Introduction

The shape memory polymers (SMPs) are nonmetallic shape memory materials (SMMs) with the ability to undergo a large recoverable deformation upon the application of an external stimulus (Ref 1, 2). SMPs have several advantages over shape memory alloy (SMAs) such as lightweight, low cost, ease in processability, high strain recoverability, tailorable switching temperature, and response to a wide range of stimuli (Ref 3). Various types of stimuli for SMPs are heat (Ref 4,5,6,7), electric field (Ref 8,9,10,11), magnetic field (Ref 12, 13), moisture (Ref 14, 15), pH value (Ref 16, 17), light (Ref 18,19,20), pressure (Ref 21), multi-stimuli (Ref 22), etc. The SMPs have several potential uses in biomedical engineering, control systems, smart textile, self-deployable structure, artificial muscles (Ref 23, 24), etc. However, their applications are restricted due to their low recovery stress (Ref 25) as well as low thermal and electrical conductivity. The desired higher recovery stress may be achieved by maximizing recoverable energy stored in the material. This recoverable energy depends mainly on polymer structure, reinforcement, crystallinity, and heat transfer mechanism. One of the effective methods to improve recovery stress and thermal properties is to reinforce virgin SMP with high-modulus and highly conductive inorganic or organic filler materials. The addition of nanofiller in very small quantity improves mechanical as well as shape memory properties of the composite. Nanomaterials are considered the most important filler for SMP due to their exceptionally superior thermal, mechanical, and electrical properties. The enhancement of shape memory effects such as shape recovery stress, recovery strain, and recovery speed in the shape memory polymers (SMPs) by introducing nanofillers has been an important research objective during the current decade. Many studies have reported the improvement in shape memory properties on addition of nanofillers which includes reinforcement of SiC (Ref 26, 27), nanoclay (Ref 28), carbon black (Ref 29), CNTs (Ref 2, 11), cellulose nanowhiskers (Ref 30), graphene (Ref 31, 32), etc. Among these fillers, carbon-based fillers are popular which include CNTs (Ref 1, 11, 33), carbon black (Ref 29, 34), carbon nanopaper (Ref 35, 36), and graphene (Ref 8, 9). Researchers have shown that a few key factors governing composite properties depend upon filler geometry, the process of mixing, and concentration of the raw materials.

Graphene drew the researcher’s attention due to its unique mechanical and electrical properties. Graphene nanoplatelets (GNPs) have intermediate geometry between graphene and graphite with 1-20 nm thickness, 1-50 µm lateral size, and specific surface area of 50-150 m2/g. Many studies have been reported related to graphene-reinforced shape memory polymer composites. The addition of functionalized GNP in polyurethane matrix improved mechanical properties and strain recovery (Ref 37). Jung et al. reported recovery stress of 1.8 MPa/cm3 on the addition of 0.1% few layer graphene in polyurethane by solvent route (Ref 38). Jin yoo et al. reported that graphene oxide-reinforced polyurethane composites exhibit 50% strain recovery under a load of 6.5 gf (Ref 39).

Recently, microwave (MW) irradiation has emerged as an important stimulus for SMP actuation. Haiyan Du et al. reported the microwave-induced shape memory effect of chemically cross-linked moist poly(vinyl alcohol) networks (Ref 40). Leng and co-workers have developed MW-induced SMP/CNT composites and showed that 99.84% of microwave energy could be absorbed by 5% addition of CNT (Ref 41). Zhang Zhou et al. investigated the use of an infrared thermal imager and digital camera to record the instantaneous shape recovery (Ref 42).

In the present work, different fractions of GNPs were dispersed uniformly in SMPU matrix by melt-blending process using micro-compounder. The application of microwave irradiation as the stimulus for shape recovery in the composites was studied. The shape recovery stress and recovery strain were determined using the thermomechanical cycle. The dynamic mechanical analysis was carried out to determine storage modulus, loss modulus, and damping factor. Their influence on shape memory effects was studied and reported. The work suggests microwave as quick and efficient stimuli for SMPs actuation as compared to conventional heating.

Experimental Details

Materials

The shape memory thermoplastic polyurethane (ether type) MM6520, in the form of pellets, were obtained from SMP Technologies Inc. Japan. Graphene nanoplatelets (GNP) 11-15 nm having a specific surface area 50-80 m2/g was obtained in the form of powder from IoLiTec nanomaterials, GmbH, Germany.

