- Email: [email protected]
The role of cellulose nanocrystals in the improvement of the shape-memory properties of castor oil-based segmented thermoplastic polyurethanes
Composites Science and Technology 92 (2014) 27–33
Contents lists available at ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
The role of cellulose nanocrystals in the improvement of the shape-memory properties of castor oil-based segmented thermoplastic polyurethanes Ainara Saralegi a, Maria Luz Gonzalez b, Angel Valea b, Arantxa Eceiza a,⇑, Maria Angeles Corcuera a,⇑ a ‘Materials + Technologies’ Group, Department of Chemical and Environmental Engineering, Polytechnic School, University of the Basque Country UPV/EHU, Pza. Europa 1, 20018 Donostia-San Sebastián, Spain b Department of Chemical and Environmental Engineering, School of Technical Industrial Engineering, University of the Basque Country UPV/EHU, Paseo Rafel Moreno ‘Pitxitxi’ 3, 48013 Bilbao, Spain
a r t i c l e
i n f o
Article history: Received 12 August 2013 Received in revised form 18 November 2013 Accepted 4 December 2013 Available online 16 December 2013 Keywords: A. Polymer–matrix composites A. Nanocomposites Polyurethane B. Fiber/matrix bond A. Smart materials
a b s t r a c t The effect of the addition of cellulose nanocrystals on segmented thermoplastic polyurethane shape-memory properties was investigated. To this end, polyurethane bionanocomposites were synthesized by in situ polymerization, adding cellulose nanocrystals in the first step of polymerization and using components from renewable resources in polyurethane formulation. Thereby, microphase separated polyurethane/cellulose nanocrystal bionanocomposites were obtained, presenting two main transitions, soft and hard phase melting temperatures. Furthermore, it was observed that due to the addition of cellulose nanocrystals, hard phase crystallinity increased. Regarding thermally-activated shape-memory properties, a significant improvement was observed in the shape recovery values with the addition of cellulose nanocrystals, being these bionanocomposites good candidates to use in applications where shape-memory properties are required. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Segmented polyurethanes are an important class of polymers that have found many applications as biomaterials due to their excellent physical properties and relatively good biocompatibility [1–3]. Technically, segmented thermoplastic polyurethanes (STPU) are block copolymers with alternating amorphous or crystalline hard and soft segments, denoted as HS and SS, respectively, that separate into microphases or domains, due to the thermodynamic incompatibility between both segments [4–7]. Therefore, manipulating their composition and choosing properly the chemical structure of the components, a wide variety of polyurethanes can be synthesized with different types of molecular architectures, specifically designed for each application [8,9]. In the last decades, the use of components from renewable resources have received steadily growing attention in polyurethane synthesis, owing to the dominant consumption patterns, the limited fossil resources and their increasing cost, as well as the public concern about climate change. Polyols derived from vegetable oils [10–14] are a good example of materials that can be used to synthesize segmented polyurethanes with a high percentage of components derived from ⇑ Corresponding authors. Tel.: +34 943017186; fax: +34 943017130. E-mail addresses: [email protected] (A. Eceiza), [email protected] (M.A. Corcuera). 0266-3538/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compscitech.2013.12.001
renewable sources. Moreover, the use of nanoreinforcements also from renewable resources, such as cellulose nanocrystals, chitin nanocrystals and bacterial cellulose [15–18], are gaining increasing interest, because they are used to improve the final properties of the synthesized nanocomposites, as well as to increase the amount of carbon from renewable resources in nanocomposites formulation. The use of segmented polyurethanes as shape-memory polymers, mainly as thermally-activated shape-memory (TASM) polymers, has received increasing attention in the last few years because of their applications in microelectromechanical systems, actuators, for self healing and health monitoring purposes, and in biomedical devices [19–21]. TASM polymers consist of two components on the molecular level: first, molecular switches, which are segments having a thermal transition at a determined temperature (Ttrans) and responsible for shape fixity by forming physical crosslinks; and second, network points that link these segments and determine the permanent shape [22–25]. Regarding segmented polyurethanes, in almost all cases, SS acts as the switching segment, and HS determines the permanent shape. The main objective of this work was the synthesis of polyurethane/cellulose nanocrystal bionanocomposites with high contents of carbon from renewable resources by in situ polymerization, and the posterior TASM characterization of the synthesized bionanocomposites, in order to study the role of cellulose nanocrystals
28
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
(CNC) in shape-memory properties. To this end, thermal and mechanical properties of the synthesized bionanocomposites were determined, as well as their TASM properties by cyclic thermomechanical tests. 2. Materials and methods 2.1. Materials Concentrated sulphuric acid (H2SO4, >98%) was purchased from Sigma–Aldrich Corporation (Missouri, United States), as well as tetrahydrofuran (THF, >99.9%) and microcrystalline cellulose (MCC) with a particle size between 100 and 150 lm. For polyurethane synthesis, 1,6-hexamethylene diisocyanate (HDI) supplied by Bayer (Leverkusen, Germany) under the trade name Desmodur H and corn-sugar based 1,3-propanediol (PD) supplied by Quimidroga SA (Barcelona, Spain) were used as hard segment. Poly(butylene sebacate) diol was used as soft segment, a difunctional semicrystalline polyester derived from castor oil (Tg = 54 °C, Tm = 67 °C and DHm = 138.26 J g1), denoted as CO2, with a molecular weight of 1900 g mol1, which has been already used and characterized in previous works [6,26]. Moreover, the quantity of carbon from renewable resources of CO2 was determined by ASTM D 6855, resulting in a value of 70%. 2.2. Isolation of cellulose nanocrystals Cellulose nanocrystals were prepared by acid hydrolysis of MCC using the following method: 20 g of MCC were dispersed in 500 mL of 64% (w/w) aqueous H2SO4 and stirred constantly for 30 min at 45 °C. The resultant suspension was subsequently diluted and washed by continuous centrifugation with deionized water. For removal of the remaining acid, nanocrystals were dialyzed against deionized water until the pH of the dialysis water stayed constant. Finally, a loose powder was obtained by freeze-drying the CNC suspension at a concentration of 0.05 mg mL1. Therefore, CNC with a rod-like morphology were isolated with an average diameter of 8.3 ± 0.9 nm and an average length of 152 ± 21 nm, resulting in an aspect ratio of 18.3 ± 2.7. 2.3. Synthesis of polyurethane bionanocomposites Polyurethane bionanocomposites were synthesized by in situ polymerization with different CNC contents and 17 wt% of HS content (with a molar ratio of 1:2:1 among components, polyol:diisocyanate:chain extender + CNC), using a two step procedure in THF solution [15]. As sketched in Fig. 1, prepolymers with different chemical structures can be formed. Although not shown in the scheme, prepolymers with isocyanate and hydroxyl ending groups from cellulose nanocrystals could form longer prepolymer structures. In the second step of polymerization, the chain extender was added, enabling the coupling of these prepolymers and forming different hard segment structures. To avoid the agglomeration of cellulose nanocrystals, the freeze-dried CNCs were redispersed in THF at a solid content of 0.5 wt%, without additives or any surface modification. First of all, the prepolymer was synthesized at 100 °C for 6 h by reacting HDI with CO2 in the presence of CNCs, previously dispersed in THF. The concentration of the resulting prepolymer was 75 mg mL1 and a condenser was used to avoid THF evaporation. Afterwards, the chain extender, 1,3-propanediol, was added in stoichiometric amount to the prepolymer, and the mixture was stirred at 100 °C for 2 h before casting and curing on a Teflon dish at 60 °C for 48 h under vacuum, in order to obtain a film. The amount of reactive hydroxyl group on cellulose nanocrystals was measured
