Page 153 - IJB-7-4
P. 153

Jaksa, et al.
           define a multi-material printhead and generate G-codes   poly-lactic  acid  (PLA)  filament  from  Fillamentum
           for multi-material  FFF-DIW print  jobs. No post-   Manufacturing (Hulín, Czech Republic) was used with the
           processing of the generated G-code files is necessary, and   E3D V6 FFF printhead in case of prints that demonstrate
           an extrusion correction factor can also be set to fine-tune   the multi-material capabilities. PLA was chosen as it is an
           dosing accuracy, when necessary. To start a print job, the   easily accessible and popular FFF material.
           generated G-code files must be uploaded to the Duet 2   2.4. Printing tests
           Wi-Fi board through the Duet Web Interface. This way,
           the printer is likely also compatible with other popular   For accurate dosing, the silicone printhead was calibrated
           slicing software, such as Cura or Simplify3D.       for the  chosen material  using a KERN PES 42002M
                                                               scale (Kern & Sohn GmbH, Balingen, Germany). After
           2.3. Materials                                      this, 15 × 15 × 10 mm silicone blocks were printed with
           The selected silicone material is a high-viscosity single-  various speeds and layer thicknesses to find reasonable
           component  condensation-crosslinking  liquid  silicone   settings for further printing  tests.  The integrity  of the
           rubber called  Elkem  AMSil 20101 (Elkem  Silicones,   printed cubes was qualitatively evaluated by observing
                                                               them and then slicing them with a blade to see if there are
           Oslo, Norway), which was used with the Viscotec DIW   any internal faults (Figure 3).
           printhead. This material is intended for cold extrusion;   Based on the calibration prints, a printing speed of
           therefore, no heating or other means of energy input is   15 mm/s and a layer thickness of 0.3 mm were chosen
           required during printing. Moreover, a 1.75 mm diameter   for further trials. Furthermore, every printed object

                                                               was left on the building platform untouched for 24 h to
                                                               ensure  sufficient  crosslinking  before  any  manipulation
                                                               or inspection. After calibration,  the system’s ability  to
                                                               print silicone objects with closed internal cavities, infill
                                                               structuring and thin walls as well as to combine silicone
                                                               DIW and thermoplastic FFF was assessed by conducting
                                                               six printing tests:
                                                               1.  In the first test, a thin-walled shell was printed based
                                                                  on the same 15 × 15 × 10 mm cuboid that was used
                                                                  for the calibrations. In this case, only two lines of
                                                                  outer contour were used, resulting in approximately
                                                                  0.7  mm shell  wall  thickness (Figure  4). This  is
                                                                  relevant to anatomic models in case of printing
                                                                  vessels or membranes, which feature thin walls.
                                                               2.  The second test involved a silicone block of the same
                                                                  dimensions as in the first test, but with 40% volume
                                                                  fraction  gyroid  infill  structuring  to  simulate  down-
           Figure 1. The modified Railcore II 300 ZL printer, extended with a   tuning (Figure 5).
           Viscotec Vipro-HEAD 3/3 extruder.

           A                        B


















           Figure  2.  The  Viscotec  Vipro-HEAD 3/3  extruder  with  custom
           Luer-compatible endpieces (A), and the original E3D V6 filament   Figure 3. A 15 × 15 × 10 mm test block that was printed after
           extruder on the opposite side of the printhead carriage (B).  calibration and cut in half after printing.

                                       International Journal of Bioprinting (2021)–Volume 7, Issue 4       149
   148   149   150   151   152   153   154   155   156   157   158