Fabrication and Characterization of Electrically Conductive 3D Printable TPU/MWCNT Filaments for Strain Sensing in Large Deformation Conditions

This study investigates the development of thermoplastic polyurethane (TPU) filaments incorporating multi‐walled carbon nanotubes (MWCNT) to enhance strain‐sensing capabilities. Various MWCNT reinforcement ratios are used to produce customized feedstock for fused filament fabrication (FFF) 3D printing. Mechanical properties and the piezoresistive response of samples printed with these multifunctional filaments are concurrently evaluated. Surface morphology and microstructural observations reveal that higher MWCNT weight percentages increase filament surface roughness and rigidity. The microstructural modifications directly influence the tensile strength and strain energy of the printed samples. The study identifies an apparent percolation threshold within the 10–12 wt.% MWCNT range, indicating the formation of a conductive network. At this threshold, higher gauge factors are achieved at lower strains. A newly introduced Electro‐Mechanical Sensitivity Ratio (ESR) parameter enables the classification of composite behaviors into two distinct zones, offering the ability to tailor self‐sensing structures with on‐demand properties. Finally, flexible structures with proven application in soft robotics and shape morphing are fabricated and tested at different loading conditions to demonstrate the potential applicability of the custom filaments produced. The results highlight a pronounced piezoresistive response and superior load‐bearing performance in the examined structures.