by titration of excessive isocyanate groups, resulting in a value of 2.1 ± 0.2 mmol g1. This polymerization method allows obtaining bionanocomposites with cellulose nanocrystals covalently linked to the polyurethane chains. Furthermore, the neat polyurethane was also synthesized following the procedure used for bionanocomposites synthesis, but without adding CNCs. Thereby, a neat polyurethane, denoted as STPU17, and bionanocomposites with different CNC contents ranging from 0.25 to 2 wt% were successfully synthesized and denoted as STPU17/CNC-X, being X the amount of CNC added (wt%). 2.4. Characterization techniques The thermal behavior of the neat polyurethane and the bionanocomposites was analyzed by differential scanning calorimetry (DSC) and dynamic mechanical analysis (DMA). On one hand, DSC scans were recorded on a Mettler Toledo 822e instrument (Ohio, United States), equipped with a robotic arm and with an electric intracooler as refrigeration unit. Samples were scanned from 60 to 220 °C at a scanning rate of 20 °C min1, operating under dry nitrogen atmosphere. Moreover, the crystallization process was followed by cooling the samples from 220 to 60 °C at a scanning rate of 10 °C min1. A second heating run was also performed, but as it did not add any extra information, it was not included in this work. On the other hand, DMA measurements of the neat polyurethane and the bionanocomposites were performed in tension mode on a DMA Q-800 (TA Instruments, Delaware, United States). A constant frequency of 1 Hz was used, and samples were scanned from 100 to 150 °C at a rate of 2 °C min1. Thermally-activated shape-memory properties of the synthesized neat polyurethane and the bionanocomposites were also characterized in tension mode on a DMA Q-800, performing cyclic tests at 60 °C. Samples were first conditioned at 60 °C for 15 min and subsequently elongated to 50% applying a force ramp of 0.1 N min1. Then, samples were cooled to 15 °C at a rate of 5 °C min1 and the stress was released maintaining the temperature at 15 °C. Finally, samples underwent the recovery process by being heated at a rate of 5 °C min1 to the recovery temperature, 60 °C. This temperature was selected as the transition temperature because above 60 °C SS crystallites melt, acting as switching segments, whereas HS crystallites restricted molecular motion, being responsible for shape recovery. Shape fixity (Rf) and shape recovery (Rr) values were determined for each thermo-mechanical cycle, according to the equations presented in literature [19,22]. The mechanical properties of the synthesized films were measured on a MTS Insight 10 instrument (MTS Systems Corporation, Minnesota, United States), equipped with a 250 N load cell. Dogbone shape specimens (22.25 mm in length, 5 mm in width and 0.7 mm in thickness) were used for tensile tests, according to ASTM D 1708-93 standard. Five replicates of each material were used to measure Young’s modulus (E), yield strength (ry), tensile strength at maximum elongation (rmax) and strain at break (eb) values, calculated from the load–displacement data, where the deformation was measured by the crosshead displacement. Moreover, the work-of-fracture in the elastic region (8% strain) and the total work-of-fracture were also measured integrating the area under stress–strain curves obtained by tensile tests. Finally, atomic force microscopy images (AFM) were obtained in tapping mode on a Nanoscope IIIa scanning probe microscope Multimode™ Digital Instrument (Bruker Corporation, Massachusetts, United States), equipped with an integrated silicon tip/cantilever having a resonance frequency of 300 kHz. Samples cross sections were prepared using a Leica EM FC6 cryo-ultramicrotome (Leica Microsystems, Wetzlar, Germany) equipped with a diamond knife and operating at 120 °C, in order to observe the dispersion of CNCs in the neat polyurethane. The measurements were performed
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
29
Fig. 1. Schematic representation of the procedure used for polyurethane bionanocomposites synthesis, showing the chemical structures of the prepolymers and final bionanocomposites, as well as the chemical structures of the main components. The molar ratio of the components was 1:2:1 for polyol:diisocyanate:chain extender + CNC.
at a scan rate of 1.0 Hz, using a scan head with a maximum range of 16 lm 16 lm. Several regions were scanned obtaining similar results.
3. Results and discussion As was explained earlier, shape memory polymers require two components or segments on the molecular level, which are switching segments (amorphous or crystalline) that fix the temporary shape, and network points (achieved by either physical interactions or chemical bonds), that determine the permanent shape. Therefore, the thermal characterization of the synthesized neat polyurethane and the bionanocomposites is very important in order to understand the TASM behavior.
On one hand, the thermal transitions obtained by DSC related to both soft and hard phases are collected in Table 1. As can be observed from the values in the table, soft phase melting temperature (TmSS) appeared around 60 °C, while hard phase melting temperature (TmHS) appeared around 125 °C. Taking into account the melting enthalpy values obtained for both soft (DHmSS) and hard (DHmHS) phases, and following Eq. (1), relative crystallinity values were obtained for each phase (vc(SS) and vc(HS)), in order to study the effect of CNC content on the thermal properties (Fig. 2) [27].
vc ¼
DH m
x DH100
ð1Þ
where DH100 is the melting enthalpy value of the corresponding phase, soft or hard, obtained for the neat polyurethane (STPU17, synthesized without CNCs) by DSC (DH100 = 51.2 J g1 to determine
30
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
Table 1 TmSS, DHmSS, TmHS, DHmHS, vc(SS) and vc(HS) values obtained from DSC measurements for the neat polyurethane and the bionanocomposites synthesized by in situ polymerization. Sample
TmSS (°C)
DHmSS (J g1)
vc(SS)
TmHS (°C)
DHmHS (J g1)
vc(HS)
STPU17 STPU17/CNC-0.25 STPU17/CNC-0.5 STPU17/CNC-0.75 STPU17/CNC-1 STPU17/CNC-2
62.5 63.9 60.8 60.9 61.2 61.2
51.2 46.9 46.4 46.6 46.4 47.2
1 0.92 0.91 0.92 0.92 0.94
120.3 122.9 122.5 124.3 127.6 128.3
8.0 9.1 10.1 9.9 13.8 13.8
1 1.14 1.27 1.32 1.74 1.76
soft phase relative crystallinities and DH100 = 8.0 J g1 to determine hard phase relative crystallinities), x is the weight fraction of the polymeric matrix material in the bionanocomposites and DHm is the experimental melting enthalpy value obtained by DSC for the bionanocomposites soft and hard phases (Table 1). Owing to the low amount of HS content in the samples (17 wt%), lower melting enthalpy values were observed for the hard phase comparing with the ones obtained for the soft phase. As can be observed in Fig. 2, hard phase relative crystallinity values increased with increasing CNC content, due to the nucleation effect of CNCs over the hard phase, already observed in other works [15,16,28], as CNCs are covalently linked to the polyurethane. Moreover, in order to corroborate the nucleation effect, cooling scans of the synthesized samples are also presented (Fig. 3). It can be clearly observed that while STPU17 did not present an exotermic peak related to the
crystallization of the hard phase (absence of an exotermic peak at 80 °C in Fig. 3), the addition of CNCs induced hard phase crystallization, observing an exotermic peak around 80 °C, corroborating in that way the nucleation effect. Nevertheless, soft phase relative crystallinity values slightly decreased with the addition of CNCs, due to the interactions between soft segments and neighbouring CNCs, as well as due to the higher crystallization degree of the hard phase which restrict soft phase crystallization. On the other hand, Fig. 4 shows the dependency of storage modulus (E0 ) and tan d as a function of temperature for the synthesized neat polyurethane and the bionanocomposites, obtained by DMA measurements. The addition of CNCs led to bionanocomposites with higher storage modulus values than the neat polyurethane, due the reinforcement effect of CNCs, as well as to the nucleation effect of the hard phase [16]. At low temperatures, a plateau appeared in storage modulus values, which presented higher E0 values with increasing CNC content, comparing with the values obtained for STPU17. At higher temperatures, a slight decrease in E0 and a maximum in tan d were observed, related to soft phase glass transition temperature (TgSS). TgSS remained nearly constant even with high CNC contents, however, the intensity of the peak decreased and broadened, indicating that less material was taking part in the transition. This was an indicative of the interactions between CNCs and soft segments, because TgSS is associated with the mobility of the amorphous soft segments, and therefore, the interactions between CNCs and soft segments restricted the mobility of these segments, as well as the covalent bonds between CNCs and polyurethane chains [29,30]. Above TgSS, E0 continued decreasing until TmSS, where a sharp decrease was observed in storage modulus values, related to soft phase melting temperature. As was also observed by DSC, the addition of CNCs slightly affects TmSS values, observing small differences in that region with the addition of CNCs. Above this temperature, in the second plateau region, bionanocomposites showed an improvement in storage modulus values, due to the effective dispersion of CNCs and the before
Fig. 3. Cooling scans for the neat polyurethane and the bionanocomposites.
Fig. 4. Storage modulus and tan d as a function of temperature for the neat polyurethane and the bionanocomposites.
Fig. 2. Relative crystallinity values for soft and hard phases, varying cellulose nanocrystal content.
