Introduction

Shape memory polymers (SMPs) represent a class of smart polymer materials that can memorize temporary shapes and recover their permanent shapes with respect to external stimuli [1–3]. These include thermoplastic [4–6] and thermosetting types of SMPs [7]. Compared with thermoplastic SMPs, thermosetting types have some advantages by greater thermal and chemical stability [8], higher shape recovery stress, higher shape fixity ratio, and better processability. Thermosetting SMPs, especially epoxy resin, have wide application prospects, such as robot hand [9] and spacecrafts [10]. Diglycidyl ether of bisphenol A (DGEBA) represents the most reported epoxy precursors for shape memory epoxy resins in the world, while DGEBA is largely dependent on petroleum-based raw material, bisphenol A, which is deemed to be toxic to living organisms [11]. Therefore, sustainable epoxy resins with non-toxicity should be developed as alternatives for DGEBA.

Recent years have witnessed rapid developments in bio-based polymers. Up to now, a large quantity of renewable materials, such as starch [12], cellulose [13], chitosan [14], lignin [15], vegetable oil [16], and rosin acid [17, 18], have been employed to synthesize bio-based polymeric materials. Among which, rosin which contains more than 90 % rosin acids with rigid structure similar to petroleum-based cycloaliphatic or aromatic compounds is an abundantly available natural product, with a global annual production of about 1.2 million tons [19]. Thus, many efforts are made to develop epoxy resins from rosin acids to acquire a suitable alternative to DGEBA [20–22], and our previous works proved that rosin-based epoxy and curing agents exhibit better or similar thermal and mechanical properties compared with their petroleum-based counterparts [23–25]. Moreover, rosin ring was demonstrated to have a significant influence on the shape memory properties of polymer in our previous work [26]. Rosin ring was incorporated into the main chain of polyurethanes, which resulted in excellent shape memory properties with recoverable strain of 960 %, nearly 2.5 times of the highest reported value (400 %), and shape recovery of 85 % [26].

In this paper, a rosin-based shape memory epoxy resin was synthesized. The study on the structure–property relationships was reported by tests on the epoxy monomer (DGEAPA, as shown in Scheme 1) other than rosin-based epoxy oligomers or mixtures [27–29]. For comparison, the petroleum-based counterpart containing planar benzene ring, diglycidyl ester of terephthalic acid, was also synthesized as the control. Following the syntheses, the epoxies were cured with poly(propylene glycol)-bis(2-aminopropyl ether) (D230), and their thermal and mechanical properties, especially the shape memory properties were systematically investigated. The results in this study might provide us new insight to develop the epoxy with good shape memory properties.

Scheme 1
scheme 1

Chemical structures of DGT, DGEAPA, and D230

Full size image

Experimental

Materials

Rosin acid (RA), acrylic acid (AA), benzyltriethylammonium chloride (TEBAC), cetyltrimethyl ammonium chloride (CTMAC), and poly(propylene glycol)-bis (2-aminopropyl ether) (D230) were purchased from Aladdin Reagent, China. Sodium hydroxide, petroleum ether, anhydrous ether, epichlorohydrin (ECH), dichloromethane, and terephthalic acid (TPA) were obtained from Sinopharm Chemical Reagent Co., China. All the chemicals were used as received.

Synthesis of acrylic pimaric acid (APA)

The reaction components included 15.1 g (0.05 mol) rosin acid, 4.4 g (0.06 mol) acrylic acid, and 0.1 g hydroquinone (as a polymerization inhibitor), which were added into a 250 mL three-necked flask equipped with a reflux condenser, a magnetic stirrer, and a N2 stream. The mixture was heated at 140 °C for 0.5 h, 160 °C for 0.5 h, and 180 °C for another 9 h under the protection of nitrogen. After that, the gaseous unreacted acrylic acid was removed by N2. Finally, the reaction was stopped and cooled to room temperature. The mixture was dissolved in anhydrous ether and precipitated in excess petroleum ether. The pale yellow precipitate was collected by filtration and washed three times with de-ionized water. The pale yellow crude product was recrystallized from acetone to obtain the pure APA which is a white, solid, and powder material weighing 10.5 g (yield: 56 %). The synthetic scheme is shown in Scheme 2.

