Description:

The invention pertains to hydrogelators and to elastic hydrogels with thermo- and water-responsive shape-memory properties in its dehydration state.

Both hydrogelators and supramolecular hydrogels are known.

A supramolecular hydrogel is a colloidal material composed of gelator molecules (low molecular weight organic compound) and water molecules. The gelator molecules are arranged in such a way that they form a mesh that traps the water molecules in the interstitial spaces. Such small organic molecules (low molecular weight) capable of gelling aqueous solvents are called hydrogelators. Gels can be classified as either chemical or physical. Chemical gels form a molecular mesh by covalent bonds, whereas physical gels form as a result of non-covalent interactions. The majority of physical gels are made from polymeric molecules, but

non-polymeric molecules can also self-assemble to form aggregates that gelate solvent. Many non-polymeric physical gels have been reported that are capable of forming gels in various organic solvents. Polymeric gels are common but

nonpolymeric physical hydrogels are rare. Studies on supramolecular gels have been of considerable interest in recent years because of the many potential applications like tissue engineering, nanotechnology, used as biomaterials, vehicles for controlled drug release, pollution control etc. Despite the recent advances, a priori design of a gelator for the gelation of water molecules has remained a challenging task.

Supramolecular hydrogels are formed by the self-assembly of a small amount of hydrogelators in water through non-covalent interactions such as hydrogen bonding, ττ-π stacking, van der Waals interaction, coordination and so on. These non-covalent interactions endue supramolecular hydrogels with fascinating physical properties. However, these reversible non-covalent interactions often make the formed hydrogels brittle, and not suitable for the tissue engineering applications. Others are more elastic, but are then not stable in time. They show self-healing, creep etc. due to the dynamic nature of the physical bonds.

It is thus a great challenge to prepare a supramolecular hydrogel that is highly elastic with excellent mechanical properties, while the morphology is so stable that no exchange of the supramolecular units can take place. Preferentially, the materials is a nanoscopic phase-separated material .

At the same time, shape-memory polymers (SMPs) are an emerging class of active polymers. They have the capability of changing their shapes from temporary shape to permanent shape upon exposed to an appropriate external stimulus.

Biocompatible synthetic polymers with shape memory (SM) behavior hold tremendous promise for critically important applications in the biomedical industry, including sutures and implantable stents.

In recent years, various SMPs have been designed and synthesized. And have enabled diverse biomedical applications including Vascular Stents, Clot Removal microactuator, dialysis needles, soft-tissue reconstruction, drug carrier and release systems and so on.

These polymers must be strong, durable and resilient, preferably under many conditions, including physiological (e.g., hydrated, elevated temperature, cell presence). Fulfillment of the mechanical requirements for a broad range of demanding circumstances requires exploitation of several characteristics in designing polymeric systems intended for SM behavior. This includes (1 ) an easily accessible transition between soft and hard states within a matrix and (2) mechanically stable physical or chemical crosslinks that establish a permanent shape. There exist multitude options for matrix components, with which the desired transition temperature or other response can be specifically tailored. Shape memory polymers (SMPs) are a subset of resilient materials, where the crucial feature is the control over the chain conformations and dynamics under various conditions. After macroscopic deformation from an equilibrium (i.e., permanent) shape, the temporary shape should be ideally retained until a controlled stimulus (e.g., heat, light, solvent exposure) induces recovery to the original geometry. The permanent molecular structure typically consists of a deformable soft matrix with either physical or chemical crosslinks defining an equilibrium conformation. Shape recovery is driven by the entropic gain from chain relaxation, which is accelerated appreciably above T _g or T _m .

Thermally induced shape recovery is often achieved with a semicrystaNine polymer matrix comprising the switching domain, where the recovery transition temperature (T _tr ) corresponds to crystallite melting. The crystalline regions provide a

mechanically robust scaffold for retaining the strained temporary shape at T < T _m =

For biomedical applications, predominant focus has been on poly(ethylene oxide) (PEO) (ref) and poly(E-caprolactone) (PCL) (ref) with melting temperatures (T _m ) of 50-60 °C and 65 °C, respectively, with poly(l-lactide) sometimes used as a hard, physically crosslinked phase. However, great effort has been exerted to modify and enhance the mechanical performance of these characteristically brittle materials, ultimately targeting products that require flexibility and ductility in addition to the implicit biocompatibility and cargo delivery capacity in certain instances. The properties have been modified with chemical crosslinks, copolymerizations, etc., and pendent hydrogen bonding sites, for example. Some designs contain specific, stimuli responsive degradable segments that encapsulate and subsequently release cargo, aimed at controlled drug delivery capabilities. The crosslink density can be systematically varied in an effort to tune the ultimate modulus and failure elongations, which cumulatively reflect the toughness of the material. Combining toughness with high recovery has been a formidable challenge, with some recent success accomplished through complex molecular makeup.

