High-Temperature Shape Memory Behavior of Semicrystalline Polyamide Thermosets.

We have explored semicrystalline poly(decamethylene terephthalamide) (PA 10T) based thermosets as single-component high-temperature (>200 °C) shape memory polymers (SMPs). The PA 10T thermosets were prepared from reactive thermoplastic precursors. Reactive phenylethynyl (PE) functionalities were either attached at the chain termini or placed as side groups along the polymer main chain. The shape fixation and recovery performance of the thermoset films were investigated using a rheometer in torsion mode. By controlling the Mn of the reactive oligomers, or the PE concentration of the PE side-group functionalized copolyamides, we were able to design dual-shape memory PA 10T thermosets with a broad recovery temperature range of 227-285 °C. The thermosets based on the 1000 g mol-1 reactive PE precursor and the copolyamide with 15 mol % PE side groups show the highest fixation rate (99%) and recovery rate (≥90%). High temperature triple-shape memory behavior can be achieved as well when we use the melt transition ( Tm ≥ 200 °C) and the glass transition ( Tg = ∼125 °C) as the two switches. The recovery rate of the two recovery steps are highly dependent on the crystallinity of the thermosets and vary within a wide range of 74%-139% and 40-82% for the two steps, respectively. Reversible shape memory events could also be demonstrated when we perform a forward and backward deformation in a triple shape memory cycle. We also studied the angular recovery velocity as a function of temperature, which provides a thermokinematic picture of the shape recovery process and helps to program for desired shape memory behavior.

Introduction

Interest in shape memory pan class="Chemical">polymers (SMPs) has grown rapidly since the 1980s.[1−5] Typical applications are heat-shrink tubing, temperature sensors and actuators, biomedical and surgical materials, and aerospace devices.[6,7] Thermoresponsive SMPs are the most investigated systems that polymers are able to adopt a temporary shape upon deformation and to revert back to the permanent shape upon exposure to heat.[7,8] They are generally composed of a polymer network to maintain the permanent shape and a reversible switch responsible for the shape fixation and recovery.[8−11] The glass transition temperature (Tg) and melting temperature (Tm) are the two most important thermal transitions for thermoresponsive SMPs. Polyurethanes, polyesters, and (methyl)acrylate-based polymer networks have been investigated as SMPs based on the glass transition. Cross-linked semicrystalline networks or (multi)block copolymer systems have been developed to design Tm-based SMPs, such as polyolefins, polyethers, and polyesters.[6] However, most SMPs exhibit a switching temperature lower than 100 °C, which may not meet the requirements for high temperature aerospace, automotive, or electronic applications.

Recently, several examples of high-temperature (>200 °C) pan class="Chemical">SMPs have been reported.[12−21] Vaia and co-workers synthesized an amorphous fluorinated polyimide with a shape recovery temperature of 220 °C.[20] This was the first example of a single-component high-temperature SMP. However, this fluorinated polyimide can only be obtained as thin films, which excludes the possibility to produce complex shapes using melt processing methods. Weiss et al. have introduced ionic moieties into poly(ether ether ketone) (PEEK) leading to a Tg-based SMP with a switching temperature close to 200 °C, where the exact switching temperature depends on the metal counterions used.[14] In a later publication, this system was further developed into a Tm-based SMP by incorporation of sodium oleate, which displays a higher switching temperature of 230–240 °C due to the melting of sodium oleate.[15]

In contrast to the dual SMPs mentioned above, pan class="Chemical">triple SMPs featuring two independent temporary shapes were first reported by Lendlein’s group.[22] They require two programming steps and show two recovery steps. Triple-shape memory effects can be realized in a polymer material possessing two distinct thermal transitions[15,22−24] or one single broad transition such as a broad glass transition.[16,25] High temperature triple SMPs based on a single-component thermotropic liquid crystalline poly(esterimide) thermosets have recently been reported by Guan et al. The polymer shows two glass transition temperatures at ∼120 and ∼200 °C, both of which can be used as switches for triple SMPs.[21]

In previous publications we reported on the synthesis and (thermo)mechanical properties of semipan class="Chemical">aromatic polyamide thermosets based on poly(decamethylene terephthalamide) (PA 10T).[26,27] We showed that this class of polymers can be processed into semicrystalline thermoset films where the degree of crystallinity can be controlled. The melting/crystallization of the crystalline phase and the covalent network can be used as the high-temperature switch and permanent scaffold, respectively, for a dual shape memory effect (SME). More sophisticated triple and one-way reversible SME can be designed with the glass transition as the second switch. To the best of our knowledge, this is the first demonstration of a high-temperature SMP based on a single-component semicrystalline polyamide thermoset.