Sample Preparation

Shape memory polyurethane granules and GNPs were first dried in a vacuum oven at 105 °C for 6 h. A Thermo HAAKE MiniLab II micro-compounder with conical twin screw (model PolyLab OS) was used for melt mixing of SMPU and GNPs at a rotor speed of 60 rpm and mixing temperature of 210 °C for 5 min. Different compositions of SMPU with 0.2, 0.4, 0.6, and 0.8 phr GNP were prepared in batches of 6 g each in micro-compounder. Each 6-g batches of the mix was extruded in the form of a strand. These strands were chopped to 5-mm size and reprocessed in micro-compounder to obtain a uniform composition. The compounded material was injection-molded to test specimen using a micro-injection-molding machine, having process parameters as temperature 210 °C, injection pressure 620 bar, injection time 10 s, post-pressure 600 bar for 20 s, and mold temperature 60 °C. The procedure for the sample preparation is described schematically in Fig. 1.

Fig. 1
figure 1

Schematic description of sample preparation

Full size image

All samples were annealed in a vacuum oven at 110 °C for 10 h and subsequently at room temperature for 12 h. Nanocomposite samples containing GNPs were prepared with a filler concentration of x up to 0.8 phr and were labeled as xGPU. The virgin polymer was labeled as SMPU.

Shape Memory Behavior Evaluation

Shape memory behavior was evaluated using Tinius Olsen25kt UTM attached to the environment chamber. The test specimen was characterized according to a progressive stretch–relax–stretch (PSRS) scheme of shape memory creation procedure (Ref 1). A total of 100% elongation was achieved in three subcycles, sequentially. In the first subcycle, the sample was elongated by 50%, second subcycle by 75%, and third subcycle by 100%. Each cycle involved following steps: (i) heating the sample to a temperature Th which is higher than the glass transition temperature Tg, (ii) stretching to programmed level, (iii) bringing down the temperature below Tg without relaxing the deformation strain, (iv) removal of imposed strain (clamps, etc.) and allowing the sample to relax and attain a fixed length. In the second subcycle, the sample was heated to Th, stretched to the next programmed level, cooled down to room temperature and after that relaxed by declamping at ambient temperature. In the third subcycle, the sample was brought to Th, clamped and stretched to the programmed maximum level as decided earlier, i.e., 100% of the original length of the sample. After that sample was once again brought to room temperature in clamped condition and was unclamped. Now specimen is ready with new temporary shape and can be tested for recovery stress.

Microwave-Induced Shape Recovery

Microwave-induced shape recovery was carried out in an open-glass vessel on a modified microwave oven model 2001 ETB with rotating tray and a power source 230 V. The shape recovery was measured using a digital camera and an infrared thermal imager (FLIR E5). The specimen was tested at 120 W microwave power and 2.45 GHz frequency. The distance between the permatron and the specimen was 30 cm. Each test specimen is pre-deformed in temporary shape at 70 °C and cooled to 30 °C to fix the temporary shape.

Dynamical Mechanical Analysis

Dynamical mechanical analysis (DMA) is a technique for measuring the viscoelastic properties of the material. It is known that shape memory properties are governed by the modulus of material below and above the transition temperature. This experiment allows to determine the material’s response by the application of the temperature and dynamic load. The dynamic mechanical properties of the samples have been determined using “Dynamic Mechanical Analyzer” DMS 6100 by Hitachi Instruments. The samples were injection-molded, with dimensions 40 × 10 × 1 mm3. Test conditions were as follows: the measurement method was the tensile mode, loading frequency 1 Hz, heating rate 20 °C/min, for a temperature interval of 30-85 °C.

Microstructural Characterization

Morphological studies were conducted using FESEM of M/s Nova NanoSEM 430. The cryogenically fractured surfaces were observed after gold coating. The transmission electron microscope (TEM) observations were performed on a JEOL JEM-1400 with an acceleration voltage of 100 kV. Before the observation, the samples were ultramicrotome into ultrathin films of about ~ 150 nm thickness in liquid nitrogen by an ultramicrotome (Leica FC-6) equipped with a diamond knife.

