Silk Fibroin Cryogel Building Adaptive Organohydrogels with Switching Mechanics and Viscoelasticity

: In contrast to synthetic gels, their biological counterparts such as cells and tissues have synergistic biphasic components containing both hydrophilic and lyophilic phases, providing them some special abilities including adaptive biomechanics and freezing tolerance. Hydrogels containing both hydrophilic and lyophilic phases, referred to as organohydrogels (OHGs), are capable of mimicking the biological systems, and they might have great potential in various applications. Here, we present a facile strategy to obtain adaptive OHGs with tunable and programmable mechanics and viscoelasticity. We utilize a hydrophilic cryogel sca ﬀ old as the continuous phase of OHGs, while the pores of the sca ﬀ old act as the reaction loci for the formation of organogel microinclusions. Thus, we ﬁ rst prepared mechanically robust cryogels based on silk ﬁ broin (SF) via cryogelation reactions at − 18 ° C. The cryogels with 94% porosity containing interconnected μ m-sized pores were then immersed in an ethanolic solution of acrylic acid (AAc), n -octadecyl acrylate (C18A), N , N ′ -methylenebis(acrylamide), and a free-radical initiator. Polymerization reactions conducted within the pores of the cryogels lead to mechanically strong adaptive OHGs consisting of a SF sca ﬀ old containing semi-crystalline poly(AAc-co - C18A) organogel microinclusions. The mechanical strength of OHGs is much higher than that of their components due to the signi ﬁ cant energy dissipation in the OHG networks. Depending on the amount of the crystallizable C18A monomer units, the melting temperature T m and the degree of crystallinity of OHGs could be varied between 49 and 54 ° C and 1.3 and13%, respectively. The crystallinity created in OHGs provided them switchable mechanics and viscoelasticity in response to a temperature change between below and above T m . All OHGs exhibited shape-memory function with a shape-recovery ratio of more than 92%. The strategy developed here to obtain high-strength smart OHGs is suitable for a wide variety of combinations of hydrophilic sca ﬀ olds and organogels.


INTRODUCTION
In contrast to synthetic gels, their biological counterparts such as cells and tissues have synergistic biphasic components containing both hydrophilic and lyophilic phases simultaneously, which showed up during billions of years by evolutionary natural selection. 1 Such coexistence of the antagonistic features in the body provides some special abilities including adaptive biomechanics and freezing tolerance. For instance, echinoderms such as sea cucumber and starfish possessing mutable collagenous tissue, which enables a nervous system-induced stiffness altering, can regulate their body mechanics as soft and hard modes, when they are at rest and threatened, respectively. 2−4 Moreover, arctic wood frogs containing specialized proteins called antifreeze proteins, which possess both hydrophilic and lyophilic fragments, can live in harsh and cold environments without freezing. 5,6 These evolutionary adaptations of the biological systems have always been a model for scientists to develop new-generation synthetic smart materials. Such materials containing both hydrophilic and lyophilic phases can be referred to as OHGs, and they might have great potential in various applications including sensors of humidity, 7 strain, 8−11 and temperature, 12 microfluidic and optical devices, 13,14 signal transmission, 15 freeze-tolerant materials, 16−18 and soft robotics. 19 One generally used strategy to prepare an OHG is dispersing a discontinuous organic (lyophilic) phase in a continuous aqueous phase to obtain an oil-in-water emulsion. 4,20 These kinds of systems are thermodynamically unstable emulsion systems, and their kinetic stabilization needs some special stabilizers such as amphiphilic copolymers, nanoparticles, or surfactants. 19,21,22 Recently, OHGs prepared in the absence of a stabilizer were also reported, in which bi-step manufacturing techniques were used. For instance, Gao et al. synthesized freeze-tolerant OHGs after UV irradiation of a system formed via the infusion of a lyophilic phase to a pre-synthesized and dehydrated poly(N,N-dimethylacrylamide) gel network. 23 In a similar manner, anti-freezing compounds including glycerol, glycol, and sorbitol were introduced into a pre-obtained Caalginate/poly(acrylamide) gel network in order to fabricate non-freezable materials. 17 However, these kinds of studies are mainly focused on the freeze-tolerant properties of the materials rather than the switchable mechanical and viscoelastic abilities. In addition, the first network of these OHGs is classical hydrogels possessing insufficient mechanical properties due to poor energy dissipation. 24,25 Here, we present a facile strategy to obtain adaptive OHGs with tunable and programmable mechanics and viscoelasticity. Instead of dispersing the lyophilic phase in a continuous aqueous phase, we utilize a macroporous hydrophilic cryogel scaffold as the continuous phase, while the pores of the scaffold act as the reaction loci for the formation of microinclusions, as schematically illustrated in Scheme 1.
Cryogelation, that is, gelation conducted below the freezing point of the reaction system is a very efficient method for the preparation of tough macroporous materials having an interconnected open porous architecture. 26−30 In contrast to the classical hydrogels, the cryogels possess permanent pores independent of the drying conditions, and exhibit extraordinary mechanical properties such as superior compressibility (>95%) and fatigue resistance in addition to rapid water absorption/desorption capability occurred within seconds. 29,31−39 These abilities provide them an advantage over the traditional hydrogels, especially for load bearing applications.