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
31
mentioned covalent bonds between CNCs and hard segments. Finally, at high temperatures, a marked decrease of E0 values was observed, owing to the disruption of hard crystalline domains. The addition of CNCs led to bionanocomposites with slightly higher thermo-mechanical stability, due to the nucleation effect of the hard phase observed by DSC. Taking into account the results obtained by both DSC and DMA measurements, a phase separation was observed in the synthesized neat polyurethane and bionanocomposites, with two main transitions, soft and hard phase melting temperatures. Therefore, regarding thermally-activated shape-memory properties, SS will be responsible for shape fixity, acting as the switching segment, while HS will be responsible for shape recovery, determining the permanent shape. In order to assess the thermally-activated shape-memory properties, thermo-mechanical cyclic tensile tests were performed at 60 °C, as explained in the experimental section. Fig. 5 presents the shape fixity and shape recovery values obtained in the first two thermo-mechanical cycles for the neat polyurethane and the bionanocomposites. As can be observed in Fig. 5a, shape fixity values remained nearly constant and did not present any considerable variation from the first to the second thermomechanical cycle, and also, the addition of CNCs did not affect significantly shape fixity values. Taking into account that the shape-memory properties were measured at 60 °C, soft segment acted as the switching segment, being responsible for the temporary shape, which is measured by shape fixity values. Therefore, as DSC and DMA measurements revealed, the addition of CNCs did not affect soft phase melting temperature values, and that is
why Rf values remained nearly constant even with high amounts of CNCs. Furthermore, as was also observed in literature [31,32], the capacity of the SSs to act as a switching segment is retained for several cycles, and that is why there were not observed significant changes from the first to the second thermo-mechanical cycle. Regarding the shape recovery values, it was observed that STPU17 presented low Rr values for both first and second thermo-mechanical cycles (Fig. 5b). However, with the addition of CNCs, Rr values presented a significant improvement, because CNCs acted as network points, improving shape recovery values. STPU17 has been synthesized with low HS content, and therefore, there are not enough hydrogen bonds or physical crosslinks between hard segments to restore the polymer back to the original shape [33]. Nevertheless, due to the addition of CNCs, hard phase crystallinity increased and bionanocomposites had more network points, which contributed to the improvement observed in Rr values. Moreover, during the first stretching process, plastic and non-reversible deformation occurs, and therefore, the Rr values obtained for the first cycle were lower than the ones obtained for the second cycle. Thereby, Rr values close to 100% were obtained for the second thermo-mechanical cycle. Finally, Fig. 6 presents the stress–strain–temperature curves obtained for STPU17/CNC-1 (similar results were obtained for the rest of the bionanocomposites but they are not included in this work for simplification) after the first two thermo-mechanical cycles, corroborating the cyclic shape-memory properties of the sample, because repetitive cycles were obtained giving Rr and Rf values close to 100%. In the works reported in literature for polyurethane/cellulose nanocomposites and their shape-memory properties [32,34,35], there was not observed such a significant increase in Rr values, because the neat polyurethanes studied presented quite high shape recovery values even without CNCs, and therefore the addition of CNCs did not strongly affect Rr values. However, in this work, the neat polyurethane presented good Rf values but poor Rr values (50%), and the addition of CNCs strongly improved shape recovery values, specially for the second thermo-mechanical cycle, resulting in bionanocomposites with high Rf and Rr values. Regarding the mechanical properties, Table 2 summarized the values obtained by tensile tests. As can be observed from the values in the table, the addition of CNCs led to polyurethane bionanocomposites with higher mechanical properties in the elastic region, obtaining higher E, ry and work-of-fracture at 8% strain values. However, rmax, eb and total work-of-fracture values decreased. On one hand, during tensile tests, amorphous soft segment chains
Fig. 5. Shape fixity (a) and shape recovery (b) values obtained for the first two shape-memory cycles, varying cellulose nanocrystal content.
Fig. 6. Stress–strain–temperature diagram for the several consecutive shapememory cycles obtained after the second cycle for STPU17/CNC-1.
32
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
Table 2 Mechanical properties of the neat polyurethane and the bionanocomposites. Sample
E (MPa)
ry (MPa)
rmax (MPa)
e (%)
Work-of-fracture at 8% strain (MJ m3)
Total work-of-fracture (MJ m3)
STPU17 STPU17/CNC-0.25 STPU17/CNC-0.5 STPU17/CNC-0.75 STPU17/CNC-1 STPU17/CNC-2
238.8 ± 6.7 264.4 ± 5.6 272.7 ± 7.2 297.4 ± 6.7 321.6 ± 8.3 328.4 ± 11.2
10.3 ± 0.8 10.9 ± 0.8 11.3 ± 0.7 12.5 ± 0.9 14.1 ± 1.0 15.4 ± 1.2
26.5 ± 2.3 24.0 ± 2.1 19.3 ± 1.9 17.3 ± 2.2 15.7 ± 2.0 14.9 ± 1.9
762 ± 42 690 ± 38 591 ± 39 551 ± 24 430 ± 20 352 ± 21
0.94 ± 0.05 1.37 ± 0.07 1.39 ± 0.05 1.47 ± 0.06 1.54 ± 0.05 1.76 ± 0.06
103.6 ± 10.3 80.5 ± 7.4 72.5 ± 4.8 67.9 ± 5.3 45.6 ± 4.3 33.4 ± 5.6
orient under strain, giving high tensile strength values for STPU17 sample. Nevertheless, the addition of CNCs covalently bonded to the polyurethane chain and the increase in hard phase crystallinity restricted the orientation and crystallization of the amorphous segments under strain, giving lower rmax and eb values, and consequently, decreasing the total area under stress–strain curves, which is a direct measurement of the work-of-fracture. It has to be also taken into account that in the works reported in literature for polyurethane/cellulose nanocomposites [15,16,35,36], where methylene diphenyl 4,40 -diisocyanate (MDI) or amorphous soft segments were used, the overall crystallinity of the nanocomposites was lower than the overall crystallinity observed for the bionanocomposites synthesized in this work. Therefore, soft and hard crystallites of the bionanocomposites synthesized in this work act as load-bearing components in the same way that CNCs did, being in this case the effect of CNCs less pronounced on the final mechanical properties. Comparing the increase in E values obtained for the bionanocomposites synthesized by in situ polymerization (cellulose nanocrystals are chemically linked to the polyurethane chain) with the increase obtained for the bionanocomposites prepared by solvent casting procedure in a previous work (cellulose nanocrystals are physically bonded to the polyurethane chain) [37], it was observed that the bionanocomposites synthesized by in situ polymerization present a higher improvement, as the E values increase 35%, while the bionanocomposites synthesized by solvent casting procedure present an increase of 22% (for the same CNC content). Therefore, due to the chemical bonds between polyurethane chains and CNCs formed during in situ polymerization, bionanocomposites with high mechanical properties can be obtained adding small amounts of CNCs.
Finally, Fig. 7 shows height and three-dimensional AFM images of STPU17 and STPU17/CNC-2 samples, obtained from cryoultramicrotomed sample cross section, where an effective dispersion of CNCs (whitest regions in height images, similar results were obtained for the bionanocomposites prepared by solvent casting in a previous work [37]) was observed (Fig. 7b), as well as an increase in the roughness of the bionanocomposites, comparing with the roughness observed for the neat polyurethane. The reaction between hydroxyl groups from CNC surface and isocyanate groups from polyurethane precursors, either HDI or prepolymer, resulted in polyurethane chains anchored to CNCs, improving in that way the dispersion of nanocrystals in the neat polyurethane and changing the surface morphology of the synthesized bionanocomposites [15]. 4. Conclusions Bionanocomposites with high contents of carbon from renewable resources were synthesized by in situ polymerization, adding different amounts of cellulose nanocrystals, which were covalently bonded to polyurethane chains. A castor oil-based diol and a cornsugar based chain extender were used as raw materials in order to increase the amount of carbon from renewable resources in bionanocomposite formulation. Regarding the thermal properties, two main transitions were observed by DSC and DMA, the melting temperature of soft and hard phases, which were responsible for shape fixity and shape recovery, respectively. As it was observed by thermo-mechanical cyclic test, the neat polyurethane presented poor shape-memory properties, with shape recovery values around 50%. Nevertheless, due to the addition of CNCs, hard phase crystallinity increased and more network points took part determining
Fig. 7. Height (left) and three-dimensional (right) AFM images of STPU17 (a) and STPU17/CNC-2 (b), obtained from cryo-ultramicrotomed sample cross section.
