Relation between temperature memory effect and multiple-shape memory behaviors based on polymer networks

Abstract

The temperature memory effect (TME) refers to the capability of shape memory materials to memorize the deformation temperature (T_d). In this article, the TME of multiple-shape memory polymer based on a PMMA/PEG semi-interpenetrating polymer network (PMMA/PEG semi-IPN) and a PMMA-PCL covalently crosslinked polymer co-network (PMMA-PCL CPN) are studied, and the relation between the TME and the multiple-shape memory effect (SME) is investigated. The results indicate that both the PMMA/PEG semi-IPN and PMMA-PCL CPN can show the TME in a specific range of response temperatures but not in its entire range of glass transition temperatures (T_g). This paper further confirms that broad transition temperature is the key to realizing the TME and the multiple-SME. Furthermore, we discover that the characteristic temperature of the TME corresponds to the multi-step transition temperature of the multiple-SME for the first time. Consequently, the characteristic temperature of the TME can be directly used as a multi-gradient T_d to achieve the multiple-SME. Thus it facilitates the selection of a multi-gradient T_d in the multiple-SME. The paper also reveals the internal relation of broad T_g, TME and multiple-SME of PMMA/PEG semi-IPN and PMMA-PCL CPN.

---

1. Introduction

In recent years, shape memory polymers (SMPs) have received more and more attention, due to their many potential applications such as in biomedicine, aerospace engineering, intelligent textiles, and switchable information carriers.^1–4 SMPs are a promising class of stimuli-responsive materials that have the ability to memorize temporary shapes and recover their permanent shape upon exposure to an external stimulus such as heat, light, electricity, magnetic field and chemistry.^5–14 The majority of the researches in the SMPs have been based on the traditional shape memory effect (SME) mentioned above, including novel structure designs of SMPs,^15–17 alternative shape recovery triggering mechanisms,^18–22 shape memory performance optimization^23–25 and innovative SMP applications.^26–29 In the last few years, significant progress has also been made beyond the traditional concept of polymer SME, which includes the discovery of multiple-shape memory effect (multiple-SME) that are able to memorize multiple temporary shapes in a single shape memory cycle,^30–41 two-way reversible SME^42–46 and temperature memory effect (TME).^47–52

With further advancement of the SMPs, however, many specific and complicated applications require more sophisticated SMPs with multi-transition temperature (T_trans) to memorize more than one temporary shape in a shape memory cycle, so that they can achieve the controllable recovery to meet more requirements. To some extent, the ultimate potential for SMPs is limited by the number of temporary shapes it can memorize in each shape memory cycle and the ability to tune the shape memory transition temperature for the targeted applications. The recently introduced TME might be a promising approach to reach the above mentioned goal. In conventional SMPs systems, the switching temperature (T_sw) is correlated to a thermal transition of the switching domains, which is typically varied by adjusting the polymer's molecular structure (e.g., change of the switching segment length or chemical composition), and requires synthesis of a new material. Recent studies^47–52 have shown that, the discovery of temperature-memory polymers (TMP) enables the realization of various response temperatures with the same polymer material, without requiring synthesis of a new material. TME was first discovered and proposed by Philippe Poulin in 2007.⁴⁷ It refers to the capability of a SMP to memorize the temperature at which it has been deformed. That is, for SMPs with temperature-memory capacity, they are capable of memorizing not just the strain, but also the thermomechanical history. Therefore, TMP can meet some requirement of specific and complicated applications. Moreover, the TMP are of academic and technological significance, which is a novel expansion of the conventional SMPs.

The broad thermal transition range (ΔT_trans) key to the tunable multi-SME has also been found responsible for the so-called TMP, which refers to the capability of a SMP to memorize the deformation temperature (T_d). Xie discovered that an amorphous polymer called Nafion displays both TME and multi-SME in its broad glass transition temperature range (ΔT_g).⁴⁸ Lendlein et al. reported the TME could be realized in a crystallizable polymer network with a broad melting temperature range (ΔT_m).⁵⁰ Currently, tunable multiple-SME has been reported,³⁴ however, the relationship between multi-SME and TME, and its mechanism has been researched less. Huang proposed a theoretical mechanical framework, and explained the SME and TME based on the viscoelasticity theory, which links the broad ΔT_trans to both the tunable multi-SME and TME. Unfortunately, he did not research the relationship between multi-SME and TME.⁵²