In the present study, we first prepared mechanically robust cryogels based on methacrylated silk fibroin (SF) and N,Ndimethylacrylamide (DMAA) as a spacer via cryogelation reactions at −18°C (Scheme 1a,c). The cryogels with 94% porosity containing interconnected μm-sized pores were then immersed in an ethanolic solution of the hydrophilic acrylic acid (AAc) monomer, crystallizable hydrophobic n-octadecyl acrylate (C18A) comonomer, N,N′-methylenebis(acrylamide) (BAAm) cross-linker, and a free-radical initiator (Scheme 1b). As typical for many cryogels, 29 SF cryogels do not swell in this solution by volume, but their mass significantly increases due to the filling of their pores with the solution. Free-radical polymerization reactions conducted within the pores of the cryogels lead to mechanically strong adaptive OHGs consisting of a SF scaffold containing semi-crystalline poly(AAc-co-C18A) organogel microinclusions (Scheme 1c,d). As will be seen below, the mechanical strength of OHGs is much higher than that of their components due to the significant energy dissipation in the OHG networks. The melting temperature, T m , and the degree of crystallinity of OHGs could be varied between 49 and 54°C and 1.3 and13%, respectively, by varying the amount of the crystallizable C18A units in the organogel component. The crystallinity created in OHGs provided them switchable mechanics and viscoelasticity in response to a temperature change between below and above T m . What is more, all OHGs exhibited shape-memory function with a shape-recovery ratio of more than 92%.

RESULTS AND DISCUSSION
In the following, we first discuss the preparation and characteristics of SF cryogels forming the continuous phase of the OHGs. The copolymerization of AAc and C18A monomers within the macropores of the cryogels and the properties of OHGs containing organogel microinclusions are then discussed in the following sections.
2.1. SF Cryogel Scaffold of OHGs. A SF-based cryogel scaffold used as the skeleton of OHGs was prepared from an aqueous solution of methacrylated SF (meth-SF) with a methacrylation degree of 14% and DMAA monomer as a spacer (Scheme 1a). 1,4-Butanediol diglycidyl ether (BDDE) was the cross-linker for SF, while the ammonium persulfate (APS) and N,N,N′,N′-tetramethylethylenediamine (TEMED) redox system was the initiator for DMAA. The synthesis conditions were the same as those of the SF hydrogels we reported before, 40 except that the reaction temperature was reduced from 50 to −18°C to provide cryogenic conditions. In the preliminary experiments, the concentrations of meth-SF and DMAA were varied over a wide range to optimize the cryogel properties suitable for OHG preparation, that is, the highest weight swelling ratio, porosity and gel fraction, and the largest pore size, together with good mechanical properties (Figures S1−S3). The cryogel scaffold with the optimum properties could be obtained from an aqueous solution of 5 w/ v % meth-SF and 2 w/v % DMAA, that is, the pore walls of the cryogel are composed of 71 wt % SF and the rest is poly(DMAA) (PDMAA) which is a biocompatible polymer (Scheme 1c). 41−43 The extraction tests in water revealed that the gel fraction W g is 0.84 ± 0.02, indicating that 84% of SF and PDMAA are incorporated into the cross-linked cryogel network ( Figure S1). The conformation analysis of SF of the cryogel could not be conducted by both Fourier transform infrared (FTIR) and X-ray diffraction (XRD) techniques due to the overlapping peaks of PDMAA and SF. 40 Therefore, we carried out FTIR measurements on the cryogel prepared under the same experimental conditions but in the absence of DMAA. Figure 1a,b shows FTIR spectra of meth-SF before (a) and after (b) cryogelation. The symbols represent the original data, while the solid and dashed curves represent the fitted spectra and hidden peaks at 1620, 1640, 1660, and 1698 cm −1 due to the β-sheets, random coil, α-helix, and β-turn configurations, respectively. 44−46 Calculations reveal that the content of β-sheets increases from 15 to 32% after cryogelation reactions, which is responsible for the formation of a waterinsoluble scaffold ( Figure S4). We should note that the β-sheet content of the present SF cryogels is lower than that of the corresponding SF hydrogels formed at 50°C. 40,46 This is due to the decreasing reaction temperature impeding conformation transition in fibroin. However, when compared with the cryogels based on native SF, 34 they exhibit a similar β-sheet content (33 ± 2%).