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
the permanent shape, causing an improvement in shape-memory properties and obtaining shape recovery values close to 100% for the second thermo-mechanical cycle. Therefore, these bionanocomposites are good candidates to use as smart materials, where thermally-activated shape memory properties are required. Acknowledgements Financial support from the Basque Government in the frame of Grupos Consolidados (IT-776-13) is gratefully acknowledged. Additionally, A. S. thanks Eusko Jaurlaritza/Gobierno Vasco for ‘Programas de becas para formación y perfeccionamiento de personal investigador’ (BFI-09-167). Moreover, technical support provided by SGIker (UPV/EHU, MINECO, GV/EJ, ESF) is gratefully acknowledged. References [1] Vermette P, Griesser HJ, Laroche G, Guidoin R. Biomedical applications of polyurethanes (Tissue Engineering Intelligence Unit). Austin, Texas: Landes Bioscience; 2001. [2] Osman AF, Edwards GA, Schiller TL, Andriani Y, Jack KS, Morrow IC, et al. Structure–property relationships in biomedical thermoplastic polyurethane nanocomposites. Macromolecules 2012;45:198–210. [3] Guelcher SA. Biodegradable polyurethanes: synthesis and applications in regenerative medicine. Tissue Eng Part B 2008;14:3–17. [4] Choi T, Weksler J, Padsalgikar A, Runt J. Influence of soft segment composition on phase-separated microstructure of polydimethylsiloxane-based segmented polyurethane copolymers. Polymer 2009;50:2320–7. [5] Ferna´ndez-d’Arlas B, Ramos JA, Saralegi A, Corcuera M, Mondragon I, Eceiza A. Molecular engineering of elastic and strong supertough polyurethanes. Macromolecules 2012;45:3436–43. [6] Saralegi A, Rueda L, Fernández-d’Arlas B, Mondragon I, Eceiza A, Corcuera MA. Thermoplastic polyurethanes from renewable resources: effect of soft segment chemical structure and molecular weight on morphology and final properties. Polym Int 2013;62:106–15. [7] Fernández d’Arlas B, Rueda L, de la Caba K, Mondragon I, Eceiza A. Microdomain composition and properties differences of biodegradable polyurethanes based on MDI and HDI. Polym Eng Sci 2008;48:519–29. [8] Cavaco LI, Melo JA. Polyurethane: properties, structure and applications. London, United Kingdom: Nova Science Publishers; 2012. [9] Thomson T. Polyurethanes as specialty chemicals: principles and applications. Boca Raton, Florida: CRC Press; 2004. [10] Petrovic ZS. Polyurethanes from vegetable oils. Polym Rev 2008;48:109–55. [11] Desroches M, Escouvois M, Auvergne R, Caillol S, Boutevin B. From vegetable oils to polyurethanes: synthetic routes to polyols and main industrial products. Polym Rev 2012;52:38–79. [12] Zlatanic A, Lava C, Zhang W, Petrovic ZS. Effect of structure on properties of polyols and polyurethanes based on different vegetable oils. J Polym Sci, Part B: Polym Phys 2004;42:809–19. [13] Lligadas G, Ronda JC, Galia M, Cádiz V. Plant oils as platform chemicals for polyurethane synthesis: current state-of-the-art. Biomacromolecules 2010;11:2825–35. [14] Hojabri L, Kong X, Narine SS. Fatty acid-derived diisocyanate and biobased polyurethane produced from vegetable oil: synthesis, polymerization, and characterization. Biomacromolecules 2009;10:884–91. [15] Rueda L, Saralegi A, Fernández-d’Arlas B, Zhou Q, Alonso-Varona A, Berglund LA, et al. In situ polymerization and characterization of elastomeric polyurethane-cellulose nanocrystal nanocomposites, Cell response evaluation. Cellulose 2013;20:1819–28.
33
[16] Rueda L, Saralegui A, Fernández d’Arlas B, Zhou Q, Berglund LA, Corcuera MA, et al. Cellulose nanocrystals/polyurethane nanocomposites. Study from the viewpoint of microphase separated structure. Carbohydr Polym 2013;92:751–7. [17] Chen SH, Tsao CT, Chang CH, Wu YM, Liu ZW, Lin CP, et al. Synthesis and characterization of thermal-responsive chitin-based polyurethane copolymer as a smart material. Carbohydr Polym 2012;88:1483–7. [18] Juntaro J, Ummartyotin S, Sain M, Manuspiya H. Bacterial cellulose reinforced polyurethane-based resin nanocomposite: a study of how ethanol and processing pressure affect physical, mechanical and dielectric properties. Carbohydr Polym 2012;87:2464–9. [19] Ratna D, Karger-Kocsis J. Recent advances in shape memory polymers and composites: a review. J Mater Sci 2008;43:254–69. [20] Metzger MF, Wilson TS, Schumann D, Matthews DL, Maitland DJ. Mechanical properties of mechanical actuator for treating ischemic stroke. Biomed Microdevices 2002;4:89–96. [21] Lendlein A, Langer R. Biodegradable, elastic shape-memory polymers for potential biomedical applications. Science 2002;296:1673–6. [22] Lendlein A, Kelch S. Shape memory polymers. Angew Chem Int Ed 2002;41:2034–57. [23] Corcuera MA, Saralegi A, Fernández-d’Arlas B, Mondragon I, Eceiza A. Shape memory polyurethanes based on polyols derived from renewable resources. Macromol Symp 2012;321–322:197–201. [24] Gu X, Mather PT. Entanglement-based shape memory polyurethanes: synthesis and characterization. Polymer 2012;53:5924–34. [25] Wang W, Jin Y, Ping P, Chen X, Jing X, Su Z. Structure evolution in segmented poly(ester urethane) in shape-memory process. Macromolecules 2010;43:2942–7. [26] Saralegi A, Etxeberria A, Fernández-d’Arlas B, Mondragon I, Eceiza A, Corcuera MA. Effect of H12MDI isomer composition on mechanical and physicochemical properties of polyurethanes based on amorphous and semicrystalline soft segments. Polym Bull 2013;70:2193–210. [27] Peponi L, Navarro-Baena I, Sonseca A, Gimenez E, Marcos-Fernandez A, Kenny JM. Synthesis and characterization of PCL-PLLA polyurethane with shape memory behaviour. Eur Polym J 2012;49:893–903. [28] Pei A, Zhou Q, Berglund LA. Functionalized cellulose nanocrystals as biobased nucleation agents in poly(l-lactide) (PLLA)-crystallization and mechanical property effects. Compos Sci Technol 2010;70:815–21. [29] Wu Q, Henriksson M, Liu X, Berglund LA. A high strength nanocomposite based on microcrystalline cellulose and polyurethane. Biomacromolecules 2007;8:3687–92. [30] Tsagaropoulos G, Eisenberg A. Dynamic mechanical study of the factors affecting the two glass transition behavior of filled polymers. Similarities and differences with random ionomers. Macromolecules 1995;28:6067–77. [31] Ping P, Wang W, Chen X, Jing X. Poly(e-caprolactone) polyurethane and its shape-memory property. Biomacromolecules 2005;6:587–92. [32] Auad ML, Contos VS, Nutt S, Aranguren MI, Marcovich NE. Characterization of nanocellulose-reinforces shape memory polyurethanes. Polym Int 2008;57:651–9. [33] Lee BS, Chun BC, Chung YC, Sul KI, Cho JW. Structure and thermomechanical properties of polyurethane block copolymers with shape memory effect. Macromolecules 2001;34:6431–7. [34] Richardson TB, Mosiewicki MA, Uzunpinar C, Marcovich NE, Aranguren MI, Kilinc-Balci F, et al. Study of nanoreinforced shape memory polymers processed by casting and extrusion. Polym Compos 2011;32:455–63. [35] Auad ML, Mosiewicki MA, Richardson T, Aranguren MI, Marcovich NE. Nanocomposites made from cellulose nanocrystals and tailored segmented polyurethanes. J Appl Polym Sci 2010;115:1215–25. [36] Pei A, Malho JM, Ruokolainen J, Zhou Q, Berglund LA. Strong nanocomposite reinforcement effects in polyurethane elastomer with low volume fraction of cellulose nanocrystals. Macromolecules 2011;44:4422–7. [37] Saralegi A, Rueda L, Martin L, Arbelaiz A, Eceiza A, Corcuera MA. From elastomeric to rigid polyurethane/cellulose nanocrystal bionanocomposites. Compos Sci Technol 2013;88:39–47.