In our early work,^53,54 we discovered that both poly(methyl methacrylate)/poly(ethylene glycol) semi-interpenetrating polymer network (PMMA/PEG semi-IPN) and poly(methyl methacrylate)-poly(ε-caprolactone) covalently crosslinked polymer co-network (PMMA-PCL CPN) containing broad ΔT_g could display multiple-SME (including triple, quadruple and even quintuple). Specifically, they can be tailored on demand with much larger degree of adaptability by simply selecting suitable temperatures. Herein we will explore the TME of multiple-SMPs based on our early research about PMMA/PEG semi-IPN and PMMA-PCL CPN containing broad ΔT_g, and research the relation between TME and multiple-SME, which is necessary to have a better understanding of both effect for further exploring the potential of the materials.

2. Experimental section

2.1. Materials

PMMA/PEG semi-IPN and PMMA-PCL CPN were prepared by the same methods as our previous studies.^42,43 Based on our early research, the linear PEG content in PMMA/PEG semi-IPN was 35 wt%, PCLDMA content in PMMA-PCL CPN was 50 wt%.

2.2. Characterization

2.2.1. Dynamic Mechanical Analysis (DMA). DMA experiments were carried out in tensile loading mode using a DMA Q800 (TA instruments). The samples were cut into rectangular slabs with dimensions of 30 × 5 × 1 mm and the tension film mode was used with an amplitude of 20 μm, a frequency of 1 Hz. The samples were thermally equilibrated at −50 °C for 5 minutes and then heated to 150 °C at a heating rate of 3 °C min⁻¹. T_g was determined from the tan δ curves.

2.2.2. Shape memory characterization. Shape memory characterization was performed on a DMA Q800 (TA instruments) under a controlled force mode. Each test cycle consisted of a shape memory creation procedure (SMCP), where the temporary shape was created, and a recovery module under stress free conditions. (1) SMCP: during step 1 in the SMCP, the deformation force was applied to a target strain value at T_d and the force was kept constant during the subsequent cooling stage. Afterwards, the force was unloaded, resulting in the fixed strain representing the temporary shape. (2) Recovery under stress free conditions: before the strain recovery experiments, the samples were maintained under stress free condition for five minutes. All strain recovery experiments were carried out under stress free conditions. The strain recovery was performed in a staged heating fashion. The unloaded sample was heated to the respective recovery temperature at a heating rate of 3 °C min⁻¹. The shape memory properties were quantified by shape fixity ratio (R_f) and shape recovery ratio (R_r) based on the eqn (1) and (2):⁵⁵Here ε_u(N), ε_l(N) and ε_p(N) represented the strain after unloading, the maximum strain at σ = σ_m after cooling to T_low, and the strain after recovery in the Nth cycle, respectively.

2.2.3. Temperature memory effect characterization. The instantaneous recovery speed V_r were calculated using:⁴⁸Here the is the temperature derivative of strain directly obtained from the strain recovery data using the Universal Analyzer software (TA instruments).

3. Results and discussion

In our early work, we found that both PMMA/PEG semi-IPN and PMMA-PCL CPN exhibited considerably a broad ΔT_g in their dynamic mechanical analysis (DMA) curves.^53,54 As shown in Fig. 1, PMMA/PEG semi-IPN possessed a broad ΔT_g from 45 to 125 °C, and PMMA-PCL CPN possessed a broad ΔT_g from −10 to 90 °C. The broad glass transition region, which is called alpha transition, is believed to be the essential basis for its multi-SME.

Fig. 1 Tan δ-temperature curves of PMMA/PEG semi-IPN and PMMA-PCL CPN.

The TME observed in the stress recovery experiments can be called the stress based TME, and the characteristic temperature is determined at the maximum of the recovery stress (σ_max)T_σ,max under constant strain conditions. This effect can also be established in stress free strain recovery experiments, i.e., strain based TME.⁴⁸ Here the TME is established under such a condition. The switching temperature (T_sw) can be determined as inflection point of the strain-temperature recovery curve, in general, T_sw represents the temperature of maximum recovery speed (V_r,max), which is obtained under stress free recovery conditions.