Introduction
][8][9][10] The novelty of such structures involves the ability of the materials to sense and adapt to external stimuli in real time.Among the ongoing innovations in smart materials, electrically conductive polymer composites (ECPC) combined with conductive nanofillers have emerged as an exciting field of research, harnessing mechanical robustness and electrical conductivity. [11,12]olymer matrices composed of either thermoplastics or thermosets typically exhibit insulating electrical properties.However, their conductivity can be improved by incorporating electrically conductive carbon-based fillers. [13,14]When embedded into a polymer matrix, the conductive filler element fosters a percolated conductive network. [15]For ECPC materials, the conductivity percolation threshold is important, as it demands a specific volume of nanofillers to introduce conductivity to the otherwise insulating polymer matrix host. [16]This threshold is influenced by the morphology and distribution of the nanofillers and processing techniques, affecting their dispersion and agglomeration. [17,18]The inclusion of functional fillers has the potential to increase electrical conductivity and piezo-resistive behavior.The piezo-resistive trait, characterized by the change in electrical resistance during mechanical deformation, is crucial for strain sensing applications, especially in monitoring large deformations. [19]Typically, the percolation threshold ranges from 0.01 to 10 wt.% of nanofillers, which allows fine-tuning piezoresistivity, stretchability, and strength in ECPC materials. [20,21][24][25][26] Metallic fillers can yield superior electrical conductivity compared with carbon-based materials. [27]However, composites incorporating metallic constituents require a greater volume of fillers to attain similar conductivity levels, affecting and limiting processability and their practicality across various applications. [28]At the same time, achieving the threshold conductivity to transition the polymer matrix from insulator to conductor is considered challenging for carbon black and graphene nanoplatelet fillers due to the limited physical network of the fillers associated with their lower aspect ratio (i.e., lengthto-diameter ratio). [29,30]On the contrary, multi-walled carbon nanotubes (MWCNTs) achieve lower percolation thresholds and enhanced electrical conductivity due to their high aspect ratio, forming extensive conductive networks at significantly lower loadings. [31]MWCNT loadings in ECPC materials, ranging from 2.0 to 7.1 wt.%, have been shown to exhibit piezo-resistive behaviors with higher gauge factors (change of resistance with strain). [32,33][36][37] TPU has distinct favorable properties due to its excellent mechanical attributes, such as high elongation, flexibility, and durability. [10,38,39][41][42] In addition, the resilience of TPU to abrasion, wear, and environmental conditions makes it well-suited for developing strain sensors that must endure significant deformation with multifunctional capabilities. [33,43]everal fabrication techniques, such as melt mixing, solution blending, and in situ polymerization, have been reported to integrate nanofillers in polymer matrices. [12,44,45]The nanofiller dispersion within the polymer host during fabrication influences the morphological, mechanical, and electrical properties of ECPC composites.Melt mixing, for instance, is straightforward but often causes inadequate dispersion of MWCNTs, which is highly influenced by processing parameters like screw speed, mixing time, and temperature. [36]Proper filler dispersion necessitates precise control of processing parameters to balance conductivity with the final mechanical and functional attributes of the composite. [46,47]Nonetheless, achieving uniform MWCNT dispersion often leads to poor electrical conductivity due to a weakly percolated MWCNT network. [48,49]Similarly, solution blending facilitates better dispersion with good reproducibility and reproducibility. [50]Moreover, carbon nanotubes (CNTs) tend to entangle due to strong van der Waals attractions at higher concentrations, posing challenges in their dispersion within the polymer matrix. [11,51]A blend of conductive fillers in the polymer matrix can mitigate these adverse effects. [52,53]For example, Liu et al. [54] observed a synergistic effect when CNT and graphene were used together as co-fillers, enhancing the strain sensing capabilities of ECPC materials up to a gauge factor (GF) of 35.78 at 30% strain.Despite these enhancements, introducing a co-filler requires precise control over the distribution and concentration of each filler type to preserve the electrical conductivity and mechanical properties, adding complexity to the fabrication process.Further, a major challenge of this method lies in the requirement of large amounts of organic solvents, which is impractical in largescale processing. [45,55,56]onsidering the emergence of additive manufacturing (3D printing) technologies, the interest in multifunctional composite filaments has increased drastically in the past few years.The fabrication of TPU/MWCNT composite filaments for fused filament fabrication (FFF) 3D printing can actualize the desired electro-mechanical properties with a comparatively simple fabrication process.The optimization of printing parameters in translating filament to FFF printed structures is essential to prevent nozzle clogging from nanofiller agglomeration and poor surface finish. [57]Achieving the optimal balance of nanofiller size and distribution, along with controlled printing conditions (i.e., temperature, speed, and print-bed temperature), is critical for successful extrusion and deposition.As a matrix material, TPU outshines other polymers by offering sufficient flexibility during spooling while ensuring that the final printed structure retains adequate mechanical strength, flexibility, and thermal stability. [58,59]he present study reports the fabrication and characterization of electrically conductive 3D printable TPU/MWCNT filaments for strain sensing under large deformation conditions.A thorough investigation of the microstructure-property relationships in these composites is carried out and reported.The experimental studies include comprehensive characterization of the fabricated TPU/MWCNT composite filaments by examining the influence of MWCNT content on the electro-mechanical properties and demonstrating their potential in strain sensing applications under extensive deformation conditions.The ability to fine-tune the electro-mechanical properties through FFF fabrication is further explored by developing flexible auxetic structures as a potential self-sensing structure for various applications, including soft robotic actuation, as reported in the work of Pagliocca et al. [60]

