Highly Stretchable, Self-Healable Elastomers from Hydrogen-Bonded Interpolymer Complex (HIPC) and Their Use as Sensitive, Stable Electric Skin.

There is a growing interest in developing stretchable strain sensors to quantify the large mechanical deformation and strain associated with the activities for a wide range of species. Herein, we constructed elastomeric, healable hydrogen-bonded interpolymer complex (HIPC) rubberlike film by complexation of hydrogen-bond (H-bond)-donating poly(acrylic acid) (PAA) and H-bond-accepting poly(ethylene oxide) (PEO) (or poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) (F108)). All HIPC elastomers prepared from varied PAA/PEO (or PAA/F108) ratios are healable elastomers with high extensibility (with the highest strain of 1400%). Recovery of all films can automatically occur or be accelerated by externally added water droplet. The stress- and strain healing efficiencies (ησ and ηε) of the water-assisting healed PAA/F108 blends are as high as 99%. Furthermore, stretchable and healable conductor films were fabricated from silver nanowire-printed (Ag-p) and the single-walled carbon nanotube-blended (SW-b) conductor films, respectively. The healable Ag-p conductor film is an ultrasensitive strain sensor, exhibiting large electric resistance variation when stretched. In contrast, the healable SW-b film is an ultrastable strain sensor with reversible resistance strain response over 200 stretching release cycles within a high strain range of 500%. Therefore, this study provides a new and flexible HIPC strategy for the fabrication of stretchable, ultrasensitive, and stable self-healing electrode materials.

Introduction

Stretchable, wearable, flexible, and human-friendly soft electronic devices have boosted the development of various smart devices,[1,2] such as sensors,[3] light-emitting diodes (LEDs),[4] lithium-ion batteries,[5] and supercapacitors.[6] Stretchable electrodes with high toughness play a crucial role in integrating various devices together to support functions under large mechanical deformations.[7] During the past few years, several representative strain sensors using carbon nanotube (CNT),[8,9] metal/semiconductor,[10,11] graphene,[12,13] and conductive polymer[14] as conductive components combining with elastomeric substrates have been successfully fabricated. Among them, silver nanowires (AgNWs) were frequently used as conductor for flexible electronics, particularly for high-performance strain sensors[15] with great electrical and mechanical properties. In general, AgNWs were deposited on large-area flexible and stretchable substrates using solution-coating techniques such as printing, spin coating, dip coating, etc. However, the AgNW-deposited electrodes[16] generally lack self-healing ability due to the poor adhesion between AgNWs and substrates. Moreover, CNT were often applied in the strain sensor as conductor, in which the main sensing mechanism is the change in resistance at the CNT/CNT junctions when the network is stretched. In general, CNTs were deposited on flexible substrate as a separate layer, resulting in electrodes that can only be subjected to small perturbation (a strain below 5%).[17−19] Failure of the electrodes is also attributed to the poor interfacial adhesion between CNT and substrate. Overall, the AgNW- and CNT-deposited electrodes have several limitations with regard to the substrates, such as poor interfacial adhesion, limitation of elongation, and lack of self-healing ability. Therefore, it is of interest to address these limitations in stretchable conductors. Especially for stretchable and self-healable composite electrodes usually prepared by mixing a stretchable self-healing polymer matrix and a conductive material,[20−22] fabricating skin-inspired elastomer conductors that are stretchable and self-healable with good interfacial adhesion and with matched modulus between electrode and human tissue is highly challenging.[23]

Self-healing materials have attracted increasing attention recently. In general, the automatic self-healing process may involve an irreversible or a reversible mechanism. The irreversible system has the common inherent drawback that it can heal only once. Therefore, more recent research works have been concentrated on reversible system that can repeatedly break and heal upon external stimuli.[24] Reversible system can be based on either covalent interaction, such as Diels–Alder and retro-Diels–Alder,[25,26] or noncovalent interactions, such as H-bonding,[27−29] ionic bonding,[30] metal coordination,[31] or π–π stacking.[32] Because of its flexible and dynamic nature, H-bond is one of the most well-known physical interaction forces applied for the preparation of reversible self-healing materials. Several previous examples illustrated the use of H-bond interactions in constructing self-healing materials. Poly(vinyl alcohol)[33] (PVA) hydrogel, from the freezing/thawing method, can self-heal at room temperature without any stimulus (e.g., heating). However, the soft PVA hydrogel can only be stretched to 400% before fracture. A facile one-step approach was used to construct self-healing antifogging films based on partly cross-linked PVA and poly(acrylic acid) (PAA). Abundant free hydroxyl groups in the cross-linked films contribute to the antifogging property.[34] However, the inferior mechanical properties of the cross-linked films limited their use as stretchable electric components. To sum up, hydrogels showed smart intrinsic self-healing characteristics toward wounds caused by external forces, which is attributed to sufficient free groups at the scratched interfaces to reform H-bonds across the interfaces and a sufficient chain mobility that is indispensable for chain diffusion across the interfaces and association to reform H-bonds. Therefore, we attempted to use two polymers with the respective H-bond-donating and -accepting groups to generate hydrogen-bonded interpolymer complex (HIPC)[35] elastomers, which was traditionally used in the solution state as drug carriers,[36,37] with self-healing property. The stretchable, self-healable HIPC elastomers can be further used as the flexible substrate of an electric sensor.