The weight (q w ) and volume swelling ratios (q v ) of the cryogel in water were determined as 13 ± 1 and 1.0 ± 0.1, respectively ( Figure S1). Thus, there is no volume change of the cryogel when immersed in water, while its mass increases 13-fold, indicating that the gel swelling is due to the filling of the pores by water. In the following, we will use the term "wet cryogel" instead of "swollen cryogel" as no swelling of the SF network occurs. The nonswelling of the cryogel by volume is due to the cryoconcentration phenomenon producing a very high concentration of SF and PDMAA in the unfrozen domains of the cryogelation reactions forming the pore walls of the cryogel. 29 The total porosity P and the pore volume V p of the cryogel were measured as 93.7 ± 0.5% and 11.5 mL·g −1 , respectively, indicating a highly porous structure with interconnected open  pores. Scanning electron microscopy (SEM) and confocal microscopy measurements indeed revealed a macroporous structure of the cryogel scaffold with interconnected pores (Figure 1c−e). The scaffold exhibits a honeycomb morphology consisting of quasi-spherical pores. The sphere-equivalent diameter of the pores varies between 23 and 73 μm, with an average diameter of 44 ± 12 μm (Figure 1f). Figure 2a shows compressive stress−strain curves of the wet cryogel and the corresponding SF hydrogel formed at 50°C. 40 The inset of the figure is a semi-logarithmic plot of the data to highlight the low-strain regime. Although both the cryogel and hydrogel sustain 96% compression, the wet cryogel exhibits a Young's modulus E of 90 ± 20 kPa and a fracture stress σ f of 6.1 ± 0.8 MPa, which are around 15 and 8 times higher than those of the hydrogel (6.1 ± 0.5 kPa and 0.8 ± 0.2 MPa, respectively). Moreover, the stress−strain curve of the cryogel exhibits a plateau-like regime below 60% strain, during which the stress only slightly increases with increasing strain (inset of Figure 2a). We can explain this regime with the deformation of the pore walls under strain during which water in the cryogel is forced out of the pores so that the cryogel easily deforms. Moreover, the steep increase of the stress−strain curve above 60% strain is attributed to the compression of the nearly nonporous cryogel network with completely squeezed pore walls.
This explanation was indeed observed visually, as shown in Figure 2b. Here, a cryogel specimen in the wet state is compressed up to 80% and then unloaded to zero strain. Because the weight swelling ratio q w of the cryogel is 13 ± 1, the water content of the specimen is around 92 wt %, the rest being the SF/PDMAA network forming the pore walls. It is seen that this large amount of water in the cryogel is forced out of the pores due to its incompressibility during compression. Interestingly, during unloading, water expelled from the pores is absorbed again, and the sample recovers its original shape within seconds without damage. In contrast to the cryogel, the hydrogel underwent an irreversible (permanent) deformation until it fractures at 95 ± 2% compression. The observed behavior of the cryogel associated with the flowing-out and flowing-in water through the pores due to the elastic pore walls is known as poroelasticity and is also observed in soft tissues, such as articular cartilage, providing their biomechanical functions including load bearing ability. 47−49 The reversible flowing-out and flowing-in water through the pores of the cryogels were also demonstrated by rheological tests, as detailed in the Supporting Information text and Figure S5. Concluding this section, the mechanically robust SF cryogel component of OHGs presented here not only exhibits a high porosity and a macroporous structure with interconnected μmsized pores, but also its pore walls are elastic withstanding large deformations without a permanent damage. These extraordinary material properties are due to the cryoconcentration phenomenon occurring at the beginning of the cryogelation reactions. 28,29,34 Thus, during freezing of the reaction system at −18°C, the reactants are expelled from the ice crystals and accumulate in the unfrozen domains to form a highly concentrated solution where the reaction occurs. As a consequence, these unfrozen domains forming the pore walls of the cryogel consist of a high concentration SF and PDMAA, exhibiting elasticity and mechanical strength.
2.2. Formation of OHGs. The OHGs were prepared in two steps, namely, (i) immersing freeze-dried cryogel specimens in an ethanolic solution of AAc and C18A monomers, N,N′-methylenebis(acrylamide) (BAAm) cross-linker, and α,α′-azoisobutyronitrile (AIBN) initiator and (ii) free radical copolymerization reactions at 50°C within the pores of the cryogel. The total monomer concentration (C o ) in the ethanolic solution was fixed at 41 w/v % throughout this study, while the mole fraction of C18A, denoted by x C18A , was varied between 0.10 and 0.30. The concentrations of BAAm and AIBN were 1 mol % with respect to the monomers. Similar to the swelling behavior of the cryogel in water, the cryogel volume in the monomer solution remained unchanged, while its weight significantly increased due to the filling of the pores with the solution. The weight swelling ratio of the cryogels in the monomer solution, denoted by q w,o , slightly decreased from 11 to 10 with increasing C18A content of the solution ( Figure  3a).