We have investigated the dual-SME of PMMA/PEG semi-IPN at the deformation temperature T_d = 50 °C, 70 °C and 90 °C in the early research.⁵³ The result showed that the curves for dual-R_r shifted to the right with increasing the T_d, that is, stretching the sample at higher temperature led to shape recovering at correspondingly higher recovery temperature (T_r). This is a consequence of the TME. An inflexion region could be seen in the strain recovery curves of PMMA/PEG semi-IPN. The temperature of this inflexion point was close to T_d. As a result, we found PMMA/PEG semi-IPN possessed the so-called TME within its broad ΔT_g. It is suggested that the thermomechanical properties of material such as T_sw can be easily tailored by changing the applied T_d instead of synthesizing a new material with alternative material composition and structure. Based on the TME, we could realize the adjustable multiple-SME.

In the experiment of our early research about multiple-SME of PMMA/PEG semi-IPN, we found that not all the temperature in the ΔT_g could be used in the multiple-SME, as the shape could not be fixed at some temperature.⁵³ Can the PMMA/PEG semi-IPN show TME in its entire range of T_g? Accordingly, we began by evaluating the dual-SME of PMMA/PEG semi-IPN under a stress controlled deformation and stress free recovery condition at different T_d (T_d = 50, 60, 70, 80, 90, 100 °C). The sample was elongated to a target strain at respective T_d within the broad ΔT_g, then it was cooled down to 20 °C while keeping the external load constant to retain the deformation. After releasing the load, the temporary shape was fixed. When reheating the sample in a continuous manner, it gradually restored their initial shapes.

A representative dual-shape memory cycle (T_d = 80 °C) obtained under such conditions is presented in Fig. 2. Similar dual-shape memory cycles (see ESI Fig. S1†) can be constructed for other temperatures and used to calculate the shape fixity (R_f) and shape recovery (R_r). The result shows that the R_f and R_r for PMMA/PEG semi-IPN are above 85% when T_d was above the onset of the glass transition, which indicates that the IPN is an effective architecture for high performance dual-SME.

Fig. 2 A representative dual-shape memory cycle of PMMA/PEG semi-IPN at T_d = 80 °C. R_f = 97.0%, R_r = 97.7%.

The fact that PMMA/PEG semi-IPN exhibits excellent dual-shape memory properties in a broad ΔT_g (roughly 45–125 °C) makes it a suitable candidate to explore the strain based TME. To do so, all strain recovery experiments hereafter were conducted under a continuous linear temperature ramping mode. As such, the strain recovery curves of PMMA/PEG semi-IPN at different T_d (T_d = 50, 60, 70, 80, 90, 100 °C) are shown in Fig. 3a. In low temperature region, the temperature of inflexion point was close to the temperature where it was deformed. When the temperature exceeded 80 °C, the dual-R_r curves did not shift to the right with increasing the T_d, that is, the TME no longer exists.

Fig. 3 Shape recovery behavior for PMMA/PEG semi-IPN deformed at different T_d: 50, 60, 70, 80, 90, 100 °C. (a) Evolution of strain recovery rate (R_r). (b) Evolution of instantaneous strain recovery speed (V_r).

Data in strain recovery process were further used to calculate the instantaneous V_r using the eqn (3) shown in the experimental section. Since V_r are based on the temperature derivatives of the strains, they are, in principle, more sensitive than strain to reflect changes occurred in the recovery process.⁴⁸ We should note that the instantaneous V_r here represents the recovery speed at a certain temperature point, which is different from an average recovery speed that has been used to evaluate the recovery speed across a temperature range. As mentioned earlier, the temperature corresponding to the V_r,max is sometimes referred to as response or recovery temperatures (T_r).

To further improve the phenomena, we researched the V_r in the recovery process at different T_d. Data in Fig. 3a were further used to calculate the instantaneous V_r using the eqn (3) shown in the experimental section. The instantaneous V_r thus obtained were shown in Fig. 3b. In the respective curves obtained at different T_d, a peak (i.e., maximum V_r) can be observed at the corresponding T_d. The data presented in Fig. 3b clearly demonstrate a strong correlation between the applied T_d and the resulting T_r. As illustrated in Fig. 3b, in the low temperature region, when the deformation occurs at a certain T_d within the ΔT_g, the subsequent stress free strain recovery experiment shows that the peak of V_r appears at a temperature roughly the same as the corresponding T_d.

