Morphology of 3.25 wt% CNT/TPU cubic with various infill patterns and densities. Figure 1(a) shows the surface slicing images of the 3.25 wt% CNT/TPU cubic with various infill patterns and densities. At an infill density of 0%, the inner space at 0 N was empty. It became denser as the ratio increased from 20–80%. The ZG was deposited with layers divided by lines crossing in two directions perpendicular to each other. In the case of TR, the layers were divided into triangles and stacked in the orthogonal direction. In the case of ZG and TR, the line spacing narrowed with an increase in the infill density. HN were deposited using the same hexagonal layers. For HN, the size of the hexagonal units gradually decreased with an increase in the infill density.
The morphologies of 3.25 wt% CNT/TPU cubic with various infill patterns and densities are shown in Fig. 1(b). In FFF 3D printing, the layers are stacked and constructed, and each layer was partially bonded to the polymeric filaments. The bonding was induced by thermal energy as the filament melted. The bonding phenomenon can be explained by sintering, and the degree of bonding varies depending on the surface tension and viscosity. Further, the bonded part differs based on the infill pattern and density [36, 37]. The bonding between the layers was very good for the 3.25 wt% CNT/TPU cubic. The TPU exhibited excellent elasticity and adhesion[38]. For TR, the 3D infill type was the most complex. Among the three patterns, HN, which is a 3DF infill type, showed the most porous space.
Actual time and weight of 3.25 wt% CNT/TPU cubic with various infill patterns and densities. Figure 2(a) and (b) show the actual printing time and weight of a 3.25 wt% CNT/TPU cubic with various infill patterns and densities. For 0 N, the printing time was 7 min 18 s. When the infill density was 20%, the printing times were 8 min 51 s, 9 min 12 s, and 9 min 6 s for 20ZG, 20TR, and 20HN, respectively. At 50% infill density, it took 10 min 17 s, 12 min 34 s, and 10 min 23 s for 50ZG, 50TR, and 50HN, respectively. At 80% infill density, 80ZG, 80TR, and 80HN took 11 min 34 s, 14 m 25 s, and 12 m 34 s, respectively. Printing time was in the order TR > HN > ZG; the time increased with an increase in the infill density. For ZG, the printing time was the shortest because the nozzle path was divided into continuous lines. However, The TR with the longest printing time had the most complicated movement path because the nozzle moved while dividing the layer into triangles. The weights were confirmed to be 0.39 ± 0.01 g, 58 ± 0.01 g, and 0.76 ± 0.05 g for 20ZG, 20TR, and 20HN, respectively, when the infill density is 20%. At an infill density of 50%, 50ZG, 50TR, and 50HN were 0.58 ± 0.01 g, 0.50 ± 0.00 g, and 0.50 ± 0.01 g, respectively. For an infill density of 80%, 80ZG, 80TR, and 80HN were identified as 0.76 ± 0.05 g, 0.75 ± 0.07 g, and 0.65 ± 0.02 g, respectively. The actual weight is larger in the order ZG > TR > HN, and it increases with an increase in infill density. As confirmed by the slicing image and morphology, ZG had the largest weight because of the highest density of the internal space as the lines dividing the layer were located the closest. In contrast, HN had the smallest weight because it had the most space inside.
Crystallization of 3.25 wt% CNT/TPU cubic with various infill patterns and densities. Figure 3 shows the XRD of 3.25 wt% CNT/TPU with various infill patterns and densities. The results of XRD pattern of the cubic with various infill patterns and densities indicated that similar patterns of the cubic were presented peak at 2θ = 19.5° and small peak at 2θ = 26.0° and 43.0°. The peak at 2θ = 19.5° is attributed to the influence of TPU. The hard-domain crystalline phase of TPU showed a typical diffraction peak around 2θ = 19.5°, and the amorphous region showed a wide scattering region. Peaks at 2θ = 26.0° and 43.0° appeared because of the effect of the CNT. The peak observed at 2θ = 26.0° originates from the ordered arrangement of the concentric cylinders of graphitic carbon. Peaks around 2θ = 43.0° were attributed to the graphitic plane and small amount of catalyst particles remaining inside the MWCNT walls[39–42]. Therefore, the crystallographic structure of TPU did not vary with the infill conditions.
Compressive property of 3.25 wt% CNT/TPU cubic with various infill patterns and densities. Figure 4(a) shows the compressive strain–stress (S–S) curves of the 3.25 wt% CNT/TPU cubic with various infill patterns and densities. For the 0N sample, the strength was very small and had a yield point of 35%. At an infill density of 20%, a yield point with approximately 30% compressive elongation was observed. This indicates that the initial modulus was superior to that of ZG; however, HN at 50% compression was the best. At an infill density of 50%, ZG exhibited the highest strength, whereas HN exhibits the lowest. The yield point was observed at a compressive elongation of approximately 35%. No yield point was observed at an infill density of 80%; the strength of TR was the highest. The compressive strengths of the samples increased with the infill density.