After immersing the cryogel specimens in the reaction solution for 10 min, the solution containing cryogels were transferred into several plastic syringes, and free radical copolymerization reactions were conducted at 50°C for 24 h. In this way, SF scaffolds containing poly(AAc-co-C18A) organogel in their pores were obtained. They were then immersed in ethanol to extract soluble species for one week by replacing it every second day and then immersed in water to reach swelling equilibrium. The gel fraction W g was found to be 0.96 ± 0.02 for the OHG formed at x C18A = 0.10, while at higher C18A contents, all the monomers were completely incorporated into the pores of the cryogel scaffold. The weight fractions of C18A units and poly(AAc-co-C18A) organogel in OHGs with respect to dry states, represented by w C18A and w organogel , respectively, were calculated as w w m where m C18A is the weight fraction of C18A in the monomer mixture. As seen in Figure 3b, the weight fraction of C18A (w C18A ) in OHGs increases from 26 to 52% as x C18A is increased from 0.10 to 0.30, while the organogel weight fraction remains almost constant at 79%. The high amount of the organogel introduced in the scaffold is due to its large pore volume acting as the reaction loci for the formation of microinclusions. Figure 3c,d shows the weight (q w ) and volume swelling ratios (q v ) of OHGs in water (c) and ethanol (d). For comparison, the data of the starting cryogel scaffold are also shown by the open symbols. As compared to the scaffold, the weight swelling ratio q w of OHGs in water significantly decreases, while their volume swelling ratio q v slightly increases. The increase in q v is attributed to the hydrophilicity of AAc units of the organogel in the OHG, while decreasing q w reflects decreasing porosity. Because ethanol is a better solvent for poly(AAc-co-C18A) as compared to SF, 50 OHGs attain higher volumes and masses when immersed in ethanol. Figure  3e,f compares the total porosity P (e) and the pore volume V p (f) of the starting scaffold (open symbols) with OHGs (filled symbols). Both P and V p decrease from 93.7 ± 0.5 to 40 ± 2% and from 11.5 to 0.36−0.57 mL·g −1 %, respectively, after incorporation of the organogel in the pores of the cryogel scaffold reflecting pore filling with the organogel component. Figure 4 shows SEM images of the starting scaffold (left) and the OHG prepared at x C18A = 0.30 (right). The honeycomb pore morphology of the scaffold can also be seen after incorporation of the organogel in the pores. Indeed, the size distribution of the pores estimated from the images remained unchanged at 44 ± 12 μm. However, a closer examination of the SEM images reveals that as compared to the scaffold, the deepness of the pore holes disappears in the OHG, as shown in the zoomed-in images in the bottom panel of Figure 4. Similar SEM images were also obtained for OHGs prepared at various x C18A between 0.10 and 0.30 ( Figure S6). Increasing C18A content of the monomer solution decreased the deepness of the pore holes, as revealed by the decrease of the white area fraction in the images, while the pore diameters remained constant ( Figures S6 and S7). To visualize the distribution of the organogel and scaffold components in the OHGs, confocal microscopy measurements were conducted on an OHG specimen formed at x C18A = 0.30 after staining its organogel and scaffold components red and green, respectively. Figures 5a−c and S8a−c show confocal microscopy images of the scaffold (a) and organogel phases (b), and the OHG (c). Some black areas are seen in Figure 5c, indicating the existence of the unfilled pores in the OHG. Moreover, although the combination of the green scaffold and red organogel should predominantly result in a yellow color for the OHG, it appears yellowish orange, revealing the presence of an excess amount of the organogel component in the OHG. Nevertheless, the images reflect that the organogel component distributes over the whole scaffold to produce an OHG consisting of the SF scaffold with μm-sized pores partially filled with poly(AAc-co-C18A) organogel.
2.3. Thermal, Viscoelastic, and Mechanical Properties of OHGs. Figure 6a shows differential scanning calorimetry (DSC) thermograms of the starting cryogel and OHGs prepared at various x C18A , as indicated. As expected, no peak appears in the thermogram of the SF cryogel, while OHGs exhibit a melting peak that becomes sharper and shifts to a higher temperature with increasing C18A content. The melting temperature T m of OHG slightly increases from 49 ± 1 to 54 ± 1°C as x C18A is increased from 0.10 to 0.30. Upon cooling from above to below T m , no crystallization peak appeared in the DSC scan for OHG with x C18A = 0.10, which is attributed to the steric hindrance of the cryogel scaffold to the reformation of crystalline domains ( Figure S9). For OHGs with a higher x C18A , the crystallization temperature T cry increases from 40 ± 1 to 43 ± 1°C with increasing x C18A . Figure 6b shows the degree of crystallinity f cry of OHGs, which is the fraction of C18A units forming side chain crystals, plotted against x C18A . f cry increases from 0.013 ± 0.001 to 0.13 ± 0.01 with increasing x C18A from 0.10 to 0.30, indicating that 1.3 to 13% of C18A units in the pores of the scaffold form crystalline domains. The f cry values are much smaller than those reported for semicrystalline hydrogels based on AAc and C18A ACS Applied Polymer Materials pubs.acs.org/acsapm Article (10−35%), 51,52 which is attributed to the confinement of the poly(AAc-co-C18A) organogel within the pores of OHG, limiting the alignment of C18 side chains. Figure 6c shows XRD patterns of OHGs at various x C18A together with their components, namely, SF cryogel and poly(AAc-co-C18A) organogel with x C18A = 0.30. In accordance with the DSC results, XRD patterns of OHGs and organogel exhibit a distinct peak at 21.6°, corresponding to a d-spacing of 0.41 nm, that is, the distance between two aligned C18 side chains perpendicular to the same polymer backbone. 