Challenges exist, however, mainly with how to incorporate the shape-memory properties, mechanical requirements, biocompatibility and biodegradability into one material to fulfil the multi-dimensional requirements of biomaterials. Summary of the invention

Accordingly, the present invention discloses a newly designed hydrogelator obtained by adding a nitrogen containing heterocyclic organic compound that is substituted with at least one ureido-group, preferably only one ureido-group, to a poly(alkylene glycol) and wherein the ureido-group and the poly(alkylene glycol) are spaced by a hydrophobic linker or spacer, preferably an alkyl spacer.

The nitrogen containing heterocyclic organic compound can be selected at will as long as at least one H-atom on one of the carbon atoms in the ring can be substituted by the ureido group, i.e. -NH-CO-NH-group. Typically, (a substituted) pyridine (also referred to as azine or azabenzene) can be applied. Also substituted derivatives of pyridine, such as 2,2'-bipyridine and 1 ,10-phenanthroline can be used.

More preferred as nitrogen containing heterocyclic organic compounds are those with two or more nitrogen atoms in the ring, such as diazines, triazines, tetraazines, pentaazines or even hexaazines.

Preferably diazines or triazines are selected. From the triazines the 1 ,3,5-triazine (with alternating carbon and nitrogen atoms in the ring) or its derivatives are preferred.

Most preferred are diazines. A suitable diazine can also be selected at will as long as at least one H-atom on one of the carbon atoms in the ring can be substituted by the ureido group, i.e. -NH-CO-NH-group. Also derivatives of diazines can be used, such as condensed diazines, e.g. purines (purine or adenine).

Also preferred are the diazine derivatives cytosine (C), thymine (T), and uracil (U).

From the three basic forms of diazine, i.e. pyrimidine (1 ,3-diazine or m-diazine), pyradizine (1 ,2-diazine or ortho-diazine) and pyrazine (1 ,4-diazine or o-diazine) pyrimidine-structure is preferred, even more preferred is pyrimidinone, sometimes also referred to as "pyrimidine" in the literature), such as 4-pyrimidinone. Most preferably, the hydrogelator contains as the substituted diazine a 2-ureido-4[1 H]-pyrimidinone or its respective tautomer 2-ureido-4[3H]-pyrimidinone (both referred to as UPy). The structures can schematically be shown as follows:

2-ureido-4[1 H]-pyrimidinone 2-ureido-4[3H]-pyrimidinone

Preferably, the hydrogelator of the present invention contains an ureido-group that is chain extended. In further preferred embodiments the hydrogelator according to the invention contains poly(ethylene glycol) as the poly(alkylene glycol) moiety. The term "poly(alkylene glycol) also encompasses copolymers (random or block) of two or more different alkylene glycols, such as a block-copolymer of poly(ethylene glycol) and poly(propylene glycol). Multiple block-copolymers, such as

tribock-copolymers can also be used successfully. An example is poly(ethylene glycol)-poly(propylene glycol)-poly(ethylene glycol). Such a triblock-copolymer is commercially available, e.g. under the trademark Pluronic® from BASF AG.

In yet another preferred embodiment, the ureido-group and the poly(alkylene glycol) are spaced by a hydrophobic linker, such as an alkyl spacer spacer, preferably an oligomethylene, even more preferred an alkyl spacer that is (-CH ₂ -)i 2 -

Preferably, the poly(alkylene glycol) moiety is diamine terminated. In such case the hydrogelator, can be obtained by adding a chain-extended

2-ureido-[1 H]-pyrimidin-4-one to a diamine terminated poly(ethylene glycol).

Thus, the hydrogelator is obtained by adding a diisocyanate of formula (I)

OCN-C6Hi2-UPy-C6Hi2-NCO (I) to an amino-telechelic polyethylene oxide flanked with oligomethylene spacers of formula (II): H ₂ N-Ci2H2 ₄ (OC2H ) _x Ci2H2 ₄ NH2 (II), wherein x is the polymerization degree of the polyethylene oxide.

X can be varied from 10 to 300, preferably from 50 to 250 and more preferably form 200 to 250, such as 227.

The obtained present hydrogelator forms a strong and elastic supramolecular hydrogel in its hydration or swelling state, and shows both reversible thermo and water responsive shape-memory properties in its dehydration state.

Description of the preferred embodiments

In the following, the preparation, resilient mechanical response, and shape-memory behavior of PEO-based copolymers containing a triple or higher, preferably quadruple, hydrogen-bonding motif in the backbone, which constitute physical (i.e., reversible) crosslinks with uniquely dynamic association properties is described.

The preferred features include a strong, non-covalent interaction within domains comprising nanoscopic confluences of long PEO segments. The dynamic bonding among the aggregates promotes processability, in contrast to chemical crosslinks in conventional thermoset materials. Additionally the strength could foreseeably be adjusted by implementing different lengths/composition of PEO. This could be achieved in a straightforward manner owing to the modular synthetic protocol employed. The H-bonding motif allows strong, phase segregated segments to act as physical crosslinks at relatively low composition. This feature translates to low crosslink density, ultimately displaying impressively large elongation-recovery profiles with minimal diminishment of modulus compared with pristine PEO. Most instances of thermoplastic SMPs rely on high T _m or high T _g physical crosslinking components, which require relatively high composition to maximize the strength.