Herein, we reported the design and preparation of a series of HIPC elastomers, wherein PAA was mixed with different amounts of poly(ethylene oxide) (PEO) or F108 (Scheme ) to prepare stretchable and healable elastomers as flexible substrate for the AgNW and single-walled carbon nanotube (SWCNT) conductors. The constituent and composition of the respective PAA/PEO and PAA/F108 systems determine the mechanical and healing properties of the elastomers, which will be primarily discussed. On the basis of the high extensibility (1400%) and superior healing property (both ησ and ηε reach 99%), PAA/F108 1/9 was chosen as the flexible component to prepare the respective AgNW-printed (Ag-p) and SWCNT-blended (SW-b) electric sensors for study. In the presence of hard AgNWs (or SWCNT) component, the Ag-p (or SW-b) sensor is nevertheless a self-healable elastomer capable of detecting bending and stretching motions and exhibits a high stretchability (>1000%). Moreover, the Ag-p and SW-b conductor films were constructed in different configurations, that is, AgNWs of Ag-p film are deposited as a layer separate from the polymer substrate, whereas SWCNTs of SW-b film are homogenously mixed with the polymer substrate. Configurations of the conductor films thus determined their role as a stable and reliable strain sensor. Recovery of the conductivity of the strained Ag-p after being released to the relaxed state can only be attained within small strain (<50%) range. In contrast, recovery of the electrical property of the relaxed SW-b sensor can be achieved even after large stretching deformation (500%). Overall, the present methodology developed paves the way for practical applications of highly stretchable self-healing strain electronic devices.

Scheme 1

Preparations and the Inter- and Self-Associated H-Bonding of the Self-Healing HIPC Blends of PAA/PEO x/y and PAA/F108 x/y and Configuration Difference between Ag-p and SW-b Conductor Films

Results and Discussion

Qualitative Evaluation of the Self-Healing HIPC Blends

PAA/PEO x/y and PAA/F108 x/y (x/y refers to the molar ratio between H-bond-donating carboxylic acid and H-bond-accepting oxygen atom of ethylene oxide unit) blends were prepared and their self-healing behavior can qualitatively be evaluated from the three experiments illustrated in Scheme .

Scheme 2

Qualitative Evaluations of Self-Healing Properties for PAA/PEO 1/9 and PAA/F108 1/9 Films

Under conditions of automatic recovery (top left), water-assisting recovery (top right), and extension of the healed specimen (bottom).

First, rubberlike films of PAA/PEO 1/9 and PAA/F108 1/9 were cut by a knife (top left) and the cut marks of the rejoined PAA/PEO 1/9 and PAA/F108 1/9 films became invisible at 90 and 60 min after the cuts, respectively. The inherent soft poly(propylene oxide) (PPO) block of F108 is responsible for the superior healing efficiency of PAA/F108 1/9, which will be discussed later.

Second, the externally added water droplet was found to accelerate the healing process. As illustrated in Scheme (top right), the wounds covered by water droplet completely healed after the vaporization of the water droplet in few seconds. In this case, the water molecules act to mobilize the hydrophilic PEO and PAA chains in the fracture interphase to close vicinities to have intimate H-bond interactions, therefore healing the fracture interphase with accelerated rate. A similar experimental result was found in H-bonded films[38] derived from complexation of PEO and tannic acid. Instead of adding water droplets, the fractured PEO/tannic acid sample was exposed to humid environment to accelerate the healing.