Fig. 4 displays the relation between T_r and T_d. As shown in Fig. 4, when the temperature was below 80 °C, the temperature corresponding to V_r,max were found to be in the range of the applied T_d, T_r increased systematically with rising T_d, that is, the corresponding T_r would depend on its initial T_d. So the PMMA/PEG semi-IPN was able to memorize its thermomechanical conditions, which demonstrates the successful temperature memory functionalization of the PMMA/PEG semi-IPN. However, the corresponding T_r did not increase with the increasing of T_d as the temperature exceeded 80 °C, that is, PMMA/PEG semi-IPN could not possess TME in the high temperature part of the transition temperature region.

Fig. 4 The relationship between response temperature (T_r) and deformation temperature (T_d) of PMMA/PEG semi-IPN.

The previous studies about multiple-SME of PMMA/PEG semi-IPN have shown that PMMA/PEG semi-IPN could only memory four shapes in the whole T_g range.⁵³ The temporary shape could not be fixed once the temperature exceeded 80 °C. The above research indicates that semi-IPN no longer possessed TME at temperature higher than 80 °C, although the broad ΔT_g was from 45 to 125 °C. As shown in Fig. 3 and 4, the strain recovery will occur at almost same temperature when T_d exceeded 80 °C, which confirms that TME plays a crucial role in multiple-SME. The result shows that it was based on the TME that PMMA/PEG semi-IPN could realize the adjustable multiple-SME, which means that multiple-SME could not be realized in the temperature range where the TME could not be achieved. Moreover, the realization of multiple-SME of PMMA/PEG semi-IPN is actually based on the characteristic temperature (T_c) of TME. As shown in Table 1, T_d used to realize quintuple-SME in our previous research was 110, 70, 50 °C, and the T_c for realizing TME was 90, 70, 50 °C. Although the first T_d (110 °C and 90 °C) was different, the response temperature was almost same. The reason is that the strain recovery will occur at almost the same temperature when T_d exceeded 80 °C.

Table 1 The characteristic temperature (T_c) for realizing TME and T_d for realizing quintuple-SME of PMMA/PEG semi-IPN

Tc (°C)	Td (°C)
90	110
70	70
50	50

---

The realization of multiple-SME is actually based on multiple phase structure or broad ΔT_trans (ΔT_g or ΔT_m). The selection of T_d for multiple-SME based on multiple phase structure is simple, as T_trans of each phase can be used as T_d directly. However, for multiple-SMPs based on broad ΔT_trans, the selection of T_d is uncertain. It requires us to choose different temperature in the broad ΔT_trans to try. As a result, there are no rules to follow. Here we find the T_d for realizing multiple-SME of PMMA/PEG semi-IPN is actually based on the T_c of TME. As a result, the T_c of TME can be directly used as multi-gradient T_d to achieve multiple-SME, thereby simplifying the selection of T_d in the multiple-SME.

In order to prove the above mentioned relation between TME and multiple-SME is not coincidental, that is, it is not belong to some special materials, we further investigated the PMMA-PCL CPN which possessed multiple-SME. In our previous research, we discovered the PMMA-PCL CPN containing broad ΔT_g exhibited TME, that is, when T_d was 0, 20, 40 °C, the curves for dual-R_r shifted to the right with increasing the T_d.⁵⁴ It is means stretching the sample at higher temperature led to shape recovering at correspondingly higher recovery temperature T_r. We also found the versatile multiple-SME from dual to quadruple can be adjusted simply by shifting the corresponding T_d on demands.

Accordingly, we firstly evaluated the dual-shape memory performance of PMMA-PCL CPN under a stress controlled deformation and stress free recovery condition at different T_d (T_d = 0, 10, 20, 30,40, 50 °C) as mentioned above. It should be noted that, the shape fixing temperature (T_f) for all samples was −20 °C. Fig. 5 shows a representative dual-shape memory cycle (T_d = 30 °C) obtained under such conditions. Other dual-shape memory cycles at different T_d are shown in Fig. S2.† It can be concluded from the quantitative results that the IPN is an effective architecture for high performance dual-SME.