Experimental Section

Materials

The syntheses of phenylethynyl (PE) end-cappan class="Chemical">ped PA 10T oligomers (Scheme A) and PE side-group-functionalized copolyamides (Scheme B) have been described in detail elsewhere.[26,27] In order to obtain a covalently cross-linked polyamide network, the reactive precursors were thermally cured. A standard melt compression technique was used to prepare the thermoset films according to the following procedure: The precursor polymers were ground into a fine powder using a mortar and pestle. The powder was placed between two metal plates lined with Kapton film, and this stack was consolidated in a Joos hot-press using a predetermined temperature program and a 5 kN force. The temperature program was set to heat to 350 °C at 5 °C min–1, hold for 15 min, and cooled to 50 °C at 20 °C min–1. The obtained cured films were used for further characterization.

Scheme 1

Structures of Precursors via (A) Reactive End-Group Approach and (B) Reactive Side-Group Approach

The film samples prepared from PA 10T reactive oligomers (Mn of 1000 and 3000 g mol–1) are denoted as pan class="Chemical">PE-1K and PE-3K. TPE-5, TPE-10, and TPE-15 represent the film samples prepared from reactive TPE-copolyamides with 5, 10, and 15 mol % of reactive TPE comonomer, respectively. A PA 10T thermoplastic sample (Mn = 7500 g mol–1) was used as the reference in this paper and denoted as Ref. The 1H NMR spectra of the synthesized precursors (Supporting Information, Figure S1) confirms that the concentration of reactive functionalities, either as end-caps or as side groups, is consistent with the feed ratio of the monomers.

Characterization

1H NMR span class="Chemical">pectra were recorded on a 400 MHz Bruker WM-400 at 25 °C using trifluoroacetic acid-d as solvent. DSC was conducted on a PerkinElmer Sapphire DSC under a nitrogen atmosphere at a heating/cooling rate of 20 °C min–1. DMTA was performed on a PerkinElmer Diamond DMTA with film samples (0.2–0.3 mm thick) at a heating rate of 2 °C min–1 under a nitrogen atmosphere. Data were collected at a frequency of 1 Hz.

The SME was characterized in a cyclic torsion mode as illustrated in Figure . Compared to the traditional extension or bending tests,[28] torsion tests involve nonhomogeneous strains and stresses in the cross section of a rectangular bar and enable the SMPs to reach large pan class="Disease">deformations with moderate strains, which are believed to be more representative of practical applications of SMPs.[29−32]

Figure 1

Illustration of a shape deformation, fixation, and recovery cycle of the dual SME in torsion mode.

The shape programming and recovery cycles were pan class="Chemical">performed using a Thermofisher Haake MARS III rheometer equipped with a solid clamp geometry under a N2 atmosphere. Rectangular thin films with a width of ∼2.5 mm and thickness of ∼0.25 mm were loaded into the setup with a length of 15 mm between the clamps. The samples were deformed in a torsion mode at a constant strain rate of 0.1% s–1 equivalent to a rotation speed of 3.4° s–1.

The procedure of one cycle for a dual SME test includes the following steps: (1) heat the sample to the programming temperature (Tprog); (2) rotate the sample to the predetermined angle (ϕd); (3) keep the angle constant, cool to the fixation tempan class="Chemical">perature (Tf) at 10 °C min–1 and stabilize for 10 min; (4) remove the stress; (5) heat the sample to Tprog at 10 °C min–1 in a stress-free condition followed by an isothermal hold at Tprog for 10 min to stabilize. This cycle was repeated multiple times to characterize the reproducibility of the shape memory performance. The rotation angle of the sample was monitored and recorded during the whole test. For triple SME measurements, the sample was consecutively deformed in two steps in one programming cycle. Two different Tprogs were used, one being above Tm and the other between Tg and Tm.

