Light, Strong, and Ductile Architectures Achieved by Silk Fiber "Welding" Processing.

Light, strong, and ductile materials (LSDMs) are desired in many emerging fields, such as biomedicine, aerospace industries, and structural engineering materials. However, producing such materials remains a significant challenge because their structures cannot confer the desired mechanical properties. In this study, we developed a silk fiber "welding" strategy to construct bioinspired LSDMs. The key to the welding process is to etch the surface of silk fiber through a partial dissolution process. The dissolved silk proteins further serve as welding materials or glues to bond the silk fibers together. Remarkably, these silk-LSDMs are not only lightweight (with the densities of around 0.28 g cm-3) but also strong and tough. Their compression strength reaches up to 13.8 ± 3.4 MPa, which is higher than those of most natural and engineered porous materials. These favorable structural and mechanical characteristics, together with outstanding biocompatibility of silk proteins, render these silk-LSDMs applicable in regenerated engineered tissues and water treatment materials.

Introduction

Materials with lightweight, high mechanical strength, and ductility are much needed in multiple fields such as biomedicine,[1,2] aerospace industries,[3] and structural engineering materials.[3−7] However, designing and fabrication of materials with these favorable features is still challenging.[5−7] For example, many inorganic materials (e.g., glasses and ceramic) are light and strong, but they are often brittle.[2] Ductile materials, such as rubbers, on the other hand, are generally soft. In this regard, natural lightweight materials can inspire the design of light, strong, and ductile materials (LSDMs). Natural materials such as bird feathers, bonds, and woods are typical LSDMs that universally underpin biological functionalities.[8,9] Feathers, beaks, and bones of birds are prime examples.[10] Other examples are animal horns,[11] wood,[12] sucker ring teeth,[13] and arthropod exoskeletons.[14] All of them are designed to be lightweight and durable. Their lightweight, high mechanical strength, and ductility stem from their structures.[15−17] For instance, foam, porous, and hollow structures are common in a variety of natural materials, such as bird feathers, quills of hedgehog, and echidnas.[8−10,18−21] The uniqueness of these structures is that they consist of thin and strong frameworks and porous foam fillers. In such systems, the framework contributes stiffness and strength to resist compression, bending, and torsion. The inner foam, instead, plays an essential role in regulating the local buckling behavior, as they can withstand high tensile or compressive deformation.[20] The synergetic effects between frameworks and foams result in an optimal balance in weight, strength, and ductility, although most of those materials contain a single component. For example, feathers, quills, and animal horns are formed from keratin, and sucker ring teeth just contain one component of silk-like proteins.[18−25]

Inspired by natural lightweight materials, freeze casting techniques, such as unidirectional and bidirectional freezing techniques, have been developed to create LSDMs.[26] Different biomimetic lightweight structures, such as nacrelike foams,[26] woodlike cellular structures,[27] and polar bear hairlike porous structure,[28] have been created to achieve a variety of functionalities, such as mechanical reinforcement and temperature regulation. The freeze casting techniques generate these lightweight structures through controlling the nucleation and growth of ice crystals under dual temperature gradients, thereby producing large-scale aligned, lamellar, and porous structures.[28] However, the materials used for freeze casting are either suspension of powders or polymer solutions. Hence, they are unable to directly construct composite architectures that consist of both framework and foam fillers. To create these hierarchical composite structures, different techniques are required.[6,29] For example, most recently, unidirectional freezing has been integrated with thermocuring approach to fabricate Resol cellular materials to mimic the hierarchical cellular structure of wood.[27]

In this study, we develop a silk fiber “welding” strategy to construct bioinspired LSDMs. The center of the welding processing is to etch the surface of silk fiber without destroying the hierarchical structure existing in silk fiber through a partial dissolution process. The dissolved silk proteins further serve as welding materials or glues to bond the silk fibers together. After the welding process, the fibers could maintain their original length, and their loose structures are incorporated into a three-dimensional (3D) continuous network that is filled by silk protein foams, featuring self-enhanced composite structures. All these structures are made only from silk proteins. As with natural composite structures, these silk-LSDMs (S-LSDMs) are optimized in strength, toughness, and lightweight. For example, their compression strength reaches up to 13.8 ± 3.4 MPa, a value higher than those of most natural and engineered lightweight materials. These favorable structural and mechanical characteristics, together with outstanding biocompatibility of silk proteins, render these S-LSDMs advantageous in multiple applications, such as regenerated engineered tissues and water treatment materials.

