SHAPE MEMORY BEHAVIOR OF CARBON NANOTUBE FILLED SEGMENTED POLYURETHANE COMPOSITES UNDER VARIABLE STRESSES
ABSTRACT
Composites of the segmented polyurethane (SPU) elastomer with different loadings of multiwall carbon nanotubes (MWCNTs) were prepared. Atomic force microscopy and X-ray diffraction were used to study the surface topography and structural aspects of the samples. Thermogravimetric analysis was performed in order to estimate the thermal stability of the composite samples. Two sets of cyclic thermomechanical experiments, both at sub-ambient conditions, were conducted for the evaluation of shape memory performance of the samples using modulated thermomechanical analysis. The first set of experiments was designed to estimate the shape memory performance under stress free or unconstraint conditions, while the second set was planned for the constraint conditions by application of variable stresses. The experimental parameters were so chosen that differences in shape memory response of matrix and composites were clearly distinguished and demonstrated. It was found that MWCNTs strongly affect the shape memory performance of the SPU and reduce the recovery speed under unconstraint conditions. While under constraint conditions, the extent of recovery and recovery speed of composites was found to depend on the applied load and nanotube concentration in the composites.
INTRODUCTION
Thermoplastic segmented polyurethane (SPU), consisting of alternating soft and hard segments, is an important material for shape memory applications because of enhanced thermal and thermo-oxidative stability, excellent sub-ambient flexibility, and easy processbility.1–4 In addition to the above, SPU is also a material of choice for critical smart applications in the defense and aerospace sectors due to excellent shape fixity and recovery on exposure to diverse external stimuli such as temperature, light, magnetic flux density, and so on.5–8 The high elastic deformation, large recoverable strain, low cost of processing, and excellent chemical properties of SPU make it more appealing than other materials.9 Despite these merits, SPU has poor strength, leading to low recovery force, thus impeding its widespread use in critical areas. However, the strength and functional properties of SPU were augmented by reinforcing it with different nano sized fillers like nano ferromagnetic particles, nano ceramics, metal nanopowders, nanoclays, carbon nanotubes, nano silica, and so on.10–15 Among different types of nanofillers, carbonaceous nanofillers like single wall carbon nanotubes (SWCNTs) and multiwall carbon nanotubes (MWCNTs) are potentially important due to their high aspect ratio, easy surface modification procedures, and scope of retention of large aspect ratio during normal polymer processing methods.16–20 Carbon nanotubes were also used in other polymers to impart specific functionality and enhancement of properties.21–23
Most of the works that report on the shape memory performance of the SPU are centered on stress–strain types of measurements using a universal testing machine in uniaxial extension mode. The compression mode known for circumventing, crazing, and minimizing the strain localization and thus maximizing the deformation capability of materials is scarcely reported.24–31 In addition, some researchers have used dynamic mechanical analysis and nanoindentation experiments for evaluating the shape memory performance of the polymers.32,33 The limitations of the above test procedures were discussed, and use of modulated thermomechanical analysis (MTMA) was suggested as an alternative in our recent publication.3
The SPU and its carbon nanotube based composites were examined extensively for shape memory performance under stress free or unconstraint conditions by holding the samples at a constant elevated temperature or applying continuous temperature increasing conditions until strain is completely recovered.16–20 However, low temperature shape memory behavior of SPUs is important from a defense perspective for use in protective garments in extremely cold weather regions, like Siachen in India (winter temperature ∼−40°C) and similar climates. Additionally, in most published works, single programming procedures were applied to the SPU and its carbon nanotubes containing composites, and resulting responses were discussed under stress free recovery or unconstraint conditions.24–33 The constraint recovery conditions where a force is applied during the recovery process are of practical importance for many applications but seldom reported. The constraint recovery behavior of SPU and its MWCNT based composites on application of different constraint forces is yet to be fully explored. The data on extent of shape recovery under different constraint conditions are often desirable for designing the product for use in a wide variety of areas. Further, recovery speed, which describes the time scale of the response and depends on many factors, including thermal environment, thermal conductivity of nanofillers, and heating/cooling rates, is an essential parameter in design of the product but seldom reported in the literature.24 To achieve desirable results it is, therefore, essential that thermomechanical conditions should be so chosen that molecular mobility is sufficiently activated while long range chain slippages are avoided within the time scale of the experimentation. In view of the above, this study aims at understanding the effects of MWCNT loading on the shape memory performance of the SPU under both stress free or unconstraint and constraint conditions. Efforts were also made to assess the effects of various programming parameters on the shape memory behavior of SPU and its carbon nanotube based composites with a quantitative estimation of recovery speed under varied applied stress conditions. Three different concentrations of the MWCNTs, namely, 1, 3, and 5 wt.% with respect to SPU, were selected for the preparation of the composites. Atomic force microscopy (AFM) and wide angle X-ray diffraction (WXRD) were used to assess the surface topography and nature of interactions of the MWCNTs with SPU. The effects of the nanotubes on the thermal stability and degradation profile of the SPU were also investigated. The shape memory characteristics of the samples were evaluated using MTMA, and comparison was made among various composites vis-à-vis a neat sample.