The HIPC blends are all with high healing efficiency, which can be primarily evaluated from the third experiments (bottom). The fractured elastomeric films were cut, rejoined, and left untouched for 30 min to generate healed specimen for test. The results suggest that the healed interphases should be very strong because they can sustain an extension for at least 10 and 14 times the initial lengths of the PAA/PEO 1/9 and PAA/F108 1/9 films (Movies S1 and S2), respectively. Besides the healing efficiency, PAA/F108 1/9 is also a better elastomer with higher fracture strain compared to PAA/PEO 1/9.

Fourier Transform Infrared (FTIR) Analysis

As illustrated in Scheme , the possible H-bond interactions in the elastomer complexes can be classified into two modes, self-association between two carboxylic acids of PAA and interassociation between carboxylic acid of PAA and oxygen atoms of PEO (or F108). In this respect, FTIR analysis (Figure ) is a powerful tool for analyzing the possible H-bond interactions involved in the blends. As reported previously, conversion of self-associated carboxylic acids[39−42] into interassociated carboxylic acids gradually shifted the carbonyl stretching band to longer wavenumber. Therefore, the increase of PEO (or F108) content in the blends shifts the carbonyl stretching band of PEO/PAA (or PEO/F108) from 1718 to 1730 cm–1 (or 1732 cm–1). The carbonyl stretching band in the corresponding spectral region can be deconvoluted into self- and interassociated bands. The result (Table S1) coincides with the trend that fraction of interassociated carbonyl absorption increases with the increase of PEO (or F108) content in the blend. Enhanced interassociated carbonyl absorption refers to the prevalent interpolymer interactions of the HIPC blends.

Figure 1

FTIR spectra of (a) PAA/PEO x/y and (b) PAA/F108 x/y blends (magnified carbonyl stretching bands were shown in the bottom panels).

FTIR analysis also provides information regarding the major morphological change of the blended elastomers. Semicrystalline PEO[43] exhibits two characteristic crystalline peaks at 1343 and 1360 cm–1 (Figure ), respectively. In contrast, no such crystalline peaks are present in the spectra of the blends, which contain only broad band, representative of the amorphous PEO segments in the elastomers, over the corresponding absorption region. In the blends, PEO segments must interassociate with PAA chains intimately, leading to the destruction of the regular crystalline packing (Figure S4). As a matter of fact, all HIPC blends are amorphous according to their transparent appearances (Figure S5) and the differential scanning calorimetry (DSC) results provided below.

Figure 2

Transformations of the characteristic sharp crystalline peaks of PEO segments chains to the broad, amorphous peaks in the blends of (a) PAA/PEO x/y and (b) PAA/F108 x/y blends.

Differential Scanning Calorimetry (DSC) Analysis

DSC analysis provides information regarding the amorphous nature of the blends (Figure ). Before DSC scans, all blended samples needed to be vacuum-dried to remove the absorbed moisture. No thermal transition can be detected for all blended samples without preliminary drying. As being semicrystalline, PEO and F108 exhibit the characteristic melting transitions at 64 and 60 °C, respectively. The crystalline structure of PEO (or F108) in the blends was nevertheless destructed since all blends exhibit only amorphous glass transition (Tg) in their DSC thermograms. As consistent with FTIR analysis, preferable interassociated H-bond interactions between PAA and PEO (or F108) resulted in miscible, amorphous blends with single, detected Tg. For pure PAA, its characteristic Tg at 114 °C is no more preserved in the HIPC blends. The preferable interassociation in the blends results in the gradual shift of the Tg to lower temperature when the PEO (or F108) content in the blends increase. Maximum loads of PEO and F108 in PAA/PEO 1/9 and PAA/F108 1/9 result in the large shifts of Tgs to 51 and 42 °C, respectively. The resolved Tgs of PAA/F108 x/y are all lower than those of PAA/PEO x/y with the same x/y ratio, which suggests that PAA/F108 blends are softer elastomers with higher chain mobility compared to PAA/PEO blends.

Figure 3

DSC curves of (a) pure PEO, PAA, and PAA/PEO x/y and (b) pure F108, PAA, and PAA/F108 x/y blends (heating rate = 10 °C min–1).