Results and Discussion

Figure a sketches the five-step welding processes to produce S-LSDMs. First, the degummed Bombyx mori (B. mori) silk fibers were immersed into hexafluoroisopropanol (HFIP) solution with the weight ratio of 1:20–1:30 and incubated at 60 °C, the boiling point of HFIP, for 2–14 days. Here, HFIP is chosen because it can partially dissolve the B. mori silk fibers.[30−32] Based on our previous studies and attempts in this work, HFIP can dissolve B. mori silk fibers from the surface mildly.[30,31] For example, when B. mori silk fibers were immersed into the HFIP solution with the weight ratio of 1:20–1:30 and incubated at 60 °C, their basic structure is mostly maintained after incubating for 2 days. However, after 14 days of incubation, the silk fibers are dissolved into silk microfibrils.[30] Therefore, incubating for 2–14 days brings about different degrees of dissolution while preserving the basic hierarchical structure in silk fibers (Figure S1), which is very important for the mechanical properties of silk materials. During this process, the silk fibers were partially dissolved and formed sticky surfaces. The dissolved silk proteins, meanwhile, stabilized by HFIP and appeared as a uniform solution phase. The resultant silk/HFIP composite systems were then frozen at −20 °C overnight. At standard pressure, the melting point of HFIP is −4 °C, thus a liquid–solid transition occurs in this step. This transition can induce silk proteins to form a certain amount of β-sheet (17 ± 3%, evaluated by the deconvolution of Fourier transform infrared (FTIR) spectrum[33]).

Figure 1

Silk fiber welding processing for the fabrication of S-LSDMs. (a) Schematic of the five-step route to produce S-LSDMs. (b–d) Photographs to illustrate that the thawed silk/HFIP mixtures incubated for 14 days can be directly squeezed out from a needle into the media of water bath (b), clay aqueous solution (b), and air (c).

After a thawing process at room temperature, the resultant silk/HFIP composites could be directly employed to construct S-LSDMs. As illustrated in Figure b–d, silk/HFIP composites incubated for 14 days could be directly squeezed out from a needle into a water bath (Figure b) or a clay aqueous solution (Figure c). When these thawed silk/HFIP composites were extruded into a water bath, they solidified into continuous fibers owing to the rapid solvent exchange between HFIP and water. These as-spun fibers can be directly taken out of the water solution without breakage (Movie S1). The clay aqueous solution, instead, is a useful medium to produce intricate 3D constructions. As a prototype, Figure c shows that the letter “M” can be created along the X- and Z-axis. More remarkably, the thawed silk/HFIP composites even can directly extrude and shape in the air environment (Figure d); therefore, they can be used as ink to print any desired structures through 3D printing.[30,34]

Next, the solid silk/HFIP composites created in the previous step were immersed to the 75 vol % ethanol aqueous solution for inducing the conformation transition of silk proteins from water-soluble random coil to the water-insoluble β-sheet structure. Most HFIP solvents can be removed in this process through solvent exchange among ethanol, water, and HFIP. However, to ensure that the residual solvents are completely removed, an adequate washing process is often required. This treatment was performed by washing composites with deionized water at room temperature or at boiling condition (∼100 °C). At last, a freeze-drying process was performed to obtain dried S-LSDMs.