Results and Discussion

Melting and Thermomechanical Properties

In general, either the Tg or the Tm transition of a pan class="Chemical">polymer can be applied to trigger a dual-shape memory event; however, to achieve high-temperature triple-shape memory behavior in a single-component polymer is challenging, as two high-temperature temporary networks with distinct rubbery plateaus are required.[15,22−24] One solution to realize this type of structure is to introduce moderate cross-links into a high-Tg semicrystalline polymer. In our previous work, we have explored a semiaromatic polyamide, PA 10T (Tg = ∼125 °C, Tm = ∼315 °C), as the base polymer to prepare semicrystalline polyamide thermosets. Phenylethynyl (PE) shows high cure-temperature of 330–370 °C and thus is able to provide a proper processing window for the PA 10T-based precursors. Therefore, we have synthesized a reactive PE-based end-cap and a reactive PE-based comonomer, which were copolymerized with terephthalic acid and 1,10-diaminodecane to yield curable PE end-capped oligomers and PE side-group functionalized copolymers. The polyamide thermosets were obtained after a subsequent thermal cure at a temperature of 350 °C for 15 min. PE-1K and PE-3K refer to the cured end-capped oligomers with Mn = 1000 and 3000 g mol–1, respectively. TPE-5, TPE-10, and TPE-15 represent the cured side-group functionalized copolyamides with 5, 10, and 15 mol % PE side groups, respectively.

The melting curves and properties of a thermoplastic pan class="Chemical">PA 10T reference polymer (Ref) and the resultant polyamide thermosets are shown in Figure and Table . The thermoset samples after cure are semicrystalline showing a Tm of 227–288 °C and ΔHm of 8–33 J g–1. These values are much lower than the Tm and the ΔHm of the reference polymer (Ref) (318 °C and 82 J g–1), which means the crystallizability of the polymer chains in the thermosets is strongly suppressed.

Figure 2

DSC heating scans (first heat) of (A) reference polymer (Ref), PE-3K, and PE-1K and (B) reference polymer (Ref), TPE-5, TPE-10, and TPE-15. The asterisk refers to melting peaks (N2 atmosphere and heating rate of 20 °C min–1).

Table 1

Thermal Properties of the Thermoplastic Reference Polymer (Ref) and the Cured Thermoset Samples

sample	Tma (°C)	ΔHm (J g–1)	Tgb (°C)	cross-linking densityc (kmol m–3)
Ref	318	82	127
PE-3K	283	33	127	0.68
PE-1K	244	8	129	0.72
TPE-5	285	33	123	0.94
TPE-10	250	27	123	1.79
TPE-15	227	20	125	2.34

Tm refers to the max of the melting peak as observed in DSC experiments.

Tg refers to the max of E″ as observed in DMTA experiments.

Cross-linking density (υ) was calculated using , where E′ is the storage modulus of cured films at 350 °C (T = 623 K) and R the universal gas constant (8.314 J K–1 mol–1).

DSC heating scans (first heat) of (A) reference polymer (Ref), pan class="Chemical">PE-3K, and PE-1K and (B) reference polymer (Ref), TPE-5, TPE-10, and TPE-15. The asterisk refers to melting peaks (N2 atmosphere and heating rate of 20 °C min–1).

Tm refers to the max of the melting peak as observed in DSC expan class="Chemical">periments.

Tg refers to the max of E″ as observed in pan class="Chemical">DMTA experiments.

Figure shows the thermomechanical behavior of the reference polymer (Ref) and the thermosets. The Ref film exhibits a pan class="Chemical">Tg at 127 °C, but this film fails at 297 °C because it has reached the melting point (Tm). In contrast to the reference polymer (Ref), the thermoset samples show two plateau regions (Tg – Tm and >Tm) in the DMTA profiles. The second plateau of the thermosets is stable up to ∼400 °C, which confirms the presence of a network structure. The Tgs of both thermoset samples remain virtually unchanged (123–129 °C) when compared to that of the reference polymer (Ref).

Figure 3

DMTA of (A) reference polymer (Ref), PE-1K, and PE-3K films and (B) reference polymer (Ref), TPE-5, TPE-10, and TPE-15 films. Heating rate 2 °C min–1 under N2 atmosphere and a frequency of 1 Hz. The insets show E″ at the glass transition temperature.