52,53 This peak becomes sharper as x C18A is increased, reflecting increasing degree of crystallinity. Moreover, the broad peak at 20°that appears in the cryogel and becomes a shoulder to the main peak of OHG with x C18A = 0.10 is typical for PDMAA existing in the cryogel network. 54 The crystallinity created in OHGs provided them switchable mechanics and viscoelasticity in response to a change in the temperature between below and above T m . Figure 6d shows the variations of the storage modulus G′ and loss factor tan δ (= G″/G′ where G″ is the loss modulus) of an OHG specimen formed at x C18A = 0.30 during heating from 25 to 65°C at a rate of 4°C·min −1 . The storage modulus G′ decreases sevenfold (from 2 to 0.3 MPa) upon heating above T m , while tan δ remains above 0.1, as typical for a so-called weak hydrogel. 55 Because tan δ represents the ratio of the dissipated energy to the stored energy under the sinusoidal deformation, this reveals that the melting of C18 crystals in OHG decreases the solid-like character of OHG, while its dissipative properties remain almost preserved. Figure 6e−g shows G′, G″ , and tan δ of the same OHG at 25 (e) and 65°C (f), and after cooling back to 25°C (g), as a function of the angular frequency ω. G′ exhibits a power law increase with ω, G′ ∼ ω n at both below and above T m , with the exponent n slightly increasing by heating above T m from 0.036 ± 0.006 to 0.089 ± 0.001, indicating the more viscous character of the OHG above T m as compared to that below T m . Moreover, cooling the OHG specimen from 65 to 25°C recovers both its initial modulus G′ and the exponent n of the frequency dependence. This shows that OHGs have reversibly switchable viscoelasticity in response to a change in the temperature between below and above T m .  The mechanical performances of OHGs were determined by uniaxial compression tests at 23 ± 2°C. Figure 7a compares the stress−strain curve of the OHG formed at x C18A = 0.30 with its poly(AAc-co-C18A) organogel and SF cryogel components, where the shaded areas under the curves represent their toughness W, that is, the energy to break. The toughness W of the OHG is 6 ± 1 MJ·m −3 , which is 3and 18-fold higher than that of the organogel and cryogel components, respectively (2.3 ± 0.4 and 0.33 ± 0.04 MJ·m −3 ), reflecting that the combination of the OHG components in a single gel results in a significant toughness improvement (inset of Figure 7a). Figure 7b,c presents stress−strain curves of OHGs formed at various x C18A and the starting SF cryogel and their mechanical parameters. The inset of (b) represents a semi-logarithmic plot of the same data up to 50% strain. A significant enhancement in all mechanical properties is observable after incorporation of the organogel component into the pores of the cryogel scaffold. For instance, the modulus E of the cryogel increases from 90 ± 20 kPa to 16 ± 3 MPa at x C18A = 0.30, indicating 180-fold increase in its stiffness. Increasing C18A content of the organogel also increases the modulus, fractures stress, and toughness, while deformation at break slightly decreases but still remains above 85%. This increase is attributed to the C18A units forming alkyl crystals in μm-sized pores of the cryogel. Moreover, the increase in the plateau stress of OHGs with increasing x C18A also reflects increasing stability of their pores filled with the organogels (inset of Figure 7b). We should note that the compressive mechanical properties of OHGs are much higher than those of the cryogels 29, 34, 35, 37− 39 and OHGs reported before. [7][8][9][10][14][15][16]18,19,21 The dramatic increase in the mechanical strength of OHGs as compared to the cryogel and organogel components and their toughening mechanism are attributed to the significant energy dissipation in OHG networks. Thus, the brittle organogel component of the OHG in the dispersed phase acts as a sacrificial bond and fractures at a low strain, 40% (Figure 7a), by dissipating energy, while the continuous SF scaffold phase keeps the OHG together. This mechanism is similar to that of double-network hydrogels composed of tough and ductile components. 56,57 To verify the toughening mechanism of OHGs, successive loading−unloading cycles were conducted up to a maximum compressive strain ε max of 80% with a waiting time of 5 min between cycles. Figure 8a,b shows eight successive cycles with increasing ε max from 10 to 80% (a) and five successive cycles up to a fixed    Figure S10). The cyclic stress−strain curves of both OHG and cryogel show different paths of the loading and unloading curves, represented by the solid and dotted curves, respectively. The insets of Figure 8a,b shows the hysteresis energy U hys , that is, the area covered by the loading and unloading curves, for the OHG (filled symbols) and the cryogel (open symbols) plotted against ε max (a) and cycle number (b). It is seen that U hys for OHG is an order of magnitude higher than that of the cryogel. Because U hys is a measure of energy dissipation during a mechanical cycle, it reveals that a much larger amount of energy is dissipated from OHG as compared to the cryogel during each cycle. Assuming that the viscoelastic effect in stress−strain curves is negligibly small, the extra energy dissipation is due to the microscopic fracture of the organogel by dissipating energy before the catastrophic failure of the OHG structure. Thus, the toughening of the SF cryogel occurs due to the introduction of the brittle organogel component acting as sacrificial bonds and hence energy dissipative structure.