EXPERIMENTAL
Materials used
The SPU used was thermoplastic polyurethane (DP 9386) procured from Bayer Material Science, AG, Leverkusen, Germany. The SPU was derived from the polytetramethylene glycol [PTMG], diphenyl methane diisocynate [MDI], and 1, 4 butanediol (BD) as a chain extender. The MWCNTs were procured from Sigma Aldrich, Bangalore, India. The solvent tetrahydrofuran (THF), analytical grade, was obtained from Qualigens, Mumbai, India. The above materials were used as received without any further purification.
Preparation of the polyurethane composites
An 8% solution of the SPU in THF was prepared in four separate round bottom flasks. Then 1, 3, and 5 wt.% of MWCNTs (with respect to SPU) were separately mixed in THF and sonicated for 15 min to minimize the agglomeration. Thereafter, they were added individually into the abovementioned flasks separately containing SPU solution and stirred for 15 min followed by sonication for another 10 min to make homogenous dispersion of MWCNTs in SPU. The mixtures were then poured onto separate Petri dishes covered with aluminum foil for evaporation to dryness at room temperature and then kept in vacuum oven at 40°C until constant weight was obtained. The composite samples were named by prefixing PU before different wt.% of MWCNTs present in them, namely, PUCNT1, PUCNT3, and PUCNT5. In addition to the above, neat SPU was also treated in the same way, and samples were designated as PU.
Characterization techniques used
Surface topographic images were collected using AFM (Model-MFP-3D-BIO) of Asylum Research, Goleta, CA, USA. The instrument has a silicon probe with force constant of 26 N/m, and a frequency of 158 kHz was used for image collection in tapping mode. The images were collected for the scanning area of 1000 nm × 1000 nm. WXRD data were collected using a D8 Advance diffractometer (Cu Kα radiation, λ =1.54056 nm) of Bruker, Rheinstetten, Germany. The tube voltage and the tube current were 40 kV and 40 mA, respectively, and a scanning range of 2θ = 5°–40° was selected for the study.
The cyclic shape memory experiments on the samples were conducted using MTMA (Model- Q400EM) of TA instruments, New Castle, DE, USA, with N2 purge flow of 100 mL/min maintained throughout the experiments. The specimens had film with dimensions 8 × 4 × 0.6 mm. Two sets of MTMA experiments with different programming parameters were configured and performed on each sample. The first set of experiments corresponded to free or unconstraint recovery of the samples, while the second set was dedicated to constraint recovery under applied stress.
In the first set of experiments, the following sequence of programming was applied for the collection of the experimental data,
(1) Initial temperature 10.00°C, (2) force 0.250 N, (3) ramp 10.00°C/min to −80.00°C, (4) force 0.001 N, (5) isothermal for 5.00 min, (6) ramp 10.00°C/min to 20.00°C, and (7) end of method.
The minimum force at sequence no. 4 was applied for the proper functioning of the measurement system. The contact of the probe to the sample would be lost if the force at sequence no. 4 is made zero
In the second set of experiments, the sequence of operating parameters applied were
(1) Initial temperature −10.00°C, (2) force 0.250 N, (3) ramp 10.00°C/min to −80.00°C, (4) force 0.001 N, (5) isothermal for 5.00 min, (6) force 0.020 N, (7) ramp 10.00°C/min to 20.00°C; and (8) end of method
The three separate constraint forces, namely, 0.020 N to 0.030 N to 0.050 N, were selected at sequence no. 6 during the constraint recovery process.
In all experiments, the sample was kept beneath the probe of the MTMA system and temperature was equilibrated at an initial value of either 10.00°C or −10.00°C for deformation under compression. Once the sample reached equilibrium temperature, the required force mentioned at sequence no. 2 was instantaneously applied, and the sample was cooled concomitantly at a cooling rate of 10°C/ min to reach the temperature of −80°C. Once the final temperature of −80°C was reached, the load was removed and the sample was kept isothermally there for 5 min. This completes the shape fixing step. After the isothermal period, the sample was again heated at heating rate of 10°C/min to reach the temperature of 20°C. This step was designated as a shape recovery process. During all these steps the dimensional change of sample both with respect to temperature and time was measured.
