Comparing design to imaged volumes
X-ray CT was deemed as the most appropriate 3D imaging technique for
visualizing whole gyroids due to the high resolutions achievable without
requiring sample embedding or destructive sectioning. Figure 1a displays
three printed gyroids atop a sample holder in preparation for sequential
scanning. Each gyroid design uses identical structurally repeating
units, which results in a bed porosity of 50% and provides mechanical
strength in addition to providing uniform flow paths both axially and
radially . Three gyroid designs with specified feature sizes of 500
µm, 300 µm and 200 µm were
designed and 3D printed for comparison. CAD renders available in Figure
2a-c display the visual comparison between designs, where both the
material and void phases are equally scaled in size.
Conventional bioseparation structures at a comparative length scale such
as packed bed chromatography columns exhibit heterogeneity due to
non-uniform particle sizes and wall effects, which have been visualized
and characterized in a previous study . By specifying a regular feature
shape and size that can then be accurately fabricated through 3D
printing, more uniform structural characteristics are theoretically
achievable. By X-ray CT imaging 3D printed materials then direct
comparisons can be made to the original designs, with a 3D
reconstruction of each feature size available in Figure 2d-f.
The imaged gyroids in each case display the feature sizes as specified
from CAD files throughout the 3D geometry. The capability to design and
fabricate intricate structures with confidence enables physical
characteristics to be specifically tailored to the needs of the
bioprocess product to be purified. Figure 2d-f does display some
artefacts that suggest that fabrication presents some fidelity
challenges, particularly as feature size moves closer to the achievable
3D printing resolution. Some over-curing of the DLP resin was observed
that results in a reduced porous channel diameter, especially in the
internal regions of the 3D printed model.
Imaging complex geometries at the microscale can result in reduced
signal to noise ratios closer to the center of each sample. Printing
smaller samples may reduce this concern whilst improving pixel size from
5 µm achieved here, however producing gyroids with smaller diameters was
found to exhibit structural integrity issues during the drying process
required for effective X-ray CT acquisition.
In order to examine 3D printing fidelity in detail, internal tomography
was performed on a 500 µm feature at an improved pixel size of 2 µm. By
focusing on a single unit within the overall gyroid structure then more
characteristics such as surface roughness could be visualized. As can be
seen in the material phase of Figure 3a horizontal striations 50 µm
apart are visible, with these being the interface between printed
layers. Material and voidage phases were binarised in Figure 3b.
The imaged dataset was aligned in the same position as the original
design file. Figure 3c displays a 3D render of the solid-fluid interface
and Figure 3d overlaying the design and imaged material. Reductions to
the designed feature size may become more problematic at the individual
unit scale, comparable to an individual chromatography bead within a
packed bed. However it may be favorable to increase the surface area
available for fluid to interface with printed material whilst
maintaining an ordered flow pattern. Individual features may get closer
in size to conventional chromatography beads at improved printing
resolutions, which will enable direct comparisons between the two
approaches at this length scale.