Fig. 5 A representative dual-shape memory cycle of PMMA-PCL CPN at T_d = 30 °C. R_f = 97.9%, R_r = 97.3%.

The strain–temperature recovery curves of PMMA-PCL CPN at different T_d (T_d = 0, 10, 20, 30, 40, 50 °C) are presented in Fig. 6a. In the low temperature region, increase of the T_d resulted in the dual-R_r curves shifting to the right (the higher temperature region), that is, the corresponding T_r would depend on its initial T_d. Therefore the PMMA-PCL CPN was able to memorize its thermomechanical conditions. Whereas, once the temperature exceeded 30 °C, the dual-R_r curves did not shift to the right with increasing the T_d, that is, the TME no longer exists. To further improve the phenomena, we make further investigation on the strain recovery speed at different T_d, as shown in Fig. 6b. T_r is referred as the temperature corresponding to the V_r,max. The relation between T_r and T_d is shown in Fig. 7. When the temperature was below 30 °C, T_r corresponding to V_r,max was found to increase almost linearly with rising T_d, which demonstrates the successful temperature memory capacity of the PMMA-PCL CPN. However, the corresponding T_r did not increase with the increasing of T_d as the temperature exceeded 30 °C, that is, PMMA-PCL CPN could not possess TME in the high temperature region of the broad ΔT_g.

Fig. 6 Shape recovery behavior for PMMA-PCL CPN deformed at different T_d: 0, 10, 20, 30, 40, 50 °C. (a) Evolution of strain recovery rate (R_r). (b) Evolution of instantaneous strain recovery speed (V_r).

Fig. 7 The relationship between response temperature (T_r) and deformation temperature (T_d) of PMMA-PCL CPN.

In the research about quadruple-SME of PMMA-PCL CPN, we found that the temporary shape could not be kept in the high temperature region of ΔT_g, such as the shape could not be fixed once the temperature exceeded 30 °C.⁵⁴ The above research indicates that the PMMA-PCL CPN no longer possessed TME at temperature higher than 30 °C, although the ΔT_g was from −10 to 90 °C. The strain recovery will occur at almost same temperature when T_d exceeded 30 °C, which confirms that TME plays a crucial role in multiple-SME. Furthermore, the realization of multiple-SME of PMMA-PCL CPN is based on the T_c of TME. For instance, as shown in Table 2, T_d for realizing multiple-SME in our previous research was 80, 20, 0 °C, and the T_c for realizing TME was 40, 20, 0 °C. Although the T_d (80 °C and 40 °C) was different, the T_r was almost same. The reason is that the CPN no longer possessed TME at high temperature region, and the strain recovery will occur at almost the same temperature when T_d exceeded 30 °C.

Table 2 The T_c for realizing TME and T_d for realizing quintuple-SME of PMMA-PCL CPN

Tc (°C)	Td (°C)
40	80
20	20
0	0

---

The above two examples show that the tunable multi-SME and TME rely both on the broad △T_trans is not a coincidence. The temperature memory capability of such polymers is a result of the fact that the temporary shape is solely fixed by the polymer chains, which are in the rubbery state at T_d. The so established temporary shape remains stable until the fixing domains start to get flexible, when the materials are heated to T_d again and in this way allowing the recovery of the original shape. In fact, the concept of elemental memory unit (EMU) used to interpret the tunable multi-SME can be readily applied to explain the TME.^34,56 According to the concept, only the EMUs with T_trans below T_d can be activated for the strain fixing function. In the corresponding recovery process, 100% of EMUs are re-activated for recovery only when the temperature reaches T_d. Consequently, a peak of recovery speed is observed. Such a mechanistic interpretation is consistent with the theoretical mechanical framework proposed by Sun et al.,⁵² which also links the continuous broad thermal transition to both the tunable multi-SME and TME.