The thermogravimetric analysis (TGA) data were collected in an inert argon environment using a modulated thermogravimetric analyzer (MTGA) (Model-Q500) of TA Instruments. The 60 mL/min and 40 mL/min purge flow of argon was maintained in the furnace and balance assembly of the instrument, respectively.
All the systems were calibrated as per the standard procedures of the original equipment manufacturer before performing the experiments.
RESULTS AND DISCUSSION
AFM analysis
The phase morphology of the MWCNTs dispersed in the SPU matrix was assessed using a scanning electron microscope and reported recently.4 In this paper, two- and three-dimensional (2D and 3D) surface topographic images of the samples were collected in the tapping mode of AFM and are presented in Figure 1a–d. The uniform distribution of the MWCNTs inside the SPU matrix is reaffirmed, which is a prerequisite for the preparation of high performance composites. The topographic images of neat PU can be seen in 2D and 3D images of Figure 1a, where lighter regions correspond to more elastic soft segment domains, whereas the darker areas are indicative of harder blocks.34 The neat sample surface is covered with a thin layer of smooth soft phase, since the soft segment is more elastic compared with the hard segment and consequently protrudes to higher elevations, as can be seen in the 2D and 3D views of Figure 1a. The aggregated hard domains with the spherical structures are also observed. However, a well-defined spherical order is not observed because of secondary ordering in which hard domains approach each other and coalesce.35,36 Figure 1b,c indicates that MWCNTs are distributed both in soft and hard phases. However, the extent of protrusions of the soft segment, as observed in neat SPU, is diminished due to interaction between MWCNTs and PU matrix.34,37 The effects of the inclusion of the MWCNTs on the crystallization characteristics of hard segment of the PU were already reported by us.4
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
X-ray analysis
Figure 2 illustrates the WXRD pattern of the composite samples in the 2θ range of 5° to 40°. A diffused diffraction peak centered at 20° can be noted for all composite samples. This wide diffraction peak is assigned to a characteristic of MDI/BD hard segment reflection, indicating some aggregated hard segments domain resulting from micro phase separation and a lack of ordered crystal structure in these samples.38 The intensity of the halo increases and peaks, becoming shaper with increasing content of the MWCNTs, while location of the halo is not shifted appreciably with increasing concentration of the MWCNTs. The WXRD patterns of the composites have another sharp prominent diffraction peak appearing at about 2θ = 26.06°, and intensity of this peak increases with increased concentration of the MWCNTs. This peak is associated with the crystal plane (0 0 2) of the crystallite structure of graphitic carbon, which may be present as impurities in the MWCNTs, as evident from the increase in intensity upon higher loading of the MWCNTs.34,39 These results are in agreement with the AFM observations.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
Shape memory characteristics
Analysis of Transitions Temperature
The glass transition temperature (Tg) corresponding to soft segments was determined using TMA, and the graph is presented in Figure 3a. The effect of the inclusion of the MWCNTs on the Tg of the neat sample PU was reported earlier in detail,4 where it was found that Tg shows a positive temperature shift on inclusion of MWCNTs, and the magnitude of this shift depends on the concentration of the MWCNTs, for example, Tg of PU is increased by about 19°C on addition of the 5 wt.% of the MWCNTs. The enhancement in Tg value of SPU on addition of MWCNTs was ascribed to high modulus of carbon nanotubes along with interaction between SPU and MWCNTs via Π–Π stacking.40
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
Thermal Stability
Figure 3b represents the overlay of the TGA plots of the neat as well as composite samples. It can be seen that temperature corresponding to 5 wt.% degradation of the samples shows positive shift on increasing the MWCNT loading. The maximum thermal stability corresponding to 5 wt.% weight loss is obtained for the sample PUCNT5 with an overall improvement of about 20°C in comparison with the neat sample. The higher thermal of the composite samples is the result of better polymer–filler interaction, improvement in Tg, coupled with more tortuous paths offered to the decomposed products and volatiles evolved during heating in TGA.