The HIPC blends discussed in this study belong to a class of “physical cross-linked network”, within which the interassociated H-bonded PEO segments act as the junction points and the rest of the mobile PEO segments are the bridge units. The increase of PEO content in the blends increase the fraction of mobile bridge units, which leads to the lowering of the detected Tg value. For PAA/F108 system, the PPO segments of F108 are immune from the restriction of H-bonding and are more mobile units compared to the PEO segments. Thus, incorporation of PPO chains in PAA/F108 results in soft blends with lower Tg compared to PAA/PEO. Soft chains with high mobility are beneficial for the diffusion of the ruptured segments to near vicinities, redistributing and activating the interassociated H-bonds to heal the cut or fractured interphases. With this prospect, PAA/F108 blends should represent a better healing system compared to PAA/PEO blends, a point that can be verified by the following mechanical analyses.

Dynamic Mechanical Analysis (DMA)

The mechanical elastic properties of the rubberlike films were primarily analyzed by DMA. The storage modulus (E′) of the PAA/PEO and PAA/F108 films was given as a function of frequency (0–5 Hz) in Figure a,b, respectively. The results from both systems suggest that we can tune the film property by adjusting the content of the soft PEO (or F108) chain in the blends. A previous study[44] indicated that stiffness and strength of triblock copolymers decrease with increase of the soft block content. Similarly, softer blends with higher soft PEO (or F108) content should exhibit lower E′. Comparatively, PAA/F108 x/y films all exhibit lower E′ than PAA/PEO x/y films of the same x/y ratio; therefore, PAA/F108 represent a softer system compared to PAA/PEO. All of the results are correlated with the DSC analysis. Again, soft chains in the blends favor the segmental diffusion and facilitate the healing of the cut and fracture surface. The mobile PPO chain of F108 makes PAA/F108 a better healing system compared to PAA/PEO, which can be further verified by rheology analysis.

Figure 4

Storage modulus (E′) of (a) PAA/PEO x/y and (b) PAA/F108 x/y blends from DMA.

Rheology Analysis

In general, rheological measurement provides variation of the time scale involved in the dynamic relaxation process of the polymer network. Rheology analysis of the present systems should help evaluate the healing process of the cross-linked HIPC blends. The resultant elastic (G′) and viscous (G″) moduli of PAA/PEO and PAA/F108, at frequency ranging from 0.05 to 126 rad s–1, are given in Figure a,b, respectively. The moduli G′ and G″ of PAA/F108 x/y (x/y = 1/1 and 1/9) blends are all lower than those of PAA/PEO blends of the same x/y ratio, which is correlated with the DMA result that PAA/F108 is a softer system compared to PAA/PEO system. Moreover, blends with higher PEO (or F108) content are also softer than blends with lower PEO (or F108) content.

Figure 5

Rheological behavior of (a) PAA/PEO x/y and (b) PAA/F108 x/y blends (x/y = 1/1 and 1/9).

The resolved G′ and G″ curves of the rubberlike films all follow similar paths. At low frequency, G″ is greater than G′, which refers to liquidlike character, while at higher frequency, G′ exceeds G″, which corresponds to gel-like behavior. For all tested films, transformation from a low-frequency liquidlike to a high-frequency gel-like states occurs at the point when G″ intersected G′ (inset, Figure ). The moduli of PAA/PEO 1/9 intersected at frequency ωc ∼ 8.83 rad s–1, corresponding to a relaxation time of 1/ωc ∼ 0.113 s, and PAA/F108 1/9 intersected at a frequency ωc ∼ 10.18 rad s–1, corresponding to a fast relaxation time of 1/ωc ∼ 0.1 s. After the intersecting point, both curves flattened at higher frequencies and entered into the rubbery plateau region. Compared to blends with 1/9 ratio, PAA/PEO 1/1 and PAA/F108 1/1 blends all relaxed with longer relaxation times of 0.44 and 0.16 s, respectively. A previous study on a self-healing ionic network system[45] concluded that materials with short relaxation time are better in chain mobility and are therefore superior in healing efficiency than materials with long relaxation time. In accordance with the above results, PAA/F108 blends represent better healing system compared to PAA/PEO blends. HIPC blends with higher PEO (or F108) content are also superior in healing property to blends with lower PEO (or F108) content.