As such, the multifunctionality of these materials imparted by the dynamic crosslink interactions, the hydrophilicity of the matrix and molecular shielding to prevent crosslink disruption, and the inherent biocompatibility by demonstrating cell viability and mechanical properties in a hydrogel form is emphasized. Finally, the viability of these materials is considered in the biomedical industry based on the characteristics uncovered here.

Brief description of the drawings

Fig. 1 a/b shows diagrams of rheology (a) and tensile tests (b) of the hydrogel film at an elongation rate of 60 mm/min at room temperature.

Fig. 2 a/b shows a scheme of the thermo and water responsive shape recovery (a) and a diagram from a tensile test at 60 °C and 60 mm/min elongation rate (b) of the dehydrated CEUPy polymer film.

Fig. 3 shows DSC curves of the original and stretched dehydrated

hydrogel films.

Fig. 4 a/b shows diagrams of the temperature dependence of the G' and G" of the dehydrated (a) and hydrogel (b) films.

Fig. 5 shows DSC curves of the dehydrated and hydrogel films with different water ratios.

Fig. 6 shows a schematic representation to make a shape-memory tube and photos of the prepared tubes in all its forms.

Fig. 7 shows PEO-UPy (co)polymers with telechelelic (1 ) and segmented multiblock (2) architectures and (3) complementary quadruple hydrogen bonding interaction between two UPy segments.

Fig. 8 shows (a) DSC thermograms for the dry polymer 2 and hydrogels at various concentrations (2nd heating; 10 °C min ^"1 ) and photographs of (b) melt pressed film of dry polymer 2 and (c) hydrogel with 85 wt % water.

Fig. 9 displays (a) temperature dependent storage modulus (Ε') and loss modulus (E") for chain extended UPy-PEO polymer 2 (1 Hz; 3 °C min ^"1 ); (b) Tensile testing results for pristine, dry polymer 2 at ambient temperature with an inset to show yielding and onset of cold-drawing; (c) Cyclic tensile testing at 70 °C (5 cycles each having e _ma x = 100 %) and the inset illuminating the determination of recovery ratios. Fig. 10 shows cyclic stress-strain results with (a) maximum strain (£ _ma x) of 100% and (b) £ _ma x of 300% having 5 cycles each and (c) recovery ratio (R _r ) for each cycle (N).

Fig. 11 displays shape memory behavior observed in response to (a) thermally activated phase transition (Tt _ran s = 7 ^" _m , _PE o) and (b) water activated phase transition during hydrogel formation.

Detailed description of the embodiments

One chosen molecular design is shown below and involved two components: a diisocyanate UPy (HDI-UPy-HDI) synthon and a diamine-terminated poly(ethylene glycol) (DAPEG). The synthesis of a preferred embodiment of the chain extended UPy (CEUPy) hydrogelator from HDI-UPy-HDI and diamine-terminated PEG with a molecular weight of about 10 kg mol ^"1 can be described by the following reaction equation:

CEUPy polymer

The storage moduli, G', and loss moduli, G" of the fully swelled hydrogel, which contains 90 % water, as function of angle frequency at a fixed strain (1 .0 %) are shown in Figure 1 a. The sample has a single plateau region in its dynamic moduli. The G' value has a substantial elastic response and is always higher than the G" over the entire range of frequency. The typical G' value is about 20 kPa, which is close to the desired modulus for soft tissue engineering and living soft tissues (on the order of 1000 Pa). At the same time, as shown in Fig. 1 b, the present hydrogel also shows excellent elastic properties and is able to withstand strains of up 550 % before breaking. It is noteworthy that, although the UPy-UPy association constant in water is low, the hydrophobic shielding of this bond in the hydrophobic spacers (Ci ₂ H ₂₄ ) of the polymer and the lateral urea hydrogen bonding motifs make this binding much stronger. And so the chain-extended UPy polymer of the present invention exhibited excellent mechanical and elastic properties.

The invention is also directed to a hydrogel comprising the addition reaction product of a diisocyanate of formula (I) OCN-C6Hi2-UPy-C6Hi2-NCO (I) and an amino-telechelic polyethylene oxide flanked with oligomethylene spacers of formula (II): H ₂ N-Ci2H2 ₄ (OC2H4)xCi2H ₂₄ NH2 (II), wherein x is the polymerization degree of the polyethylenoxide. X can be varied from 10 to 300, preferably from 50 to 250 and more preferably form 200 to 250, such as 227.

It is intriguing that the dehydrated hydrogel film has both thermo and water responsive shape memory properties. As shown in Figure 2a, the permanent shape was stretched at 60 °C and fixed at room temperature to yield temporary shape 1 , which could be further twisted at room temperature to yield temporary shape 2. Upon reheating to 60 °C or immersing in water, temporary shape 2 (or temporary shape 1 ) could recover to its permanent shape. At the same time, the dehydrated hydrogel film is able to withstand strains of up 1000 % before breaking at 60 °C and 60 mm/min elongation rate (Figure 2b).