Scheme 2
scheme 2

Synthesis procedure of DGEAPA from RA

Full size image

1H NMR (400 MHz, CDCl3, δ ppm): 5.37 (s, H), 2.50 (s, H), 2.31 (m, 2H), and 0.5–2.0 (m, 28H). FTIR (cm−1): 1162, 1276, 1459, 1693, 2869, and 2947.

Synthesis of diglycidyl ester of acrylic pimaric acid (DGEAPA)

The components, including 10.5 g (0.028 mol) APA, 69 g (0.75 mol) epichlorohydrin, and 0.0842 g benzyltriethylammonium chloride (TEBAC, as a phase transfer catalyst), were entered into a 100 mL three-necked round flask equipped with a reflux condenser and a magnetic stirrer. The mixture was stirred at 115 °C for 2 h. Then, the reaction was cooled to 60 °C, and 2.98 g (0.074 mol) sodium hydride and 4.200 g (0.074 mol) were added. After that, the reaction was continued at 60 °C for another 3 h. Finally, the mixture was cooled to room temperature and the solid precipitate was removed via filtration. The filtrate was diluted with 100 mL dichloromethane (CH2Cl2) and washed with 60 mL of de-ionized water three times. The CH2Cl2 layer was dried with anhydrous magnesium sulfate and concentrated using a rotary evaporator, giving 11.9 g (yield: 88 %) of yellowish liquid. The synthetic process is presented in Scheme 2.

1H NMR (400 Hz, CDCl3, δ ppm): 5.37 (s, H), 4.43 (d, H), 4.1 (t, H), 3.95 (m, H), 3.85 (m, H), 3.20 (m, 2H), 2.85 (m, 2H), 2.65 (m, 2H), 2.50 (s, H), 2.31 (m, 2H), and 0.5–2.0 (m, 28H). FTIR (cm−1): 840, 910, 1146, 1273, 1460, 1726, 2870, and 2940.

Synthesis of diglycidyl ester of terephthalic acid (DGT)

The reaction components, including 30.2 g (0.18 mol) of terephthalic acid, 273 g (2.95 mol) of epichlorohydrin, and 1.51 g of cetyltrimethyl ammonium chloride (CTMAC, as a phase transfer catalyst), were put into a 500 mL three-necked flask equipped with a magnetic stirrer and a reflux condenser. The mixture was stirred at 90 °C for 3 h and then cooled to room temperature and 45 g aqueous solution of sodium hydroxide was added. Finally, the reaction was heated to 30 °C for 6 h. The mixture was washed with deionized water three times. The organic part was concentrated by a rotary evaporator. At the end, the white power weighing 39 g (yield: 77 %) was obtained. The synthetic route is presented in Scheme 3.

Scheme 3
scheme 3

Synthesis procedure of DGT from TPA

Full size image

1H NMR (400 Hz, CDCl3, δ ppm): 8.15 (d, 4H), 4.71 (m, 2H), 4.20 (m, 2H), 3.39 (m, 2H), 2.92 (t, 2H), and 2.76 (m, 2H). FTIR (cm−1): 824, 982, 1460, 1074, 1117, 1276, 1502, 1577, 1720, 2956, and 3064.

Preparation of cured epoxy resins

Epoxy monomers (DGEAPA or DGT) and poly(propylene glycol)-bis(2-aminopropyl ether) (D230) with the desired ratios are shown in Table 1. Epoxy monomer was cured with curing agent (D230) in a 1:1 equivalent ratio. An appropriate amount of CH2Cl2 was added to dissolve the curing agent and epoxy before it was stirred for 20 min at room temperature to form a homogeneous mixture. The mixture was transferred into a vacuum oven at 60 °C for 2 h to remove CH2Cl2 and then poured into a stainless steel mold to be cured at 100 °C for 2 h, 130 °C for 2 h, and 170 °C for 4 h. All the samples were cured under the same conditions and removed from the stainless steel mold for dynamic mechanical analysis, and thermal and mechanical property tests.