Results and Discussion

Shape Memory Behavior

Stress–strain and recovery strain plots of shape memory polyurethane containing the various fractions of graphene nanoplatelets have been generated using the thermomechanical cycle method. Figure 2(a), (b), and (c) shows plots for SMPU, 0.4 GPU, and 0.8 GPU as per the PSRS method, (Ref 2). PSRS scheme has been employed successfully in improving the recovery stress of GNP/SMPU thermoplastic composite. In the present study, 100% strain deformation was achieved progressively at 65 °C in the air-circulated environment chamber. Figure 2(a) shows the stress–strain curve of three stretching and three recovery steps of SMPU samples. The crosshead speed of 2 mm/min was maintained for stretching the samples. After each stretching step, samples were cooled under the clamped condition to 30 °C and kept at hold for 10 min to fix the imposed strain. The sample was unclamped, and the final deformation was measured after 10 min to obtain the shape fixity ratio. The sample was reclamped, and the stress developed with increased temperature up to 65 °C was recorded. The strain imposed was relaxed by moving crosshead at a speed of 2 mm/min in the downward direction until the recovery force generated by material became 0 N. The sample was unclamped and free recovery was allowed for 10 min at the same temperature and the length of the sample was measured. After the third stretching of the maximum 100% strain, a tensile stress of SMPU was 2.07 MPa. Similarly, other samples, namely 0.2 GPU, 0.4 GPU, 0.6 GPU, and 0.8 GPU, were studied. In the study of shape memory effect of material, strain recovery is an important parameter that describes the part of strain that can be recovered from strain imposed by stretching.

Fig. 2
figure 2

(a) Stretch and recovery curves of SMPU at the different strain levels. (b) Stretch and recovery curves of 0.4 GPU at the different strain levels. (c) Stretch and recovery curves of 0.8 GPU at the different strain levels

Full size image

In the present study, constrained as well as unconstrained (free) strain recovery is calculated by using the following relationship, \(R_{\text{r }} (\% ) = \frac{{ \in_{\text{r}} }}{{ \in_{\text{d}} }} \times 100\), where \(\in_{\text{r}}\) is the recovered strain and \(\in_{\text{d}}\) is the maximum deformed strain. The experimental value of constrained strain recovery was 72.5% and free strain recovery was 94.2% for SMPU, and similarly, the values 0.2 GPU, 0.4 GPU, 0.6 GPU, and 0.8 GPU were studied and are reported in Table 1. The total recovery percentage was more in pure SMPU as compared to composites; however, there were insignificant changes in constrained recovery. The GNPs in SMPU act as a physical constraint that could decrease the mobility of the hard segment chain, and leads to longer relaxation time, and results in decreased strain recovery. The stress–strain and recovery stress–strain plots were generated as shown in Fig. 2(a), (b), and (c), for SMPU, 0.4 GPU, and 0.8 GPU, respectively. The tensile stress developed in SMPU, 0.2 GPU, 0.4 GPU, 0.6 GPU, and 0.8 GPU were found to be 2.1, 4.64, 5.14, 5.11, and 5.02 MPa, respectively, though stress–strain curves for 0.2 GPU and 0.8 GPU are not shown here. On addition of GNPs in SMPU and corresponding increment in tensile stress indicates good bonding and uniform dispersion of GNPs in SMPU. The recovery stresses were 1.87, 3.09, 3.9, 4.1, and 3.87 MPa for SMPU, 0.2 GPU, 0.4 GPU, 0.6 GPU, and 0.8 GPU, respectively. It was observed that SMPU exhibits lower recovery stress as compared to composites. The recovery stress of virgin SMPU was approximately doubled on the addition of 0.6 phr GNPs. The detailed findings are summarized in Table 1. The data indicate that recovery stress and tensile stress of composites increased significantly as compared to virgin shape memory polyurethane. The increase in tensile strength at 100% strain leads to an increase in area under the stress–strain curve. This area represents the strain energy stored during stretching provides the driving force for strain recovery on the release of stress in the rubbery state of the thermoplastic. The addition of GNPs increases the area under the stress–strain curve to improve shape recovery properties. In Fig. 2(a), tensile stress and recovery stress are not much affected with progressive step 1 to step 3 in the case of SMPU, but in the case of composites tensile stress and recovery stress, both increased in each progressive steps as shown in Fig. 2(b) and (c). This improvement in tensile and recovery stress is due to the increase in elastic modulus of the polymer by the addition of graphene in the composite. More elastic strain energy is stored during stretching of specimens in the form of internal stress due to the reinforcement of high-modulus GNPs. GNPs having a high specific area helps in phase separation of a hard and soft segment of segmented polyurethane, which may attribute to improved shape memory effect. Shape fixity (Rf) shows the ability of the material to hold its temporary shape and defined as the ratio of the strain fixed (\(\in_{\text{f}}\)) to maximum deformation strain (\(\in_{\text{d}}\)), \(R_{\text{f }} (\% ) = \frac{{ \in_{\text{f}} }}{{ \in_{\text{d}} }} \times 100\). The variation in shape fixity for different amounts of GNP-reinforced SMPU at 50, 75, and 100% strain is shown in Fig. 3. The shape fixity values were 92.5, 94.3, 95, 97, and 98% for SMPU, 0.2 GPU, 0.4 GPU, 0.6 GPU, and 0.8 GPU, respectively, at 100% strain. The above data clearly show that the shape fixity was increased on the addition of the GNP content in the SMPU matrix.