SMPs often consist of two segments/phases; one of them is a fixed phase and the other is a reversible or switching one. Thus their shape recovery effect is always accompanied with the phase transition of these domains. For the present chain-extended UPy SMP, differential scanning calorimetry (DSC) experiments revealed only melting transitions of the poly(ethylene glycol) (PEG) part in the temperature range from 10 °C to 150 °C with heating rate of 10 °C/min (as shown in Figure 3). And there were almost no differences between the original dry hydrogel film and the elongated film. This means the stretching process at high temperature did not influence the crystal behaviour of the PEG part. Temperature dependence rheology experiments of the dehydrated hydrogel film and fully swelled hydrogel were also conducted. As shown in Figure 4a, both the G' and G" dramatically decreased over a sharp temperature range (50-60 °C). This is the same temperature range where melting of the PEG part and shape recovery were observed. While, in the case of the fully water swelled hydrogel, G' and G" only showed little decrease in the same temperature range (Figure 4b). At the same time, in both cases, the G' values were always higher than G" within the whole temperature range (20-90 °C), suggesting the system was still in solid state even at high temperature and the strong interactions between the present UPy groups.

The above DSC and rheology experiments revealed the thermo-responsive shape-memory phenomena should be contributed to the crystallization transitions of the PEG chains. In other words, the crystallization domains of the PEG part functioned as the temporary physical cross-linkers and fixed the elongated shape; whereas, the aggregated UPy hard domains acted as the permanent physical cross-linkers and held the permanent shape. Cryo-TEM analysis of the hydrogel shows that the UPy domains are spheres that are connected by the PEG chains to have an average distance of the UPy domains of 25 nanometer.

Without wishing to be bound by any theory it is believed that the mechanism of the water-responsive shape-memory phenomena is due to complete transformation of the crystalline PEG domains to solution had occurred following water sorption. DSC experiments of the dehydrated hydrogel film and hydrogel films with different water ratios were thus performed. As shown in Figure 5, the crystallization transition temperature decreased from -60 °C to -23 °C with increasing water ratio from 0 wt % to 31 wt %, respectively. And the sample with 42 wt % water exhibited complete loss of the phase transition peak, suggesting complete transformation of the crystalline PEG domains to solution had occurred following water sorption. Thus, the water swelling process of the present SMP reduced and deconstructed the crystalline domains, which contributed to the water-responsive shape recovery. It is noteworthy that the present SMP is a non-covalently cross-linked SMP, which yields the advantages of easy process thereafter. Rather than the chemically cross-linked ones, which cannot be reshaped after processing.

It is also different from the thermo-moisture responsive polyurethane SMP, which is a glass transition adjusted SMP. The present SMP is a crystallization

transformation tuneable thermo-water responsive one. And in general,

crystallization transition is preferred over the glass transition as melting is comparatively a shaper transition. Additionally, the UPy based supramolecular polymers are eminently suitable for producing bioactive materials owing to their low-temperature processability, favourable degradation and biocompatibility properties. Thus the present CEUPy polymer almost fulfils all the multi-dimensional requirements of SMPs and hydrogels used as biomaterials.

The invention will be further elucidated by the following examples, which by no means should be limiting the scope of the invention.

Amphiphilic copolymers containing biocompatible hydrophilic poly(ethylene glycol) (PEG) midsegments flanked with hydrophobic oligomeric methylene spacers and UPy end-caps capable of quadruple hydrogen bonding were previously prepared (Figure 7, 1 ). The telechelic polymers effectively gelled upon hydration, and showed remarkably dynamic gelation characteristics that depended largely on temperature and pH. The intimately connected structural features and mechanical response are driven in part by the inherent thermodynamic incompatibility between the different chain segments in concert with the strong tendency for

self-complementary aggregation of the UPy motif.

To this end, chain extended copolymers were prepared with the UPy moiety incorporated directly in the backbone (Figure 7, 2). The potential implications of this molecular architecture was gauged from the analogous chain extended

poly(ethylene-s-butene) with benzene tricarboxamide (BTA) groups in the backbone, which exhibited drastically enhanced toughness compared with the telechelic analog. Hydrogels formed from chain extended UPy-PEO copolymers with a structure closely related to polymer 2 show impressive pressure-induced self-healing properties.

The modular synthetic features employed in this invention allow for the

straightforward tailoring of solvophilicity and ductility by adjusting the composition through judicious choice of macromolecular building blocks. This has routinely been accomplished by extending the connectivity [e.g., multiblock copolymers (AB)n]. Extension of telechelic UPy-ABA-UPy copolymers to a segmented multiblock architecture -(UPy-ABA) _n - is demonstrated with various lengths of oligomeric methylene A-segments and different components for the B-segments (e.g., poly(ethylene-butene), poly(ethylene glycol). The increased chain length substantially enhanced the strength, ductility and stability in water for PEG based copolymers compared with the telechelic analogs having nearly identical composition. Self-healing was demonstrated by heat treatment, owing to the dynamic, thermally reversible UPy-dimerization; single component self-healing polymers represent a simplified approach in comparison to more complicated multi-component blends.

