adsorbent from the biopharmaceutical sector within a glass chip. The design allows direct visualisation of events in the bed using 3D confocal microscopy and fluorescently labelled proteins. The channels within a glass chip were fabricated using photolithography and isotropic etching. The design includes a 1 cm long microfluidic column where compressible, porous agarose beads keystoned and packed. This type of microfluidic system allowed for the visualisation and study of; bead arrangement (packing), fluid flow, and adsorption within the packed bed.

fInitial studies with the work the 3 dimensional images produced of the packed bed were used to analyse the degree of packing. These pictures showed that the height (150 μm) of the column was such that 2 bead layers were formed, as the depth of the column was double the diameter of an average bead. The voidage of the system, was calculated from the confocal microscopy pictures, to be 0.55, demonstrating that there are more gaps present in the system than in a standard column. This value is greater than is seen in a standard column (approximately 0.4). which has a voidage of approximately 0.4.

Further studies used the system to visualise flow and adsorption in the bed. This was done by generating breakthrough curves. Fluorescently labelled lysozyme was pumped into the channel and the breakthrough was recorded using a fluorescent imaging system. The system allowed the visualisation of flow through the chromatography bed, displaying unconventional flow profiles. A high degree of channelling occurred through the packed system implied that the profile deviated from plug flow conditions. Images of the breakthrough front exiting the packed bed will be shown. Useful breakthrough curves were achieved using 4 linear velocities from 60 ? 270 cm h^-1. Capacities were calculated from the breakthrough curves and compared to previously published data as well as experimental data performed on a 30 ml column at laboratory scale. Maximum capacities in microfluidic system were recorded in the range of 100 ? 140 mg ml^-1 matrix which was found to be close to values published previously in the literature for laboratory scale columns. This represented a reduction in scale from 500 ? 20000 fold.

The microfluidic ion exchange column was also used to separate 3 fluorescently labelled proteins (based on the separation of egg white) using the same fluorescent imaging system as before. At the laboratory scale, lysozyme passes straight through the column, while ovalbumin and conalbumin bind to the chromatography matrix and are eluted using a salt gradient. The same linear velocities as before were used. Offline MALDI Mass Spectrometry analysis demonstrated that the separation occurs at the microscale in the same manner.

Further work will explore the use of these microfluidic columns to predict the performance of larger scale columns to help facilitate more rapid bioprocess development.