Filament Fabrication
The first step in fabricating property-tailorable 3D printing feedstock was to create filaments with the desired multi-walled carbon nanotube (MWCNT) weight ratios.Commercially available thermoplastic polyurethane, TPU (Ninjaflex 85A, PA, USA), was acquired and broken into smaller 1-2 mm pellets using a hand crank grinder.Pelletized TPU was mixed with the desired MWCNT ratio and stored in a desiccator for one hour.The MWC-NTs (10-30 μm length, 2.1 g cm −3 true density) were purchased from Nanostructured & Amorphous Materials, Inc. (TX, USA) as loose powder.Figure 1 shows a series of SEM micrographs of the as-received MWCNTs, showing a high-magnification view of the entangled nanotubes.The average size of the powder particles used to fabricate composite filaments was ≈50 μm.
The desiccated mixture was melted in a vacuum oven to form a visually uniform composite.The melted phase was cooled to room temperature inside the oven.Upon solidification, the composite was manually cut into smaller pieces and then reground using the same hand crack grinder.The two-step grinding process resulted in a more uniform distribution of MWCNTs and better reproducibility of the properties, discussed in detail in the following sections.Composite filaments with 0 (control samples) to 15 wt.%MWCNT contents were produced.In the control case (i.e., 0 wt.%MWCNT), the same fabrication procedure, including initial melting and double grinding, was followed.The final product of the double grinding process was fed into a filament extruder (Filabot EX2, VT, USA) equipped with a speed-controlled spooler.The extrusion temperature was set to 190 °C for all extruded filaments, producing TPU-MWCNT composite filaments with nominal diameters of 1.75 mm.A fan cooling system was used to control the spooling temperature to ensure no adhesion between the spooled composite filaments.Figure 2 shows the schematic of the filament fabrication process as well as representative macro and microscale images of a composite filament (10 wt.% MWCNT).The MWCNTs were found in the form of well-dispersed ≈1-10 μm aggregates inside the extruded com-posite filaments.The fused filament fabrication 3D printing process led to a more effective dispersion of MWCNT aggregates, forming a continuous network of nanofillers.The extruded composite spools were stored and dried in a filament drier (Sunlu S1 Plus, CA, USA) at 40 °C for at least 6 h before printing.The TPU-MWCNT filaments fabricated in this work included composites with 0 to 15 wt.% reinforcement ratio at 1-2 wt.% intervals.

3D Printing and Electro-Mechanical Characterizations
The TPU-MWCNT filaments were used for 3D printing of tensile test specimens for electro-mechanical characterizations.ASTM D638 dog-bone shaped tensile test specimens were printed with a thickness of ≈0.4 mm (three layers through thickness) with a 0.4 mm brass nozzle in an Ender 3 Pro machine.As shown in Figure 3a, the printing direction was selected to be parallel to the tensile loading axis.The optimal (continuous printing with minimal clogging and cross-sectional area variations in the deposit) hot-end temperature of the extruder was determined and set as 220 °C for composite filaments with 11 wt.% or less of MWC-NTs.For higher weight ratios, a hot end temperature of 240 °C was used due to the effects MWCNTs on the melting and rheological behaviors of the composite filaments.The bed temperature of 65 °C and print speed of 50 mm s −1 were used for all printings.Six tensile samples were prepared and tested for each MWCNT weight ratio.The stress-strain behavior of the samples was characterized in uniaxial tensile tests performed using a Shimadzu ACS-X frame equipped with a 10 kN load cell.Tensile tests were performed in displacement-control mode at a constant crosshead speed of 5 mm min −1 .Six samples per MWCNT loading were tested.
All tensile tests were accompanied by in situ electrical resistance measurements.Resistance of the tensile samples was measured before and during the mechanical tests to delineate the interdependence of electrical and mechanical behaviors.The electrical resistance of the tensile specimens was measured using  the four-point probe and source meter (Keithley 2450).Four electrodes were along the gauge length of the tensile specimen to measure the electric resistance as a function of tensile strains.
Several metrics were used to evaluate the mechanical, electrical, and co-dependent electro-mechanical behaviors of the 3D-printed samples in this work.Strain-at-failure and tensile strength (maximum engineering stress) were characterized for samples with various MWCNTs weight ratios.The specific (volume normalized) strain energy, W, of the tensile test pieces was also evaluated as the area below the stress-strain curves, i.e.
where,  and  denote engineering stress and strain, respectively, and  f is the strain at failure.To enable a rational comparison of the electro-mechanical response in this work, instantaneous gauge factors, GF, were determined using: [61] where, R 0 and R denote the initial (strain free,  = 0) and instantaneous resistance values measured at strain , respectively.Notably, R 0 and R() are functions of the MWCNT weight content.The variation of R with strain was measured over tensile strains ranging from 0 to 90% of strain-at-failure (i.e., 0.9  f ) in each MWCNT weight content.
The intended application of the multifunctional 3D printing feedstock is integration in electro-mechanical applications such as soft robotics and wearable electronics submitted to large deformations.Therefore, a multi-parameter electro-mechanical metric was defined to associate the mechanical and electrical properties of the materials, which is termed ESR herein.ESR is the ratio of failure strain,  f , to the product of tensile strength,  TS , and resistance, R (at 90% elongation), i.e.
A maximized ESR ratio indicates high strain-at-failure, low resistance (equivalent to high electrical conductivity), and low mechanical strength.