DMTA of (A) reference pan class="Chemical">polymer (Ref), PE-1K, and PE-3K films and (B) reference polymer (Ref), TPE-5, TPE-10, and TPE-15 films. Heating rate 2 °C min–1 under N2 atmosphere and a frequency of 1 Hz. The insets show E″ at the glass transition temperature.

The molecular structure of the resultant semicrystalline thermosets is depicted in Figure . PE groups combine at the cure tempan class="Chemical">perature resulting in cross-links with multiple after-cure chemical structures depending on cure temperature, time, and PE concentration. We attempted to investigate the cure chemistry using Raman and FTIR spectroscopy.[33,34] However, Raman spectroscopy failed due to a strong fluorescence background, and FTIR cannot detect the acetylene bond of PE because of IR insensitivity. Hence, direct evidence is not available to confirm the chemical change of the PE functionalities. The DMTA results of the cured samples show a rubbery plateau above Tm, and this plateau is stable up to ∼400 °C, which confirms that cross-linking has taken place during cure. The two-plateau thermal behavior enables triple-shape memory properties by taking both Tg and Tm transitions as the switches. The Tm transition can be solely used as a switch to design a high-temperature (>200 °C) dual-shape memory polymer.

Figure 4

Molecular representation of the semicrystalline PA 10T thermosets. Crystalline polymer (dark blue) is embedded in an amorphous matrix (light blue), and the red dots represent covalent cross-link points.

Molecular representation of the semicrystalline PA 10T thermosets. Crystalline pan class="Chemical">polymer (dark blue) is embedded in an amorphous matrix (light blue), and the red dots represent covalent cross-link points.

Dual-Shape Memory Behavior

Because of the semicrystalline nature of the synthesized polyamide thermosets, their melting and crystallization processes can be used as the thermoresponsive switch for dual-shapan class="Chemical">pe memory behavior. Three consecutive deformation, fixation, and recovery cycles in a torsion mode were conducted for each sample to test the shape memory performance over multiple cycles.

Shape fixation rate (Rf) and shapan class="Chemical">pe recovery rate (Rr) are the most important parameters to characterize the shape memory performance.[35]Rf describes how accurately the temporary shape can be fixed, and Rr quantifies the ability of the polymer to memorize its permanent shape. When performing the measurements in torsion mode, Rf and Rr can be calculated using eqs and 2.where φd, φf, and φr denote the rotational angle after deformation, at the fixed temporary shape at Tf, and after recovery, respectively. The instantaneous recovery velocity Vr can be calculated as the time derivative of the angle as shown in eq .Plotting Vr against temperature reveals the temperature range corresponding to the shape recovery process.[21] This provides a clear thermokinematic view of the shape recovery and helps to program for desired SME.

The thermoplastic reference film (ref) cannot be deformed above the melting temperature because of the absence of a rubbery plateau, thus the glass transition was used as the switch for the SME. As shown in Figure A, the sample was deformed at 220 °C, which lies between pan class="Chemical">Tg (127 °C) and Tm (318 °C), and the shape was fixed at 100 °C where the sample is in a glassy state. Therefore, the crystalline domains act as the scaffold, and the deformation of the polymer takes place in the amorphous state.

Figure 5

Three consecutive dual-shape memory cycles for (A) reference polymer (Ref), (B) PE-3K, and (C) PE-1K. Shape recovery velocity as a function of temperature in the second cycle for (D) reference polymer (Ref), (E) PE-3K, and (F) PE-1K. Cooling/heating rate 10 °C min–1 and N2 atmosphere.

Three consecutive dual-shape memory cycles for (A) reference pan class="Chemical">polymer (Ref), (B) PE-3K, and (C) PE-1K. Shape recovery velocity as a function of temperature in the second cycle for (D) reference polymer (Ref), (E) PE-3K, and (F) PE-1K. Cooling/heating rate 10 °C min–1 and N2 atmosphere.

In contrast to the reference polymer (Ref), pan class="Chemical">PE-3K and PE-1K thermoset films are stable up to ∼400 °C in DMTA experiments and exhibit two plateau regions (Tg – Tm and >Tm), as shown in Figure . This allows for deforming the film samples in the second plateau region (>Tm). Figures B and 5C show that PE-3K and PE-1K were deformed at 315 and 275 °C, respectively, which is about 30 °C higher than their Tms (283 and 244 °C). At this temperature the covalent network acts as the scaffold, and the temporary shape is fixed by crystallization and vitrification at the fixation step. At the recovery step, the angle shows a slight decrease when the sample passes through the glass transition, which is associated with the release of the stress trapped in the mobile amorphous region. A subsequent slow increase in the angle is observed, which can be explained by thermal expansion.[36] When the sample starts melting, the angle recovers rapidly due to the fast release of the trapped stress.