Maximal potential from the strength embodied in the UPy aggregates can be reached in strongly hydrophobic (e.g., non-protic) environments. The same principles apply to other supramolecular polymers that rely on hydrogen bonding for self-assembly, such as benzene-1 ,3,5-tricarboxamides. The amido-functionality should be essentially shielded from undesirable hydrophilic/solvophilic segments in hybrid multi-component systems— most prominently so in PEG based hydrogel systems. The mechanical performance rests largely on retention of the physical crosslinks across a range of temperatures and water content.

To this end, the mechanical properties of polymer 2 (Fig. 7, 2) both in bulk and hydrogel states were examined. This was built by a modular strategy combining a pseudo-symmetric OCN-C6Hi ₂ -UPy-C6Hi ₂ -NCO diisocyanate with an

amino-telechelic PEO flanked with oligomethylene spacers—

NH ₂ Ci ₂ H ₂₄ (OC ₂ H ₄ ) ₂₂ 7Ci ₂ H ₂₄ NH ₂ (M _n ~10 kg/mol; PDI ~1 .04). The targeted end-group stoichiometry was 1 :1 during the step-growth chain extension, ultimately providing samples with drastically larger average size, as reflected in a gel permeation chromatogram.

Thermal analysis and water uptake

Calculated by ΦΡΕΟ = AH _m /(AH _m ⁰ -wp _E o) where w = weight fraction of PEO in the sample and AH _m ° (heat of fusion at 100% crystal I in ity) is taken as 196.6 J g ^"1 from: Qiu, W.; Wunderlich, B. Thermochim. Acta 2006, 448, 136-146.

The melting transition of the PEO block in 2 (T _m ,p _E o) occurs with a maximum in the heat flow profile at 56 °C and a melting enthalpy (AH _m = 95 J g ^"1 ) corresponding to a normalized crystalline PEO fraction φ _χ = 0.53 (Figure 8a). The values are expectedly lower than pristine PEO with M _n = 10 kg mol ^"1 , as the confined chain architecture impedes crystallization (T _m = 62 °C; AH _m = 165-175 J g ^"1 ; φ _χ = 0.85- 0.90). Likewise, crystallites within polymer 2 melt with notably less endothermicity than the telechelic polymer 1 (T _m = 56 °C; AH _m = 118-125 J g ^"1 ; φ _χ = 0.66-0.70), suggesting that less chain mobility occurs within the chain extended sample, owing to the segmented macromolecular architecture. This is consistent with higher fraction of tethered PEO bridges between hydrophobic pockets in polymer 2 compared with polymer 1 .

The large proportion of PEO in both polymers 1 and 2 lend to extensive water uptake. Hydrophobic segments phase-separate from the hydrophilic PEO and thereby a physically crosslinked system arises at relatively low polymer

concentrations. However, in the limit of infinite dilution, polymer 1 will ultimately be molecularly dissolved at ambient temperature; dissolution is accelerated at elevated temperature. However, a film consisting of polymer 2 does not dissolve molecularly in water at 25 °C; there was no measurable mass loss after soaking the films at concentration <0.1 mg/mL for longer than 30 days. The water is absorbed into the PEO domains, gradually dislodging increasing amounts of lattice-organized chain segments from crystalline regions with increasing water content, as evident from DSC (Figure 8a). Eventually (ca. 42 wt % water), the material is essentially completely amorphous. Soaking the polymer in very dilute solutions leads to an eventual equilibrium degree of swelling and a transparent sample with 85 wt % water and 86% volume increase (Figure 8b,c). The prevented dissolution presumably arises from the physical crosslinks, which are maintained in the hydrophilic hydrogel environment despite the reliance on hydrogen bonding between UPy groups. This suggests that the hydrophobic methylene spacers effectively shield the UPy aggregation such that disruption/dissociation caused by the hydrated matrix is prevented. This observation is consistent with a microphase separated, physically crosslinked system as opposed to a compositionally homogeneous sample merely having very high molecular mass. In fact, PEO with molar mass of 5,000,000 g mol ^"1 is still water soluble; stable hydrogels are not formed (at >80 wt% water) despite the astronomical chain length (> 100,000 repeating units).

Nanoscale morphology

The remarkable mechanical stability of polymer 2 in water suggests a finely tuned interplay among the water, the PEO, the UPy and the oligo-methylene segments. Transmission electron microscopy (TEM) allowed a glimpse of the nanoscale morphology in both bulk and hydrogel states, ultimately revealing that microphase separation persists across a broad spectrum of conditions. TEM micrographs of dry polymer 2 show a predominantly laminar morphology composed of alternating crystalline and amorphous PEO segments, consistent with the high crystalline content indicated by DSC and wide-angle x-ray scattering. The lamellar thickness measured by TEM is approximately 15 nm, consistent with previously measured periodicities for pure PEO of similar molecular weight.