Contents lists available at ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
The role of cellulose nanocrystals in the improvement of the shape-memory properties of castor oil-based segmented thermoplastic polyurethanes Ainara Saralegi a, Maria Luz Gonzalez b, Angel Valea b, Arantxa Eceiza a,⇑, Maria Angeles Corcuera a,⇑ a ‘Materials + Technologies’ Group, Department of Chemical and Environmental Engineering, Polytechnic School, University of the Basque Country UPV/EHU, Pza. Europa 1, 20018 Donostia-San Sebastián, Spain b Department of Chemical and Environmental Engineering, School of Technical Industrial Engineering, University of the Basque Country UPV/EHU, Paseo Rafel Moreno ‘Pitxitxi’ 3, 48013 Bilbao, Spain
a r t i c l e
i n f o
Article history: Received 12 August 2013 Received in revised form 18 November 2013 Accepted 4 December 2013 Available online 16 December 2013 Keywords: A. Polymer–matrix composites A. Nanocomposites Polyurethane B. Fiber/matrix bond A. Smart materials
a b s t r a c t The effect of the addition of cellulose nanocrystals on segmented thermoplastic polyurethane shape-memory properties was investigated. To this end, polyurethane bionanocomposites were synthesized by in situ polymerization, adding cellulose nanocrystals in the first step of polymerization and using components from renewable resources in polyurethane formulation. Thereby, microphase separated polyurethane/cellulose nanocrystal bionanocomposites were obtained, presenting two main transitions, soft and hard phase melting temperatures. Furthermore, it was observed that due to the addition of cellulose nanocrystals, hard phase crystallinity increased. Regarding thermally-activated shape-memory properties, a significant improvement was observed in the shape recovery values with the addition of cellulose nanocrystals, being these bionanocomposites good candidates to use in applications where shape-memory properties are required. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Segmented polyurethanes are an important class of polymers that have found many applications as biomaterials due to their excellent physical properties and relatively good biocompatibility [1–3]. Technically, segmented thermoplastic polyurethanes (STPU) are block copolymers with alternating amorphous or crystalline hard and soft segments, denoted as HS and SS, respectively, that separate into microphases or domains, due to the thermodynamic incompatibility between both segments [4–7]. Therefore, manipulating their composition and choosing properly the chemical structure of the components, a wide variety of polyurethanes can be synthesized with different types of molecular architectures, specifically designed for each application [8,9]. In the last decades, the use of components from renewable resources have received steadily growing attention in polyurethane synthesis, owing to the dominant consumption patterns, the limited fossil resources and their increasing cost, as well as the public concern about climate change. Polyols derived from vegetable oils [10–14] are a good example of materials that can be used to synthesize segmented polyurethanes with a high percentage of components derived from ⇑ Corresponding authors. Tel.: +34 943017186; fax: +34 943017130. E-mail addresses: [email protected] (A. Eceiza), [email protected] (M.A. Corcuera). 0266-3538/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compscitech.2013.12.001
renewable sources. Moreover, the use of nanoreinforcements also from renewable resources, such as cellulose nanocrystals, chitin nanocrystals and bacterial cellulose [15–18], are gaining increasing interest, because they are used to improve the final properties of the synthesized nanocomposites, as well as to increase the amount of carbon from renewable resources in nanocomposites formulation. The use of segmented polyurethanes as shape-memory polymers, mainly as thermally-activated shape-memory (TASM) polymers, has received increasing attention in the last few years because of their applications in microelectromechanical systems, actuators, for self healing and health monitoring purposes, and in biomedical devices [19–21]. TASM polymers consist of two components on the molecular level: first, molecular switches, which are segments having a thermal transition at a determined temperature (Ttrans) and responsible for shape fixity by forming physical crosslinks; and second, network points that link these segments and determine the permanent shape [22–25]. Regarding segmented polyurethanes, in almost all cases, SS acts as the switching segment, and HS determines the permanent shape. The main objective of this work was the synthesis of polyurethane/cellulose nanocrystal bionanocomposites with high contents of carbon from renewable resources by in situ polymerization, and the posterior TASM characterization of the synthesized bionanocomposites, in order to study the role of cellulose nanocrystals
28
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
(CNC) in shape-memory properties. To this end, thermal and mechanical properties of the synthesized bionanocomposites were determined, as well as their TASM properties by cyclic thermomechanical tests. 2. Materials and methods 2.1. Materials Concentrated sulphuric acid (H2SO4, >98%) was purchased from Sigma–Aldrich Corporation (Missouri, United States), as well as tetrahydrofuran (THF, >99.9%) and microcrystalline cellulose (MCC) with a particle size between 100 and 150 lm. For polyurethane synthesis, 1,6-hexamethylene diisocyanate (HDI) supplied by Bayer (Leverkusen, Germany) under the trade name Desmodur H and corn-sugar based 1,3-propanediol (PD) supplied by Quimidroga SA (Barcelona, Spain) were used as hard segment. Poly(butylene sebacate) diol was used as soft segment, a difunctional semicrystalline polyester derived from castor oil (Tg = 54 °C, Tm = 67 °C and DHm = 138.26 J g1), denoted as CO2, with a molecular weight of 1900 g mol1, which has been already used and characterized in previous works [6,26]. Moreover, the quantity of carbon from renewable resources of CO2 was determined by ASTM D 6855, resulting in a value of 70%. 2.2. Isolation of cellulose nanocrystals Cellulose nanocrystals were prepared by acid hydrolysis of MCC using the following method: 20 g of MCC were dispersed in 500 mL of 64% (w/w) aqueous H2SO4 and stirred constantly for 30 min at 45 °C. The resultant suspension was subsequently diluted and washed by continuous centrifugation with deionized water. For removal of the remaining acid, nanocrystals were dialyzed against deionized water until the pH of the dialysis water stayed constant. Finally, a loose powder was obtained by freeze-drying the CNC suspension at a concentration of 0.05 mg mL1. Therefore, CNC with a rod-like morphology were isolated with an average diameter of 8.3 ± 0.9 nm and an average length of 152 ± 21 nm, resulting in an aspect ratio of 18.3 ± 2.7. 2.3. Synthesis of polyurethane bionanocomposites Polyurethane bionanocomposites were synthesized by in situ polymerization with different CNC contents and 17 wt% of HS content (with a molar ratio of 1:2:1 among components, polyol:diisocyanate:chain extender + CNC), using a two step procedure in THF solution [15]. As sketched in Fig. 1, prepolymers with different chemical structures can be formed. Although not shown in the scheme, prepolymers with isocyanate and hydroxyl ending groups from cellulose nanocrystals could form longer prepolymer structures. In the second step of polymerization, the chain extender was added, enabling the coupling of these prepolymers and forming different hard segment structures. To avoid the agglomeration of cellulose nanocrystals, the freeze-dried CNCs were redispersed in THF at a solid content of 0.5 wt%, without additives or any surface modification. First of all, the prepolymer was synthesized at 100 °C for 6 h by reacting HDI with CO2 in the presence of CNCs, previously dispersed in THF. The concentration of the resulting prepolymer was 75 mg mL1 and a condenser was used to avoid THF evaporation. Afterwards, the chain extender, 1,3-propanediol, was added in stoichiometric amount to the prepolymer, and the mixture was stirred at 100 °C for 2 h before casting and curing on a Teflon dish at 60 °C for 48 h under vacuum, in order to obtain a film. The amount of reactive hydroxyl group on cellulose nanocrystals was measured