Fabrication and Testing of Self-Sensing Flexible Structures
[67] In this study, a unit cell design characterized by rectangular, center-symmetric cells was employed, also previously investigated by the authors. [60,64]The perforations in this design induce auxetic behaviors when stretched in the axial direction. [68,69]  the horizontal and vertical dimensions of the perforations, respectively.The horizontal distance between voids, denoted as s, characterizes the width of the vertical cell ligaments, and the total length, L, of each cell is consistently maintained at 10 mm.Incorporating these geometric parameters, two dimensionless parameters were utilized to define the unit cell geometry: the aspect ratio (AR) and inter-cell spacing (IS).
The selected design (see Figure 4a), with an AR = 3 and an IS = 0.1, is known for its effective auxetic response.The final design is composed of a pattern of 3 × 7 repeated unit cells with rectangular perforations and 10 mm solid gripping tabs at both ends of the sample to allow tensile load transfer to the structure during mechanical tests.
Two structures with 0 and 13 wt.%MWCNTs were fabricated using the previously discussed FFF process and printing parameters.Structural behaviors were evaluated using a Shimadzu AGS-X 100 kN load cell (Figure 4b).Tensile tests were conducted with 5% and 10% tensile strain during one full loading-unloading cycle at a 3 mm min −1 deformation rate.Concurrently, the piezoresistive response was analyzed using the same four-point probe setup discussed above.

Microstructure of TPU/MWCNT Filaments and 3D Printed Samples
Figure 5 shows the surface morphology of composite TPU filaments with various MWCNT weight ratios.The surface characteristics of the extruded filaments show distinct differences.The control (TPU-0 wt.% MWCNT) filament has a relatively smooth surface with shallow striations along the extrusion direction, possibly due to the roughness of the extrusion nozzle since similar imprints are evident on the remaining filaments.Increasing the MWCNT weight content leads to discernible surface features, e.g., circumferentially oriented surface.The 15 wt.% composite filaments exhibit slight bambooing.Generally, all these surface imperfections are byproducts of the filament extrusion process, [70] which are secondary artifacts herein since these filaments will be further processed during the 3D printing process.
Tensile samples were printed from composite filaments and used for electro-mechanical characterizations, as shown in Figure 6 (e.g., TPU-10 wt.% MWCNT) at different magnifications.The parallel lines in Figure 6a show the surface morphology of the printed tensile test sample, indicating overlapped roads that are conducive to an interconnected network of electrically conductive TPU-MWCNTs.High magnification SEM micrographs shown in Figure 6b confirm the presence of well-dispersed MWCNT-rich areas in the pristine printed tensile samples.

Mechanical Response of TPU/MWCNT Composite Samples
Tensile samples were characterized by several key performance metrics based on their tensile stress-strain response.Figure 7a shows typical stress-strain curves obtained from mechanical testing of neat and composite TPU filaments with 7 and 15 wt.%MWCNTs.An increase in the weight content of MWCNTs significantly affects the mechanical performance metrics, including the tensile strength (Figure 7b) and failure strain (Figure 7c).Strain-at-failure exhibits a more substantial decrease as the MWCNT weight ratio increases.Nonetheless, the variability in the strain-at-failure values (characterized by the standard deviation of the key performance metrics between five independent measurements) suggests highly repeatable results.Hence, the functional filler distribution in the filament feedstock is consistent.The highly nonlinear variation of stress-strain response with MWCNT weight ratio translated into the specific strain en-ergy metric (Figure 7d).The specific strain energy data obtained herein can be extremely useful in designing highly flexible selfsensing structures, especially those with applications in cyclic loading conditions. [61]