Figure shows the robust recovery behavior of PAA/PEO 1/9 and PAA/F108 films by rheological measurements. The rubberlike films were subjected to oscillatory force with alternatively changing amplitudes. At a low strain of 1%, G′ is greater than G″, and both moduli do not change over time, which implies that both films remained intact under small oscillatory strain. When the applied oscillatory strain is high at 200%, G′ drops rapidly and becomes lower than G″ due to the breakage of physical network. When the applied strain was returned to 1%, the sample immediately recovered to the initial gel-like state (i.e., G′ > G″) with the same G′ and G″ values as the first experimental run. Therefore, the disrupted network films were completely recovered, which confirmed the healing ability of PAA/PEO 1/9 and PAA/F108 1/9. Disruption and recovery of the gel properties under different oscillatory shears can be repeatedly conducted for at least three times without causing any change in the film’s oscillatory property. In contrast, the G′ and G″ values of PAA/PEO 1/1 and PAA/F108 1/1 deviated from the first experimental runs (Figure S2) under repeated oscillatory strain. Films with higher PEO (or F108) content are therefore better in recovery property than films with lower PEO (or F108) content.

Figure 6

Storage modulus (G′) and loss modulus (G″) of (a) PAA/PEO 1/9 and (b) PAA/F108 1/9 films under a continuous strain sweep with alternating oscillation strains of 1 and 200%, respectively.

Tensile Test

To further evaluate the self-healing ability of the blend films, tensile tests were performed in both the virgin (uncut) and healed PAA/PEO 1/9 (Figure a) and PAA/F108 1/9 (Figure b) films, respectively. The virgin PAA/PEO 1/9 and PAA/F108 1/9 films are all good elastomers with a fracture strain (εf) of 1100% at a fracture stress (σf) of ∼0.45 MPa and with εf of 1400% at a σf of ∼0.15 MPa, respectively. To the best of our knowledge, εf (1400%) of PAA/F108 1/9 is the highest strain value reported so far for the self-healing polymeric materials.[46−48] Benefiting from the reversible H-bond interactions, the healed PAA/PEO 1/9 and PAA/F108 1/9 films are also good elastomers with a common feature that both σf and εf values increased with healing time. If the automatic healing process was allowed to proceed for a long period, i.e., 24 h, the healed PAA/PEO 1/9 and PAA/F108 1/9 actually approached the initial, virgin states according to the resolved values of εf = 1000% and σf = 0.39 MPa for the healed PAA/PEO 1/9, and εf = 1309% and σf = 0.15 MPa for the healed PAA/F108 1/9.

Figure 7

Stress–strain curves of virgin and self-healing films of (a) PAA/PEO 1/9 and (b) PAA/F108 1/9; stress healing efficiencies of (c) PAA/PEO x/y and (d) PAA/F108 x/y; and strain healing efficiencies of (e) PAA/PEO x/y and (f) PAA/F108 x/y at various healing times.

Performances of the healed films can be easily evaluated from the healing efficiencies ηε and ησ of the healed PAA/PEO x/y (Figure c,d) and PAA/F108 x/y (Figure e,f), respectively. In general, a long healing time is beneficial for the healing efficiencies and films with higher PEO (or F108) content have higher healing efficiencies compared to films with lower PEO (or F108) content. Moreover, PAA/F108 x/y is superior in healing efficiency, with the resolved ηε and ησ values higher than those of PAA/PEO x/y. The calculated ησ and ηε are 86 and 90% for PAA/PEO 1/9, and 95 and 94% for PAA/F108, respectively. All of these results are correlated with the comment that softer (or more flexible) chains are beneficial for segmental diffusion of the H-bonding groups and therefore facilitate the recovery of the healed films.

It is also interesting to examine the effect of water droplet regarding the accelerated water-assisting healing process illustrated in Scheme . Primarily, water droplet was dripped on the cut surfaces of the sample film and the surfaces were rejoined and left untouched for different times before the tensile analysis. Representative stress–strain curves from the water-assisting healed films of PAA/PEO 1/9 (Figure a) and PAA/F108 1/9 (Figure b) illustrate the beneficial role of healing time for the tensile property of healed films. The calculated healing efficiencies from the water-assisting healing process were compared to those from automatic healing in Figure . It is then clear that water droplets are beneficial for the recovery of the damaged films. Healing of the blends were accelerated by water droplet according to the resolved ησ and ηε: 88 and 98% for PAA/PEO 1/9, and 99 and 99% for PAA/F108, respectively.

Figure 8

Stress–strain curves of virgin and water-assisting healed films of (a) PAA/PEO 1/9 and (b) PAA/F108 1/9; stress healing efficiencies of virgin and water-assisting healed films of (c) PAA/PEO 1/9 and (d) PAA/F108 1/9; and strain healing efficiencies of virgin and water-assisting healed films of (e) PAA/PEO 1/9 and (f) PAA/F108 1/9 at various healing times.

