Infrared light actuated shape memory effects in crystalline polyurethane/graphene chemical hybrids

Polycaprolactone diol (PCL diol, M_n = 4000) was purchased from Capa (Perstorp Chemicals, Korea). It was dried and degassed at 80 ° C under vacuum for 3 h before use. 4,4'-methylenedicyclohexyl diisocyanate (H₁₂MDI, Aldrich), 2-hydroxyethyl acrylate (HEA, Aldrich), dibutyltin dilaurate (DBTDL, Aldrich), phenyl glyoxylic acid methyl ester (Darocur MBF, Ciba Specialty Chemicals), and allyl isocyanate (Aldrich) were used as received. Graphene oxide was kindly prepared and donated from the lab of Professor Han Mo Jeong, University of Ulsan.

The process of functionalization of the GO was performed according to the literature procedure [26, 27]. GO (1.2 g) was loaded into a 500 ml round-bottomed flask equipped with a stirrer under nitrogen. Anhydrous DMF (120 ml) was added to create an inhomogeneous suspension. Allyl isocyanate (3.99 g) was then added and the mixture was stirred for three days before the reaction mixture was poured into methylene chloride (2.4 l) to coagulate the product, which was filtered, washed and dried by lyophilization.

The overall reaction scheme for synthesizing the composite is shown in scheme 1. PCL diol having a molecular weight of 4000 g mol⁻¹ was first reacted with H₁₂MDI in dimethylformamide (DMF) at 60 ° C for 4 h using a 500 ml four-necked flask equipped with a mechanical stirrer, thermometer and condenser with a drying tube. High molecular weight PCL was employed to make the crystalline soft segment of PU. The NCO terminated PU was end capped with HEA at 55 ° C. This reaction yielded vinyl terminated prepolymer having theoretical molecular weights of about 9000 determined from the [NCO]/[OH] index. The composition of the PCL 4000/H₁₂MDI/HEA was fixed at 2/3/2 (by mol).

Zoom In Zoom Out Reset image size

Formulations of the reactants are given in table 1. The progress of the reaction was monitored via FT-IR measurements (figure 1), where the absorption peak of NCO at 2270 cm⁻¹ completely disappeared upon capping it with HEA. The HEA capped PU and iGO dissolved in DMF were poured into the flask. A photoinitiator was then added to this dispersion, and stirred for another 2 h before the dispersion was cast on a polyethylene film and partially dried. Then the film was cured using a UV lamp (365 nm, 8W, Crosslink) for 2 h at atmosphere.

Zoom In Zoom Out Reset image size

The HEA capping reaction of the H₁₂MDI terminated prepolymer was monitored by FT-IR spectroscopy using the absorption peak of the isocyanate group. ATR-FT-IR (attenuated total reflectance Fourier transform infrared, Thermo Scientific, Nicolet 6500 FT-IR) spectroscopy was used to check the reaction between the vinyl groups of the iGO and HEA capped PU. The NIR/IR spectra were taken in the 1000–3000 nm range using a FT-UV–visible–IR spectrometer, VERTEX 80.

AR-XPS (angle-resolved x-ray photoelectron spectroscopy) and standard XPS measurements were done using a VG-Scientific ESCALAB 250 spectrometer with an Al Kα x-ray source. Thermal analyses were carried out with a differential scanning calorimeter (DSC, Seiko DSC 220) over the temperature range from −30 to 100 ° C at a heating rate of 5 ° C min⁻¹. The state of particle dispersion was studied with the scanning electron microscope (SEM).

For isothermal crystallization the sample was first heated to 80 ° C and kept at that temperature for 20 min to completely erase the thermal history; this was followed by quenching to 23 ° C at which temperature the crystallization was carried out. Isothermal crystallization was done to ensure that the crystallization was completed during the cooling step of cyclic loading and unloading.

The tensile properties of the hybrid were measured at room temperature with a Universal Testing Machine (UTM, Lloyd LRX plus) at a crosshead speed of 500 mm min⁻¹ using specimens prepared according to ASTM D-1822.

The shape memory properties were characterized using a UTM with a heating/cooling chamber attached. The sample was first heated to T_m + 20°C and uniaxially stretched to a maximum strain (ε_m) of 100%; this was followed by cooling and unloading at T_m −20 ° C. A long enough cooling time was provided to assure crystallization. Upon unloading, a small part of the strain (ε_m − ε_u) was instantaneously recovered, leaving an unload strain (ε_u). Then, the sample was reheated to T_m + 20 °C to recover the strain, leaving a substantial amount of permanent strain (ε_p). These three steps complete one thermomechanical cycle. Shape fixity (R_f) and shape recovery ratios (R_r) for the cycle are defined as follows [28]:

$$\begin{eqnarray} &&\%\text{shape fixity}=\frac{{\varepsilon }_{\mathrm{u}}}{{\varepsilon }_{\mathrm{m}}}\times 1 0 0 \end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray} &&\%\text{shape recovery}=\frac{{\varepsilon }_{\mathrm{r}}}{{\varepsilon }_{\mathrm{m}}}\times 1 0 0 \end{eqnarray} \tag{ 2 }$$

where ε_r = ε_m − ε_p is the recovered strain.

For infrared absorption, an Hg–Xe lamp was used as the light source, positioned 20 cm away from the samples. The power density delivered to the sample was ∼25 mW cm⁻² as measured using a light intensity meter. Samples were first elongated to 100% at 65 ° C, and this was followed by cooling to 23 ° C to fix the deformation. The cooling time was typically 15 min. Finally, the deformation was recovered by exposing the sample to the light. In all cases, more than three samples were tested, from which the mean and standard deviation were calculated.

