Ultrastretchable and wearable conductive multifilament enabled by buckled polypyrrole structure in parallel

Stretchable conductive fibers have attracted much attention due to their potential use in wearable electronics. However, the ultra-high strain insensitive conductivity is hindered by mechanical mismatch in Young’s modulus and failure of stretchable structures under large deformation. This challenge is addressed with a conductive and elastic multifilament made of the polyurethane monofilaments that are surface-coated with buckled polypyrrole (PPy) of which flexibility is improved by sodium sulfosalicylate. Such parallel conductive monofilaments with PPy buckling on surface reduce the influence of cracks in the conductive coating on the overall conductivity, displaying an ultra-high strain insensitive behavior (quality factor Q = 10.9). Remarkably, various complex forms of wearable electronic textiles made by this conductive multifilament maintain the strain-insensitive behavior of the original multifilament, even upon the large deformation of human joint. This multifilament with wrinkled PPy has attractive advantages in the application of super-stretched wearable electronic devices.


INTRODUCTION
In recent years, wearable electronic textiles such as sensors 1,2 , energy harvesting and storage devices 3,4 , memory devices 5 , displays 6 and heaters 7 have attracted an increasing interest from both research and industrial communities. The high-speed growth of wearable electronic textiles places a high demand on flexible and stretchable electronic circuit to ensure the stable and lossless transmission of electrical signals under a large mechanical stretch 8,9 . Various technological methods including helical coil structure 10,11 , "twining spring" structure 12 , layered structure 13 , and buckled structure 8,14 have been developed to fabricate the stretchable electronic circuit toward designing multifunctional fibers and yarns. The most commonly used strategy relies on the geometrical constructions based on coiled spring principles, enabling stretchable electronics of high performance 10,11 . Despite the effectiveness of the method, disadvantages such as the multidimensional structures that often require complex bonding materials and bonding technology, and the non-planar movement of the coils during stretching, instigated the search for other alternative strategies 15 . The conductive fibrous composites with a spirally layered structure formed by rolling the films only provide an unsatisfactory solution for their low strain-insensitive behavior (quality factor Q = 0.57 at 356%) 13 . Therefore, a fiber-based stretchable electronic circuit with easy integration and high strain-insensitive behavior is still extremely desirable for wearable electronic textiles.
Buckling structures have been used to overcome spatial constraints while maintaining a constant conductivity in various flexible devices 16 . The apparent modulus of the conducting elements can be significantly lowered by forming the buckled structures, e.g., curled, wavy and wrinkled, on surface of the elastic substrate, guaranteeing the stretchability of the conductive fiberbased circuit 17 . For example, a worm-shaped conductive multifilament was prepared by firstly pre-stretching a polyurethane (PU) multifilament core and then coating it with a layer of graphene 8 . Such a structure enabled a negligible resistant change (about 0.26) up to 300% deformation. However, an increase in electrical resistance was observed at the initial stage of stretching due to the separation of the graphene wrinkles, a useful property for sensors 18,19 but undesirable for stretchable electronic circuits. Furthermore, although the buckled structure has been proved being able to effectively alleviate the mechanical mismatching between the conducting elements and substrate, a high interfacial interaction is always required to enhance the cycle performance 17 . Thus, it is critical to select the conducting elements for different substrates in order to obtain a superior interfacial interaction.
The conductive elements used in stretchable electronics mainly include metal nanowires 20 , carbon nanomaterials 8 , and conductive polymers 21,22 . Some of the most commonly used coating methods for the first two include spraying and dip-coating. The damage and debonding at the interface during stretching were reported, showing the necessity to further increase the interfacial interactions between the conductive elements and the substrate 17,23 . In contrast, a strong interfacial interaction between conductive polymers and substrates can be achieved by a simple in-situ polymerization method. On the other hand, conducting polymers are more compatible with elastomeric substrates in mechanical properties than with metal nanowires and carbon nanomaterials 17 . Among conductive polymers, polypyrrole (PPy) has attracted a great attention because of its lightweight, low cost, good electrical conductivity, and good biocompatibility 24 . However, the mechanical mismatch in Young's modulus, fracture tensile strain and tensile strength between PPy and an elastomeric substrate still exists even though they are all polymers. Consequently, a flexible (i.e., low Young's modulus) PPy coating is essential to achieve a superior stretchability.