2.4. Shape-Memory Behavior of OHGs. All OHGs reported here exhibited shape-memory function induced by heating above the melting temperature T m of their crystalline domains. For instance, the images in Figure 9a−c show the shape-fixity and shape-recovery behavior of an OHG specimen with x C18A = 0.30. Its permanent shape is rod shape (a). Heating above T m (65°C), deforming to a horseshoe-like temporary shape, followed by cooling below T m (25°C), fixes its temporary shape (b). The specimen returns to its initial shape by increasing the temperature above T m (c). The shapememory behavior of OHGs is attributed to the coexistence of C18 crystals in the microinclusions and the cryogel matrix forming their skeleton. Thus, C18 crystals confined within the pores act as switching segments to fix the temporary shape, while the continuous cryogel component determines its permanent shape due to the entropic elasticity of the pore walls. Figure 9d,e shows the shape fixity ratio R f and the temperature dependent shape-recovery ratio R r of OHGs, respectively, calculated as where θ d , θ f , and θ r are the deformed angle under load, fixed angle after unloading, and recovered angle upon heating, respectively. Figure 9d reveals that the R f ratio significantly increases with increasing x C18A of OHGs, that is, with increasing degree of crystallinity. For instance, R f increases from 29 to 99.5% with increasing x C18A , indicating that the OHG at x C18A = 0.30 is able to fix its temporary shape almost completely. Moreover, all OHGs exhibit a shape-recovery ratio R r of more than 92% at 70°C, reflecting their switchable mechanical properties in response to temperature (Figure 9e). Figure 9. (a−c) Shape-fixity and shape-recovery behavior of an OHG specimen with x C18A = 0.30. θ f and θ r are the fixed angle after unloading, and the recovered angle at the temperature indicated, respectively. (d,e) Shape-fixity ratio R f (d) and the temperature-dependent recovery ratio R r (e) for OHGs with various x C18A .

CONCLUSIONS
Soft materials containing both hydrophilic and lyophilic phases, namely, OHGs, have great potential in various applications including sensors, microfluidics and optical devices, signal transmission, freeze-tolerant materials, and soft robotics. Here, we presented a facile strategy to obtain adaptive OHGs with tunable and programmable mechanics and viscoelasticity. We utilize the hydrophilic cryogel scaffold as the continuous phase, while the pores of the scaffold act as the reaction loci for the formation of microinclusions. Thus, the pores are filled with a monomer, crystallizable hydrophobic comonomer, chemical cross-linker, and an initiator, and then, free-radical polymerization is conducted within the pores of the cryogel to fabricate OHGs containing hydrophobic microinclusions. In the first step of OHG preparation, mechanically robust cryogels based on meth-SF and PDMAA were prepared via cryogelation reactions at −18°C. The cryogels with 94% porosity containing interconnected μmsized pores were then immersed in an ethanolic solution of AAc, C18A, BAAm, and AIBN initiator. Polymerization reactions conducted within the pores of the cryogels lead to mechanically strong adaptive OHGs consisting of the SF scaffold containing semi-crystalline poly(AAc-co-C18A) organogel microinclusions. A significant enhancement in all mechanical properties was observed after incorporation of the organogel component into the pores of the cryogel scaffold. The modulus E of the cryogel increases from 90 ± 20 kPa to 16 ± 3 MPa at x C18A = 0.30, indicating 180-fold increase in its stiffness. Simultaneously, the fracture stress σ f increases 5.4-fold (from 6.1 ± 0.8 to 33 ± 1 MPa), while the compression ratio at break remains almost constant. The dramatic increase in the mechanical strength of OHGs as compared to their components is attributed to the significant energy dissipation in OHG networks. Thus, the brittle organogel component in the pores, acting as a sacrificial bond, fractures at a low strain by dissipating energy, while the continuous SF scaffold phase keeps the OHG together. This toughening mechanism was also verified by successive loading−unloading cycles conducted up to a maximum compressive strain below failure. Increasing C18A content of the organogel also increases both the modulus and fracture stress, while deformation at break slightly decreases but still remains above 85%. This increase is attributed to the C18A units forming alkyl crystals in μm-sized pores of the cryogel. The melting temperature T m and the degree of crystallinity of OHGs could be varied between 49 and 54°C and 1.3−13%, respectively, by varying the amount of the crystallizable C18A units in the organogel component. The crystallinity created in OHGs provided them switchable mechanics and viscoelasticity, in response to a temperature change between below and above T m . All OHGs exhibited shape-memory function with a shape-recovery ratio of more than 92%. Thus, OHGs presented here are a good candidate in various applications including biomedical, tissue engineering, aerospace, and electronics. The strategy developed here to obtain highstrength smart OHGs is suitable for a wide variety of combinations of hydrophilic scaffolds and organogels.