The shape-memory pan class="Chemical">performance of the first few cycles is usually not representative, as can be seen in Figure B,C. This is generally attributed to residual strain from the processing history of the sample.[13,35]Table shows the fixation and recovery results of the first three cycles of the reference thermoplastic polymer (Ref), the PE-3K, and PE-1K thermoset films. The PE-3K film shows a very low Rr of 36% in the first cycle compared to the following cycles (Rr = 72% and 76%), whereas the Rr of the PE-1K film shows a medium change over cycles (Rr = 81%, 90%, and 88%). Such significant difference can originate from the different cross-linking densities of these two samples. PE-3K has a lower cross-linking density and therefore longer polymer chains between cross-links, which is more likely to store residual strain. The recovery of both samples cannot reach 100% because the irreversible chain-segment orientation and the relaxation effect in the polymer network partly dissipate the stored entropy.[20]

Table 2

Dual-Shape Fixation and Recovery Results of the Reference Thermoplastic Polymer (Ref), PE-3K, and PE-1K Thermoset Films

			cycle 1		cycle 2		cycle 3
sample	Tproga (°C)	Trb (°C)	Rf (%)	Rr (%)	Rf (%)	Rr (%)	Rf (%)	Rr (%)
Ref	220	210	89	86	89	82	89	84
PE-3K	315	285	93	36	94	72	94	76
PE-1K	275	144, 240	99	81	98	90	95	88

Tprog refers to the programming temperature.

Tr refers to the temperature at the maximum recovery velocity.

The shape recovery velocity of the second cycle is plotted as a function of tempan class="Chemical">perature as shown in Figure D–F. The reference polymer (Ref) shows a low velocity (<20° min–1) through the recovery process because the shape recovery is triggered by the activation of the amorphous chain segments, which are strongly restricted by the crystalline domains. A low recovery velocity is generally observed in Tg-based shape-memory polymers.[8] In contrast, PE-3K shows a quick recovery in the temperature range of 260–310 °C and reaches a maximum Vr of ∼60° min–1 at 285 °C, which is close to its Tm (283 °C). The shape recovery is triggered by melting of the crystalline domains, which is a relatively fast process.

It is worthy to note that the PE-1K film was deformed at 275 °C in one step; however, this pan class="Chemical">polymer shows two distinct recovery steps with a maximum recovery velocity of 20–30° min–1 at around 144 and 240 °C, respectively. The first recovery takes place when the sample passes through the glass transition and reaches an Rr of ∼42% at 200 °C. The second recovery is triggered by the melting of the crystalline domains. These two recovery steps indicate that the crystalline domains in PE-1K are not able to form a penetrating scaffold throughout the sample and thus cannot completely lock the amorphous chains. This allows the sample to partially recover at the Tg – Tm range. These results demonstrate that the degree of crystallinity, which is reflected in the ΔHm values in Table , strongly affects the shape memory behavior of our semicrystalline PAs.

The dual-shape memory behavior of the pan class="Chemical">TPE-based thermosets were investigated using the same method. Table shows the shape fixation and recovery results for these samples. All samples reveal good shape fixation with Rf values of 93–99%. The TPE-15 samples with the highest cross-linking density shown in Table exhibit excellent recovery efficiency (Rr > 90%), whereas the other samples show Rr values of 62–66% in the first cycle and 78–82% in the second and third cycles.

Table 3

Dual-Shape Fixation and Recovery Results of the TPE Thermoset Films

			cycle 1		cycle 2		cycle 3
sample	Tproga (°C)	Trb (°C)	Rf (%)	Rr (%)	Rf (%)	Rr (%)	Rf (%)	Rr (%)
TPE-5	315	277	93	62	93	82	94	82
TPE-10	295	242	98	66	96	78	96	79
TPE-15	275	140, 227	99	93	99	93	97	96

Tprog refers to the programming temperature.

Tr refers to the temperature at the maximum recovery velocity.