Cryogenic conditions were exploited to probe the morphological evolution with increasing water content, during which the crystalline lamellae are rendered amorphous. The contrast from the dense fibrous structure associated with the crystallites diminishes in intensity as water is absorbed and the PEO phase become homogeneous. Ultimately upon complete transition to a hydrogel state cryo-TEM micrographs reveal a nearly homogenous light matrix consisting of the amorphous hydrated PEO. There are also darker spots with higher electron density, which can be attributed to small compartments that contain UPy-UPy dimers flanked by the dense oligo-methylene segments. The hydrophobic segments effectively shield the water from disrupting the hydrogen bonding, upon which UPy dimerization depends. Somewhat unexpectedly, the hydrophobic segments appear to be dispersed as small spherical compartments with approximately 2-5 nm diameter, indicated by high resolution imaging. Further investigation using

3-dimensional electron microtomography corroborates both the suggested dry and hydrated morphologies.

The snapshot micrograph of the dry polymer is consistent with the fibrous structure attributed to PEO crystalline lamellae. The white spots dispersed in a black matrix further corroborate the microphase separated hydrogel with hydrophobic spherical domains dispersed in a matrix of PEO-water.

This morphology must contain a large portion of bridged PEO segments connecting the dispersed hydrophobic compartments as a natural consequence of the molecular architecture. The ramifications of this highly interpenetrating network are expressed partially by the remarkable stability of the hydrogel.

The suggested physical network would presumably have profound influence on the mechanical performance, and implicate that resilient materials should be realized in the hydrogel state. Furthermore, the performance of the materials and stability of the physical crosslinks in various temperature regimes is of great interest, both in the pristine and hydrated states.

Mechanical Performance

Dynamic mechanical analysis (DMA) shows a precipitous decrease (2 orders of magnitude) in both storage and loss moduli (E' and E", respectively) over a narrow temperature range of 50-60 °C for polymer 2 (Figure 9a). The indicated T _tra ns corresponds well with T _m p _E o measured by DSC. The transition occurs over a relatively narrow temperature range, which is typical of SMPs that employ semicrystalline switching domains. The associated contributions to the mechanical response indicate predominantly elastic behavior (E'»E") over the entire temperature range (20-140 °C). The hydrophobic domains with strongly

self-associated UPy units and high proportions of interconnected PEO network (bridging) contribute to the elastic behavior well above T _m , PEO, with E' maintained in the range 1-3 MPa from 140-70 °C, marking the rubbery plateau. This is directly comparable to the modulus observed in SMPs comprising poly(methylacrylate), poly(methyl methacrylate), and poly(isobornyl acrylate) mixtures. However, this single-component system offers appealing conceptual simplicity for achieving a similar rubbery modulus.

The contrasting behavior in the different temperature regimes for polymer 2 was explored with uniaxial elongation, which reveals a strong ductile material at ambient temperature (Figure 9b; Young's modulus, E = 234 ± 6 MPa) Chain extended polymer 2 exhibits profoundly different bulk mechanical characteristics than bifunctional polymer 1 , implicating the importance of the chain architecture in accessing increased toughness. For example, uniaxial elongation caused nearly immediate fracture at ambient temperature for polymer 1 . The brittle nature is reflected by the low tensile toughness (TT) value (TT = 0.20 ± 0.02 MJ m ^"3 ) compared with the impressively ductile polymer 2 (TT = 193 ± 12 MJ nrf ³ ).

Maximum elongation before fracture was consistently more than 10 fold the original length of polymer 2, with average yield strength (a _y id) of 13 MPa (Figure 9b). The strong incompatibility of the two components (UPy vs. PEO) allows very low hydrophobic content (w _PE o = 0.92), whereby the modulus is only modestly depressed compared with pristine PEO (5,000 kg mol ^"1 ; E = 285 ± 7 MPa).

Representative stress-strain curve for polymer 2 at ambient temperature highlights the various responses associated with the macroscopic deformation mechanism (Figure 9b). Region I represents the linear viscoelastic regime, which is

characterized by a steep incline that reflects the relatively high Young's modulus. The sample yields and begins to neck at approximately 10% strain, marked by a decrease in the stress resulting from smaller cross-sectional area than the original, undeformed guage (Region II). The sample is subsequently cold drawn (Region III) as the neck propagates throughout the guage length, during which stress is essentially constant. The strength continuously increases after the neck has completely propagated the guage, whereby the sample undergoes strain hardening with a remarkable stress increase (Region IV) until failure at -1000% strain (Point V). The cumulative results from the tensile test suggest a remarkably tough material despite the relatively low UPy-hydrophobic content.

The chain extended polymer 2 exhibits elastic behavior above T _m, p _E o- Resilience was probed with cyclic tensile testing at elevated temperature (T = 70 °C > T _m p _E o) (Figure 9c). The sample became transparent upon heating, consistent with an amorphous microphase separated morphology. Uniaxial extension was applied to 100% strain followed by returning to the original guage length. This cycle was repeated five times and the strain-recovery ratio (R _r ) was calculated for each cycle (equation 1 ).