Table 1 Compositions for the different systems
Full size table

Characterization

Fourier transform infrared (FTIR) spectra were recorded with Nicolet 6700 FTIR, scanning from 4000 to 400 cm−1.

1H NMR spectra were recorded with a Bruker 400 MHz Avance III spectrometer. It was conducted at room temperature with the solvent of deuterated chloroform (CDCl3).

The differential scanning calorimetry (DSC) measurement was carried out using a Mettler-Toledo TGA/DSCIunder nitrogen atmosphere. The sample weighed about 7 mg and the test was conducted from 0 to 130 °C at a heating rate of 5 °C min−1.

The thermogravimetric analyses (TGA)’ measurement was performed on a Mettler-Toledo TGA/DSC1 thermogravimetric analyzer (Switzerland) under a nitrogen atmosphere at a scanning rate of 10 °C min−1 from 50 to 600 °C. Samples each weighing about 5 mg were used for the tests.

The mechanical properties of the cured epoxy resins were analyzed by Instron 5567 Universal Mechanical Testing Machine at a speed of 10 mm according to GB/T 1040.2-2006. Tensile tests at room temperature were carried out using dumbbell-shaped specimens with gauge length of 12 mm. For this test, at least five specimens were measured for each sample to obtain an average value.

DMA was performed on a Mettler-Toledo DMA/SDTA 861e using a tensile mode at a frequency of 1 Hz. The samples were scanned from 25 to 180 °C at a heating rate of 5 °C min−1.

Rectangular specimens (75 × 7.1 × 1.6 mm3) were applied to evaluate the shape memory properties of the cured epoxy resins. The shape memory test was carried out according to the following steps: (1) the specimens were heated to a certain temperature (T g +30 °C); (2) the heated specimens were bent into “U shapes” and they were cooled down at room temperature for fixing the shape; (3) the U-shaped specimens were reheated to recover their original shapes at T g +30 °C and the shape recovery process was observed. The maximum angle of bending was tested as θ max. The initial deformation angle of the fixed sample was measured as angle θ fixed. The deformation angle after recovering process was measured as angle θ final.

The shape fixity (R f ) and shape recovery (R r ) were calculated using the following equations:

$$ Rf = [\theta_{\text{fixed}}/\theta_{ \hbox{max} }] \times 100\;\% $$
(1)
$$ Rr = [(\theta_{\text{fixed}} - \theta_{\text{final}})/\theta_{ \hbox{max} }] \times 100\;\% . $$
(2)

Results and discussion

Chemical characterization of DGEAPA and DGT

The chemical structures of DGT and DGEAPA were determined by FTIR and 1H NMR. The FTIR spectra of APA, DGEAPA, and DGT are presented in Fig. 1. In the FTIR spectrum of DGT, the characteristic peaks arising from C–O stretching vibration and C=O stretching vibration appear at 1276 and 1720 cm−1, respectively, the peaks at 2956 cm−1 belong to the stretching of C–H in oxirane ring, the peaks at 3060 and 1502 cm−1 correspond to the stretching absorptions of C–H and C=C in benzene ring, and the peaks at 1074, 1276, and 860 cm−1 belong to the C–O–C in oxirane ring. In the FTIR spectrum of APA, the peak at 1693 cm−1 represents the C=O stretching vibration of –COOH. After the reaction between APA and ECH, as shown in the FTIR spectrum of DGEAPA, ester bonds were formed which can be seen from the disappearance of the characteristic peak of C=O from –COOH and appearance of the peak at 1726 cm−1 corresponding to the stretching absorptions of C=O from ester bonds. Meanwhile, the peaks at 910 and 840 cm−1 belong to the characteristic absorptions of oxirane ring.