Conductivity Films

With the superior mechanical properties, PAA/F108 1/9 was selected as the base material to construct the healable and stretchable conductor sensor. Primarily, AgNWs was used as the conductor to build up Ag-p sensor film. Scanning electron microscopy (SEM) image of the as-printed AgNWs (Figure a) was found to be randomly stacked and distributed on the surface of PAA/F108 1/9 film, forming a network structure whose junctions led to the conductivity of the film. Here, besides elongated nanowires in straight geometry, we also observed some curly AgNWs in the network structure. The adhesion between the deposited AgNWs layer and the rubberlike film should be enough to sustain the bending motion (Figure b) for more than 100 times without any visual delamination of the deposited AgNWs layer. The flexible Ag-p conductor film was also applied as a self-healing conductor in a circuit (Figure c).[49,50] The flexible conductor film is conductive enough to transmit electricity in a circuit to an LED. The LED immediately went off after the conductor film was cut into two pieces. However, by bringing the two separated pieces into contact, the conductor film was recovered to result in the light-up of LED. Moreover, the conductance of the flexible Ag-p film could be maintained when the sample was bent (Figure d). The above result suggests that this flexible Ag-p conductor can be applied as a wearable sensor capable of detecting bending motion.

Figure 9

(a) SEM image of the as-printed AgNWs on the top of PAA/F108 1/9 elastomer; (b) flexible Ag-p conductor film with bending, and a series of photos demonstrating that the conductance of the healed Ag-p films (c) after being cut and (d) after being bent can be well restored to result in the light-up of integrated LED again.

The Ag-p conductor film was attached onto a glove (Figure a) for the detection of finger joint bending. When the finger bent, the resistance gradually increases from ∼30 Ω to a maximum of ∼70 Ω. The reversible change of the resistance from ∼30 to ∼70 Ω, corresponding to the bending or releasing of the finger, is clearly visible over the cycles. The electric resistance ratio (R/R0) (Figure b) of the Ag-p conductor film is rather constant about 100% over the 100 bending–releasing cycles. Therefore, the Ag-p sensor can be employed for accurate motion detection of finger joints and could be used as electric skin for robots.

Figure 10

(a) Resistance variation of the bent Ag-p conductor film over different bending times. (b) Electric resistance ratio (R/R0) and resistance changes of Ag-p conductor films with strains (c) 800% and (d) 50%.

The possibility of using Ag-p conductor film as a self-healing strain sensor was then explored. The Ag-p conductor film can be stretched up to ∼800% without visual fracture of the film. Regarding the good conductivity of AgNWs, the initial resistance of the conductor film is low (60–70 Ω, Figure c). Afterward, the resistance of the film exhibited a sudden jump at a strain of 100% and then increased monotonously to attain a final value of 1.68 × 107 Ω at a strain of 800%. For Ag-p, applying a strain beyond 100% caused the failure of resistance recovery. The Ag-p film nevertheless exhibits reliable response to small strain (Figure d), that is, a small strain of 50% contributes to little change of the conductive path and therefore, the resistance of the restored film is the same with the initial value. Displacement of AgNWs and decrease of conductive junction points[51] were suggested to be responsible for the failure of AgNW-printed polyurethane at a strain <100%. However, besides these two possible causes present in the AgNWs layer, the failure mode in the interfacial zone should not be overlooked. As the strain increases, certain AgNWs in the interfacial zone would detach from the polymer substrate, resulting in loss of some conductive paths and the increase of resistance. Extra interfacial zone therefore adds potential failure mode to the Ag-p conductor film.

In contrast to the separate AgNWs layer in Ag-p, the SWCNT conductors of SW-b were directly blended with the HIPC elastomers, forming a miscible system with high sustainability to large strain deformation. The SEM image taken from the vacuum-dried SW-b film is shown in Figure a, which exhibits the main feature that the polymer-coated SWCNTs were interconnected with each other and were surrounded by the polymer matrix. As the polymer-coated SWCNTs are intimately and homogeneously connected with the elastomeric polymer matrix, the SW-b conductor film is therefore highly stretchable in sustaining high strain. A large strain of 1000% caused a little deviation of the resistance value when the stretched film was relaxed back to the initial length (Figure b). With a lower applied strain of 500%, the relaxed SW-b film exhibited the complete recovery of the resistance to the initial value (Figure b, inset) before stretch. Therefore, SW-b is a more reliable stain sensor compared to Ag-p. The healable SW-b film is a ultrastable strain sensor as it exhibited a reversible resistance strain response over 200 stretching release cycles within a high strain range of 500%. Moreover, the electric resistance ratio (R/R0) (Figure c) of the SW-b conductor film is quite stable under the bending motion, which remains about 100% over the 150 bending release cycles. The R/R0 values of the sensor remain almost constant with minor fluctuations within 10% in the first 200 cycles strain test (Figure d) (between 0 and 500% strain). We thereby suggest that the SW-b conductor film is an ultrastable sensor under bending motion and high strain.