by titration of excessive isocyanate groups, resulting in a value of 2.1 ± 0.2 mmol g1. This polymerization method allows obtaining bionanocomposites with cellulose nanocrystals covalently linked to the polyurethane chains. Furthermore, the neat polyurethane was also synthesized following the procedure used for bionanocomposites synthesis, but without adding CNCs. Thereby, a neat polyurethane, denoted as STPU17, and bionanocomposites with different CNC contents ranging from 0.25 to 2 wt% were successfully synthesized and denoted as STPU17/CNC-X, being X the amount of CNC added (wt%). 2.4. Characterization techniques The thermal behavior of the neat polyurethane and the bionanocomposites was analyzed by differential scanning calorimetry (DSC) and dynamic mechanical analysis (DMA). On one hand, DSC scans were recorded on a Mettler Toledo 822e instrument (Ohio, United States), equipped with a robotic arm and with an electric intracooler as refrigeration unit. Samples were scanned from 60 to 220 °C at a scanning rate of 20 °C min1, operating under dry nitrogen atmosphere. Moreover, the crystallization process was followed by cooling the samples from 220 to 60 °C at a scanning rate of 10 °C min1. A second heating run was also performed, but as it did not add any extra information, it was not included in this work. On the other hand, DMA measurements of the neat polyurethane and the bionanocomposites were performed in tension mode on a DMA Q-800 (TA Instruments, Delaware, United States). A constant frequency of 1 Hz was used, and samples were scanned from 100 to 150 °C at a rate of 2 °C min1. Thermally-activated shape-memory properties of the synthesized neat polyurethane and the bionanocomposites were also characterized in tension mode on a DMA Q-800, performing cyclic tests at 60 °C. Samples were first conditioned at 60 °C for 15 min and subsequently elongated to 50% applying a force ramp of 0.1 N min1. Then, samples were cooled to 15 °C at a rate of 5 °C min1 and the stress was released maintaining the temperature at 15 °C. Finally, samples underwent the recovery process by being heated at a rate of 5 °C min1 to the recovery temperature, 60 °C. This temperature was selected as the transition temperature because above 60 °C SS crystallites melt, acting as switching segments, whereas HS crystallites restricted molecular motion, being responsible for shape recovery. Shape fixity (Rf) and shape recovery (Rr) values were determined for each thermo-mechanical cycle, according to the equations presented in literature [19,22]. The mechanical properties of the synthesized films were measured on a MTS Insight 10 instrument (MTS Systems Corporation, Minnesota, United States), equipped with a 250 N load cell. Dogbone shape specimens (22.25 mm in length, 5 mm in width and 0.7 mm in thickness) were used for tensile tests, according to ASTM D 1708-93 standard. Five replicates of each material were used to measure Young’s modulus (E), yield strength (ry), tensile strength at maximum elongation (rmax) and strain at break (eb) values, calculated from the load–displacement data, where the deformation was measured by the crosshead displacement. Moreover, the work-of-fracture in the elastic region (8% strain) and the total work-of-fracture were also measured integrating the area under stress–strain curves obtained by tensile tests. Finally, atomic force microscopy images (AFM) were obtained in tapping mode on a Nanoscope IIIa scanning probe microscope Multimode™ Digital Instrument (Bruker Corporation, Massachusetts, United States), equipped with an integrated silicon tip/cantilever having a resonance frequency of 300 kHz. Samples cross sections were prepared using a Leica EM FC6 cryo-ultramicrotome (Leica Microsystems, Wetzlar, Germany) equipped with a diamond knife and operating at 120 °C, in order to observe the dispersion of CNCs in the neat polyurethane. The measurements were performed
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
29
Fig. 1. Schematic representation of the procedure used for polyurethane bionanocomposites synthesis, showing the chemical structures of the prepolymers and final bionanocomposites, as well as the chemical structures of the main components. The molar ratio of the components was 1:2:1 for polyol:diisocyanate:chain extender + CNC.
at a scan rate of 1.0 Hz, using a scan head with a maximum range of 16 lm 16 lm. Several regions were scanned obtaining similar results.
3. Results and discussion As was explained earlier, shape memory polymers require two components or segments on the molecular level, which are switching segments (amorphous or crystalline) that fix the temporary shape, and network points (achieved by either physical interactions or chemical bonds), that determine the permanent shape. Therefore, the thermal characterization of the synthesized neat polyurethane and the bionanocomposites is very important in order to understand the TASM behavior.
On one hand, the thermal transitions obtained by DSC related to both soft and hard phases are collected in Table 1. As can be observed from the values in the table, soft phase melting temperature (TmSS) appeared around 60 °C, while hard phase melting temperature (TmHS) appeared around 125 °C. Taking into account the melting enthalpy values obtained for both soft (DHmSS) and hard (DHmHS) phases, and following Eq. (1), relative crystallinity values were obtained for each phase (vc(SS) and vc(HS)), in order to study the effect of CNC content on the thermal properties (Fig. 2) [27].
vc ¼
DH m
x DH100
ð1Þ
where DH100 is the melting enthalpy value of the corresponding phase, soft or hard, obtained for the neat polyurethane (STPU17, synthesized without CNCs) by DSC (DH100 = 51.2 J g1 to determine
30
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
Table 1 TmSS, DHmSS, TmHS, DHmHS, vc(SS) and vc(HS) values obtained from DSC measurements for the neat polyurethane and the bionanocomposites synthesized by in situ polymerization. Sample
TmSS (°C)
DHmSS (J g1)
vc(SS)
TmHS (°C)
DHmHS (J g1)
vc(HS)
STPU17 STPU17/CNC-0.25 STPU17/CNC-0.5 STPU17/CNC-0.75 STPU17/CNC-1 STPU17/CNC-2
62.5 63.9 60.8 60.9 61.2 61.2
51.2 46.9 46.4 46.6 46.4 47.2
1 0.92 0.91 0.92 0.92 0.94
120.3 122.9 122.5 124.3 127.6 128.3
8.0 9.1 10.1 9.9 13.8 13.8
1 1.14 1.27 1.32 1.74 1.76
soft phase relative crystallinities and DH100 = 8.0 J g1 to determine hard phase relative crystallinities), x is the weight fraction of the polymeric matrix material in the bionanocomposites and DHm is the experimental melting enthalpy value obtained by DSC for the bionanocomposites soft and hard phases (Table 1). Owing to the low amount of HS content in the samples (17 wt%), lower melting enthalpy values were observed for the hard phase comparing with the ones obtained for the soft phase. As can be observed in Fig. 2, hard phase relative crystallinity values increased with increasing CNC content, due to the nucleation effect of CNCs over the hard phase, already observed in other works [15,16,28], as CNCs are covalently linked to the polyurethane. Moreover, in order to corroborate the nucleation effect, cooling scans of the synthesized samples are also presented (Fig. 3). It can be clearly observed that while STPU17 did not present an exotermic peak related to the
crystallization of the hard phase (absence of an exotermic peak at 80 °C in Fig. 3), the addition of CNCs induced hard phase crystallization, observing an exotermic peak around 80 °C, corroborating in that way the nucleation effect. Nevertheless, soft phase relative crystallinity values slightly decreased with the addition of CNCs, due to the interactions between soft segments and neighbouring CNCs, as well as due to the higher crystallization degree of the hard phase which restrict soft phase crystallization. On the other hand, Fig. 4 shows the dependency of storage modulus (E0 ) and tan d as a function of temperature for the synthesized neat polyurethane and the bionanocomposites, obtained by DMA measurements. The addition of CNCs led to bionanocomposites with higher storage modulus values than the neat polyurethane, due the reinforcement effect of CNCs, as well as to the nucleation effect of the hard phase [16]. At low temperatures, a plateau appeared in storage modulus values, which presented higher E0 values with increasing CNC content, comparing with the values obtained for STPU17. At higher temperatures, a slight decrease in E0 and a maximum in tan d were observed, related to soft phase glass transition temperature (TgSS). TgSS remained nearly constant even with high CNC contents, however, the intensity of the peak decreased and broadened, indicating that less material was taking part in the transition. This was an indicative of the interactions between CNCs and soft segments, because TgSS is associated with the mobility of the amorphous soft segments, and therefore, the interactions between CNCs and soft segments restricted the mobility of these segments, as well as the covalent bonds between CNCs and polyurethane chains [29,30]. Above TgSS, E0 continued decreasing until TmSS, where a sharp decrease was observed in storage modulus values, related to soft phase melting temperature. As was also observed by DSC, the addition of CNCs slightly affects TmSS values, observing small differences in that region with the addition of CNCs. Above this temperature, in the second plateau region, bionanocomposites showed an improvement in storage modulus values, due to the effective dispersion of CNCs and the before
Fig. 3. Cooling scans for the neat polyurethane and the bionanocomposites.
Fig. 4. Storage modulus and tan d as a function of temperature for the neat polyurethane and the bionanocomposites.
Fig. 2. Relative crystallinity values for soft and hard phases, varying cellulose nanocrystal content.