Electrical and Electro-Mechanical Behavior of TPU/MWCNT Composites
The electrical resistance was characterized before (i.e., unstrained) and during mechanical deformation.Figure 8 shows the variation of electrical resistance with the weight ratio of MWCNTs for three strain conditions.The strain-free results elucidate the profound effect of increasing the MWCNT content on the resistance of the printed samples, evidencing a significant decrease in resistance potentially due to the percolation of the conductive long carbon nanotubes.For example, the resistance at 15 wt.%MWCNTs is three orders of magnitude lower than the 5 wt.% counterpart.Furthermore, these results demonstrate a significant resistance drop in the 10-12 wt.% MWCNTs, which is substantially higher than those reported in previous works, [70] and the theoretically predicted values. [71]Generally, the trends observed in the resistance-filler weight content are typical of those reported in previous studies, substantiating the validity of comparative characterizations presented herein.Interesting to note are the non-monotonic trends that indicate a local resistance peak at 11 and 14 wt.%MWCNT loading.The reason for such anomalies was not fully explored here but possibly attributed to the manufacturing imperfections discussed above that give rise to the distribution of the rupture in the CNT connectivity throughout the samples.
Gauge factors, GF, (ratio of the relative resistance change to the strain) of the 3D printed composites were determined as a function of MWCNT weight ratios and various strains.Figure 9 depicts the GF-strain data obtained over a [0-0.35]strain range for all TPU-MWCNT composites examined herein.In general, the higher-weight content of MWCNTs resulted in larger gauge factors at smaller strains.While low-weight contents of the functional filler have lower gage factors, they remain sensitive and functional at larger tensile strains.Consistent with the findings in Novikov et al., [72] the highly nonlinear correlations between GF Figure 8. Variation of electrical resistance with filler weight content measured in tensile samples 3D printed from TPU-MWCNT composite filaments at strain free, 0.5 f , and 0.9 f ( f denotes strain-at-failure).
and tensile strain can be expressed as a power law, identified by regression and included in Figure 9.
Results shown in Figure 9 indicate GF values >20 at strains as high as 10%, which is considerable compared to previously reported GFs on similar composites.The results from Figure 9 Figure 9. Variation of gauge factor with strain within the full strain range for all TPU/MWCNT composites, fitted with a power law function that describes the highly nonlinear relationship.
are collated with gauge previously reported values in Figure 10.Since the literature data report gauge factors at different strains than those used herein, all GFs shown in Figure 10 are normalized by the corresponding strains (or maximum strain) at which the measurements have been conducted.Notably, the collated data in Figure 10 are obtained from TPU-MWCNT composites manufactured by various methods, including compression molding, [54,[73][74][75] twin-screw extrusion, [76] melt spinning, [32] melt blending and extrusion, [77] non-solvent induced phase separation (NIPS), [78] and solution casting. [79]The results in Figure 10 point to 1) higher CNT content than those reported before and 2) the 3D samples produced herein exhibit higher electro-mechanical sensitivity required for soft robotics and wearable electronics applications.

Self-Sensing Structures with Customizable Electro-Mechanical Properties
The ability to attain relatively large GFs at given strain ranges indicates that self-sensing structures can be fabricated from the current customizable feedstock by readily available FFF printers.Figure 11 summarizes the combined electro-mechanical properties in terms of the ESR parameter defined earlier.The ESR parameter inherits the uncertainty and nonlinear variation in mechanical and resistance values.However, the trend revealed in Figure 11 suggests two distinct remarks.First, the ESR curve can be divided into two zones, separated at 11 wt.%MWCNT loading.The curve parts that fall on either side of this transitional filler content show a nearly constant but distinct ESR value, e.g., representing an electro-mechanical switch.Second, the ESR ratios for filler loadings between 0 and 10 wt.% show at least two orders of magnitude lower values than those within the 12-15 wt.% range.The two zones identified in Figure 11 are of practical importance as they guide the design and development of self-sensing structures with desirable electro-mechanical behaviors.

