*2.3. CerAM VPP*

As bioceramic material, the same calcined HAp was used as for the Freeze Foams. HAp (content: 40 vol.-%) was dispersed in a fluid (polyethylene glycol, Sigma-Aldrich, now Merck KGaA, Darmstadt, Germany) with a dispersing agent (BYK-Chemie), and various monomers were used a as binder (a mixture of acrylic resins) and a photoiniator (combination of a camphor derivate and an amine). A planetary centrifugal high-speed vacuum mixer (Thinky ARV310, Thinky Corporation, Tokyo, Japan) was used for the stepwise (three times 5 min at 2000 rpm) preparation of the suspensions. Following CerAM VPP manufacturing, the column geometry as defined in Section 2.2 was chosen.

#### *2.4. Mold Filling, Hybridization and Part Characterization*

For achieving the hybrid parts, cylindrical rubber molds were made, in which the CerAM VPP parts fit in very closely. In the first step, the ceramic suspension filled the molds. Then the column structure was pushed into the cavity. All molds were then transferred to the freeze dryer and foamed at once. A crucial step in creating the structural hybrids is the adjustment of the shrinkage of the two different structures. At the beginning, different suspensions were developed having different contents of water, rheological modifier and binder. With those suspensions, Freeze Foams were manufactured, which shrank between 30–46% (determined by thermo dilatometry DIL 402 C/7/G Netzsch-Gerätebau, Selb, Wunsiedel, Germany). The VPP-manufactured column structure shrank

around 30% altogether. However, it was found that the VPP part shrank around 5.4% at the beginning of the heating process (the debinding), whereas the foam did not. As a result, the foams would have shrunk onto the VPP part, leading to a part failure. To compensate for the overall shrinkage, the VPP columns were pre-sintered prior to being used for in situ Freeze Foaming. This pre-sintering amounted to a shrinkage of around 5% volume. Thus, carefully adjusted for shrinkage, the hybrid parts were sintered at 1250 K (+50 K overheating effect) for 1.5 h. Afterwards, the column-including foams were dismantled and evaluated regarding porosity and microstructure. It must be noted that, after sintering, the initial hydroxyapatite was changed to β-TCP. Among many other references reporting about the transition of HA to TCP during heat treatment, similar Freeze Foams with the same initial HA powder were analyzed via XRD in a previous work [22], showing the HAp to TCP transformation. For microstructure analysis, the resulting Freeze Foams were characterized by SEM (Ultra 55, Co. Carl Zeiss, Oberkochen, Germany). By measuring the height and diameter of three different foam positions of manufactured Freeze Foams and deriving the average, geometrical porosities were calculated according to (*P* = porosity, *ρ*th = theoretical density, *ρ*bulk = bulk density):

$$P = 1 - \left(\rho\_{\text{th}} / \rho\_{\text{bulk}}\right) \tag{2}$$

In addition, porosity was determined via a foam structure analysis tool based on computer tomographic images of the manufactured parts. The allocation of that 3D volumetric pore morphology information (foam cell size) was managed using VGStudio Max v3.0 (Volume Graphics GmbH, Heidelberg, Germany). For X-ray computed tomography, a CT-Compact (Procon X-ray, max. 150 kV power) was used. The universal electromechanical testing machine Instron 8562 (Norwood, MA, USA) was used for the compression strength tests (load cell 10 kN, 1-Taster). To determine the surface roughness, the hybrid foam samples and the Cerasorb M (Kleinostheim, Germany) control were examined using a KEYENCE 3D Laser Scanning Microscope VK-X210 (Keyence, Osaka, Japan). The surface roughness (Sa) was determined using KEYENCE VK analysis software version 3.5.0.0. At least 3 different samples of each were analyzed. Five different positions were determined for each sample. The specimens were measured using 400x magnification.