The mesostructures of S-LSDMs were first examined by scanning electron microscopy (SEM). Figure a shows the cross section of S-LSDM after welding process with incubating time of 2 days and cutting off in liquid nitrogen, revealing the two-phase composite structure of S-LSDMs. They consisted of both silk fiber (colored red) and highly porous silk protein foam (colored blue), and the interfaces between silk fiber and silk proteins were firmly bonded. High-resolution SEM images (Figure b) further revealed that these nanopores were highly homogeneous with pore diameters ranging from hundreds of nanometers to microns. To observe longitudinal morphology of silk fibers in S-LSDMs after the welding process of incubating silk fibers for different days, we tore S-LSDMs by hand followed by SEM characterization. As shown in Figure c, the welding process releases microfibrils and forms silk fibroin (SF) foams while maintaining some hierarchical fiber structure of the original silk. With increased incubating time, less fiber structure can be observed in S-LSDMs (Figure S1), and S-LSDMs are almost occupied by silk foams, indicating that the diameter of the silk fiber and the volume fraction of the undissolved silk fiber would decrease with increase in incubating time. Synchrotron Fourier transform infrared spectroscopy microspectroscoy (S-microFTIR) was further used to determine the β-sheet distribution in S-LSDMs based on the method in our further paper[33] (Figure d). In these experiments, S-LSDMs were first sliced into thin sections with a thickness of 10 μm by freezing microtome and then transferred onto a barium fluoride tablet. The dried S-LSDM sections were tested in transmission mode with an individual pixel size of about 5 μm × 5 μm and mapped over a 100 μm × 90 μm field of view. The FTIR spectra were collected in the mid-infrared range of 650–4000 cm–1 at a resolution of 8 cm–1 with 64 co-added scans. Then, we deconvoluted the amide I bands of the spectrum extracted from each pixel. The detailed deconvolution method is described in Figure S2, and the β-sheet content was calculated according to the area of the peak, which represented the β-sheet conformation divided by the total area of amide I. The calculated β-sheet content of each pixel was then reintegrated to the two-dimensional contour image (Figure e). The representative spectra extracted from blue (SF foam) and red (silk fiber) regions are also shown in Figure f for reference. Both the β-sheet content contour image and the associated original spectra confirm that the silk proteins in silk foam indeed have a certain amount of β-sheet (10–30%), although their content is less than that in silk fibers (30–60%). In addition, the shape of the boundary in Figure e is very similar to that observed in the microscopic image (Figure S3), which further confirms the significant difference in the β-sheet structure between the SF foam and the silk fibers in S-LSDMs.

Figure 2

Mesostructure of S-LSDMs. (a) SEM image of the cross section of S-LSDM after welding process with incubating time of 2 days. (b) SEM image of S-LSDM at silk fibroin foam region in image (a). False color was used in SEM images. (c) SEM image of the longitudinal section of S-LSDM after welding process with incubating time of 2 days and tearing by hand. (d) Schematic of the hierarchical structure of S-LSDMs. (e) FTIR mapping of the S-LSDM section. The width and height of the image is 100 and 90 μm, respectively. (f) FTIR spectra of S-LSDM textracted from the silk fiber and the SF foam region, respectively.

Figure a,b presents the appearance of S-LSDM. It is light but shows high strength and ductility, a cylindrical foam with a diameter of 22 mm and a height of 15 mm can stand on a leaf without dropping, the weight of this S-LSDM is only 0.7 g, with a density of 0.12 g cm–3, a value that is comparable with the rigid polymer foams (0.04–0.5 g cm–3),[2,8] and natural porous materials, such as proximal tibial trabecular bone (0.30 ± 0.10 g cm–3)[35] and toucan beak (about 0.1 g cm–3).[2,8] In addition, such a S-LSDM is resistant to cut (the insect image in Figure a and Movie S2) and can withstand more than 3 kg of weight (Figure b), which is more than 4000 times larger than its own weight. It can even withstand impacts generated by a 3 kg metal without being destroyed (Movie S3).

