Generative 3D design
A graphical algorithm editor (Grasshopper 3D, McNeel, USA) integrated with computer aided design (Rhinoceros 3D, 3D (Rhino) modelling tools (McNeel, Miami USA), was used to control the parameterization of the nacre inspired design. Principally this design considers the basic hexagonal unit of the hard aragonite platelets within nacre (Fig. 1a) and the organization of the modelled counterparts though a Generative Design workflow (Fig. 1b) into a nacre-like, brick-and-mortar model (Fig. 1c). The developed biomimetic approach integrates two different design steps: the first step produces the polygonal unit or tale in which the number of sides, diameter and thickness can be provided as input, the second step ensures the polygonal unit distribution within the boundaries of interest through a cascading automated sequence of CAD operations to generate the final design (Fig. 1d).
The hexagonal unit was repeated within the 3D space to fill the bulk of the final component by first covering a planar surface (in the X and Y axes) within the bounding box of the final part. Stacking additional layers of platelets in the Z axis up to the top surface of the bounding box (multi-material mesh). A ‘dog bone’ geometry was chosen for the bounding box. as a mechanical testing standard (ASTM D638) but with slight width modified to embed complete platelets within the volume. Length, width, and height of the bounding box were evaluated and divided by the corresponding dimensions of the designed platelet to obtain the number of grid cells nx and ny as shown in Fig. 1d. The grid is repeated along the Z axis as many times as allowed by the ratio between the height of the bounding box (box enclosing the object) and the height of the platelets (nz). Alternate layers in a ‘true-false’ pattern are shifted along the X axis so that the platelets belonging to the even layers, in the workflow of Fig. 1d indicated as 2*(n + 1), and the platelets belonging to the odd layers, 2*n, are always sharing the same surface area. Two offset values were introduced into the model to leave spaces in between platelets in the XY plane and Z direction. The thickness of the bulk material representing the polymer soft phase within the nacre was tuned by setting the x and y offsets representing the size of the hexagonal cell radius, and the z axis offset representing the distance between the edges of the platelets of different layers. A non-manifold merge between the bulky component and the fibres pattern creates the two regions was exported as .STL files, representing the hard phase and the complementary polymeric phase of the composite. The platelet aspect ratio was varied from 2 to 9, therefore, as the thickness was held constant to 1mm, platelets widths ranging from 2 mm to 9 mm. The WT matrix material between the platelets was held constant at a thickness of 300 µm. These platelet dimensions resulted in a reinforcement volume fraction of the composites ranging from 53–65%.
Multi-material 3D Printing
An inkjet-based 3D printer (ProJet 5500X, 3D Systems, USA) was used as the manufacturing tool for the simultaneous layered deposition of a hard-white (WT) material (VisiJet® CR-WT 200, 3D Systems, USA) and a soft black (BK) material (VisiJet® CE-BK, 3D Systems, USA) from the same 3D printer. The 3D printer resolution was set to an ultra-high definition 13 µm layer thickness (750 x 750 x 2000 Dots per Inch (DPI)) and two orthogonal orientations of the 3D printed parts were used to ascertain variability during manufacturing. Previous work has noted the errors that can exit between the design and physically 26.