Figure 11

(a) SEM image taken from the vacuum-dried SW-b film; (b) resistance recovery of SW-b conductor films with strains of 500 and 1000%; (c) electric resistance ratio of the bent SW-b film over different bending times; and (d) electric resistance ratio of the bent SW-b film over different stretching times from 0 to 500% strain.

The key parameters of several previous strain sensors,[52−59] including healable and nonhealable sensors, are listed in Table to compare with the results from the present Ag-p and SW-b sensors. In this work, the healable SW-b strain sensor that we fabricated can sustain a large strain (>1000%), which made it suitable for broad applications and exhibits a good balance between high sensitivity and stability. Among the healable sensors,[52−54] the SWCNT/PVA/borax[52] sensor is a superior strain sensor with a large ΔR/R0 ratio of 1514% in a large strain range of 1000%. However, this superior strain sensor required the use of adhesion tapes as protective layers to prevent quick vaporization of the absorbed water from the PVA/borax hydrogels. The rest of the sensors[55−59] containing carbon-based conductive fibers is either limited in stretching ability or nonhealable. Without the use of extra adhesion tapes, our SW-b conductor film is healable, highly stretchable, and exhibits stable electric response toward a high strain of 500%. Also, the Ag-p sensor film shows high sensitivity with a large ΔR/R0 (ΔR = R – R0, in which R0 is the resistance at 0% strain and R is the resistance under stretch) ratio of 100 000% in the strain range of 800%, comparable to that from gold-deposited supramolecular polymeric materials[53] sensor within a smaller strain of 400%.

Table 1

Comparison of Performance Parameters for Different Strain Sensors

materiala	ΔR/R0b (%)	strain range (%)	self-healingc	released	ref
SWCNT/PVA/borax	1200	1000	√	√	(52)
SPM-2 Au	100 000	400	√	×	(53)
CuNW/PU	300	50	√	√	(54)
STEP microfiber	450	40	×	×	(55)
CNT-fiber	350	900	×	×	(56)
aligned MWCNT	1150	100	×	√	(57)
CNT-PDMS	300	120	×	√	(58)
SWCNT/PDMS	150	60	×	×	(59)
Ag-p film	100 000	800	√	√	present work
	1700	50	√	√
SW-b film	100	1000	√	√
	50	500	√	√

SPM Au: gold-deposited supramolecular polymeric materials; CuNW: copper nanowire; STEP: stretchable tubular elastomeric piezoresistive; PDMS: poly(dimethylsiloxane); MWCNT; multiwalled carbon nanotube.

(R – R0)/R0 = ΔR/R0, where R0 is the resistance at 0% strain and R is the resistance under stretch.

Healable (√) or nonhealable (×).

Resistance of the released conductor films being measured (√) or not (×).

Conclusions

HIPC blends of PAA/PEO x/y and PAA/F108 x/y are all elastomers with self-healing properties. The results suggest that blends with higher x/y (e.g., 1/9) ratio are softer materials with higher stretchability and healing efficiencies than blends with lower x/y ratio. With the built-in soft PPO block of F108, the PAA/F108 x/y blends are also better system with higher stretchability and healing efficiency than PAA/PEO x/y blends. Among all blends, PAA/F108 1/9 exhibits the best results with the resolved εf of 1400% at a σf of ∼0.15 MPa, and ησ and ηε of the healed blend of 95 and 94%, respectively. PAA/F108 1/9 was further incorporated with AgNWs and SWCNT to construct Ag-p and SW-b conductor films, respectively. Although inferior in the large strain stability, the highly conductive Ag-p film is nevertheless ultrasensitive with a large ΔR/R0 ratio of 100 000% in the strain range of 800%. In contrast, the SW-b film is an ultrastable strain sensor, showing a stable R/R0 ratio over 200 stretching release cycles in the strain range of 500%. The low cost of this HIPC strategy makes it a potential commercial route for fabricating flexible, healable conductor sensors.