Figure 4(b) shows the compressive property of the 3.25 wt% CNT/TPU cubic with various infill patterns and densities. The initial moduli were 5.76, 9.13, and 13.94 MPa for 20ZG, HN, and 80HN, respectively. They were 3.61, 6.06, and 12.79 MPa for 20TR, 50TR, and 80TR, respectively. Thus, the highest compressive property for ZG was at 20% infill density, and the highest for HN was at 50% and 80% infill densities. The TR pattern exhibited the lowest initial modulus. At 50% compression, the compressive stresses were 0.79, 2.59, and 10.12 MPa for 20HN, 50ZG, and 80TR, respectively. For toughness, 0 was 0.09 J and ZG was 0.26 J, 0.87 J, and 1.84 J; TR was 0.18 J, 0.84 J, and 1.99 J; and HN was 0.27 J, 0.67 J, and 1.81 J, as infill density increased from 0–80%. Therefore, the strength of the sample increased with an increase in the infill density. For the infill pattern, TR exhibited the best performance, and HN was the toughest. In the case of TR, the rate of increase in the compressive strength was the greatest with an increase in the infill density. During the FFF process, the movement path of the nozzle was determined based on the infill conditions. The movement path of the nozzle determines the internal shape and affects the physical properties[36, 37, 43]. Therefore, as confirmed by the slicing image and morphology, the TR pattern, which was stacked with the most complex layers, showed more layer-bonded parts than those of ZG and HN. Therefore, the strength increased with an increase in the infill density. HN showed the most elastic and toughest performance with an increase in the infill density because it was porous inside compared to ZG and TR.
Electrical conductivity property of 3.25 wt% CNT/TPU cubic with various infill patterns and densities. Figure 5 shows the electrical properties of the 3.25 wt% CNT/TPU cubic with various infill patterns and densities. For the 0N sample, the current value was more than twice as large as the others at 28.27 mA. The current values of 20% infill density was 14.50, 15.52, and 16.09 mA for 20ZG, 20TR, and 20HN. For 50% infill density, the current values at 50 V are 13.87, 12.23, and 12.63 mA for 50ZG, 50TR, and 50HN, respectively. For an 80% infill density, the values of current at 50 V were 13.63, 6.48, and 14.70 mA for 80ZG, 80TR, and 80HN, respectively. The current tended to decrease with an increase in the infill density.
For the infill pattern at each infill density, 20HN, 50ZG, and 80HN showed the highest conductivities at 16.09 mA, 13.87 mA, and 14.7 mA, respectively. The conductivity of HN was excellent because regular hexagonal layers were stacked in the same layer. In contrast, the conductivity of TR decreased rapidly with an increase in infill density. Electrical properties can be analyzed in terms of the slicing image and morphology. Conductivity decreases with an increase in the length of the sample. Further, conductivity is proportional to length and inversely proportional to the cross-sectional area, and therefore, conductivity decreases with an increase in length[44]. The moving path of the output nozzle also increased with an increase in infill density. In addition, the nozzle movement path increased and the output time was the longest because the TR was formed with the most complex inner layer. If there were many moving paths of the nozzle, the path through which the current flowed was the longest; therefore, the conductivity decreased.
Electrical heating property of 3.25 wt% CNT/TPU cubic with various infill patterns and densities. Figure 6 shows the electrical heating properties of the 3.25 wt% CNT/TPU cubic with various infill patterns and densities. When 50 V was applied for 5 min, the surface temperature of 0 N was measured at 23.2 ℃. For a 20% infill density, the surface temperatures were 24.1, 24.9, and 29.7 ℃ for 20ZG, 20TR, and 20HN, respectively. The surface temperatures for a 50% infill density were 33.1, 31.8, and 32.8 ℃ for 50ZG, 50TR, and 50HN, respectively. For the 80% infill density, they were 41.2, 29.8, and 40.9 ℃ for 80ZG, 80TR, and 80HN, respectively.
The heating temperatures were high and in the order 80% > 50% > 20% > 0%. The higher the infill density, the higher were the electrical heating temperatures. 20HN at a 20% infill density and 50ZG and 80ZG at 50% and 80% infill densities have the highest electrical heating temperatures of 29.7 ℃, 33.1℃, and 41.2 ℃. The exothermic temperature increased with an increase in the infill density. The electric heating property had the highest temperature in the 80% infill density and the ZG infill pattern with the highest internal density was attributed to the influence of the slicing image and morphology. In contrast, the conductivity of 80TR decreased rapidly, and therefore, no heat was generated. In addition, with an increase in the infill density, there was a decrease in conductivity and an increase in surface temperature; this was confirmed by resistance heating. Resistance heating can be confirmed by the increase in the CNT content with an increase in the infill density. When a voltage is applied, current flows in the CNT portion, and electrons move and change in an orderly manner and collide with each other, resulting in heat generation[44, 45].
3DP CNT/TPU orientation models with various infill patterns and densities. Figure 7(a) shows the 3.25 wt% CNT/TPU orientation model of the outer walls and cubic lateral side. Figure 7(b)–(d) show a CNT/TPU slicing image with an infill density of 80%, the nozzle movement path, and a schematic of the CNT/TPU orientation among the HN infill patterns with the best performance in this study. The orientations of the CNT and TPU were stacked in an identical manner to the morphology based on the moving path of the nozzle. The electrical and electrical-heating characteristics can be understood based on an orientation diagram. The same layers were successively deposited for HN. The CNTs were placed in the same direction when each part was viewed in detail; thus, when voltage was applied, electricity could flow without being significantly hindered. The nozzle movements and schematics of different fill patterns and fill densities are shown in Supplementary Table S1 in Supplementary information. ZG had a higher density than the other patterns, and in the case of TR, the obstacles to electrical flow increased because of its complex structure. Therefore, manufacturing a soft sensor HN in which the same layers are stacked can be considered the most suitable infill pattern. In addition, the CNT content increased with increasing density.