Figure shows the dual-shape memory cycles and the recovery velocity of the pan class="Chemical">TPE thermoset films. Similar to the PE-3K sample, TPE-5 and TPE-10 exhibit one recovery step in the temperature ranges of 260–300 and 230–270 °C, respectively. Interestingly, the TPE-15 film also displays two recovery steps, which is similar to that of the PE-1K film. The two recovery steps originate from the low degree of crystallinity in TPE-15 and PE-1K, where the amorphous chain segments cannot be completely fixed when the temperature is above Tg. TPE-5, TPE-10, and PE-3K, on the other hand, show higher degrees of crystallinity, and this prevents the Tg-induced shape recovery.

Figure 6

Three consecutive dual-shape memory cycles for the TPE series (A) TPE-5, (B) TPE-10, and (C) TPE-15. Shape recovery velocity as a function of temperature in the second cycle for the TPE series (D) TPE-5, (E) TPE-10, and (F) TPE-15. Cooling and heating rate 10 °C min–1 and N2 atmosphere.

Three consecutive dual-shape memory cycles for the pan class="Chemical">TPE series (A) TPE-5, (B) TPE-10, and (C) TPE-15. Shape recovery velocity as a function of temperature in the second cycle for the TPE series (D) TPE-5, (E) TPE-10, and (F) TPE-15. Cooling and heating rate 10 °C min–1 and N2 atmosphere.

The difference between these samples clearly indicates that the shape memory behavior strongly depan class="Chemical">pends on the degree of crystallinity of the semicrystalline thermosets. By changing the concentration of the PE side groups in the reactive copolyamides, the cross-linking density of the thermosets can be controlled, consequently leading to adjustable crystallinity and shape memory behavior.

Triple-Shape Memory Behavior

Based on our study above, two reversible processes, the glass transition and the melting process, can both act as the switches to trigger shape recovery. Unlike the traditional high-tempan class="Chemical">perature triple-SMP composed of multiple components, we were able to design a single-component triple-SMP using both switches to achieve two distinct recovery processes.

Figure A shows the three consecutive triple-shape memory cycles for pan class="Chemical">PE-3K. The film sample was first heated up to 315 °C, which is above the Tm (283 °C) and rotated by 90° from the original shape A (φA) to a temporary shape B (φB). The sample was cooled to 200 °C, a temperature between Tm and Tg, to fix the shape B. A second rotation of 90° was then applied to reach the temporary shape C (φC). The sample was subsequently cooled to 60 °C, which is below the Tg of 127 °C, to fix the shape C. The external stress was then removed, leading to the final fixed temporary shape (φf). A shape fixation rate (Rf) of ∼90% was calculated using eq .

Figure 7

Three consecutive triple-shape memory cycles for the PE series (A) PE-3K and (B) PE-1K. Shape recovery velocity as a function of temperature in the second cycle for the PE series (C) PE-3K and (D) PE-1K. Cooling and heating rate 10 °C min–1 and N2 atmosphere.

Three consecutive triple-shape memory cycles for the pan class="Chemical">PE series (A) PE-3K and (B) PE-1K. Shape recovery velocity as a function of temperature in the second cycle for the PE series (C) PE-3K and (D) PE-1K. Cooling and heating rate 10 °C min–1 and N2 atmosphere.

In the subsequent recovery step, the sample was heated up to 315 °C in a stress-free state to monitor the recovery of the rotational angle. The shape recovery was accomplished in two distinct steps, as shown in Figure A. This means the thermoset can memorize two temporary shapan class="Chemical">pes in one single shape memory cycle. The shape recovery rate Rr(C→B) and Rr(B→A) were calculated based on eqs and 6, respectively.[21]where φA, φB, and φC denote the rotational angle of shape A (φA = 0°), shape B (φB = 90°), and shape C (φC = 180°); φB/rec and φA/rec are the rotational angles of the first and second recovered shapes, respectively. Figure B shows the triple-shape memory cycles of PE-1K. This polymer exhibits a lower Tm (244 °C) compared to PE-3K (283 °C); therefore, lower programming temperatures (275 and 150 °C) were used.

