Page 323 - IJB-10-2
P. 323

International Journal of Bioprinting                                Continuous gradient TPMS bone scaffold


















                                   Figure 1. (a) C = 0, G surface; (b) C ≠ 0, G surface; (c) segregation phenomenon.


               In this study, G and P cells were utilized as the subjects   The goal is to obtain a TPMS porous structure with
            of research. The grasshopper parametric modeling   continuously changing porosity. In Equation IV, the linear
            platform  was employed to design a continuous gradient   function for C with respect to the parameter z is defined as:
            porous structure that is suitable for bone reconstruction.           C   z               (IV)
                                                         35
            Within  the  grasshopper  environment,  the  input implicit
            function can be transformed into a visual grid structure.   In this  equation, C is the parameter  used  to control
            By iterating through the grid structure multiple times, a   the porosity, while  z represents the Z-direction porosity
            grid surface with fewer defects can be obtained, ultimately   adjustment value of the TPMS structure. The value range
            leading to the development of a TPMS model for     for z is set to be [0,2.5π]. Figure 2 illustrates the variation
            additive manufacturing.                            of porosity and z per unit volume of the porous structure.
                                                               The porosity changes linearly along the Z-axis as z changes.
               Within a given cubic space measuring 20 mm on each   Similarly, if the parameter C is given a linear function of x
            side, different TPMS porous structures can be obtained by   or y, the same rule applies to the X-axis or Y-axis.
            adjusting the parameter C of the implicit function based
            on  the  existing  parameters.  Taking  the  G  surface  as  an   To obtain a TPMS model with a continuous porosity
            example, when C is set to 0, the porosity of the unit cell   gradient and good structure, the porosity range is
            remains the same (refer to Figure 1a). However, when C is   limited to 30–80%. If the porosity is too high, partition
            not equal to 0, the porosity between adjacent units becomes   phenomena may occur, while a porosity that is too low
            distinct (as shown in Figure 1b). Similar rules apply to other   may result in poor 3D printing quality. Within the limited
            TPMS structures. Increasing the absolute value of C leads   porosity range, different Z values are selected to determine
            to an increase in porosity, but if the absolute value becomes   the corresponding unit cell porosity. A graph of unit cell
            too large, partition phenomena may occur, resulting in an   porosity versus Z values is then plotted. By fitting the data,
            incomplete TPMS structure (see Figure 1c).         a functional relationship between the porosity parameter
                                                               C and z is obtained:
               To achieve a tunable pore gradient in the TPMS structure,
            a linear function is introduced for the parameter C.             C  0 1954.   z  0 3124.   (V)
                                                                  In this study, six models of G and P surfaces were
                                                               designed.  The  fitting  curve  in  Figure  2  shows  that  the
                                                               porosity of the six models is controlled at approximately
                                                               65%. The implicit function expressions of the two models
                                                               are as follows:
                                                                     f(x, y, z)  = sin(ωx)cos(ωy) + sin(ωz)cos(ωx)
                                                                               + sin(ωy)cos(ωz) = 0.1954*z
                                                                               + 0.3124                   (VI)

                                                                       f(x, y, z)  = cos(ωx) + cos(ωy) + cos(ωz)
                                                                             = 0.1954*z + 0.3124         (VII)

                                                                  The difference between the different models of the two
                                                               surfaces is achieved by controlling the periodic parameter
                      Figure 2. Linear fitting of porosity change  ω. The periodic parameters for each model are set as


            Volume 10 Issue 2 (2024)                       315                                doi: 10.36922/ijb.2306
   318   319   320   321   322   323   324   325   326   327   328