Mechanical and Piezo-Resistive Responses of Self-Sensing Flexible Structures
To demonstrate the practical outcomes of this research, mechanical and piezo-resistive responses of custom-designed self-sensing flexible structures were concurrently evaluated.Figure 12a illustrates the stress response of neat TPU and TPU-13 wt.% MWCNT composites used to fabricate the auxetic structures discussed in Section 2.3 when subjected to 5% and 10% cyclic (full loading-unloading) strains.A higher stress peak observed at 10% strain compared to 5% strain indicates the strain hardening response of the structures, irrespective of their MWCNT weight ratio.During unloading, the neat TPU exhibited negative stresses at the end of each cycle, a behavior that occurs due to the viscoelastic nature of the material. [80]his phenomenon is likely due to the hysteresis in the stressstrain response, where the energy dissipation during the loading phase is not fully recovered upon unloading as the material undergoes a relaxation process. [81,82]In contrast, the TPU-13 wt.% MWCNT structure exhibited a distinct behavior under cyclic deformation, initially displaying a similar stress profile followed by a higher stress peak compared to the neat TPU at 5% and 10% strain levels.This enhancement in stress response occurred due to the reinforcing effect of the MWCNT filler, which likely restricts the polymer chain mobility, leading to a higher apparent stiffness. [83,84]Furthermore, the TPU-MWCNT composite samples completed their cyclic loops without displaying negative stress during the unloading half cycle, indicating reduced viscoelasticity.Such characteristics make MWCNT-reinforced composites potential candidates for self-sensing applications with quick response time.
Figure 12b shows the temporal and strain-dependent variations in the electrical resistance of the TPU-13 wt.% MWCNT composite.During the first cycle, a significant increase in resistance is observed, corresponding to the elongation of the sample under remote tensile loads.The stretching of the conductive MWCNT network within the TPU matrix increases the distance between conductive pathways and, consequently, the overall electrical resistance of the material. [12,54,85]The resistance decreases upon unloading, suggesting a reversible piezo-resistive behavior resulting from the reversibility of the conductive MWCNT pathway.Upon applying a higher strain of 10% in the second cycle, the resistance rises to values than those at 5% strain.The higher resistance suggests that the conductive MWCNT network experiences greater disruption under larger extensive deformations.Moreover, structural changes can lead to the variation of conductive path lengths, also affecting the global resistance characteristics of the structure.88] The higher resistance level at 10% strain indicates the examined potential effectiveness in measuring larger strains, although the fact that the resistance does not return to zero upon unloading from 10% strain implies the occurrence of permanent rupture of the MWCNT network.Regardless, integrating MWCNT filler content into an otherwise nonconductive TPU network, coupled with the tailored unit cell design, is anticipated to yield structures in potential applications requiring mechanical flexibility and electrical conductivity.Finally, it is worth noting that several other factors have been identified as parameters potentially affecting the electro-mechanical performance of the structures discussed herein.The most significant of these contributing factors are the number of cycles applied (and their resulting irrecoverable strain and/or resistance change), the loading rate (as it affects the polymer chain mobility), and temperature (hot and cold).[91][92] However, the interplay between the aforementioned parameters and the electro-mechanical properties of TPU-MWCNT composites is rarely discussed.