Table 1 Shape fixity, shape recovery, maximum tensile, and recovery stress of different specimen
Full size table
Fig. 3
figure 3

Shape fixity of GNP-reinforced SMPU composite for different deformation sequences

Full size image

The recovered strain was reduced due to the hindered effect of incompressible GNPs. As we cool SMPU in the deformed state under the constrained condition to fix the deformation below the glass transition temperature, the soft segment of polymer shifts to a glassy state. When a constraint is released, the polymer chain, which is having certain mobility tries to release their remaining elastic strain and results in lower shape fixity. In the case of composites, GNPs act as a restriction so it reduces elastic strain and contributes to increased shape fixity. Figure 3 indicates an increase in shape fixity with an increase in deformation from 50 to 100%. It may be attributed that with the increase in strain, there will be some plastic deformation that hinders elastic recovery during shape fixing. Figure 4(a) indicates the thermomechanical cycle at 100% strain, which reveals the increase in tensile stress due to the addition of GNPs in SMPU as compared to virgin SMPU. An increase in stress was observed for composites due to the increase in stiffness of composites with the loading of high-modulus GNPs. Figure 4(b) shows the energy absorption which is calculated by area under the stress–strain curve of Fig. 4(a) for each composition. The addition of high-modulus GNPs increased energy absorption of the polymer composites, which results in increased recovery stress shown in Fig. 4(c).

Fig. 4
figure 4

(a) Stretching and recovery curve at 100% strain for different composition of composites. (b) Energy absorbed in stretching for different composites. (c) Maximum recovery stress at 100% strain for SMPU, 0.2 GPU, 0.4 GPU, 0.6 GPU, and 0.8 GPU

Full size image

To study the effect of temperature on induced stresses of SMP composites, the clamped temporary shapes of the composites were heated to 65 °C with a constant heating rate of 2 °C/min and the corresponding values of stresses were recorded. The developed samples SMPU, 0.2 GPU, 0.4 GPU, and 0.6 GPU were stretched previously to 100% strain and were allowed to fix a temporary shape. The induced stress versus temperature curves is shown in Fig. 5. The stress induced at 65 °C in SMPU, 0.2 GPU, 0.4 GPU, 0.6 GPU, and 0.8 GPU were 1.87, 3.09, 3.9, 4.1, and 3.87 MPa, respectively. These curves show two temperatures range (I) 40-55 °C and (II) 55-65 °C. SMPU reached its maximum recovery stress at a lower temperature as compared to the reinforced composite. In the case of the composite, the transition temperature is shifted toward the higher side on the addition of GNPs, which can be confirmed from DMA results.

