*2.3. Manufacturing*

Scaffolds were manufactured from Ti-6Al-4V powder in ARCAM A2 EBM machine by ARCAM, EBM (Mölnlycke, Sweden) operating under firmware version 3.2 using standard Themes without modification of the default parameter settings. All samples were manufactured using 50 μm layer thickness and process temperature of 720 ◦C.

In beam-based structures, PBF AM specimens can be manufactured using different beam energy application strategies, influencing, in particular, melt pool dimensions, solidification rate and final material microstructure. Modern PBF AM machines are commonly

operated using parameter sets provided by the manufacturer. An ARCAM "Theme" is a set of settings incorporating beam scanning strategy and process parameters. The standard Themes provided by the EBM machine manufacturer ARCAM EBM are called "Melt", "Net", and "Wafer". The "Melt" Theme is optimized for manufacturing of solid structural elements. The "Net" Theme is designed for the manufacturing of the beam-based porous structures. The "Wafer" Theme is designed for manufacturing of support structures (essential elements of the EBM process), where wafer-thin surfaces are required [28]. These supports are commonly used for stabilizing overhanging elements, and as spacers between the solid elements and the base plate. Such supports are removed after manufacturing. So, the "Wafer" Theme was designed to produce very thin structures to guarantee a small amount of waste material. Structures manufactured using Net and Wafer Themes are designed as zero-thickness geometry, and the cross-section of the elements is defined by the beam energy and its deposition rate.

The "Melt" Theme is more complicated than the "Wafer" one. The first path of the beam (called first contour) is shifted out from the CAD-defined element periphery by a certain value called the first contour offset (CO1). Next, the beam performs a second contour scan with a second contour offset (CO2) moving slightly inward from the CADdefined cross-section periphery. After that, the beam melts the area enclosed by these two contours (hatching) using different types and strategies of the raster motion. Corresponding contours are performed by continuous beam spot motion over the whole periphery length, or in so-called 'MultibeamTM mode'. In the latter case, the beam moves only through the short sector of the contour and 'jumps' away to melt another sector, repeating the operation many times to cover all needed contours. The main purpose for selecting singleor double- contours in continuous or MultibeamTM mode is the optimization of the process to obtain the smallest possible roughness of side surfaces of the components. It is clear that when using any mode for the manufacturing of lightweight and porous structures, careful parameter selection is needed to guarantee that the resulting element cross-sections are as close to the CAD design as possible. When the elements with a small cross-sectional area are EBM-manufactured using Melt Theme, automated file preprocessing can reduce the number of contours (in an automatic fashion) until only the hatch is left, minimizing the increase in the dimensions of the manufactured elements.

The "Wafer" Theme, traditionally, is mainly used for the plane sheet supports with relatively small surface areas; our experience shows that it is also effective when used for quite complicated structures. Slicer software created a line pattern for WT based on zero-thickness gyroid model for each slice, while for MT, it created areas for melting based on 200 μm thickness gyroid model. Depending on the parameter settings, thin structures made with Wafer Theme can have a certain amount of through-holes. However, it is clear that application of the Wafer Theme for EBM of sheet-based lightweight and porous structures can bring significant improvement of the mechanical properties of the resulting components.

Specimens manufactured using Melt Theme and Wafer Theme are further referred to as MT and WT correspondingly. Examples of design structures and manufactured EBM WT and MT samples are presented in Figure 1.

In the build file, WT samples were placed directly on the base plate and were oriented with the side walls normal to the build direction. MT models were oriented in such a way that their cross-section diagonal was aligned with the build direction.

After manufacturing, the specimens were removed from the powder bed and separated from the base plate in a standard ARCAM Powder Recovery System. All specimens were subjected to compressed airflow for more than 10 min, thereby ensuring the removal of the powder particles loosely connected to the surfaces.

WT specimens were manufactured with turned off MultibeamTM contour mode (continuous beam path), beam current of I = 5 mA, and scanning velocity of v = 1000 mm/s. Three specimens were manufactured for each of tension and compression tests. The fixation heads for all tension specimens were manufactured using Melt Theme with default

parameter settings, Figure 1f, h. Equivalent parameters of the manufactured compression specimens are given in Table 1.



The surface morphology was studied by scanning electron microscopy Quanta 200 3D (FEI, Eindhoven, The Netherlands).