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
31
mentioned covalent bonds between CNCs and hard segments. Finally, at high temperatures, a marked decrease of E0 values was observed, owing to the disruption of hard crystalline domains. The addition of CNCs led to bionanocomposites with slightly higher thermo-mechanical stability, due to the nucleation effect of the hard phase observed by DSC. Taking into account the results obtained by both DSC and DMA measurements, a phase separation was observed in the synthesized neat polyurethane and bionanocomposites, with two main transitions, soft and hard phase melting temperatures. Therefore, regarding thermally-activated shape-memory properties, SS will be responsible for shape fixity, acting as the switching segment, while HS will be responsible for shape recovery, determining the permanent shape. In order to assess the thermally-activated shape-memory properties, thermo-mechanical cyclic tensile tests were performed at 60 °C, as explained in the experimental section. Fig. 5 presents the shape fixity and shape recovery values obtained in the first two thermo-mechanical cycles for the neat polyurethane and the bionanocomposites. As can be observed in Fig. 5a, shape fixity values remained nearly constant and did not present any considerable variation from the first to the second thermomechanical cycle, and also, the addition of CNCs did not affect significantly shape fixity values. Taking into account that the shape-memory properties were measured at 60 °C, soft segment acted as the switching segment, being responsible for the temporary shape, which is measured by shape fixity values. Therefore, as DSC and DMA measurements revealed, the addition of CNCs did not affect soft phase melting temperature values, and that is
why Rf values remained nearly constant even with high amounts of CNCs. Furthermore, as was also observed in literature [31,32], the capacity of the SSs to act as a switching segment is retained for several cycles, and that is why there were not observed significant changes from the first to the second thermo-mechanical cycle. Regarding the shape recovery values, it was observed that STPU17 presented low Rr values for both first and second thermo-mechanical cycles (Fig. 5b). However, with the addition of CNCs, Rr values presented a significant improvement, because CNCs acted as network points, improving shape recovery values. STPU17 has been synthesized with low HS content, and therefore, there are not enough hydrogen bonds or physical crosslinks between hard segments to restore the polymer back to the original shape [33]. Nevertheless, due to the addition of CNCs, hard phase crystallinity increased and bionanocomposites had more network points, which contributed to the improvement observed in Rr values. Moreover, during the first stretching process, plastic and non-reversible deformation occurs, and therefore, the Rr values obtained for the first cycle were lower than the ones obtained for the second cycle. Thereby, Rr values close to 100% were obtained for the second thermo-mechanical cycle. Finally, Fig. 6 presents the stress–strain–temperature curves obtained for STPU17/CNC-1 (similar results were obtained for the rest of the bionanocomposites but they are not included in this work for simplification) after the first two thermo-mechanical cycles, corroborating the cyclic shape-memory properties of the sample, because repetitive cycles were obtained giving Rr and Rf values close to 100%. In the works reported in literature for polyurethane/cellulose nanocomposites and their shape-memory properties [32,34,35], there was not observed such a significant increase in Rr values, because the neat polyurethanes studied presented quite high shape recovery values even without CNCs, and therefore the addition of CNCs did not strongly affect Rr values. However, in this work, the neat polyurethane presented good Rf values but poor Rr values (50%), and the addition of CNCs strongly improved shape recovery values, specially for the second thermo-mechanical cycle, resulting in bionanocomposites with high Rf and Rr values. Regarding the mechanical properties, Table 2 summarized the values obtained by tensile tests. As can be observed from the values in the table, the addition of CNCs led to polyurethane bionanocomposites with higher mechanical properties in the elastic region, obtaining higher E, ry and work-of-fracture at 8% strain values. However, rmax, eb and total work-of-fracture values decreased. On one hand, during tensile tests, amorphous soft segment chains
Fig. 5. Shape fixity (a) and shape recovery (b) values obtained for the first two shape-memory cycles, varying cellulose nanocrystal content.
Fig. 6. Stress–strain–temperature diagram for the several consecutive shapememory cycles obtained after the second cycle for STPU17/CNC-1.
32
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
Table 2 Mechanical properties of the neat polyurethane and the bionanocomposites. Sample
E (MPa)
ry (MPa)
rmax (MPa)
e (%)
Work-of-fracture at 8% strain (MJ m3)
Total work-of-fracture (MJ m3)
STPU17 STPU17/CNC-0.25 STPU17/CNC-0.5 STPU17/CNC-0.75 STPU17/CNC-1 STPU17/CNC-2
238.8 ± 6.7 264.4 ± 5.6 272.7 ± 7.2 297.4 ± 6.7 321.6 ± 8.3 328.4 ± 11.2
10.3 ± 0.8 10.9 ± 0.8 11.3 ± 0.7 12.5 ± 0.9 14.1 ± 1.0 15.4 ± 1.2
26.5 ± 2.3 24.0 ± 2.1 19.3 ± 1.9 17.3 ± 2.2 15.7 ± 2.0 14.9 ± 1.9
762 ± 42 690 ± 38 591 ± 39 551 ± 24 430 ± 20 352 ± 21
0.94 ± 0.05 1.37 ± 0.07 1.39 ± 0.05 1.47 ± 0.06 1.54 ± 0.05 1.76 ± 0.06
103.6 ± 10.3 80.5 ± 7.4 72.5 ± 4.8 67.9 ± 5.3 45.6 ± 4.3 33.4 ± 5.6
orient under strain, giving high tensile strength values for STPU17 sample. Nevertheless, the addition of CNCs covalently bonded to the polyurethane chain and the increase in hard phase crystallinity restricted the orientation and crystallization of the amorphous segments under strain, giving lower rmax and eb values, and consequently, decreasing the total area under stress–strain curves, which is a direct measurement of the work-of-fracture. It has to be also taken into account that in the works reported in literature for polyurethane/cellulose nanocomposites [15,16,35,36], where methylene diphenyl 4,40 -diisocyanate (MDI) or amorphous soft segments were used, the overall crystallinity of the nanocomposites was lower than the overall crystallinity observed for the bionanocomposites synthesized in this work. Therefore, soft and hard crystallites of the bionanocomposites synthesized in this work act as load-bearing components in the same way that CNCs did, being in this case the effect of CNCs less pronounced on the final mechanical properties. Comparing the increase in E values obtained for the bionanocomposites synthesized by in situ polymerization (cellulose nanocrystals are chemically linked to the polyurethane chain) with the increase obtained for the bionanocomposites prepared by solvent casting procedure in a previous work (cellulose nanocrystals are physically bonded to the polyurethane chain) [37], it was observed that the bionanocomposites synthesized by in situ polymerization present a higher improvement, as the E values increase 35%, while the bionanocomposites synthesized by solvent casting procedure present an increase of 22% (for the same CNC content). Therefore, due to the chemical bonds between polyurethane chains and CNCs formed during in situ polymerization, bionanocomposites with high mechanical properties can be obtained adding small amounts of CNCs.
Finally, Fig. 7 shows height and three-dimensional AFM images of STPU17 and STPU17/CNC-2 samples, obtained from cryoultramicrotomed sample cross section, where an effective dispersion of CNCs (whitest regions in height images, similar results were obtained for the bionanocomposites prepared by solvent casting in a previous work [37]) was observed (Fig. 7b), as well as an increase in the roughness of the bionanocomposites, comparing with the roughness observed for the neat polyurethane. The reaction between hydroxyl groups from CNC surface and isocyanate groups from polyurethane precursors, either HDI or prepolymer, resulted in polyurethane chains anchored to CNCs, improving in that way the dispersion of nanocrystals in the neat polyurethane and changing the surface morphology of the synthesized bionanocomposites [15]. 4. Conclusions Bionanocomposites with high contents of carbon from renewable resources were synthesized by in situ polymerization, adding different amounts of cellulose nanocrystals, which were covalently bonded to polyurethane chains. A castor oil-based diol and a cornsugar based chain extender were used as raw materials in order to increase the amount of carbon from renewable resources in bionanocomposite formulation. Regarding the thermal properties, two main transitions were observed by DSC and DMA, the melting temperature of soft and hard phases, which were responsible for shape fixity and shape recovery, respectively. As it was observed by thermo-mechanical cyclic test, the neat polyurethane presented poor shape-memory properties, with shape recovery values around 50%. Nevertheless, due to the addition of CNCs, hard phase crystallinity increased and more network points took part determining
Fig. 7. Height (left) and three-dimensional (right) AFM images of STPU17 (a) and STPU17/CNC-2 (b), obtained from cryo-ultramicrotomed sample cross section.