Fig. 1
figure 1

FTIR spectra of APA, DGEAPA, and DGT

Full size image

To further identify their structures, the 1H NMR spectra of APA, DGEAPA, and DGT are also shown in Fig. 2. The spectrum of DGT (Fig. 2a) exhibits the peaks of protons Ha, Ha’, Hb, He, Hf, and Hf of oxirane ring, the peaks of protons Hc, Hc’, Hd, and Hd, of CH2 next to the oxirane ring, and the peak of protons Hg on the benzene ring. In the spectrum of APA (Fig. 2b), the peak at 5.4 ppm is ascribed to the proton Hb attached to the unsaturated carbon on the rosin ring, the peaks at 2.2–2.6 ppm represent the protons Ha, Hc, and Hd of CH next to the C=C bond on the rosin ring and Hd of CH next to –COOH. In the spectrum of DGEAPA (Fig. 2c), there appear characteristic peaks of protons Ha, Ha’, Hb, Hd, Hd, and He on oxirane ring at 2.6–2.7, 2.8–2.9, and 3.1–3.3 ppm, and the peaks for protons Hc, Hc’, Hf, and Hf of CH2 next to oxirane ring at 3.8–4.5 ppm. In addition, the integrated area ratios of the characteristic signals for protons in each 1H NMR spectra were all in good agreement with the theoretical values. These results indicate that the target epoxy monomers were successfully synthesized.

Fig. 2
figure 2

1H NMR spectra of: a DGT, b APA, and c DGEAPA

Full size image

The cross-link density of the cured epoxy resins

The compositions of different epoxy systems are summarized in Table 1. The average molecular weights (M c ) can be used to reflect the cross-link density of the cured epoxy resins. The higher the M c , the lower would be the cross-link density. The average molecular weights (M c ) between cross-link points of different samples (listed in Table 1) are calculated using the following equation [30]

$$ Mc = (M_{\text{DGEAPA}} \times n_{\text{DGEAPA}} + M_{\text{DGT}} \times n_{\text{DGT}} + Md230 \times nd230)/nd230 $$
(3)

where M and n represent the molar mass and molarity of the corresponding component in the epoxy systems, respectively. As seen from Table 1, the M c values of the epoxy systems increase with the increase of DGEAPA content, since DGEAPA has molecular weight of 486 higher than that of DGT with a molecular weight of 278. This means that the cross-link density of the cured epoxy resins decreases with the incremental content of DGEAPA.

Mechanical properties of the cured epoxy resins

The tensile stress–strain curves for the cured epoxy resins are shown in Fig. 3 and the data for their tensile properties are listed in Table 2. Obviously, the stress of all cured epoxy systems presents linear increase with strain until final rupture without yielding. EP1 exhibits higher tensile strength and tensile modulus than EP0, and the tensile strength and tensile modulus increase with the incremental content DGEAPA content. This might be due to the rigid rosin ring of DEEAPA. In general, a polymer which contains more rigid units are usually more brittle [31], and this may be interpreted to lower elongation-at-break of EP1 compared with EP0 and EP0.5.

Fig. 3
figure 3

Tensile curves for different cured epoxy resins

Full size image
Table 2 Mechanical properties of the cured epoxy resins
Full size table

Dynamic mechanical properties of the cured epoxy resins

The dynamic mechanical properties of the cured epoxies were investigated by DMA and the curves are shown in Figs. 4 and 5. All the samples showed relatively constant storage modulus in their respective glassy and rubbery regions and the difference between the glassy modulus and rubbery modulus for each individual sample has been in 2–3 orders of magnitude (Fig. 4). Figure 5 indicates the loss tangent (tan delta) for different samples as a function of temperature. The peak temperature of the tan delta curves was recorded as the glass transition temperature (T g ) of the cured samples, the data are shown in Table 3. Clearly, the T g of the cured epoxy systems increased with the increasing content of DGEAPA. The T g of cross-linked polymers has close ties with their cross-link density and the chain segment structure [32–34]. Theoretically, a higher cross-link density or more rigid chain segment will produce a higher T g . The neat DGEAPA system (EP 1) has lower cross-link density but has much more rigid chain segment structure (rosin ring) than the neat DGT system (EP 0). By introducing DGEAPA, the impact of greater chain rigidity of the cross-linked networks on T g exceeds the impact of lowered cross-link density, corresponding to the increasing T g of cross-linked networks with the increase in DGEAPA content. The same trend in T g values was obtained by DSC measurement, as shown in Table 3.