Experimental Section

Materials

The chemical compounds used are PAA (Mw ∼ 100 000, 35 wt % in H2O, Sigma-Aldrich), PEO (Mw ∼ 14 000, Sigma-Aldrich), Pluronic F108 (Mw ∼ 14 600 with 82.5 wt % of PEO, Sigma-Aldrich), poly(vinylpyrrolidone) (PVP, Mw ∼ 58 000, Sigma-Aldrich), ethylene glycol (EG, Alfa Aeser), AgNO3 (Sigma-Aldrich), iron(III) chloride (Acros), and SWCNT-2012 (length <2 μm, purity >97%, Golden Innovation Business Co., Ltd.). The glass substrates used in this work was cleaned by sonication in 2-propanol/water (v/v = 1/1) for 20 min and washed with ultrapure water three times before drying in a vacuum oven. AgNWs used in this work were synthesized according to the reported method.[51]

Preparation of AgNWs

AgNWs used in this work were synthesized according to the reported method.[51] The typical experiments were described as follows: 20 mL of ethylene glycol (EG) was injected into a 50 mL three-necked flask and preheated for 30 min; silver nitrate (0.33 g) was added within 10 s while 20 mL of EG solution of poly(vinylpyrrolidone) (PVP, Mw ≈ 58 000) (0.88 g) and FeCl3 (0.0027 g) was injected dropwise by a syringe within 5 min, and then the reaction was maintained at 170 °C for 80 min. The as-obtained products were then filtered and washed by methanol several times, and the purified products were preserved in ethanol.

Preparation of HIPC Films

A calculated amount of PEO (or F108) in methanol (5 mL) was added dropwise into a stirred solution of a calculated amount of PAA in methanol (5 mL). The resultant homogeneous solution was then slowly deposited on Teflon substrate. The resultant solution mixtures were dried in open air for 1 day, rendering transparent rubberlike films for study.

Preparation of Ag-p Conductor Film

Typically, 1 mL of ethanol suspension of AgNWs (25 mg mL–1) was dropped on top of a glass substrate before drying in open air for 1 day to obtain AgNW-decorated glass substrate. A clean PAA/PEO 1/9 (or PAA/F108 1/9) film was then pressed on the AgNW-decorated glass substrate and then peeled off to result in a AgNW-printed conductive film of PAA/PEO 1/9 (or PAA/F108 1/9).

Preparation of SW-b Conductor Film

The films were prepared by adding a calculated amount of F108 in methanol (5 mL) dropwise into a stirred solution of a calculated amount of PAA in methanol (5 mL), then adding 33 wt % of SWCNT into the resultant homogeneous solution, and then allowing the solution to dry in open air for 1 day.

Characterization

FTIR spectra were recorded by a Bruker Tensor 27 FTIR spectrophotometer. The complex solution was deposited on a premade KBr pellet, and the whole assembly was dried in a vacuum oven to prepare specimen for analysis. Thermal transition of the complex was determined by a TA-Q20 differential scanning calorimeter with a scan rate of 10 °C min–1. The storage modulus (E′) of the films was determined from a PerkinElmer Instruments DMA 8000 apparatus. Sample films with dimensions of 25 mm × 10 mm × 2 mm were used for analysis. Rheology data were obtained by a TA Instruments ARES G2 rheometer equipped with a 2.5 mm parallel plate. Samples with 2 mm diameter were used. Angular frequency from 0.05 to 126 rad s–1 was collected at a strain amplitude of 0.05%, which was within the linear viscoelastic response region. The recovery property was investigated by strain films under an alternatively changing amplitude of oscillatory force at 1 Hz. Tensile test was conducted by a PRO TEST PT-1699 V tensile tester with an extension speed of 30 mm min–1. Films with dimensions of 15 mm × 5 mm × 2 mm were used for analysis. The stress healing efficiency (ησ) was calculated by the formula ησ = σf,h/σf,v, where σf,v and σf,h are the fracture stresses of the virgin and self-healed samples, respectively. The strain healing efficiency (ηε) was calculated by the formula ηε = εf,h/εf,v, where εf,v and εf,h are the fracture strain of the virgin and self-healed samples, respectively. SEM images were obtained by a JEOL JSM-6700F microscope operated at 10 kV. The resistance was measured on a Tonghui TH2829 LCR meter. Wide-angle X-ray diffraction (WXRD) spectra were obtained from a Siemens D5000 diffractometer. UV–vis absorption spectra were recorded on an Ocean Optics DT 1000 CE 376 spectrophotometer.