Herein, inspired by the structure of the electrically sensitive muscle fiber bundles, we designed a conductive and elastic fiber bundle made of the PU monofilaments that were surface-coated with buckled PPy via a simple in-situ polymerization method. Such parallel conductive monofilaments with PPy buckling on surface reduced the influence of cracks in the conductive coating on the overall conductivity, displaying an outstanding ultra-high strain insensitive behavior. Note that the flexibility of PPy was improved by adding sodium sulfosalicylate (NaSSA) as the plasticizer. This particular conductive multifilament exhibited a small diameter (0.21 mm), a good initial conductivity (238.0 S m −1 ), and a remarkable strain-insensitive performance (Q = 10.9). Even at 900% tensile strain, it still maintained a stable electrical conductivity (relative resistivity change was only 3.5) and a recovery ability. More importantly, this conductive fiber can be easily integrated into the fabric by weaving, knitting, or embroidery. This integrable, highly stretchable, ultra-fine, and highly conductive PU multifilament is highly competitive for smart wearable textiles.

RESULTS AND DISCUSSION Materials Design and Fabrication
Muscle tissue, the main contributor to a physiological movement, possesses an aligned structure composed of multiple fibers, and it is sensitive to electrical signals (Fig. 1a). Inspired by this structure, we developed the conductive multifilament based on two principles, i.e., strain insensitive conductivity under ultra-high strain and easy integration to wearable devices. For the latter requirement, undoubtedly, finer fibers will offer indubitable possibilities to be integrated into the compact and sophisticated electronic textiles. Similar to muscle fiber bundles, the conductive PU multifilament was made of multiple parallel monofilaments with the PPy buckling on surface, with the detailed fabrication process illustrated in Fig. 1b. PU x y @PPy denotes that the PU multifilament was x denier (D, one denier has a linear density of one gram mass per 9000 m of fiber) and the PPy coating was prepared under the pre-stretching strain y. Prior to in situ deposition of PPy, polydopamine (PDA) was introduced onto the surface of the pre-stretched PU multifilament to improve wettability via a DA coating-enhanced solution polymerization. An improved in situ polymerization method was adopted to obtain a homogeneous PPy coating layer on each monofilament of the pre-stretched PU multifilament 25 . After releasing the stretch, the PPy wrinkles were formed on the surface of each monofilament. It is such a structure that gives the multifilament a stable conductivity under high strain. Hence, the brightness of the LED barely changed between 0% and 900% strain (Fig. 1c). In addition, the PU 420D 200% @PPy multifilament can possess a high conductivity at a small size. Thanks to its fine diameter, the conductive multifilament can pass through a #1 sewing needle, suggesting that it can be fabricated to wearable electronic devices with the existing textile technologies (Fig. 1d).
The limited flexibility of PPy due to its rigid conjugated backbone is the biggest obstacle to its application in stretchable conductors 26 . There were cracks under a small strain on surface of the PU 210D 50% @PPy multifilament when only FeCl 3 was used as the dopant ( Supplementary Fig. 1). In this work, in addition to decreasing the apparent modulus of the PPy layer by structural design, NaSSA was used as a plasticizer to improve the intrinsic flexibility as well. The PPy films with or without NaSSA were synthesized via interfacial polymerization (more details can be found in the Supporting Information) to determine whether the incorporation of NaSSA indeed improved the mechanical property. As shown in Supplementary Fig. 2a-c, the breaking elongation of the PPy films with Cl − /SSA − as dopant (PPy FeCl3=NaSSA ) was 8.58%, which was much higher than that (3.81%) of the PPy film with Cl − as dopant (PPy FeCl3 ). A larger bending angle (51.20°) of the PPy FeCl3 film was further demonstrated by the cantilever method, confirming that the flexibility of the PPy film was enhanced by NaSSA ( Supplementary Fig. 2d). The PPy FeCl3 film exhibited incredible capacity under repeated bending at room temperature (Fig. 2a). More interestingly, this flexibility was retained in liquid nitrogen ( Fig. 2b and Supplementary Video 1). We have previously reported the pristine PPy membranes that are as soft in liquid nitrogen as at room temperature and discussed this interesting phenomenon as well 26,27 . In this work, the intrinsic properties were also believed to be the cause of flexibility of both PPy films and PPy coatings. Furthermore, it is contradictory for conductive polymers to have both high conductivity and high flexibility 28 . Excitingly, NaSSA could improve not only the flexibility of PPy but also the conductivity. The conductivity of PU@PPy increased from 70.8 ± 4.6 S m −1 to 358.2 ± 27.9 S m −1 after NaSSA was employed as dopant. To further explain the role of NaSSA in enhancing the conductivity and flexibility of PPy, the chemical analyses were explored using XPS and Raman spectrum. Sulfur (S) element was observed in the XPS spectrum of the PPy with FeCl 3 /NaSSA as dopant, confirming that NaSSA acted as counter ions (Fig. 2c). And the relative content of -NH + -(26.59%) in the PPy with FeCl 3 /NaSSA as dopant was higher than that (21.55%) in the specimen without doping of NaSSA, indicating that NaSSA could improve the conductivity of PPy by increasing the doping level (Fig. 2d). Raman spectra of the PU@PPy prepared with or without NaSSA as dopant was measured also. The spectrum ( Supplementary Fig. 3) shows the main bands of PPy at 1576 cm −1 (C = C stretching), 1374 and 1327 cm −1 (two bands of ring-stretching), 1238 cm −1 (antisymmetric C-H in plane bending), 1084 and 930 cm −1 (the bipolaron structure), and 1046 and 963 cm −1 (the polaron structure) 25 . The ratios between the bipolaron (1084 and 930 cm −1 ) and polaron (1046 and 963 cm −1 ) absorptions were used to estimate the degree of doping and these ratios are reported in Supplementary  Table 1 29 . Higher bipolaron/polaron ratios of the PU@PPy prepared with NaSSA as dopant was considered as an evidence explaining that the NaSSA dopant enhanced the conductivity of PPy also. Sulfonate and sulfonimide anions have been reported to have a charge screening effect that weakens the interactions between polymers 28 . What is particularly important is that the doping of d-sorbitol could weaken the interactions between polymer chains to improve mechanical flexibility 30 . Therefore, we suspect that the NaSSA has two effects: 1. NaSSA molecules weakened the interactions between PPy chains (Fig. 2e), and so facilitated the movement of the molecular chains leading to a reduced intrinsic modulus of PPy; 2. NaSSA dopants increased the doping level of PPy, resulting in an improved conductivity.