PE-3K shows moderate Rr(C→B) values of 67–80% in the first recovery step and Rr(B→A) values of 60–73% in the second recovery step, as listed in Table . Compared to pan class="Chemical">PE-3K, PE-1K shows much higher Rr(C→B) values of 128–138% and lower Rr(B→A) values of 40–49%. The first recovery is triggered by the glass transition and is due to the release of stress in the amorphous phase. The major recovery takes place in the first step, indicating that the crystalline phase cannot completely lock the stress in the amorphous phase. The stress, which is supposed to release in the second recovery step, is actually partially released in the first recovery step. This, together with the stress induced by the second deformation, drives the sample to reach a high recovery rate (>100%) in the first step. This result is consistent with the result we obtained from the dual-shape memory experiments and originates from the low degree of crystallinity of the PE-1K thermoset.

Table 4

Triple-Shape Fixation and Recovery Results of the PE- and TPE-Series Thermoset Films

					cycle 1			cycle 2			cycle 3
sample	Tprog1a (°C)	Tprog2a (°C)	Tr1b (°C)	Tr2b (°C)	Rf (%)	Rr(C–B) (%)	Rr(B–A) (%)	Rf (%)	Rr(C–B) (%)	Rr(B–A) (%)	Rf (%)	Rr(C–B) (%)	Rr(B–A) (%)
PE-3K	315	200	137	289	91	67	60	93	80	66	89	79	73
PE-1K	275	150	144	240	97	128	49	97	135	44	93	139	40
TPE-5	315	200	132	285	88	64	69	86	74	79	86	79	79
TPE-10	295	175	138	250	92	79	91	92	89	80	91	91	82
TPE-15	275	150	133	236	93	97	84	92	123	55	93	136	45

Tprog1 and Tprog2 refer to the first and second programming temperature, respectively.

Tr1 and Tr2 refer to the temperature at the first and second maximum recovery velocity, respectively.

Tprog1 and Tprog2 refer to the first and second programming temperature, respan class="Chemical">pectively.

Tr1 and pan class="Gene">Tr2 refer to the temperature at the first and second maximum recovery velocity, respectively.

Figures C and 7D show the shape recovery velocity of pan class="Chemical">PE-3K and PE-1K samples as a function of temperature. The first recovery step of PE-3K shows a lower Vr (13° min–1) than the second step (28° min–1) because of the limited mobility of the amorphous chain segments restricted by the crystalline domains. In contrast, the PE-1K sample shows much higher Vr (54° min–1) in the first step, which is due to its low degree of crystallinity.

The triple-shape memory cycles and recovery velocity of the pan class="Chemical">TPE thermosets are shown in Figure . The TPE-5 and TPE-10 films show moderate Rr(C→B) and Rr(B→A) (64–91%) and higher Vr values in the second recovery step, which is similar to the behavior of PE-3K, whereas the TPE-15 sample, similar to PE-1K, exhibits high Rr(C→B) (97–136%) and Vr values (38° min–1) in the first recovery step, because this polymer has the lowest degree of crystallinity.

Figure 8

Three consecutive triple-shape memory cycles for (A) TPE-5, (B) TPE-10, and (C) TPE-15. Angular velocity of shape recovery as a function of temperature in the second cycle for (D) TPE-5, (E) TPE-10, and (F) TPE-15. Cooling and heating rate 10 °C min–1 and N2 atmosphere.

Three consecutive triple-shape memory cycles for (A) pan class="Chemical">TPE-5, (B) TPE-10, and (C) TPE-15. Angular velocity of shape recovery as a function of temperature in the second cycle for (D) TPE-5, (E) TPE-10, and (F) TPE-15. Cooling and heating rate 10 °C min–1 and N2 atmosphere.

The temperatures at the first maximum recovery velocity (pan class="Gene">Tr1) of all samples are within a narrow range of 132–144 °C because the Tg of all samples is around 125 °C. However, the temperatures of the second maximum recovery velocity (Tr2), which depend on the Tm of the samples, display a broad variation between 236 and 289 °C. Therefore, the triple-shape memory behavior of the thermosets is highly tunable over a wide temperature regime.

Besides deforming the sample in one direction only, more interestingly, we can involve both forward and backward deformations in a triple-shapan class="Chemical">pe memory cycle; therefore, the corresponding temporary shapes would recover in a reverse direction. To demonstrate this, the TPE-10 thermoset sample, which shows a high Rf (>90%), Rr(C–B) (∼90%), and Rr(B–A) (∼80%), was selected as a representative example.