Fig. 5
figure 5

Variation in recovery stress with temperature for different composites

Full size image

Dynamic Mechanical Analysis

The dynamic mechanical analysis (DMA) of SMPU and GNP-reinforced SMPU composites were conducted to study storage modulus E′, loss modulus E″, and damping factor tan δ of specimens. The test was conducted in tensile mode, and changes in E′, E″, and tan δ were plotted against the temperature for different samples. Storage modulus represents the ability of a material to store elastic energy, and higher elastic energy is required to produce good shape recovery behavior. In SMPs, shape recovery is governed by high rubbery modulus and slow chain relaxation (Ref 33). The value of storage modulus is increased by the addition of GNPs in SMPU as shown in Fig. 6(a). An increase in storage modulus is due to high-modulus GNP reinforcement in the SMPU matrix. Storage modulus in the rubbery stage at 65 °C is shown in Fig. 6(b) for different GNP content. It increased from 9.25 MPa for SMPU to 15.1 MPa with GNP addition for 0.6 GPU which resulted in high recovery stress. Figure 7(a) represents the variation of loss modulus with temperature on the addition of GNPs. In these composites, mainly three kinds of frictional interaction take place: (i) between graphene platelets, (ii) graphene and polymeric chain, and (iii) inter-polymer chain. The value of loss modulus was increased with the addition of GNPs in SMPU composites. Loss modulus represents the loss in strain energy due to internal friction in polymer chains. For good shape-fixing capacity, a polymer should have a higher loss modulus. The maximum value of loss modulus shifts toward higher temperatures for composites due to the restriction in mobility of the polymeric chain. There is very little variation in loss modulus in the glassy state as compared to rubbery state on the addition of graphene platelets. There is a broadening of loss modulus curve with an increase in transition range on the addition of GNPs. It may be attributed to the increased heat transfer rate due to the addition of GNPs.

Fig. 6
figure 6

(a) Storage modulus vs. temperature for SMPU and GNP-reinforced SMPU composites. (b) Rubbery modulus at 65 °C for different composites

Full size image
Fig. 7
figure 7

(a) Loss modulus vs. temperature for SMPU and GNP-reinforced SMPU composites. (b) tan δ vs. temperature for SMPU and GNP-reinforced SMPU composites

Full size image

The damping parameter of material provides the balance between the viscous and elastic phases of polymer composites. The value of the damping factor is low, below the glass transition temperature, because the chain segments are in the frozen state due to insufficient thermal energy to cause translational and rotational motion of segments. Figure 7(b) shows that the value of the damping factor is more in the glassy state for pure SMPU as compared to GNP–SMPU composites. In the case of the rubbery state, the damping factor is more in the case of composites as compared to pure SMPU. Below glass transition temperature, polymer chain segments and GNPs are in an almost frozen state and provide good binding, so elastic energy is more in composites as compared to pure SMPU. Above glass transition temperature, the chain segments are free to flow and GNPs are also free to slide with the chain which provides more frictional losses as compared to friction between chain segment only in case of pure SMPU. The peak of tan δ represents glass transition temperature, which shifts toward higher temperatures by the addition of GNPs. The damping factor first increased by the addition of GNPs up to 0.2 phr and then decreased. The peak value of tan δ for 0.6 and 0.8 phr wt.% GNP–SMPU composites was lower than the pure SMPU. At higher GNPs, concentration increase in energy absorption capacity was more compared to the energy dissipated.

Morphological Analysis

The morphology of the pristine GNPs and GNP–SMPU composites for different composition was studied using FESEM. The FESEM was used to observe the surface morphology and dispersion of GNPs in SMPU. Figure 8(a) shows the pristine GNPs, and Fig. 8(b), (c), (d), (e) and (f) shows the cryogenic fractured surface (b) of SMPU, (c) of 0.2 GPU, (d) of 0.4 GPU, (e) of 0.6 GPU, and (f) of 0.8  GPU, respectively. The SEM image of thin flakes of GNP showed that the lateral size of GNP was found to be 2-10 µm, and flakes are present in the form of aggregates. From Fig. 8(a), it is observed that GNPs are composed of a few layers of graphene. Figure 8(b) of virgin SMPU reveals the smother-layered riverbed-type structure and polymer grains with elongated structure due to some ductility in pure SMPU. With the addition of GNPs in polymer composite, the number of polymer grains also increases as compared to pure SMPU. It may be attributed to the increase in the number of nucleation points due to the addition of GNPs. It is observed that no GNPs are directly exposed to the surface or separated from the matrix, so we can assume that they are uniformly distributed and having a good bonding with the polymer. Figure 9 shows the magnified view of the cryogenically fractured surface of composites, where the dispersion of GNPs in polyurethane can be observed easily and GNPs are encircled.