To evaluate the effect of PDA on the buckling structure, the morphology of the PU 210D 300% @PPy multifilament with or without PDA treatment was observed ( Fig. 2f-h and Supplementary  Fig. 4a-c). PPy cracks along the fiber axis were only observed on the PU 210D 300% @PPy multifilament without PDA treatment ( Fig. 2f and Supplementary Fig. 4a), indicating that the use of PDA prevented PPy from cracking. This was because that the PDA coating increased PPy deposition 31 , resulting in a thick PPy coating that was more resistant to the circumferential expansion when the filament was released. Indeed, the thickness of the PPy on PU 210D 300% @PPy was 0.80 ± 0.17 μm, while that without PDA was only 0.54 ± 0.10 μm (Supplementary Fig. 4b and Fig. 2g). Further, the height of the bucklings on the PU@PPy with PDA modification (~1.56 μm) was much larger than that without PDA (~0.25 μm), as shown in Fig. 2h, Supplementary Fig. 5, and Supplementary Fig. 4c. Such differences in shape and size of the buckling structure were also caused by the thickness of the PPy coating (more detailed analysis can be found in Supplementary Method 1). Remarkably, the SEM images of the cross-sections revealed that each monofilament of the PDA modified multifilament was coated with buckled PPy, in contrast to the incomplete PPy coating on the unmodified monofilaments ( Fig. 2g and Supplementary Fig. 4b). This is a fascinating structure that had rarely been observed and reported in literature so far.
To investigate the effect of polymerization time on the shape of the buckling, PPy was polymerized on the PU 210D 300% @PPy for different times. Compared to the smooth surface of the PU multifilament ( Supplementary Fig. 6a), the surface of each parallel  Fig. 6) on the PU 210D 300% @PPy multifilament were obtained when the polymerization time was 24 h, which was considered as the optimal reaction time.

Mechanical properties
To elucidate the effects of the buckling structure on mechanical properties, we evaluated the PU@PPy multifilament made of a different number of monofilaments, the PU multifilament, and the PU 420D y @PPy multifilament prepared under different prestretching strains ( Supplementary Fig. 9). As the number of monofilaments increased, the breaking strength of the PU x 50% @PPy multifilament gradually decreased, while the breaking elongation showed an opposite tendency ( Supplementary  Fig. 9a-c). The former was attributed to the increase in the probability of weak nodes as the number of monofilaments increases, and the latter was related to the increase in the multifilament diameter 33 . The mechanical properties of PU without PPy coating also showed a similar trend, which makes us believe that this phenomenon was caused by the structure of the PU multifilament rather than the PPy coating ( Supplementary Fig.  9d-f). The tensile properties of the PU 420D y @PPy prepared under different pre-stretching strains were also evaluated (Supplementary Fig. 9g-i). Although the elongation of the multifilament gradually decreased with the increase of the pre-stretching strain, all the multifilaments exhibited an outstanding high stretchability (>1000%) and the similar strength at break (~100 MPa), in consistent with previous studies 8 . In addition, the PU 420D 200% @PPy multifilament showed a good recoverability within 100% strain, and a small hysteresis within 200% strain that is comparable to those in the reference (Supplementary Fig. 10a) 8,34 . After 100 loading-unloading cycles under 100% strain, the recovery performance of the multifilament was hardly deteriorated (Supplementary Fig. 10b). The strong strength and highly stretchable properties of the PU 420D 200% @PPy multifilament could guarantee the availability for the manufacture of large strain wearable electronic textiles. 50% @PPy and PU 840D 50% @PPy after breaking. Scale bar is 2 mm. c Relative resistance change of PU x 50% @PPy with different diameters upon increasing strain. SEM images of the cross-section of d PU 420D 200% @PPy and e PU 420D 200% @WPU@PPy. Scale bar is: 100 μm, inset 15 μm. f Change in relative resistance of the conductive multifilament and sheath-core filament under stretching. g Working mechanism of the parallel conductive multifilament under increasing strain. h Current path (yellow arrow) along the multifilament during stretching.