The results of test I are shown in Figure A,B. The sample was programmed forward and backward by 180° in two steps and reached a final temporary shape, which is identical to the original shapan class="Chemical">pe. The subsequent heating resulted in an increase in angle in the first recovery step achieving a maximum recovery velocity at 142 °C and an Rr(C–B) of 79–91% at 213 °C. By increasing the temperature further, the sample recovered in the reverse direction reaching a maximum recovery velocity at 243 °C and an Rr(B–A) of 81–87% (Table ). These results demonstrate that our semicrystalline PA thermosets show one-way reversible shape memory behavior. The reversible shape changes in this experiment occurred without application of any external force. The shape changes were driven by internal stress of the oppositely strained networks formed in the first and second programming steps.

Figure 9

Three consecutive triple-shape memory cycles for TPE-10 and shape recovery velocity as a function of temperature in the second cycle. (A) and (B) deformation of 180° and −180° in two steps, respectively; (C) and (D) deformation of 180° and −360° in two steps, respectively.

Table 5

Triple-Shape Fixation and Recovery Results of TPE-10 Deformed in Two Opposite Directions

test	cycle no.	Δφd1a (deg)	Δφd2a (deg)	Rf (%)	Rr(C–B) (%)	Rr(B–A) (%)	Tmaxb (°C)
I	1	180	–180	97	91	81	213
	2	180	–180	98	85	81	211
	3	180	–180	99	79	87	209
II	1	180	–360	90	86	54	218
	2	180	–360	89	83	59	213
	3	180	–360	91	82	61	212

Δφd1 and Δφd2 refer to the deformation angles in the first and second deformation steps.

Tmax refers to the temperature at the maximum angle.

Three consecutive triple-shape memory cycles for pan class="Chemical">TPE-10 and shape recovery velocity as a function of temperature in the second cycle. (A) and (B) deformation of 180° and −180° in two steps, respectively; (C) and (D) deformation of 180° and −360° in two steps, respectively.

Tmax refers to the tempan class="Chemical">perature at the maximum angle.

We can adjust the second programming step from Δφd2 = −180° to Δφd2 = −360° in test II, which results in two opposite temporary shapes, as shown in Figure C,D. The sample shows a Rf of 89–91%, Rr(C–B) of 82–86%, and Rr(B–A) of 54–61% (Table ). The maximum recovery velocity occurs at 142 and 244 °C in the first and second recovery steps, respan class="Chemical">pectively, which is consistent with the results of Test I. Our results clearly demonstrate that high-temperature SME with tunable recovery directions and amplitudes can be designed based on our single-component semicrystalline polyamide thermosets.

Conclusions

We have investigated two novel series semicrystalline PA thermosets and demonstrated that they can be as single-component high-tempan class="Chemical">perature (>200 °C) shape memory polymers (SMPs). Two molecular design approaches based on reactive phenylethynyl (PE) functionalities have been explored: cross-linked semicrystalline PA films were prepared by curing reactive thermoplastic PA 10T oligomers, and films were prepared by curing reactive thermoplastic side-group functionalized copolyamides. Compared to the thermoplastic PA 10T reference polymer, the PE-3K, TPE-5, and TPE-10 thermoset films show high-temperature dual-shape memory behavior (>200 °C) when Tm is used as the switching temperature. The densely cross-linked PE-1K and TPE-15 films show the highest fixation rate (99%) and recovery rate (≥90%) and two distinct recovery steps because the degree of crystallinity is too low and cannot provide a scaffold that locks the shape at temperatures between Tg and Tm. Triple-shape memory behavior can be demonstrated when the Tg (∼125 °C) is used as the second switching temperature. The recovery rate of the two recovery steps is highly dependent on the degree of crystallinity of the thermosets and vary within a broad range of 74%–139% and 40–82% for the first and second step, respectively. One-way reversible shape memory events can also be designed when we perform forward and backward deformation in a triple shape memory cycle. We also studied the recovery velocity as a function of temperature to elaborate the thermokinematics of the shape recovery process. The use of phenylethynyl reactive functionalities allows us to control the degree of crystallinity in the final polyamide thermosets and in turn tune their shape memory behavior in terms of recovery temperature, velocity, and efficiency. We believe that the design rules presented herein will help in designing new shape memory polymers based on well-known semicrystalline polyamide chemistries.