Fig. 8
figure 8

SEM micrograph of raw GNPs (a) and cryogenic fractured surfaces (b–f) of SMPU, 0.2 GPU, 0.4 GPU, 0.6 GPU, and 0.8 GPU, respectively

Full size image
Fig. 9
figure 9

SEM micrograph of (a) 0.2 GPU, (b) 0.4 GPU, (c) 0.6 GPU, (d) 0.8  GPU; TEM micrograph of (e) 0.4 GPU, (f) 0.8 GPU

Full size image

From the TEM analysis, presence of GNPs in the polyurethane matrix was confirmed, as shown in Fig. 9(e) and (f). GNPs have a layered structure, which can be easily observed in Fig. 9(e) and (f). 0.4 GPU composite shows uniform dispersion of GNPs (Fig. 9e), whereas in the case of 0.8 GPU composite, small agglomeration of GNPs is observed. TEM image shows good interfacial interaction between the graphene layer and the SMPU matrix. Folding and wrinkling of the GNPs layer also observed in both the composites.

Microwave-induced shape recovery of developed composites was studied by using an infrared thermal imager and digital camera. Firstly, the pure SMPU sample and the developed composite was deformed in oval shape (temporary shape) above its glass transition temperature and then allowed to cool below its glass transition temperature to fix this shape. Then both the samples with an oval shape are kept in the microwave oven for studying its shape recovery property. Initially, the pure SMPU sample was kept under the microwave, which does not respond with the application of microwave, and thus, no shape recovery was found in it. The shape recovery behavior of 0.8 GPU nanocomposite with time was studied and is depicted in Fig. 10. The microwave propagation in a material depends on its dielectric and magnetic properties, and thus, GNP-reinforced SMPU shows heating response to the microwave irradiation. It may be attributed to the addition of GNPs in the polyurethane matrix which improved its dielectric and magnetic properties, and also GNPs act as heating nodes in the developed nanocomposites. The figure shows that 0.8 GNPs recovered its 80% shape in 60 s, which is faster comparing it with conventional heating.

Fig. 10
figure 10

Sequence of shape recovery of SMP composite (0.8 GPU) under microwave irradiation (Group A digital images, Group B infrared thermal images)

Full size image

GNPs have a strong influence on the properties of developed composites. Higher modulus and thermal conductivity of GNPs improved the shape memory properties of the developed composite. Recovery stress of developed composite increased significantly on the addition of GNP. The addition of GNPs makes SMPU sensitive to microwave heating and increased loading of GNPs increases the heating rate considerably. Microwave-induced shape recovery behavior strongly dependent on microwave power, frequency, and distance of the object from microwave source. Further, more study needs to do for proper understanding between the interaction of microwaves with material make suitable for remote sensing applications.

Conclusions

GNP–SMPU composites have been successfully developed by melt blending using micro-compounder. GNPs were homogeneously dispersed into the SMPU matrix with different mass fractions. DMA analysis confirmed the increase in storage modulus with an increase in GNP loading. Storage modulus in the rubbery stage was also increased with GNP loading, and a maximum value 15.1 MPa was obtained for 0.6 phr loading of GNPs in the SMPU matrix at 65 °C. Shape memory creation was attained using the PSRS scheme that consisted of three subcycles 50, 75, and 100% strain level. Recovery stress increased approximately twofold of pure SMPU at 0.6 phr loading of GNPs in the matrix. The addition of GNPs also shows significant improvement in tensile and recovery stress in each progressive step, which was 50, 75, and 100% strain, whereas this improvement was insignificant in virgin SMPU. Shape fixity increased from 92.5% for SMPU to 97.3% for 0.8 GPU at 100% strain deformation. Free strain recovery is slightly reduced from 94.2% for SMPU to 89% for 0.8 GPU. The addition of GNPs increased the glass transition temperature of the composite from 53.86 to 55.74 °C. Microwave-induced shape recovery was successfully demonstrated for GNP-reinforced SMPU composites. The addition of 0.8 phr GNPs in SMPU shows 80% shape recovery in 60 s.