Parallel structure and hypothetic model The strain dependence of relative resistance change (ΔR/R 0 ) of the PU x 50% @PPy with different numbers of parallel monofilament was further evaluated to prove the function of parallel structure (Fig.  3a-c). As shown in Fig. 3c, all the PU x 50% @PPy exhibited a similar strain-insensitive resistance change at low strain (< 50%) due to the analogous buckling-shaped PPy coating, of which the distance between the adjacent buckles was about 5 μm ( Supplementary  Fig. 11a-e). The buckled morphology is an important factor affecting the strain-insensitive behavior at low strain, as reported in previous studies 18,21 . When strain was further increased, the strain dependence of ΔR/R 0 of the PU x 50% @PPy decreased as the number of monofilaments increased. This was because even if the PPy coating on one monofilament broken, the PPy coatings on other monofilaments at the same cross-section might still be intact to allow electricity to flow through freely, thanks to the disappearing free space among monofilaments during stretching to the ultra-high strains. Obviously, this probability increased as the number of monofilaments increased. The electro-response-tostrain of the PU 840D 50% @PPy was however not further reduced because the conductive coating failed to fully penetrate deep into the thick PU multifilament, i.e., 840D, confirmed by the white filaments rather than the black PPy coating inside the PU 840D 50% @PPy and exposed after its broken (Fig. 3b).
To more intuitively demonstrate the strain-insensitive performance of the conductive monofilament in parallel (PU 420D 200% @PPy, the SEM image of its cross-section is presented in Fig. 3d), and the sheath-core filaments with buckling-shaped PPy coating (Fig. 3e), named PU 420D 200% @WPU@PPy, were prepared as a control (See the Supporting Information for details). Figure 3f evidently illustrates that the ΔR/R 0 of the PU 420D 200% @WPU@PPy multifilament was much 200% @PPy with 1D conductor reported in the literature (Ref. 8,12,[35][36][37][38][39][40][41]  In order to further explain the high strain insensibility of the PU@PPy conductor, a model was established to demonstrate how the PPy changed during stretching, as shown in Fig. 3g & h. The stretching process can be generally divided into three stages. In the first stage when the strain ε < pre-stretching strain ε p , the PPy buckling gradually expands and the resistance remains constant resistance R = R 0 , where R 0 is the initial resistance. Different from previous studies 8, 35 , our work didn't show the sharp increase in resistance caused by the reduction in contact area between adjacent buckles during stretching. An ordered buckling structure is showed by the AFM (Fig. 2h) and SEM images ( Supplementary Fig. 11 & 12) of the PU@PPy multifilament. In the second stage where ε > ε p , the PPy coating begins to show a tensile deformation, which is manifested by the decrease in the thickness and the increase in the length of the PPy coating. When Poisson effect is taken into consideration the resistance in this stage is given by Eq. (1).
where γ is the Poisson's ratio of PPy coating, and R εp is the resistance of the PU@PPy at strain ε p . More details can be found in Supplementary Method 2. Under excessive strain, cracks in PPy coating appear, hindering the flow of current. However, because the conductive monofilaments are independent each other, the cracks in one filament do not propagate to the surrounding monofilaments, and the tight junction among monofilaments under the high strain can guarantee the intact conductive path for electrical current. As shown in Supplementary Fig. 13, an equivalent resistance model is used to explain the electrical stability. The conductive multifilament can be regarded as numerous resistors in series along the longitudinal direction, and each resistor can be considered being composed by multiple parallel resistors along the perpendicular or cross-section direction (the number of the parallel resistors is the number of the conductive monofilaments). When cracks occur on some resistors, their conductivity drops dramatically; however, other resistors parallel with them still maintain conductivity. In this stage, the resistance can be calculated by Eq. (2).
where erf(x) is the error function, and q, p, ε 0, and μ are fitting parameters. More details can be found in Supplementary Method 3.