4.2. Isolation and Methacrylation of SF. The isolation and methacrylation of SF were performed as reported before. 40,58 Briefly, 10 g of Bombyx mori silkworm cocoons was cut into small pieces, and after cleaning with distilled water, they were boiled in an aqueous solution of Na 2 CO 3 (1 L, 0.02 M) for 1 h to remove sericin proteins. After five-time washing of the SF with 1 L of distilled water at 70°C for 20 min each, it was dried at room temperature for 2 days. 7 g of dried SF was then dissolved in 9.3 M LiBr solution (35 mL) at 60°C within 2 h. After complete dissolution, 2 mL of GM was added for the methacrylation of SF, and stirring was continued at 60°C for further 3 h at 500 rpm. This procedure resulted in meth-SF possessing a methacrylation degree of 14%, which was determined by 1 H-NMR as detailed in our previous work. 40 Finally, the solution was dialyzed using a dialysis tube (10,000 MWCO, Snake Skin, Pierce) against distilled water that was refreshed three times a day for 4 days. In this way, an aqueous solution of 7 w/v % meth-SF was obtained, as determined gravimetrically. In order to increase its concentration, meth-SF aqueous solution was further dialyzed against PEG-10000 solution (15 w/v %) using a dialysis tube of 3500 MWCO (Snake Skin, Pierce).
4.3. Fabrication of SF Cryogels and OHGs. The cryogels were prepared from an aqueous solution of meth-SF with a methacrylation degree of 14% and DMAA monomer as a spacer. BDDE was the cross-linker for SF while the APS-TEMED redox system was the initiator for DMAA. TEMED and APS concentrations were fixed at 0.025 v/v % and 3.51 mM, respectively, 40 while the BDDE concentration was decreased from 20 to 3 mmol·g −1 (with respect to meth-SF) with increasing amount of meth-SF to prevent early gelation of the reaction solution prior to the cryoconcentration of the reactants. The synthesis conditions were the same as those of the SF hydrogels we reported before, 40 except that the reaction temperature was reduced from 50 to −18°C to provide cryogenic conditions. To optimize the formation conditions of SF cryogels, two series of experiments were carried out: in the first series, the concentration of meth-SF was fixed at 5 w/v %, while the DMAA content was varied from 0 to 10 w/v %. In the second series, the DMAA content was fixed at 2 w/v %, and the meth-SF concentration was changed between 5 and 15 w/v %. For instance, to fabricate a cryogel at meth-SF and DMAA concentrations of 5 and 2 w/v %, respectively, 0.200 g (0.208 mL) of DMAA and 1.00 mL of BDDE were first added to 7.14 mL of aqueous solution of 7.0 w/v % meth-SF. Then, 0.10 mL of APS and 0.125 mL of TEMED were introduced from their stock solutions into this mixture, whose amount was finally completed to 10 mL by using distilled water. Finally, homogeneous reaction solutions were transferred into several plastic syringes that were immediately placed at −18°C for 24 h. After thawing, the cryogels were ejected from the syringes and placed in an excess of distilled water to remove the unreacted species. After 3 days, wet cryogels were dried using a freezedryer (Christ Alpha 2−4 LD-plus) by first freezing at −25°C for 1 day, followed by freeze-drying at −40°C/0.12 mbar vacuum and −60°C /0.011 mbar vacuum for the second and the third day, respectively. For comparison, the corresponding hydrogels at meth-SF and DMAA concentrations of 5 and 2 w/v %, respectively, were also prepared at 50°C for 24 h, as described previously. 40 To fabricate OHGs, we used a cryogel scaffold prepared at meth-SF and DMAA concentrations of 5 and 2 w/v %, respectively. The cryogel specimens were immersed in ethanolic solutions containing AAc, C18A, AIBN, and BAAm at 40°C. The total monomer (AAc + C18A) concentration was fixed at 41 w/v %, while the mole fraction of C18A, denoted by x C18A , was varied between 0.10 and 0.30. The concentrations of the AIBN initiator and BAAm cross-linker were 1 mol % with respect to the total monomers. The cryogels immediately absorbed the solution to attain a constant mass. After 10 min, organic solution-embedded cryogel specimens were transferred into several plastic syringes together with the organic solution and placed in an oven at 50°C to initiate the free radical copolymerization of AAc, C18A, and BAAm within the pores of the cryogels. After 24 h, OHGs were separated by peeling off the surrounding poly(AAc-co-C18A) organogel. The OHGs thus obtained were first placed in an excess of ethanol to remove unreacted species, and then in water at least for 1 week to replace ethanol. For comparison, poly(AAc-co-C18A) organogels with x C18A = 0.30 were fabricated by the following route: AAc (1.32 mL) and C18A (2.70 g) were first dissolved in 4 mL ethanol at 40°C. Then, AIBN and BAAm were added (both 1 mol % with respect to monomers) to this solution, and the final volume was completed to 10 mL with ethanol. Reaction solution was then transferred into several plastic syringes, and they were placed in an oven at 50°C for 24 h. All material characterization studies were performed on the water-swollen OHGs, except XRD measurements where dry powdered OHG specimens were used.
4.4. Characterization of the Cryogels and OHGs. The weight (q w ) and volume (q v ) swelling ratios of the cryogels and OHGs were calculated using the equations q w = m/m dry and q i = (D/D dry ) 3 , respectively, where m and D are the weights and the diameters of the swollen specimens and m dry and D dry have the same meaning, but for their dry states. Gel fraction W g showing the mass fraction of the reactants that incorporate into the water-insoluble gel was calculated as W g = m dry /(m 0 C), where m 0 is the mass of the as-prepared cryogel or OHG, and C is the total concentration of the reactants in the reaction solution.