Conclusion
This study presented the facile fabrication and characterization of TPU/MWCNT filaments, emphasizing their application in strain sensing under large deformation scenarios.TPU was compounded with varying MWCNT weight ratios to create filaments specifically tailored for FFF 3D printing.The extrusion and printing parameters were optimized to ensure the homogeneity of the composite material as demonstrated by SEM analyses.Higher weight ratios of MWCNT led to the alteration of surface morphology with increased roughness, rigidity, and brittleness of the composite filaments.The incorporation of MWCNT nanofillers was found to enhance the tensile strength and decrease the specific strain energy of the composites.The consistent filler distribution within the composite matrix resulted in repeatable mechanical responses.Electrical characterization identified an "apparent" percolation threshold in the 10-12 wt.% MWCNTs, indicating the minimal content of MWCNT to form a fully conductive network within the TPU matrix.The gauge factor analyses revealed that higher MWCNT contents in composites can detect strain changes with high sensitivity.A novel metric, the ESR parameter, was introduced to quantify the electro-mechanical synergy of the composites, combining both mechanical and electrical properties.This parameter highlighted two distinct zones of behavior: one featuring high elasticity and low electrical resistance and the other with increased mechanical strength and resistance.Finally, the piezo-resistive response of the TPU/MWCNT composites under cyclic loading was investigated for neat TPU and MWCNTs-filled composites in novel auxetic structures relevant to soft robotics applications.Neat TPU displayed typical viscoelastic responses and intrinsic timedependent deformation characteristics with no piezo-resistive response.On the contrary, TPU with a 13 wt.%MWCNT content showcased larger peak stresses and a strain-dependent piezoresistive behavior.Such attributes could be pivotal for the strain sensing application during large deformation cases.

Figure 1 .
Figure 1.a) SEM micrographs of the as-received MWCNTs at different magnifications.b) High magnification micrograph of the entangled MWCNTs.

Figure 2 .
Figure 2. a) Schematic representation of TPU-MWCNT filament fabrication process.b) Macro and microscopic image of a TPU-10 wt.% MWCNT filament, showing the dispersion of the filler in the TPU matrix.

Figure 3 .
Figure 3. a) A typical tensile test specimen 3D printed from TPU-MWCNT filament (shown samples printed with 10 wt.% MWCNTs).b) Test setup showing the in situ electrical resistance-strain measurements using a four-point probe attached to the tensile sample.

Figure 4 .
Figure 4. a) Schematic of the designed unit cell geometry for the self-sensing flexible structure.b) Experimental setup for testing of the 3D printed flexible structure, showing the positioning of the tensile grip and the four-point probe used for in situ electrical resistance measurements during mechanical tests.
Figure 4a illustrates a schematic representation of the unit cell geometry, where the parameters a and b define

Figure 6 .
Figure 6.SEM micrographs showing a) surface morphology of printed tensile sample fabricated from TPU-10 wt.% MWCNT composite filaments.High-magnification images in (b) confirm the presence of MWCNT-rich areas.

Figure 7 .
Figure 7. a) Typical stress-strain curves obtained for 3D printed samples of neat TPU and composite filaments with 7 and 15 wt.%MWCNTs.Variation of b) tensile strength,  TS , c) strain-at-failure,  f , and d) specific strain energy as a function of MWCNT weight ratio.

Figure 11 .
Figure 11.Variation of the ESR parameter with filler weight content for TPU-MWCNT composites.Scatter bars indicate the combined uncertainty in the three experimentally measured parameters.Zones corresponding to stretchable low-resistance and high-strength high-resistance composites are marked.The horizontal red dashed lines indicate the average ESR values in each zone.
acknowledges the financial support provided by the Advanced Materials & Manufacturing Institute at Rowan University.The research leading to this article was part of the Engineering Clinic at Rowan University.G.Y. also acknowledges the support from the Department of Defense (W911NF2310150, W911NF2210199, and N00174-23-1-0009).The SEM analyses were carried out in part at the Singh Center for Nanotechnology at the University of Pennsylvania, which is supported by the NSF National Nanotechnology Coordinated Infrastructure Program under grant NNCI-2025608.