Strain-insensitive conductivity and verification of model
Furthermore, the strain-insensitive properties of the conductive multifilament with different pre-stretching strains were studied (Fig. 4). As shown in Fig. 4a, the ΔR/R 0 of all the PU 420D y @PPy multifilament increased with the increase of tensile strain. And the rate of the increase in ΔR/R 0 declined as the pre-stretching strain heightened, which was related to the buckling shape (Supplementary Fig. 12). The electrical conductivity of the stretchable PU 420D y @PPy was further calculated with the volume of the multifilament assumed constant during stretching, as shown in Fig. 4b. With the increment in axial strain, conductivity of the PU 420D 200% @PPy multifilament increased from 238.0 S m −1 to 5304.0 S m −1 , which was consistent with the reported in the literature 8,30 . This may be caused by the synergistic effect between the expansion of the PPy buckles and the stress-induced alignment of the conductive PPy chains 30 . Remarkably, the PU 420D 200% @PPy exhibited a wide range of electro-response reversibility, no downward trend up to the ultra-high strain of 900%. Such a behavior was ascribed to the parallel structures that have the capacity to provide the additional conductive path under ultra-high strain due to the tight contact among monofilaments. The strain-insensitive performance of the PU 420D 200% @PPy and the PU 420D 300% @PPy was similar (Fig. 4a) because that the tiny free space among monofilaments under a high tensile strain made PPy difficult to penetrate into the PU multifilament (Fig. 4c), which was supported by the cross-section observation of the PU 420D y @PPy ( Supplementary Fig. 14) and by the weight percentage of PPy (Supplementary Fig. 15 and Supplementary Table 2). This trend was common to other PU@PPy of different numbers of parallel conductive monofilaments (Supplementary Fig. 16). Moreover, the initial conductivity of the PU 420D y @PPy multifilament decreased from 344.1 S m −1 to 125.4 S m −1 when pre-stretching strain increased from 0% to 300% (Fig. 4b), which also indicates that the adjacent buckles are independent, i.e., not in contact ( Supplementary Fig. 12) and the reduction of the weight percentage of PPy in the multifilament resulting from the difficulty for PPy to penetrate into the tight PU multifilament under the high pre-stretching strain (Supplementary Fig. 15 and Supplementary Table 2).
In addition, the electrical conductivity of the PU 420D y @PPy with different pre-stretching strains was evaluated by the I-V curve under 5 V on-load voltage ( Supplementary Fig. 17). All the multifilaments exhibited a linear shape, indicating that the multifilament obeys the Ohm's law. And the decrease in the slope of the curves also confirmed that the resistance of the PU 420D y @PPy multifilament increased after the pre-stretching strain treatment.
The quality factor Q = (Δl/l 0 )/(ΔR/R 0 ) was adopted to further evaluate the strain-insensitive behavior of the PU 420D 200% @PPy 14 . Compared with the 1D linear stretchable conductors reported in literature 8,12,[35][36][37][38][39][40][41] , the PU 420D 200% @PPy exhibited a much larger Q factor under ultra-strain and fineness (Fig. 4d & e, detailed data can be found in Supplementary Table 3). As shown in Fig. 4d, the Q factor of the PU 420D 200% @PPy was 10.9 at a 200% elongation, which is among the best compared with the literature. The Q then remained at 2.6 for up to 900%, demonstrating the good straininsensitive behavior of this conductive multifilament under ultrahigh strain. Remarkably, the diameter of the multifilament is only about 0.21 mm that is much smaller than any other reported conductors (Fig. 4e), offering great possibilities for its integration in smart wearable devices.
To test the above hypothetic model of the parallel conductive monofilament under increasing strain, in situ tensile test was performed using the SEM equipped with a Tensile Holder (Fig. 4f and Supplementary Videos 2-4) The theoretical resistance of the PU 420D 200% @PPy during stretching is shown in Fig. 4g, which well fitted the experiment data. The Y. Li et al. difference between theoretical and experimental results in the first stage may be due to the volume change of PPy coating under the pressure between monofilaments along the cross-section direction during the buckle expansion process.