Shape Fixity
The degree of shape fixity and extent of unconstraint shape recovery are important parameters customarily measured to evaluate shape memory performance of polymers. Figure 4 represents the shape memory experiments with respect to time for the neat as well as composite samples at an applied load of 0.25 N. At time zero, when samples were at a deformation temperature of 10°C, a load of 0.25 N was applied, and samples were cooled concomitantly at a rate of 10°C/min. At around 9 min, when samples were at −80°C, the load was removed and samples were kept isothermally. This step was designated as a cooling/shape fixing process. Thereafter the samples were heated at a rate of 10°C/min up to 20° for the recovery, and this step is called the heating/recovery process. As indicated in Figure 4 the extent of deformation on equal applied loads is comparatively smaller in cases of composite samples than in neat samples of PU. This indicates relative improvement in the dimensional stability of composite samples due to addition of MWCNTs, and the highest improvement obviously is for the sample PUCNT5. Figure 5 depicts the shape memory profile of the samples with respect to temperature. It can be inferred from the figure that as the sample is cooled to a sub-ambient temperature of −80°C with an applied load, the resultant shape of the sample is fixed. This can be explained by considering broad thermal transition or glass transition (Tg), as shown in Figure 3a, consisting of an infinite number of sharp transitions.41 These sharp transitions continuously distributed over broad glass transition temperature can be regarded as individual memory elements. During cooling under deformation these memory elements undergo orientation commensurate to reduction of the conformational entropy of the system. Therefore, energy is stored in the material due to freezing in the strained state. After the load removal the samples show very small instantaneous recovery as depicted in Figure 5. This instantaneous recovery is ascribed to a small fraction of these memory units that preserve the mobility even when the sample is cooled below the transition temperature. The entropic elasticity thus generates an instantaneous reverting force upon removal of the load. However, the magnitude of instantaneous recovery is not appreciable. From Figure 4 it can be calculated that shape fixity for neat as well as composite samples is almost 100% under the chosen strain conditions.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
Shape Recovery under Unconstraint Conditions
SPU consists of alternating soft and hard segments, and there exists a variety of interactions, such as, (1) hydrogen bonding in hard segments between carbamoyl group and carbonyl group, (2) dipole–dipole interaction between carbonyl groups of hard segments, and (3) induced dipole–dipole interaction between aromatic rings of hard segments.42 These interactions are responsible for the phase separation in SPU. On heating of the samples held at their temporary shape above the reversible phase transition temperature, there is significant destabilization of these interactions leading to generation of molecular mobility. The memory units already oriented during the shape fixity process try to regain more probable entropic state at each temperature and thereby release elastic energy stored in the sample to create a mechanical restoring force in the sample to revert back to its permanent shape. However, in spite of significant improvement in the dimensional stability of the composite samples, as evident in Figure 4, the shape recovery profile has not changed much in comparison with virgin SPU. This shows that the basic shape memory mechanism of the SPU has not changed considerably on addition of MWCNTs. This is an added advantage for the shape memory applications of the MWCNT reinforced SPU composites. A comparison of the recovery profile of the neat and composite samples, depicted in Figure 5, reveals that neat samples show sharp recovery, while composites undergo a sluggish recovery. This is perhaps because of the fact that the composites become hard (enhanced glass transitions temperature and dimensional stability) due to the reinforcement effect of the MWCNTs, leading to a slower recovery process. As depicted in Figure 5 the complete recovery of composite samples is possible under the chosen strain conditions; however, the temperature window for complete recovery is enhanced. Similar observations were reported in cases of carbon nanotube reinforced polyurethane composites earlier.43
Recovery Speed/Response Time under Unconstraint Conditions
Shape memory characteristics of SPU and its composites are not an intrinsic property but depend on structure, morphology, and selection of programming parameters like stress/strain, temperature, heating/cooling rates, and so on. A careful selection of such conditions becomes crucial to tailor shape fixity/recovery and adequate recovery speed. In addition, the temperature corresponding to maximum recovery speed was used to define shape memory switching temperature “Ts,” which is more closely related to recovery behavior than shape memory transition temperature “Ttrans.”37 The “Ttrans” is generally determined by conventional thermal analysis techniques and is not directly linked to shape memory experimentation. The instantaneous recovery speed Vr is defined as time (t) derivative of strain ɛd and presented in Eq. 1
The shape memory performance of neat as well as composite samples can be evaluated using recovery speed data, since these are of paramount practical importance; however, they are scarcely reported in the literature.41 Figures 6 and 7 illustrate the comparison of recovery speed of the neat and composite samples with respect to temperature and time, respectively. The peak recovery rate values for both the cases were calculated and are presented in Table I. The recovery speed graphs of Figure 6 and Table I show that peak recovery rate is highest for the neat SPU and goes down in order of PUCNT1 > PUCNT3 > PUCNT5 for the composites. The developed trend in the composites is in conformity with extent of interaction and nature of dimensional stability of MWCNT based composites as described above. The higher the structural rigidity of composites, the greater the hindrance in the recovery process.3 Although the recovery rate of the composite samples is retarded in comparison with the neat sample, still full recovery becomes possible within the temperature interval set for experimentation, as depicted in Figures 4 and 5. The recovery speed with respect to temperature (Figure 7) follows trends similar to recovery speed with respect to time (Figure 6).