Figure 3

Appearance and mechanical properties of S-LSDMs. (a, b) Photographs of S-LSDMs to show their lightweight and strong characteristics. (c) Compression stress–strain curves of S-LSDMs. (d) Comparison of Young’s modulus and densities of S-LSDMs with other natural and engineered lightweight materials. The Ashby plot was adapted from ref (2).[2]

The mechanical properties of S-LSDMs were quantitatively evaluated by compression measurements. As presented in Figure c and Table S1, compressive stress–strain curves present an elegant deformation without a catastrophic collapse, showing a remarkable ductile failure behavior. More specifically, the response is elastic at small strains and the loading curves show extensive stress plateaus, followed by a sharply serrated plateau, beyond the elastic regime. In addition, these S-LSDMs show an elegant trade-off between density, Young’s modulus, and strength. The compression stress of S-LSDMs after the welding process with the incubation time of 5 days reaches up to 13.8 ± 3.4 MPa, which is about 4 times larger than that reported previously for silk porous materials, such as silk nanofibrous scaffolds.[36] They are also stronger than wood[37] (with the compression stress of about 10 MPa along the parallel direction and about 3 MPa along the vertical direction), which is one of the toughest biological lightweight material, and even much stronger than most of the high-performance foams, such as polyurethane foam (0.35 MPa)[38] and foams fabricated by nanocrystalline cellulose[39] (0.25 MPa). In addition, the stiffness of S-LSDMs (incubating for 5 days) is more than five times higher than that of cork, ten times higher than that of cartilage, and three times higher than that of cancellous bone.[37]

Interestingly, no direct relations between welding periods and densities and mechanical properties were observed. For example, S-LSDMs after the welding process with the incubation time of 5 days have the highest modulus while S-LSDMs after the welding process with the incubating time of 14 days has the lowest modulus, indicating that incubating silk fibers for 2 days is enough for building silk-based LSDMs (Figure d and Table S1).

In material engineering, the common strategies for material reinforcement are to add the second or even the third component, but the enhancement of the mechanical performance of S-LSDMs is achieved by structural designs and does not depend on the additional components. Therefore, S-LSDMs can be considered as self-reinforcement materials. As with natural lightweight materials, this framework-and-foam combined composite structure is favorable for LSDMs. In this kind of construction, the hierarchical structure of silk fibers is preserved and thereby their mechanical properties can be retained. In addition, the welding process allows integrating loose fibers into a 3D framework so that all of the silk fibers in S-LSDMs can act as mechanical bearing units to bear and transfer loads. Furthermore, silk protein foam in S-LSDMs is light but can provide the ability to accommodate buckling and tolerate tensile or compressive deformation.

The synergetic effects between strong frameworks and soft foams allow S-LSDMs to achieve good balance between lightweight and tough mechanical property. In fact, B. mori silk fibers have been commonly used to fabricate strong and ductile materials, such as silk scaffolds fabricated by regenerated silk fibroin using salt leaching, freeze-drying, and gas foaming methods.[40−43] However, those methods use silk fibroin as raw materials, which lack the hierarchical structure of silk fibers. As a result, the compression strength of the resultant scaffolds could only reach up to hundreds of kilopascals (Table S2), which also reflects the high efficiency to reach self-reinforcement in this work.