Reversibility is a very important parameter for wearable electronic textiles. Figure 5a-d show the I-V curves of the PU 420D 200% @PPy at different elongations. At 200% elongation, the slope of the I-V curve was only slightly less steep with respect to that at 0% elongation (Fig. 5a), and the LED lamp linked to the filament remained very bright (Fig. 5a inset). When the elongation reached 500%, the filament resistance increased approximately twice, and the LED began to dim (Fig. 5b). Figure 5c shows that even at 900% elongation, the filament resistance could still be used as a conductor to light up the LED lamp. This confirmed that the PU 420D 200% @PPy can be reliably used as a stretchable conductor under very large strains. After recovery, the filament resistance only increased by 15% (Fig. 5d), caused by the broken PPy. Figure 5e presents the periodical change of resistance in 1500 stretch-recovery cycles with strain changed from 0% to 20%. The resistance at 0% strain only increased 3% after 1500 cycles, testifying the remarkable reversibility and stability of the PU 420D 200% @PPy. To further verify the reusability of the multifilament, the conductivity of the multifilament after wash in water was characterized. The entire wash cycle of a washing machine, lasting 34 minutes, includes four steps of warm washing, rinsing, draining and drying 42 . We simulated this process using a method similar to the color fastness washing test (GB/T 3921-2008). For simplicity, the rinsing step was replaced by a more severe warm wash step and the draining step, which did little damage to the fibers, was omitted. During a simulated washing test, therefore, the multifilament underwent a warm washing (800 rpm, 40°C, 20 min) and a drying (60°C, 15 min). As the number of washing increased, the conductivity gradually decreased because of the de-doping process of PPy; 43 however, the conductivity was still 94.6 S m −1 after 16 washes (Supplementary Fig. 18). In human body movement, stretching deformation is often accompanied by bending deformation (Fig. 5f). Therefore, resistance of the filament under both stretching and bending was also studied to better simulate human body movement. We designed a stretching-bending instrument (Fig. 5g) with an LED system to monitor the realtime change in resistance of the filament, and the results are shown in Fig. 5h. At a high bending angle of 70°, the resistance of the PU 420D 200% @PPy changed only 3.5%, while the resistance of the PU 420D 0% @PPy increased by 6441.6%. The PU 420D 200% @PPy showed an unmatchable insensitive conductivity, as evidenced by the hardly noticeable LED brightness variations ( Fig. 5f and Supplementary Video 5).

Wearable Applications
The actual conductivity is another important factor besides the quality factor that affects the application of fiber conductors to wearable circuit system. As for the application in 3D and 2D stretchable conductive materials, our small diameter conductive multifilament can be used to manufacture more compact, 200% @PPy under 20% strain. Inset shows 5 cycles at the beginning and the end of the 1500 cycles, respectively. f Schematics for dual bending (pink) and stretching (purple) deformation for human body movement. g Stretchingbending instrument for the filament. h Real-time resistance change of the filament during a stretching-bending process.
sophisticated electronic textiles with textile technology. Integrating several fibers into a strand by doubling and twisting, or integrating many fibers into a fabric by weaving, knitting, and embroidery is almost an essential step in the textile forming process. This is equivalent to connecting multiple conductors in parallel, resulting in a lower overall resistance. Although paralleling resistors is an effective way to improve circuit conductivity, it inevitably results in an increase in size. Benefiting from the small diameter of conductive multifilament in this work, the strands could still meet the integration requirements of compact, sophisticated electronics after paralleled. Taking the integration of several fibers into a strand by doubling and twisting as an example, assuming that the fiber cross-section is circular and ignoring the extrusion deformation between fibers, according to the closest packing theory, the resistance of the strand obtained by twisting 7 fibers is reduced by a factor of 7, while the diameter is only increased by a factor of three. To test our hypothesis, a strand formed by twisting 7 conductive multifilaments was fabricated. The diameter of the conducting fibers was about 0.65 mm (Supplementary Fig. 19a) and the initial resistance was about one-seventh of that of PU 420D 200% @PPy ( Supplementary Fig.  19b), which is consistent with the theoretical model. Importantly, the fineness of the strands is better than that of the previously reported fiber conductors, and the electrical conductivity is also comparable 8,12,41 . To further validate the potential application in wearable circuits, an LED system with the strands formed by twisting 16 conductive multifilaments as a circuit was designed. The LEDs could be visually observed to be lit at 3 V (Supplementary Fig. 19c), indicating negligible energy loss in the circuit.