R (N) ₌ ^e max(N) - e _res (N)

^~ €m∞(N) - e _r „(JV - 1) (1 )

where £ _ma x(N) is the maximum strain for cycle N (i.e., 100% for each cycle) and £ _r es(N) is taken as the point at which negligible stress was measured for the given cycle N. The values for £ _res are indicated on the inset of Figure 9c. The remarkable resilience above T _m, p _E o is likely a direct consequence of the morphological features indicated by TEM.

The extent of physical crosslinking retention upon hydration was also probed by cyclic tensile testing. Dogbone samples were cut from a film prepared by casting the polymer as a concentrated solution in methanol followed by extensive drying under high vacuum. The dry dogbone samples were then submerged in water for several days before performing the extension measurements. As with the elevated temperature measurements, the dogbones were subjected to uniaxial elongation to 100% strain followed by returning to the original position. Identical cycles were repeated five times, and the resilience was evaluated while measuring the forces associated with deformation (Figure 10a). The ultimate stress at 100% elongation was approximately 30* lower than the dry sample at ambient temperature, and 1 .5* lower than the dry sample at 70 °C. However, the recovery ratio was impressively 0.95 after the first cycle, indicating only minor hysteresis. (ref thermoplastic elastomer behavior) The sample exhibited nearly perfect recovery during each subsequent cycle. Excellent recovery was also observed for higher ultimate elongation (e.g., 300% in Figure 10b). After the first cycle, somewhat larger hysteresis was observed; each subsequent cycle had essentially perfect recovery (Figure 10c). The recovery profile is consistent with a material having exceptionally stable physical crosslinks that are retained in the presence of high water content and under strenuous mechanical deformation. This behavior contrasts markedly with the telechelic analog 1 as a hydrogel with 90% water.

Shape memory behavior

The cumulative thermal/mechanical profile was exploited for shape memory behavior. Polymer 2 was formed into a permanent shape by two methods, which reflects the relative ease of processability. The polymer was dissolved in methanol and cast as a film (ca. 0.5 mm thickness). Alternatively the polymer was pressed into a film at 120 °C using an aluminum mold and Teflon sheets for confinement. Either straight strips or curved "S" shapes were cut from the films, representing the permanent geometries.

The hydrophilicity of PEO lent itself toward an alternative stimulus to induce recovery to a permanent shape, which was previously demonstrated for networks having a hydrophilic matrix. However, this could only occur if the complementary hydrogen bonding between UPy groups remains intact while hydrated. The mechanical behavior of the polymers suggests that this may indeed be the case, where the 18 carbon methylene spacers are adequately long and hydrophobic enough to effectively shield the UPy groups from the water, which would detrimentally alter the resilience of the hydrogel material. Likewise, the sharp 7 _tra ns associated with PEO melting in concert with the strongly associated physical network provides an alluring opportunity for shape memory investigation. Two alternative stimuli were demonstrated: heat and water. The permanent "S"-shaped samples were heated to 70 °C. The samples were then straightened (mechanical deformation by hand), followed by cooling to ambient temperature, which effectively fixed this temporary shape upon crystallization of the matrix phase. The samples become visibly transparent upon heating, whereas the semicrystallinity causes increased opacity on cooling. The straightened temporary shape was then submerged in a non-solvent at 70 °C, and the shape recovery was monitored visually with digital photography (Figure 11 a). Thermally activated shape recovery occurred rapidly, fully reverting to the permanent shape within 10 s when the fluid was silicon oil. An alternative non-solvent, heptane, was also shown to activate the recovery process very quickly (< 5 s; see Supporting information, video V1 ). The recovery was attributed solely to the thermal contribution; the mass of the sample was equal before and after the recovery experiment, and the polymer constituents are highly solvophobic for hydrocarbons like heptane. Submergence in liquid was used to aid in visualization of the very fast recovery process.

Submerging a sample in the temporary shape in water at ambient temperature shares some behavioral qualities as thermally activated shape recovery. The temporary straightened shape (Figure 11 b, t = 0 min) was initially opaque and became gradually transparent as water was absorbed. The permanent shape was gradually recovered as the water diffused completely into the structure and crystallinity was correspondingly suppressed. Notably, the recovery process was much slower than thermally induced transition (at 70 °C), requiring more than 15 minutes to recover. As a natural consequence of the hydrophilicity, the recovery was also accompanied by swelling to the eventual equilibrium water content (ca. 90 wt % water). The diffusion of heat through the materials appears to be substantially faster than water, as expected. However, the results are consistent with a material being multiply responsive toward shape recovery. Each pathway could foreseeably lend itself to certain technologically demanding biomaterials. Methods

Thermal properties were investigated with differential scanning calorimetry (DSC) on a Perkin Elmer Differential Scanning Calorimeter Pyris 1 with Pyris 1 DSC Autosampler and Perkin Elmer CCA7 cooling element under a nitrogen

atmosphere with heating and cooling rates of 10 °C/min. Tensile tests were performed on a Zwick Z100 Universal Tensile Tester at an elongation rate of 60 mm/min with a load cell of 200 N in air under room temperature or in an oven at 60 °C. Rheology experiments were carried out on a ARES LS strain controlled rheometer equipment with a 25-mm-diameter para-plate. For the measurements of the hydrogel samples, the samples were placed between the para-plate and the platform with special care to avoid evaporation of water.