4. Conclusions

In summary, the above results indicate that both the PMMA/PEG semi-IPN and PMMA-PCL CPN can show the TME in a specific range of response temperature but not in its entire range of broad △T_trans. Furthermore, the T_c of TME is corresponding to the multi-step T_trans of multiple-SME. Consequently, the T_c of TME can be directly used as multi-gradient T_d to achieve multiple-SME, thereby simplifying the selection of T_d in the multiple-SME. This paper further confirms that broad ΔT_trans is the key of realizing TME and multiple-SME, and only in the broad ΔT_trans with TME can the tunable multiple-SME be realized. It also reveals the internal relation of broad T_g, TME and multiple-shape memory capability of PMMA/PEG semi-IPN and PMMA-PCL CPN. It is expected that the relation of TME and multiple-SME of PMMA/PEG semi-IPN and PMMA-PCL CPN will promote further understanding of multiple-SME and TME, and drive the development of multiple-SMPs for the broader potential applicability.

Acknowledgements

The authors would like to thank the National Natural Science Foundation of China (Grant no. 51173185 and 51303179) for the financial support of this research.

Notes and references

1. M. C. Serrano and G. A. Ameer, Macromol. Biosci., 2012, 12, 1156.
2. R. Vaia and J. Baur, Science, 2008, 319, 420.
3. J. L. Hu, Y. Zhu, H. H. Huang and J. Lu, Prog. Polym. Sci., 2012, 37, 1720.
4. H. J. Yoo, Y. C. Jung, N. G. Sahoo and J. W. Cho, J. Macromol. Sci., Part B: Phys., 2006, 45, 441.
5. J. Li, J. A. Viveros, M. H. Wrue and M. Anthamatten, Adv. Mater., 2007, 19, 2851.
6. T. Liu, J. Li, Y. Pan, Z. H. Zheng, X. B. Ding and Y. X. Peng, Soft Matter, 2011, 7, 1641.
7. A. Lendlein, H. Jiang, O. Junger and R. Langer, Nature, 2005, 434, 879.
8. H. Koerner, G. Price, N. A. Pearce, M. Alexander and R. A. Vaia, Nat. Mater., 2004, 3, 115.
9. X. F. Luo and P. T. Mather, Soft Matter, 2010, 6, 2146.
10. H. B. Lu, Y. J. Liu and J. S. Leng, Macromol. Mater. Eng., 2012, 297, 1138.
11. R. Mohr, K. Kratz, T. Weigel, M. Lucka-Gabor, M. Moneke and A. Lendlein, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 3540.
12. M. Y. Razzaq, M. Behl and A. Lendlein, Adv. Funct. Mater., 2012, 22, 184.
13. J. R. Kumpfer and S. J. Rowan, J. Am. Chem. Soc., 2011, 133, 12866.
14. Z. Q. Dong, Y. Cao, Q. J. Yuan, Y. F. Wang, J. H. Li, B. J. Li and S. Zhang, Macromol. Rapid Commun., 2013, 34, 867.
15. R. A. Weiss, E. Izzo and S. Mandelbaum, Macromolecules, 2008, 41, 2978.
16. A. M. Kushner, J. D. Vossler, G. A. Williams and Z. B. Guan, J. Am. Chem. Soc., 2009, 131, 8766.
17. J. Li, J. A. Viveros, M. H. Wrue and M. Anthamatten, Adv. Mater., 2007, 19, 2851.
18. A. Lendlein, H. Y. Jiang, O. Junger and R. Langer, Nature, 2005, 434, 879.
19. R. Mohr, K. Kratz, T. Weigel, M. Lucka-Gabor, M. Moneke and A. Lendlein, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 3540.
20. Q. Ge, K. Yu, Y. F. Ding and H. J. Qi, Soft Matter, 2012, 8, 11098.
21. Q. Ge, X. F. Luo, C. B. Iversen, P. T. Mather, M. L. Dunna and H. J. Qi, Soft Matter, 2013, 9, 2212.