The outer dimensions of the fabricated scaffold samples were measured by a Vernier caliper; all specimens were weighted on an Acculab ALC-210d4 (Sartorius AG, Göttingen, Germany) scale. Their calculated solid volume and measured weight were used to determine the apparent density *ρ* of the structure (presented in Table 1). Assuming the density of solid Ti-6Al-4V is equal to *ρ*0 = 4.43 g/cm3, the porosity *P* of the scaffolds in % was obtained by:

$$P = 1 - \frac{\rho}{\rho\_0} \tag{2}$$

### *2.4. X-ray Computed Tomography*

The X-ray computed tomography (XCT) measurements were performed at BAM using an XCT scanner, developed together with the company Sauerwein Systemtechnik (today RayScan Technologies GmbH, Meersburg, Germany). A microfocus X-ray tube XWT-225- SE (maximum voltage 225 kV) from X-RAY WorX GmbH (Garbsen, Germany) was used as a source. An XRD1620 (CsI scintillator, 2048 × 2048 pixel) detector from PerkinElmer Inc. (Waltham, MA, USA), with in-house built enclosure and cooling system was used. A tube voltage of 120 kV and a tube current of 120 μA were used during the scans. The voxel size was 15.3 μm. The reconstruction of 3D volumes from 2D projections was made by the software developed in BAM using a filtered back-projection algorithm [33]. The obtained raw data files were analyzed using VGStudio MAX 3.3 by Volume Graphics, Heidelberg, Germany. The STL files of the 3D models and the XCT-based reconstructions were used to conduct Nominal/Actual comparison. The wall thickness was evaluated by the sphere method.

### *2.5. Mechanical Tests*

Quasi-static uniaxial compression and tension tests were performed using a universal testing machine INSTRON 3369 and INSTRON 5582, (Instron Deutschland GmbH, Darmstadt, Germany), with a 50 kN load cell. Tests were conducted at 20 ◦C according to ISO 13314:2011 [34], and using a displacement rate of 0.5 mm/min. Through the measurement of the applied load, we calculated the stress dividing the load by the effective area of the lattice structures. The failure strain was set at 50% of the specimen height. Results for the quasi-elastic gradient [35], compressive offset stress, first maximum compressive strength, energy absorption at 50% strain (Equation (3)), and specific energy absorption (Equation (4)) were calculated following ISO 1331 standard [34].

The ISO 13314 was devoted to describing the mechanical behavior of beam-based structures. TPMS sheet-based scaffolds represent structures with more complex shapes than beam-based structures. Indeed, the term "unit cell" is obvious for beam-based structures, but not for sheet-based structure. We therefore took ISO 13314 as a reference at the stage of designing the scaffold but did not strictly follow it. In fact, our goal was to qualitatively assess the behavior of our scaffolds, not to qualify them for production. Indeed, there

are many scientific articles not strictly following the requirements of standard test or production methods [5,35–38].

The quasi-elastic gradient Eqe [4,39–43] of the porous samples is the gradient of the straight line determined within the linear deformation region at the beginning of the compressive stress-strain curve, i.e., this value is defined similarly to Young's modulus E for bulk material. Additionally, the compressive offset stress and the first maximum compressive strength for porous specimens are defined similarly to the yield stress *<sup>σ</sup>y* and the compressive strength *σmax* for bulk specimens. Yield strain was defined as 0.2% strain, and the compressive offset stress was determined accordingly. Quasi-elastic gradient *Eqe*, yield stress *<sup>σ</sup>y*, and ultimate tensile strength *σmax* were estimated for tensile specimens. The plateau stress *<sup>σ</sup>p<sup>l</sup>* is the arithmetical mean of the stress values between 20% and 40% compressive strain. The point in the stress-strain curve at which the stress is 1.3 times the plateau stress is defined as the plateau end. It can be used for the determination of energy absorption and energy absorption efficiency:

$$\mathcal{W} = \frac{1}{100} \int\_0^{\varepsilon\_0} \sigma d\varepsilon \tag{3}$$

where *W* is energy absorption per unit volume (MJ/m3), *σ* is the compressive stress (MPa), *e*0 is the upper limit of the compressive strain. The energy absorption per unit volume was calculated from the area under the stress-strain curve up to 50% strain.

The crashworthiness of a material can be expressed in terms of its specific energy absorption. The specific energy absorption (Equation (3)) *ψ* is defined as the work *W* performed per unit weight when the material is compressed in a uniaxial manner up to a specific strain. The strain of 50% [41,44] was chosen for the evaluation of specific energy absorption:

$$\psi = \frac{\mathcal{W}}{\rho}; \mathcal{W} = \int\_0^\varepsilon \sigma d\varepsilon \tag{4}$$

where *ρ* denotes the mass density, *σ* the axial stress, and *ε* the work-conjugate axial strain [45].