Materials and Preparation

All starting materials and reagents were obtained from Sigma Aldrich and used without further purification, unless otherwise specified. Chloroform was dried by 4A molecular sieve before use.

Synthesis of diamine terminated PEG (DAPEG)

CDI activated PEC Diamine terminated PEG

CDI activation of PEG 10k prepolymer

The telechelic hydroxy-terminated poly(ethylene glycol) (M _n = 10,000 g/mol, PEG10k) 24 g, 2.4 mmol) was dried by heating under high vacuum at 120 °C for 3 h. The CDI-activation was performed as described: the solid polymer was added in portions to a solution of 1 ,1 -carbonyldiimidazole (CDI; 2.8 g, 17.2 mmol) in chloroform. The mixture was stirred at 21 °C under an inert argon atmosphere for 8 h. Then the polymer was precipitated in an excess of diethylether. The product was instantly used for the next synthetic steps. Yield: 89%. ¹ H-NMR (400 MHz, CDCI ₃ ): δ = 8.16 (2H, CDI), 7.36 (2H, CDI), 7.07 (2H, CDI), 4.56 (4H, EG next to CDI), 3.84-3.45 (4nH, PEG) ppm.

Diamine-termination of PEG 10k prepolymer

CDI activated PEG10k prepolymer (20.0 g, 5.0 mmol) was dissolved in 80 ml dry chloroform. To this solution 1 ,12-dodecyldiamine (6.41 g, 32.0 mmol) was added and the mixture was stirred under an argon atmosphere at room temperature for 48 h. After verification that conversion was complete by ¹ H-NMR, 80 ml chloroform was added into the mixture and it was filtered. The filtrate was concentrated to around 50 ml and it was added dropwise into 800 ml diethyl ether under vigorous stirring. The white precipitate was collected, redissolved in 50 ml chloroform, filtered and the filtrate was precipitated again in 800 ml diethyl ether. The white precipitate was collected and dried in vacuum oven at 30 °C in presence of P2O ₅ overnight to provide the product as white solid (Yield: 93%). ¹ H-NMR (400 MHz, CDCI3): δ = 4.81 (br, 2H, urethane), 4.17 (m, 4H, next to urethane), 3.79-3.42 (4nH, PEG), 3.11 (q, ³ J = 8.0 Hz, 4H, next to urethane), 2.64 (t, ³ J = 8.0 Hz, 4H, next to amine), 2.01 (br, 4H, next to urethane), 1 .44-1 .22 (m, 40 H, CH ₂ in dodecyl spacer).

HDI-UPy-HDI was synthesized as described in the literature (Sontjens, S. H. M.; Renken, R. A. E.; van Gemert, G. M. L; Engels, T. A. P.; Bosman, A. W.; Janssen, H. M.; Govaert, L. E.; Baaijens, F. P. T. Macromolecules 2008, 41, 5703).

Synthesis of chain-extended UPy hydrogelator (CEUPy polymer)

CEUPy polymer HDI-UPy-HDI, 0.455 g in dry DMF (5 ml), was quickly added to a clear solution of DAPEG (10 g) in dry DMF (95 ml) with stirring at room temperature under N ₂ atmosphere. After 30 min the reaction mixture was heated up to 75 °C for another 5 h, resulting in a clear viscous solution. The solution was precipitated in diethylene ether (1 .5 L), yielding a white power. The power was collected and dried at 40 °C under high vacuum. Yield: 98 %. ¹ H NMR (400 MHz, CD ₃ OD) δ 4.14 (br), 3.80 (q), 3.63 (br), 3.46(t), 3.30(br), 3.09(br), 2.27(s), 1 .46(br), 1 .30(br). High temperature GPC (CHCI ₃ , 90 °C, PS standard): Mn = 51 .5 KD, Mw = 177.8 KD, PDI = 3.45.

Preparation of the shape-memory hydrogel materials

The chain-extended UPy hydrogelator is dissolved in methanol and by evaporation of the methanol the shape-memory material is formed. By evaporation of the methanol from a film it will make a film and similarly fibers, tubes and any other form can be made. For a tube see Figure 6. The material can now be transformed in any other form by stretching at room temperature for thin specimens and by elevated temperature at 60-70 °C for thicker samples followed by cooling in the new form. The old form is restored by subsequent heating (thermo-triggered shape memory) or by the swelling in water (water-triggered shape memory, by which a strong elastic hydrogel is formed with the form similar to the original sample when brought into contact with enough water.