Similarly, integrating many fibers into a fabric by weaving, knitting, and embroidery is an effective way to improve circuit conductivity. To demonstrate the feasibility of preparing compact, sophisticated electronic textiles, textile fabrics manufactured with the conductive multifilament by weaving, knitting, and embroidery, respectively (Fig. 6a-d and Supplementary Fig. 20). The rapier loom and circular knitting machines used were shown in Fig. 6a & c. Different fiber paths (loops or straight lines, Fig. 6b & d) and fiber arrangements (Supplementary Fig. 20a and Fig. 6b), different fabric shapes (sheets or tubes, Fig. 6b & d inset) and complex patterns (Supplementary Fig. 20b) can be easily produced through textile technology. The plain weave fabrics with the PU 420D 0% @PPy or PU 420D 200% @PPy as the wefts were used to evaluate the functionality of the conductive multifilament after weaving ( Fig. 6a & b) because weaving with textile forming technologies causes the greatest abrasion on the PPy coating. Supplementary Fig. 21 shows that the initial resistance of the fabric integrated by four PU 420D 200% @PPy is reduced to about a quarter of the resistance of a single fiber. Moreover, changes in resistance (R/R 0 ) during stretching, bursting and twisting were also evaluated to simulate the robust deformations when used in wearable electronic textiles. Figure 6e reveals that the R/R 0 of the fabric integrated with the PU 420D 200% @PPy was only 1.46 under 300% strain, which was much lower than that of the fabric composed of the PU 420D 0% @PPy (170.78), indicating that the fabric inherited the strain-insensitive behavior of the original conductive multifilament. The bursting process, that is, applying a vertical strain to the fabric, cannot be ignored for smart textiles. A device to apply the vertical strain was designed and the change in resistance of the fabric during the bursting process is shown in Fig. 6f. When the rod (diameter: 5 mm) perpendicular to the fabric (20 mm × 20 mm) descended to 17 mm, the resistance of the deformed fabric composed of the PU 420D 200% @PPy only increased by 30% while Fig. 6 Integration of PU 420D 200% @PPy multifilament using textile technology and its potential applications. Optical images of a rapier loom machine, b fabric with conductive multifilament obtained by weaving, c circular knitting machine and d fabric with conductive multifilament obtained by knitting. Scale bar is 0.5 mm. Resistance changes of the fabric during e stretching, f bursting, and g twisting. h Dynamic resistance changes of the fabric under multiple deformations. Application of the fabric as stretchy electronics during i, j arm bending and k, l walking. the resistance of fabric composed of the PU 420D 0% @PPy increased by 1225%. Also, as shown in Fig. 6g, the electrical resistance of the fabric composed of the PU 420D 200% @PPy displayed little dependence on torsion deformation compared to the fabric composed of the PU 420D 0% @PPy. The motion of human body is complex and often results in multiple deformations of the wearables. Therefore, we designed a device that can simultaneously generate multiple deformations such as stretching, bending and twisting to simulate the complex deformations during real human body motion (Supplementary Video 6). The resistance of the fabric under test exhibited periodic changes as a function of the periodic loading and unloading of the complex deformations (Fig. 6h). Under different frequencies and deformations, the resistance of the fabric composed of the PU 420D 200% @PPy revealed little change compared to the fabric composed of the PU 420D 0% @PPy. These results illustrate that the fabric composed of the PU 420D 200% @PPy can meet the challenge of the complex movement of human body.
Furthermore, we integrated the conductive multifilament fibers into commercial elbow and knee pads through embroidery, to verify if the electrical stability is reliable under large deformations in real human motions. The real-time changes of R/R 0 of the elbow pad ( Fig. 6i) equipped with the PU 420D 0% @PPy or the PU 420D 200% @PPy are shown in Fig. 6j. The change of the PU 420D 200% @PPy pad was negligible while that of the PU 420D 0% @PPy pad showed a good linearity and repeatability under the periodic bending of the elbow. There was almost no change in the brightness of the LED light in Fig. 6j inset, confirming the strain-insensitive conductivity of the elbow pads equipped with the PU 420D 200% @PPy. Compared with the elbow, the knee underwent a greater strain during exercise, resulting in obvious changes in resistance (Fig. 6k). Despite this, the knee pads equipped with the PU 420D 200% @PPy still showed an outstanding insensitive behavior; and the LED lights also showed constantly visible brightness during exercise ( Fig. 6l and Supplementary Video 7). Therefore, the conductive PU multifilament with the PPy buckling structure in parallel is a strong candidate to build electronic circuits in stretchable and wearable electronic fabrics.
In summary, we designed and produced a conductive PU multifilament with the PPy buckling structure in parallel by a prestretching and in situ polymerization method. The pretreatment of the PU multifilament with PDA improved the wettability so that PPy could be polymerized uniformly on surface of the monofilaments. Releasing the pre-stretch after polymerization generated a parallel buckling PPy structure. Besides, the addition of NaSSA into PPy can improve the stretchability. The highly stretchable PU 420D 200% @PPy multifilament exhibited a high initial conductivity (238.0 S m −1 ), a small diameter (0.21 mm), a high Q value (10.9) at 200% strain, and a high reversibility in electrical response in a wide range of deformation up to 900% strain (Q = 2.6). Moreover, the multifilament can be easily integrated into smart textiles through the existing textile technologies and maintain its high straininsensitive behavior during the robust human body movement. This conductive multifilament therefore shows a great potential in wearable electrical devices.

