Page 467 - IJB-10-6
P. 467

International Journal of Bioprinting                                 Stress prediction in 3D-printed scaffolds


















            Figure 1. Schematic of the designed model: (a) 0°/90° scaffold; (b) 0°/60°/120° scaffold; and (c) 0°/45°/90°/135° scaffold. All scaffolds have a size of 8 × 8 ×
            3 mm , and a layer thickness of 0.3 mm.
                3
            Table 1. Scaffold model parameters                                    ρ  
             Rotation    Size (mm )  Porosity (%) Pore size (mm)           P =  1 −  ρ 1   × 100%       (III)
                               3
             angle (°)                                                             2  
             90            8 × 8 × 3    60         1.2            where P is the desired porosity, ρ  is the density of the
                                                                                            1
             60            8 × 8 × 3    60         1.2         scaffold, i.e., the mass divided by the volume, and ρ  is the
                                                                                                        2
             45            8 × 8 × 3    60         1.3         theoretical density of the HAP material.
                                                               2.5. Scaffold morphology acquisition
            2.3. Mechanical properties                         and reconstruction
            The mechanical properties of the HAP scaffolds were   The Bio-Architect-PX device integrates optical coherence
            tested using an AI-700-S (GOTECH, China) compression   tomography (OCT) technology, which allows for non-
            testing machine. Loading was performed at a displacement   destructive monitoring of the 3D structure of printed
            rate of 5 mm/min, and scaffolds with different angles in   scaffolds during the additive manufacturing process and
            different states were tested separately. According to the   captures the 3D structural images of the entire printed
            stress–strain curve, the compressive modulus (E) and   scaffold after printing. Previous works of our team have
            compressive strength () of the scaffolds were calculated   reported on this technology and system, referring to it as
            using Equations I and II, respectively.            P-OCT technology. 32,33  In this study, a swept-source OCT
                                                               system  with  a  lateral  resolution  of  15  μm  and  an  axial
                                                               resolution of 7.2 μm was used to scan the printed scaffold
                              F − F
                          E =  1  2                     (I)    layer by layer during the printing process, and the P-OCT
                             D − D 2                           images of the entire printed scaffold were obtained using
                               1
                                                               the software integrated within the Bio-Architect-PX device
                                                               after  printing. Micro-CT, a  nondestructive 3D  imaging
                            σ =  F max                 (II)    technique, was used to perform non-destructive imaging of
                                 S                             the scaffold, obtaining internal 2D cross-sectional images.
                                                               In this study, a micro-CT (μCT-100; Scanco Medical AG,
               In these equations,  F  and  F  represent the stress   Switcerland) with a resolution of 15 μm was used to scan
                                 1
                                        2
            corresponding to any two points in the linear variation   the scaffold after it was dried and sintered. The DICOM
            stage of the stress–strain curve of the scaffold, respectively.   files of the P-OCT and micro-CT images were imported
            D  and D  represent the strain corresponding to the above   into Mimics software (Materialise, Belgium), selecting the
             1
                   2
            two points, respectively.  F max  is the force corresponding   threshold value range for image segmentation to generate
            to the fracture point of the  scaffold,  and  S is  the cross-  the 3D reconstructed model.
            sectional area of the scaffold, that is, the area under force.  2.6. Finite element analysis
            2.4. Macroscopic morphology and porosity testing   Ansys Workbench 19.2 software was used to simulate the
            The  macroscopic  structure  of  the  HAP  scaffolds  was   mechanical properties of both the theoretically designed
            observed and recorded using a digital camera. The weight   scaffolds (Figure 1) and the reconstructed scaffolds
            of the scaffolds was measured using an electronic balance,   (Figure 2).  The properties of the HAP material were
                                                                                          34
            and the length, width, and height of the scaffolds were   consistent with our previous work,  as presented in Table 2.
            measured using a vernier caliper. The porosity test was   Regarding  boundary  conditions,  to  simulate  actual
            conducted using the density method:                compression mechanics testing experiments, two identical

            Volume 10 Issue 6 (2024)                       459                                doi: 10.36922/ijb.4460
   462   463   464   465   466   467   468   469   470   471   472