Figure 1 shows the FT-IR spectra of the PU prepolymer before and after HEA capping. The absorption peak at approximately 2270 cm⁻¹ assigned to the stretching vibration of the NCO group disappeared completely while the vinyl group peak at approximately 750 cm⁻¹ appeared newly after capping the PU prepolymers with HEA.

To identify the UV curing reactions between PU and iGO, ATR-FT-IR measurements were carried out and the results are shown in figure 2. The vinyl peaks at 1630 cm⁻¹ (stretching vibration) and 807 cm⁻¹ (twist motion) disappeared upon UV curing, which indicates that the double bonds of PU and iGO are converted into single bonds through the chemical reaction between the two.

Zoom In Zoom Out Reset image size

The DSC thermograms of the hybrids are shown in figure 3 along with the crystallization temperature (T_c), melting temperature (T_m), degree of supercooling (ΔT) and heat of melting (ΔH_m) data in table 1. It is seen that T_c, T_m and ΔH_m for the PU soft segment decrease with 1 wt% iGO (G10) with a slight increase in the degree of supercooling (ΔT), and increase with 1.5 (G15) and 2 wt% (G20) with a decrease in ΔT as compared with the virgin (G00) (figure 3). It seems that the effect of iGO depends on its content. At low iGO contents, iGO acts as effective cross-linker and slows crystallization, while iGO particles act as nucleating agents at high concentration.

Zoom In Zoom Out Reset image size

For the cyclic test, a sample was stretched at T_m + 20 °C and the strain was fixed at T_m − 20 °C. A high degree of crystallization during the cooling step is essential for a high shape fixity ratio. Consequently the crystallization kinetics was studied using the Avrami equation:

$$\begin{eqnarray} &&1-\frac{{\emptyset }_{\mathrm{c}}}{{\emptyset }_{\infty }}={\mathrm{e}}^{-K{t}^{n}} \end{eqnarray} \tag{ 3 }$$

where ∅_∞ is the equilibrium crystallinity and ∅_c is the crystallinity at time t, both being determined from the DSC measurement. The Avrami exponent, n (2.04), and overall rate constant for crystallization, K (5.6 × 10⁻⁶), were determined from the Avrami plot prepared using the data for isothermal crystallization at T_m − 20 °C. The half-crystallization time is obtained with $\frac{{\emptyset }_{\mathrm{c}}}{{\emptyset }_{\infty }}=0.5$ in equation (3):

$$\begin{eqnarray} &&{t}_{\frac{1}{2}}=\left (\frac{\ln 2}{K}\right )^{\frac{1}{n}}. \end{eqnarray} \tag{ 4 }$$

Then, using the n and K data determined from the Avrami plot, the half-crystallization time is obtained. The typical half-crystallization time estimated using equation (4) was 4.6 min for G00. This implies that a cooling time over 10 min is necessary for crystallization.

The stress–strain behaviors of the cast films are shown in figure 4. The curves are the averages of five measurements, and detailed values including initial modulus, break strength, and elongation at break are shown in table 2.

Zoom In Zoom Out Reset image size

The reference PU which was cross-linked with HEA termini shows a typical yield point with elongation at break over 700%. The Young's modulus and strength are increased with G10 and G15 with a small decrease in ductility. However, G20 shows a small increase of Young's modulus and a small decrease of break strength as compared with the virgin PU. The iGO particles seemingly provide nucleating sites but disturb the chain orientation at high concentrations.

The typical cyclic loading and unloading behavior of the PU/graphene chemical hybrids operating between T_m + 20 °C and T_m − 20 °C are shown in figure 5 for the first four cycles and the detailed data are given in table 3. The shape fixity is 99% for the first cycle (N = 1) for all samples. With increasing N, G10 gives a small decrease while G15 and G20 give a small increase over the unfilled PU value. The decrease and increase are the same with the crystallinity behavior rather than the modulus. On the other hand, higher shape recovery ratios are maintained for G10 and G15 than for the unfilled case with increasing N, implying that the cross-linking is more pronounced with G10 and G15. It was experimentally found that the shape recovery was insufficient for G20, indicating that cross-linking was insufficient where iGO particles rather disturbed cross-linking reactions between the HEA termini.

Zoom In Zoom Out Reset image size

Figure 6 shows that the absorption of near-IR light by PU is significantly enhanced with iGO incorporation. Absorption increases with increasing iGO content, where the effect is more pronounced at low contents due to the fine particle dispersion. The fine particle dispersion at low content is seen from the SEM micrograph in figure 7. However, particle size increases with increasing iGO content.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 8 shows the typical optical images taken during the shape recovery process based on near-IR absorption. The shape recovery ratio for G00 is about 10% while that for G15 is about 90%, in 10 min.

Zoom In Zoom Out Reset image size

The shape recovery ratio versus irradiation time for all the hybrids is shown in figure 9 (table 4). It is obvious that the reference PU (G00) shows insufficient shape recovery upon exposure to the light along with a long induction period of about 4 min. For the hybrids, approximately one minute of induction is noted for all the samples. The initial shape recovery increases with the iGO content, while the equilibrium recovery ratio shows the opposite tendency. It is noted that the equilibrium shape recovery is about 90% for G10. The shape recovery of the semicrystalline polymer is accompanied by a morphology change from crystalline solid to random rubbery state. At the early stage of recovery via light absorption, crystalline mass is being melted and the fast morphology change is due to the melting of the crystalline soft segment of the PU. Consequently, high crystallinity and high ΔH_m should provide fast strain recovery. However, at a later stage of recovery, crystalline domains are melted and the morphology change is mainly driven by entropy elasticity where the particles disturb strain recovery. This should give a decrease in equilibrium shape recovery ratio with increase in the iGO content.

Zoom In Zoom Out Reset image size