Fabrication of Conductive PU Multifilament with Buckling Structure
About one meter of the PU multifilament under stretch (0%, 50%, 100%, 200%, or 300%) was immersed in a DA solution (2 mg mL −1 , 50 mM Tris-HCl buffer, and pH 8.5) for 5 min in ultrasonic and then reacted for 8 h. After washing, it was placed in a dish (diameter: 90 mm) containing of a FeCl 3 ·6H 2 O (0.36 M) and NaSSA (0.36 M) water solution (20 mL), and then frozen at −20°C. After, cyclohexane (19.5 mL) containing of Py (0.5 mL) was transferred onto the top of the above solution. The polymerization in situ was maintained for different times (6 h, 12 h, 18 h, 24 h, or 36 h) at 2°C. At the end of polymerization, the PU@PPy multifilament was washed several times with anhydrous ethanol and then with deionized water, followed by ultrasonic washing in water for 5 min. Finally, the strain of the PU multifilament was released to form the PU@PPy with a buckling structure. The filament was dried at room temperature.

Fabrication of Conductive Fabric
Woven fabrics with PET filaments as warp, PU and PU 420D y @PPy as weft were fabricated by a rapier loom (ASL2300, Suzhou Shuohai Electromechanical Technology Co., Ltd., Suzhou, China). Knitted fabrics made of PU and PU 420D y @PPy filaments were fabricated by a circular knitting machine (Biomedical Textiles Lab, Donghua University, Shanghai, China). Embroidery fabric is realized by hand sewing the PU 420D y @PPy filament onto a knitted fabric.

Characterizations
The morphology of the samples was observed by scanning electron microscopy (SEM, Hitachi SU8010, Tokyo, Japan) at an accelerating voltage of 5 kV. Before observation, the specimens were sputter-coated with gold (Hitachi Ion Sputter MC1000, Tokyo, Japan). The cross-section of the multifilament was obtained by breaking it in liquid nitrogen. An atomic force microscope (AFM, 5500AFM-SPM, Agilent, California, USA) was further used to analyze the multifilaments by immobilizing the specimen on a glass slide with double-sided tape. In situ tensile testing was performed using the Phenom XL equipped with a tensile holder (Phenom XL, Phenom-Scientific, Shanghai, China).
The macroscopic appearance of the fabric was obtained by a stereoscopic microscope (Nikon SMZ745T, Tokyo, Japan). A Fourier transform near-infrared spectrometer (FT-NIR, Antaris II, Thermofisher, Waltham, USA) in attenuated total reflection (ATR) mode was used to characterize the chemical composition of the samples. The surface chemical composition of samples was investigated by an X-ray photoelectron spectrometer (XPS, Escalab250xi, Thermofisher, Waltham, USA) with Al Kα radiation (λ = 8.34 Å) as the excitation source. Raman spectra were collected with a Raman micro spectrometer (inVia-Reflex, Renishaw, Gloucestershire, UK) equipped with a He-Ne + laser emitting at 633 nm.
The mechanical properties of the PPy films with a width of 5 mm were tested using a Single Fiber Strength Tester (YG-001, Wenzhou Fangyuan Instrument Co., Ltd., Wenzhou, China) at 10 mm gauge length and 10 mm min −1 crosshead speed. According to the cantilever beam theory, 20 mm of a 30 mm long PPy film was extended out of the edge of a table to form a cantilever beam. The angle formed between the suspended film and the table surface was measured to evaluate the flexibility of the PPy films. The tensile properties of the multifilament were investigated using a single fiber strength tester (LLY-06E, Laizhou Electron Instrument Co., Ltd., Laizhou, China) at 5 mm gauge length and a strain rate of 50 mm min −1 . Rough silicone sheets were fixed at both ends of the filaments to prevent sliding during the test. The cyclic stretching of the PU 420D y @PPy multifilament was tested by a Spandex elastic tester (XN-1A, Shanghai New Fiber Instrument Co., Ltd., Shanghai, China) at a distance of 50 mm and a speed of 100 mm min −1 . The diameter of the filaments was measured with a latex thickness meter (CH-12.7-BTSX, Shanghai Liuling Instrument Factory, Shanghai, China). All tests were performed under standard environmental conditions (20 ± 1°C, RH 65% ± 2%), and at least three specimens were tested for each sample.
For thermostability test, approximately 3 mg of samples was heated from 50 to 800°C at a heating rate of 20°C min −1 under a nitrogen atmosphere using a thermogravimetric analyzer (TGA8000, Thermofisher, Waltham, USA).
To characterize the washability of the PU 420D y @PPy multifilament, samples of 15 cm in length were stirred in water (40°C) with a speed of 800 r min −1 for 20 min and then dried at 60°C.
To monitor the change in resistance during deformation, the two ends of the conductive filament or the conductive fabric were connected to a conductive copper tape with the help of a rough silicone sheet. Then the conductive copper tape was connected to a source meter unit (Keithley 2450, Tektronix, Cleveland, USA) to monitor the change of the resistance during the deformation. This work is not about living individuals and does not include collection of individually identifiable private information, therefore, the Institutional Review Board approval was not a prerequisite. And we have obtained informed consent from all participants.