To determine the conformation of SF in the cryogels, attenuated total reflection (ATR)−FTIR measurements were conducted on the Agilent Technologies Cary 630 ATR−FTIR spectrophotometer, as detailed before. 34−36 In short, the Amide-I region of the spectra was deconvolved by the Gaussian model after a linear baseline correction by using PeakFit software (Version 4.12, SeaSolve Software Inc). Then, the band positions at 1620, 1640, 1660, and 1698 cm −1 , corresponding to the β-sheet, random coil, α-helix, and β-turn respectively, were fixed for curve fitting, while allowing their widths and heights to vary ( Figure S4a,b). DSC measurements were performed on the PerkinElmer DSC 4000 instrument under nitrogen atmosphere. Specimens of wet cryogels and OHGs of about 10 mg were put in aluminum pans, and then, they were scanned between 0 and 65°C. Melting enthalpies ΔH m were calculated from the DSC curves and the degree of crystallinity f cry , that is, the fraction of C18A units forming crystalline regions was determined, as reported before. 53 XRD measurements were conducted on freeze-dried cryogels, OHGs, and poly(AAc-co-C18A) organogel on the PANalytical X-Pert PRO multi-purpose diffractometer using Ni-filtered Cu Kα (λ = 0.15418 nm) radiation at 45 kV and 40 mA in the range of 2θ = 5−40°.
To estimate the porosity of the cryogels and OHGs, we used their weight (q w ) and volume swelling ratios (q v ) in water. We should note that the q w of cryogels includes water locating both in the gel and in the pores, while q v is caused by the solvation of the cross-linked polymer chains, that is, pore walls. Thus, the total porosity, that is, the volume fraction of pores P of the cryogels and OHGs can be calculated from the difference between q w and q v using the equation 59 where d 1 and ρ are the densities of water (1.0 g·mL −1 ) and SF scaffold (1.35 g·mL −1 ) or OHG (1.1 g·mL −1 ), respectively. Moreover, the total volume of the pores V p was estimated from the mass increase of the scaffold and OHGs in water at pH = 3.0. It was found that in this solution, both the scaffold and OHGs do not swell by volume, that is, q v = 1, so that the mass increase is related to the amount of the solvent within the pores. Thus, V p was calculated as V p = d 1 −1 (m/m dry − 1). SEM analysis was conducted on the FEI-QUANTA FEG 250 ESEM environmental electron microscope after coating the dry cryogels and the OHGs with platinum. To analyze the size distribution of the pores, randomly selected 50 pores from the SEM images were measured by using Image-Pro Plus 6 software, and their density distribution curves and the average sphere-equivalent pore diameters were determined. 53 To visualize the cryogel and the organogel components of the OHG, confocal laser scanning microscopy (CLSM, Nikon C2, Japan) was used after their staining with FITC and Nile Red, respectively.
Rheological measurements were conducted on a Bohlin Gemini 150 Rheometer system equipped with a Peltier device that controls the temperature. A parallel plate geometry (diameter = 20 mm) and a water trap were used. Frequency (ω) sweep tests on OHGs were conducted at both 25 and 65°C at a fixed strain amplitude γ o of 0.1%. Temperature-dependent viscoelastic properties of OHGs were determined by heating the OHG specimens from 25 to 65°C at a fixed rate of 4°C·min −1 at ω = 6.28 rad·s −1 and γ o = 0.1%.
Mechanical performances of the cryogels and OHGs were determined at 23 ± 2°C and at a test speed of 5 mm·min −1 by unidirectional compression tests performed using the Zwick-Roell universal test machine equipped with a load cell of 500 N. Young's modulus E was determined from the stress−strain curves between 2 and 4% compression. The stress was represented by its nominal value σ nom (the force per cross-sectional area of the undeformed specimen), while strain was given by ε (sample length change relative to its initial length). The fracture stress and fracture strain were calculated from the maxima of the corresponding true stress−strain curves as detailed before. 60 Fatigue resistance of the cryogels and OHGs was determined via successive cyclic tests at the same test speed with a waiting time of 5 min between the cycles. The specimens were subjected to five successive loading−unloading cycles up to a maximum strain ε max of 80%. In addition, step-wise cyclic tests were also conducted by increasing ε max from 10 to 80% in eight steps.
Shape-memory tests were performed on the OHGs via bending tests. Straight cylindrical OHG specimens (5 cm in length, 4.3 mm in diameter) were heated up to 65°C and deformed at this temperature to a horseshoe-like temporary shape with a deformed angle θ d . They were then exposed to 25°C for 5 min to fix this temporary shape under load. After unloading, fixed angles θ f of the specimens were measured to determine the shape-fixity ratios R f of OHGs. Finally, OHG specimens were gradually heated from 25 to 70°C, recovery angles θ r in each 1−3°C temperature intervals were recorded, and shape-recovery ratios R r were calculated.
Details on the cryogel properties, cyclic strain-sweep tests, FTIR, DSC, SEM data, and cyclic mechanical test results (PDF)