Two sets of 3D printed hybrid nacre composites were successfully printed (Fig. 2a). The first sample set produced hard WT material platelets parallel to the 3D printing platform (in-plane), whereas the second sample set was manufactured rotated at 90 degrees to the plane of the 3D printing platform (out-of-plane). These multi-layered composites varied in terms of reinforcement volume fraction, according to the diameter of the embedded platelets. Thus, a range of nacre inspires composites, with increasing volume fraction of the reinforcement, were obtained by setting different inputs digits in the graphical algorithm editor. The elastic modulus for WT and BK materials used for manufacturing the hybrid nacre was 364 MPa and 0.05 MPa, respectively. The platelet aspect ratio was varied to influence the pull-out mechanism within hybrid nacre when fibre aspect ratio is varied. The stress transfer from matrix to fibre was modelled (Fig. 2b and 2c) using the theory of Kelly and Tyson 27 with the assumption of a single stiff fibre interacting with a matrix being in a plastic state. Shear transfer from the matrix to the fibres was observed during the tensile testing (Fig. 2d). The shear strength at the interface τ is constant and equal to the shear strength of the matrix, obtained as follow:
\(\:{{\tau\:}}_{\text{a}\text{v}}=\frac{{{\sigma\:}}_{\text{f}}\:\text{d}}{2\:{\text{l}}_{\text{c}}}\) Eq. (1)
Where d represents the fibre diameter and σf is the fibre strength when the fibre has a critical length lc. Results from adhesive lap joint shear tests ASTM (D1002) 28, showed an average interfacial shear strength for the BK material of 1.5 [MPa]. The platelets diameter was varied from 2 to 9 mm in steps of 1 mm, whereas the thickness of the plates was kept constant at 1 mm. The soft phase was 300µm thick and then reduced to 30 µm, to appreciate the variation of the mechanical interplay between phases. The interaction scenario between BK and WT, evaluated through a preliminary shear test, was defined by an interfacial overlapping distance of 5 mm and a thickness of the compliant material of 300 µm. Similarly, the result of the tensile testing on WT material produced a σf value of 24.4 [MPa], whereas the considered thickness of the fibre was 1 mm.
Optical microscopy
Optical microscopy (Leica, Switzerland) was used to evaluate the material deposition by the 3D printer and assess the quality of the manufacturing. Optical images were taken with a 25X magnification lens and the resulting platelets width and distances among platelets were measured and verified using a 0.01-millimetre division over three different locations at the ends and centre of the 3D printed samples.
Elastic properties of composites
Elastic properties of the hybrid nacre composites were obtained following the standard ASTM D638 30. Mechanical tensile tests were performed at room temperature in a Universal Testing Machine (Z030, Zwick Roell, UK) fitted with a 30 kN static load cell. Samples were fully clamped and tested at a rate of 2.0 mm.min− 1 until failure. Sandpaper was applied onto the dog bone samples’ extremities to avoid sample slipping from the machine clamps.
Impact testing
Impacting testing was used to evaluate the energy absorbing capacity of the nacre inspires composites. A particular aim here was to assess whether the composites exhibited increased energy absorption compared to the base materials, which would evidence the transfer of toughening mechanisms from nacre to the 3D printed composites. Material testing standards for additive manufacturing of polymers, ISO 179 31, ASTM D6110 32 and DIN EN ISO 179 − 133 are relevant impact testing methods with the latter used in the work here. An impact tester (5102 pendulum impact tester, Zwick Roell, UK) was used in an unnotched Charpy configuration, at room temperature (21°C). A flatwise positioning of the 3D printed parts was adopted to mimic a normal directed, physiological impact force in shell-like materials. The produced samples (80 mm×10 mm cross-section by 7 mm height) were placed in an edgewise direction and, for variety of WT volume fraction as well as in-plane and out-of-plane configurations, five samples were tested evaluating mean and standard deviation. Base materials samples for WT and BK were tested using a pendulum energy of 5J and 1J respectively due to the varying toughness of the base materials.
In situ X-ray computed tomography (XCT)
In situ XCT was conducted using an X-ray microscope (Xradia Versa 520. Carl Zeiss Microscopy, USA) coupled with a loading device (CT5000 5KN, Deben UK Ltd). Stepwise tension was performed, and three displacement steps of 1 mm were applied; following each loading step a full tomogram was then acquired. The XCT operated at 70 kV/6 W to achieve 35 µm isotropic voxel size using a total of 1601 projections over 360°, with an exposure time of 1.5 s. The 2D X-ray projections from XCT were reconstructed to a 3D volume using the Scout and Scan Reconstructor software (Zeiss) and visualized with XRM3DViewer 1.2.8 (Zeiss).