22. K. Yu, Q. Ge and H. J. Qi, Nat. Commun., 2014, 5, 3066.
23. C. M. Yakacki, R. Shandas, D. Safranski, A. M. Ortega, K. Sassaman and K. Gall, Adv. Funct. Mater., 2008, 18, 2428.
24. I. A. Rousseau and T. Xie, J. Mater. Chem., 2010, 20, 3431.
25. C. M. Yakacki, S. Willis, C. Luders and K. Gall, Adv. Eng. Mater., 2008, 10, 112.
26. A. Lendlein and R. Langer, Science, 2002, 296, 1673.
27. J. Kunzelman, T. Chung, P. T. Mather and C. Weder, J. Mater. Chem., 2008, 18, 1082.
28. T. Xie and X. C. Xiao, Chem. Mater., 2008, 20, 2866.
29. T. Xie, X. C. Xiao, J. J. Li and R. M. Wang, Adv. Mater., 2010, 22, 4390.
30. I. Bellin, S. Kelch, R. Langer and A. Lendlein, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 18043.
31. M. Behl, I. Bellin, S. Kelch, W. Wagermaier and A. Lendlein, Adv. Funct. Mater., 2009, 19, 102.
32. J. Zotzmann, M. Behl, Y. K. Feng and A. Lendlein, Adv. Funct. Mater., 2010, 20, 3583.
33. M. Behl and A. Lendlein, J. Mater. Chem., 2010, 20, 3335.
34. T. Xie, Nature, 2010, 464, 267.
35. Y. W. Luo, Y. L. Guo, X. Gao, B. G. Li and T. Xie, Adv. Mater., 2013, 25, 743.
36. Z. W. He, N. Satarkar, T. Xie, Y. T. Cheng and J. Z. Hilt, Adv. Mater., 2011, 23, 3192.
37. S. J. Chen, J. L. Hu, C.-W. Yuen, L. Chan and H. T. Zhuo, Polym. Adv. Technol., 2010, 21, 377.
38. T. Pretsch, Smart Mater. Struct., 2010, 19, 015006.
39. T. Ware, K. Hearon, A. Lonnecker, K. L. Wooley, D. J. Maitland and W. Voit, Macromolecules, 2012, 45, 1062.
40. M. Bothe, K. Y. Mya, E. M. J. Lin, C. C. Yeo, X. H. Lu, C. B. He and T. Pretsch, Soft Matter, 2012, 8, 965.
41. Y. K. Bai, X. R. Zhang, Q. H. Wang and T. M. Wang, J. Mater. Chem. A, 2014, 2, 4771.
42. T. Chung, A. Rorno-Uribe and P. T. Mather, Macromolecules, 2008, 41, 184.
43. J. Zotzmann, M. Behl, D. Hofmann and A. Lendlein, Adv. Mater., 2010, 22, 3424.
44. S. J. Hong, W.-R. Yu and J. H. Youk, Smart Mater. Struct., 2010, 19, 035022.
45. M. Bothe and T. Pretsch, Macromol. Chem. Phys., 2012, 213, 2378.
46. S. Pandini, S. Passera, M. Messori, K. Paderni, M. Toselli, A. Gianoncelli, E. Bontempi and T. Riccò, Polymer, 2012, 53, 1915.
47. P. Miaudet, A. Derre, M. Maugey, C. Zakri, P. M. Piccione, R. Inoubli and P. Poulinet, Science, 2007, 318, 1294.
48. T. Xie, K. A. Page and S. A. Eastman, Adv. Funct. Mater., 2011, 21, 2057.
49. K. Kratz, U. Voigt and A. Lendlein, Adv. Funct. Mater., 2012, 22, 3057.
50. K. Kratz, S. A. Madbouly, W. Wagermaier and A. Lendlein, Adv. Mater., 2011, 23, 4058.
51. J. Cui, K. Kratz and A. Lendlein, Smart Mater Struct., 2010, 19, 065019.
52. L. Sun and W. M. Huang, Soft Matter, 2010, 6, 4403.
53. J. Li, T. Liu, S. Xia, Y. Pan, Z. H. Zheng, X. B. Ding and Y. X. Peng, J. Mater. Chem., 2011, 21, 12213.
54. J. Li, T. Liu, S. Xia, Y. Pan, Z. H. Zheng, X. B. Ding and Y. X. Peng, Macromol. Chem. Phys., 2012, 213, 2246.
55. W. Wagermaier, K. Kratz, M. Heuchel and A. Lendlein, Adv. Polym. Sci., 2010, 226, 97.
56. T. Xie, Polymer, 2011, 52, 4985.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental section, Fig. S1 and S2. See DOI: 10.1039/c4ra02600d

---