A. Saralegi et al. / Composites Science and Technology 92 (2014) 27–33
the permanent shape, causing an improvement in shape-memory properties and obtaining shape recovery values close to 100% for the second thermo-mechanical cycle. Therefore, these bionanocomposites are good candidates to use as smart materials, where thermally-activated shape memory properties are required. Acknowledgements Financial support from the Basque Government in the frame of Grupos Consolidados (IT-776-13) is gratefully acknowledged. Additionally, A. S. thanks Eusko Jaurlaritza/Gobierno Vasco for ‘Programas de becas para formación y perfeccionamiento de personal investigador’ (BFI-09-167). Moreover, technical support provided by SGIker (UPV/EHU, MINECO, GV/EJ, ESF) is gratefully acknowledged. References [1] Vermette P, Griesser HJ, Laroche G, Guidoin R. Biomedical applications of polyurethanes (Tissue Engineering Intelligence Unit). Austin, Texas: Landes Bioscience; 2001. [2] Osman AF, Edwards GA, Schiller TL, Andriani Y, Jack KS, Morrow IC, et al. Structure–property relationships in biomedical thermoplastic polyurethane nanocomposites. Macromolecules 2012;45:198–210. [3] Guelcher SA. Biodegradable polyurethanes: synthesis and applications in regenerative medicine. Tissue Eng Part B 2008;14:3–17. [4] Choi T, Weksler J, Padsalgikar A, Runt J. Influence of soft segment composition on phase-separated microstructure of polydimethylsiloxane-based segmented polyurethane copolymers. Polymer 2009;50:2320–7. [5] Ferna´ndez-d’Arlas B, Ramos JA, Saralegi A, Corcuera M, Mondragon I, Eceiza A. Molecular engineering of elastic and strong supertough polyurethanes. Macromolecules 2012;45:3436–43. [6] Saralegi A, Rueda L, Fernández-d’Arlas B, Mondragon I, Eceiza A, Corcuera MA. Thermoplastic polyurethanes from renewable resources: effect of soft segment chemical structure and molecular weight on morphology and final properties. Polym Int 2013;62:106–15. [7] Fernández d’Arlas B, Rueda L, de la Caba K, Mondragon I, Eceiza A. Microdomain composition and properties differences of biodegradable polyurethanes based on MDI and HDI. Polym Eng Sci 2008;48:519–29. [8] Cavaco LI, Melo JA. Polyurethane: properties, structure and applications. London, United Kingdom: Nova Science Publishers; 2012. [9] Thomson T. Polyurethanes as specialty chemicals: principles and applications. Boca Raton, Florida: CRC Press; 2004. [10] Petrovic ZS. Polyurethanes from vegetable oils. Polym Rev 2008;48:109–55. [11] Desroches M, Escouvois M, Auvergne R, Caillol S, Boutevin B. From vegetable oils to polyurethanes: synthetic routes to polyols and main industrial products. Polym Rev 2012;52:38–79. [12] Zlatanic A, Lava C, Zhang W, Petrovic ZS. Effect of structure on properties of polyols and polyurethanes based on different vegetable oils. J Polym Sci, Part B: Polym Phys 2004;42:809–19. [13] Lligadas G, Ronda JC, Galia M, Cádiz V. Plant oils as platform chemicals for polyurethane synthesis: current state-of-the-art. Biomacromolecules 2010;11:2825–35. [14] Hojabri L, Kong X, Narine SS. Fatty acid-derived diisocyanate and biobased polyurethane produced from vegetable oil: synthesis, polymerization, and characterization. Biomacromolecules 2009;10:884–91. [15] Rueda L, Saralegi A, Fernández-d’Arlas B, Zhou Q, Alonso-Varona A, Berglund LA, et al. In situ polymerization and characterization of elastomeric polyurethane-cellulose nanocrystal nanocomposites, Cell response evaluation. Cellulose 2013;20:1819–28.
33
[16] Rueda L, Saralegui A, Fernández d’Arlas B, Zhou Q, Berglund LA, Corcuera MA, et al. Cellulose nanocrystals/polyurethane nanocomposites. Study from the viewpoint of microphase separated structure. Carbohydr Polym 2013;92:751–7. [17] Chen SH, Tsao CT, Chang CH, Wu YM, Liu ZW, Lin CP, et al. Synthesis and characterization of thermal-responsive chitin-based polyurethane copolymer as a smart material. Carbohydr Polym 2012;88:1483–7. [18] Juntaro J, Ummartyotin S, Sain M, Manuspiya H. Bacterial cellulose reinforced polyurethane-based resin nanocomposite: a study of how ethanol and processing pressure affect physical, mechanical and dielectric properties. Carbohydr Polym 2012;87:2464–9. [19] Ratna D, Karger-Kocsis J. Recent advances in shape memory polymers and composites: a review. J Mater Sci 2008;43:254–69. [20] Metzger MF, Wilson TS, Schumann D, Matthews DL, Maitland DJ. Mechanical properties of mechanical actuator for treating ischemic stroke. Biomed Microdevices 2002;4:89–96. [21] Lendlein A, Langer R. Biodegradable, elastic shape-memory polymers for potential biomedical applications. Science 2002;296:1673–6. [22] Lendlein A, Kelch S. Shape memory polymers. Angew Chem Int Ed 2002;41:2034–57. [23] Corcuera MA, Saralegi A, Fernández-d’Arlas B, Mondragon I, Eceiza A. Shape memory polyurethanes based on polyols derived from renewable resources. Macromol Symp 2012;321–322:197–201. [24] Gu X, Mather PT. Entanglement-based shape memory polyurethanes: synthesis and characterization. Polymer 2012;53:5924–34. [25] Wang W, Jin Y, Ping P, Chen X, Jing X, Su Z. Structure evolution in segmented poly(ester urethane) in shape-memory process. Macromolecules 2010;43:2942–7. [26] Saralegi A, Etxeberria A, Fernández-d’Arlas B, Mondragon I, Eceiza A, Corcuera MA. Effect of H12MDI isomer composition on mechanical and physicochemical properties of polyurethanes based on amorphous and semicrystalline soft segments. Polym Bull 2013;70:2193–210. [27] Peponi L, Navarro-Baena I, Sonseca A, Gimenez E, Marcos-Fernandez A, Kenny JM. Synthesis and characterization of PCL-PLLA polyurethane with shape memory behaviour. Eur Polym J 2012;49:893–903. [28] Pei A, Zhou Q, Berglund LA. Functionalized cellulose nanocrystals as biobased nucleation agents in poly(l-lactide) (PLLA)-crystallization and mechanical property effects. Compos Sci Technol 2010;70:815–21. [29] Wu Q, Henriksson M, Liu X, Berglund LA. A high strength nanocomposite based on microcrystalline cellulose and polyurethane. Biomacromolecules 2007;8:3687–92. [30] Tsagaropoulos G, Eisenberg A. Dynamic mechanical study of the factors affecting the two glass transition behavior of filled polymers. Similarities and differences with random ionomers. Macromolecules 1995;28:6067–77. [31] Ping P, Wang W, Chen X, Jing X. Poly(e-caprolactone) polyurethane and its shape-memory property. Biomacromolecules 2005;6:587–92. [32] Auad ML, Contos VS, Nutt S, Aranguren MI, Marcovich NE. Characterization of nanocellulose-reinforces shape memory polyurethanes. Polym Int 2008;57:651–9. [33] Lee BS, Chun BC, Chung YC, Sul KI, Cho JW. Structure and thermomechanical properties of polyurethane block copolymers with shape memory effect. Macromolecules 2001;34:6431–7. [34] Richardson TB, Mosiewicki MA, Uzunpinar C, Marcovich NE, Aranguren MI, Kilinc-Balci F, et al. Study of nanoreinforced shape memory polymers processed by casting and extrusion. Polym Compos 2011;32:455–63. [35] Auad ML, Mosiewicki MA, Richardson T, Aranguren MI, Marcovich NE. Nanocomposites made from cellulose nanocrystals and tailored segmented polyurethanes. J Appl Polym Sci 2010;115:1215–25. [36] Pei A, Malho JM, Ruokolainen J, Zhou Q, Berglund LA. Strong nanocomposite reinforcement effects in polyurethane elastomer with low volume fraction of cellulose nanocrystals. Macromolecules 2011;44:4422–7. [37] Saralegi A, Rueda L, Martin L, Arbelaiz A, Eceiza A, Corcuera MA. From elastomeric to rigid polyurethane/cellulose nanocrystal bionanocomposites. Compos Sci Technol 2013;88:39–47.