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
Shape Recovery under Constraint Conditions
Figure 8a–d depicts the recovery profile of the samples when they were subjected to different constraint loads, namely, 0.02 N, 0.03 N, and 0.05 N, during the recovery process. However, all the deformations were carried out using an applied force of 0.50 N. It can be seen in Figure 8a that on application of minimum constraint force of 0.02 N, the neat sample is not able to recover appreciably; therefore, no experiments using higher constraint forces were carried out. Owing to poor strength, neat PU is not able to generate enough recovery force to counter the constraint force during the recovery process. In Figure 8b it is evident that composite sample PUCNT1 shows a higher extent of recovery in comparison with neat sample PU on application of equal constraint force. However, the extent of recovery reduces with increase of the applied constraint force. The higher recovery generated under given constraint conditions is the manifestation of the reinforcement of the SPU in the presence of the MWCNTs. Figure 8c,d indicates similar types of results with the highest extent of recovery observed for the sample PUCNT5. These results are consistent and supportive of the results on the glass transition temperature and enhanced dimensional stability given in Figures 4 and 5.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
Recovery Speed/Response Time under Constraint Conditions
The shape recovery rate and response time under different applied constraint forces are important for shape memory applications and are related to the molecular origin and thermomechanical conditions applied to elucidate the shape memory effect. Figure 9 represents the comparison of constraint recovery speed with respect to time for the neat and composite samples at an applied constraint force of 0.02 N, while Figure 10 depicts the same with respect to temperature. The graphs for other applied constraint forces are not shown; however, results corresponding to each applied constraint force were calculated and presented in Tables II and III with respect to time and temperature, respectively. The results in Table II indicate that the peak recovery rate of composites is higher than that of the neat sample. This trend is contrary to what is observed in the case of the stress free or unconstraint condition (Table I). This is because under the applied constraint force the neat sample is not able to counter the constraint force during the recovery process, while composite samples are able to do so due to MWCNT reinforcement. Concomitant increase of filler concentration improves the recovery force, and hence an enhancement in the peak recovery rate is observed. These observations are consistent with the results of the thermomechanical analysis and observations made during the shape fixity process depicted in Figures 4 and 5. The results of Table III also follow similar trends as that of recovery speed with respect to time.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83794
CONCLUSIONS
The following conclusions were drawn from the constraint and unconstraint shape memory behavior of the SPU and its MWCNT based composites:
- 1.
The neat SPU and its MWCNT based composites were prepared via the solution casting method, and analytical techniques like AFM, WXRD, TGA, and MTMA were used for characterization of these materials. In particular MTMA was found to be very effective in comprehensive estimation of the dimensional stability and shape fixity/recovery profile of samples.
- 2.
The shape fixity and shape recovery of the neat PU as well as those of the composites are almost 100% under stress free or unconstraint and chosen strain conditions. Both the time and temperature peak recovery rates of SPU and its MWCNT based composites under stress free or unconstraint conditions were reduced, and the extent of this reduction is dependent on the MWCNT loading.
- 3.
Under constraint recovery conditions, the extent of recovery is dependent on the MWCNT concentration, and it is highest for the sample PUCNT5. Under applied constraint force, higher shape fixity can be used for precise shaping of the product with excellent recovery and improved dimensional stability. These characteristics make these materials attractive for critical shape memory applications.
- 4.
Unconstraint and constraint recovery processes shows opposite trends in peak recovery rate and response time.
Topographic image of the samples (a) PU, (b) PUCNT1, (c) PUCNT3, and (d) PUNCT5.
WXRD pattern of the composites.
(a) Glass transition temperature of the PU. (b) TGA plots of the samples.
Shape fixity and recovery process of samples with respect to time.
Shape fixity and recovery process of samples with respect to temperature.
Recovery speed of samples with respect to temperature.
Recovery speed of samples with respect to time.
Constraint recovery process of samples (a) PU, (b) PUCNT1, (c) PUCNT3, (d) PUNCT5.
Comparison of constraint recovery speed of samples with respect to time.
Comparison of constraint recovery speed of samples with respect to temperature.
Contributor Notes