Fig. 4
figure 4

DMA curves for epoxy series: storage versus temperature

Full size image
Fig. 5
figure 5

DMA curves for epoxy series: tan delta versus temperature

Full size image
Table 3 Thermal properties of the cured epoxy resins
Full size table

Thermal degradation behavior of the cured epoxy resins

The thermal stability and degradation behaviors of the cured epoxies were investigated by TGA. The TGA curves are shown in Fig. 6. The values of the initial degradation temperature for 5 % weight loss (T d5 ) and 50 % weight loss (T d50 ) are listed in Table 3. The neat DGEAPA system (EP1) exhibits higher T d5 and T d50 than the neat DGT system (EP0), and the T d5 and T d50 of the cured epoxy systems increased with the increasing content of DGEAPA. These results meant that the DGEAPA systems show higher thermal stability than DGT system. This might be due to the more rigid rosin ring of DGEAPA than benzene ring of DGT, which could be reflected from the higher modulus of DGEAPA systems than the neat DGT system [35–37].

Fig. 6
figure 6

TGA curves of cured systems

Full size image

Shape memory properties

The samples were tested at T g  + 30 °C using the U-type shape memory test (the T g values were obtained from DSC). Figure 7 shows a series of photographs of the shape recovery process of EP1 (55 × 5 × 1 mm) in a 145 °C oven. Using only 75 s, bended EP1 could almost recover to the original shape. Table 4 presents the data for the shape memory properties of the cured samples from the U-type shape memory test. As can be seen, all the samples show high shape fixed ratio (R f ) and recovery ratio (R r ). The R f of the cured samples increased with the increase of DGEAPA content and the neat DGEAPA system shows the highest R f (99.4 %). While the R r and shape recovery rate of the cured samples decreased with the increase of DGEAPA content. These might be due to the steric hindrance of the rosin ring structure. The steric hindrance of rosin ring structure resulted in the lower chain mobility of DGEAPA systems than the neat DGT system [38], corresponding to the higher shape fixity and slower shape recovery of DGEAPA systems than the neat DGT system. Figures 8 and 9 show the relationships between recovery time and recovery angle and the number of shape memory cycle of the shape memory epoxy series at T g  + 30 °C. As can be seen from Fig. 8, the recovery rate decreased at the last stage from 150° to 180°. This can be explained by the reason that the main stored strain energy has already been released at the earlier stages, resulting in the relatively slow recovery rate of the cured samples at the last stage [39]. After the cyclic U-type shape memory test, the shape recovery rate of all the cured samples reduced a little bit, as shown in Fig. 9.

Fig. 7
figure 7

Shape recovery process of sample EP1 in a 145 °C oven

Full size image
Table 4 Shape fixity, shape recovery, and shape recovery time of the cured epoxy resins
Full size table
Fig. 8
figure 8

Shape recovery time of the cured epoxy resins as a function of shape recovery angle

Full size image
Fig. 9
figure 9

Shape recovery time of the cured epoxy resins as a function of the number of shape memory cycle

Full size image

Conclusion

A rosin-based epoxy monomer DGEAPA and its petroleum-based counterpart containing benzene ring DGT (as the control) were successfully synthesized. The cured rosin-based epoxy exhibited higher thermal and mechanical properties, including glass transition temperature, thermal stability, tensile strength, and modulus. Both the cured DGEAPA and DGT showed great shape memory performance. Meanwhile, the steric rosin ring structure made the cured DGEAPA systems own excellent shape recovery fixity, while a bit lower shape recovery and shape recovery rate than the cured DGT systems. Thus, the rosin-based epoxy resin has a great potential to be used as shape memory material and the non-planar rosin ring structure could be used to regulate the shape memory properties of epoxy resins.