Conclusions

In summary, we developed a silk welding approach for constructing LSDMs. HFIP, a solvent that can dissolve silk fibers from the surface mildly, was used to etch the surface of the silk fibers. The dissolved silk proteins were then welded with silk fibers to form a composite structure that consists of a silk fiber framework and silk protein foams. The resultant S-LSDMs are light, strong, and ductile, with compression mechanical properties comparable to those of most natural and advanced engineered lightweight materials. For example, the strength and modulus are even higher than the cancellous bone. The resultant S-LSDMs are controllable and the shapes of S-LSDMs can be easily tailored by containers of different sizes and shape during the welding process. S-LSDMs are self-reinforcement materials, inheriting merits of silk fibers with good biocompatibility and biodegradability. Moreover, the mechanical properties and three-dimensional architecture could be easily tailored by incubating silk fibers in HFIP for different days to make them good candidates for tissue engineering scaffold, especially as bone implant materials. Besides, the three-dimensional porous structures with pore diameters ranging from hundreds of nanometers to microns also show potential application in the field of water treatment, pollutant adsorption, and thermal insulation. In addition to fabricating self-reinforcement materials, this welding approach can be also applied to build other functional LSDMs owing in part to their versatile processability. For instance, functional components, which are soluble in HFIP, can be involved in S-LSDMs for producing functional S-LSDMs. In particular, previous works have confirmed that both graphene and carbon nanotubes can disperse into the HFIP solution.[35,36] In addition, other solvent systems, such as lithium bromide/water, ionic liquid, and formic acid/calcium chloride, which have the ability to dissolve silk fibers, are also expected to be used for the welding of silk fibers. This aim can be achieved by regulating the mass ratio of silk fiber and solvent, or the concentration of the salts because the solubility of silk fibers in these solvent systems are directly associated with these dissolution parameters. Overall, this silk welding method represents a rational, low-cost, and highly efficient strategy for the design and preparation of LSDMs, showing the possibility to being transferred to other material systems.

Experimental Section

Preparation of Silk/HFIP Mixtures

Twenty grams of B. mori silkworm cocoon silk fibers was degummed by boiling in 4 L of NaHCO3 (Aladdin, China) aqueous solution with concentration of 0.5% (w/w). The degummed process were performed twice to thoroughly remove sericin. The degummed silk fibers were then washed with distilled water and allowed to dry in air at room temperature. Next, 1 g of degummed silk fibers were loosened into random state, followed by being inserted into 20 mL glass bottle. Then, 18 mL of HFIP (Aladdin, China) was added into the glass bottle before sealing the glass bottle. The silk/HFIP mixture was then incubated in airtight containers at 60 °C for 2–14 days. All of these steps should be conducted in a chemical hood with the necessary precautions, as HFIP is a toxic solvent.

Preparation of S-LSDMs

The incubated silk fibers were placed into a refrigerator at −20 °C in sealed glass bottle overnight. Then, the freezing silk fibers/HFIP mixtures were rapidly added into 75 vol % ethanol (Titan Technology, China) overnight for solvent exchange and foam formation. Finally, the silk foams were transferred into distilled water overnight and followed by freeze-drying process for 2 days.

Mechanical Testing of S-LSDMs

The mechanical properties of S-LSDMs were tested in the compression mode at 25 °C and 45% relative humidity (RH) at a compression rate of 1 mm min–1 by an Instron 5966 machine (Instron, Norwood). First, S-LSDMs were cut into segments with height of ∼5 mm. The initial height of the samples was measured with a caliper at zero load point (the point at which the samples were in contact with the compress disk but very little force exerted on it).

Synchrotron-FTIR Microspectroscopy of S-LSDMs

The experiments were performed at BL01B in the National Synchrotron Radiation Laboratory (NSRL, Hefei, China). The instrumental details are presented in our previous paper.[37] S-LSDM was first sliced into thin sections of 10 μm thickness by freezing microtome (Leica CM3050 S) and then transferred onto barium fluoride tablet. The dried S-LSDM section was mapped in the transmission mode with an individual pixel size of about 5 μm × 5 μm over a 100 μm × 90 μm field of view using OPUS 7.5 (Bruker). The FTIR spectra were collected in the mid-infrared range of 650–4000 cm–1 at a resolution of 8 cm–1 with 64 co-added scans. Determining the distribution of the secondary structure content within S-FTIR through the FTIR spectra was calculated by deconvolution of amide I bands using PeakFit 4.12. The numbers and positions of peaks were defined from the results of second derivatives spectra and fixed during the deconvolution process. As in our previous studies, a Gaussian model was selected for the band shape, and the bandwidth was automatically adjusted by the software.

Characterization

The morphologies of S-LSDMs were characterized by SEM (JSM7800F, JEOL, Japan) at an acceleration voltage of 5 kV. To prevent electrical charging, all specimens were coated with a 5 nm thick Au layer before observation.