Structural analysis of the delignified wood
Wood, as a natural polymer composite material, has formed hierarchical pores structure along the longitudinal direction due to the transport requirements of nutrients and water. In addition, during the later stage of wood cell wall formation, a large amount of lignin is synthesized and deposited in the intercellular region between porous parenchyma cells, forming a tightly bound interface through H-bonds and covalent bonds35. Therefore, in order to realize the topological transformation of wood cell wall, it is necessary to obtain wood with low Young's modulus. To this end, we selectively removed most of the lignin while maintaining the integrity of the cellulose fibrils. Next, the strong permeability of DMSO was used on the amorphous region of cellulose to achieve partial pre-dissociation of cellulose fibrils, which promoted the penetration of moisture into cellulose skeleton, thus weakening the cohesion between adjacent cellulose fibrils36. Then, the hydrogen bonding interactions between cellulose chains were transformed from cellulose-cellulose to cellulose-water-cellulose37. Natural moisture evaporation process driven the re-formation of H-bonds between cellulose chains and allowed capillary forces to induce deformation and efficient aggregation of cellulose fibrils (Fig. 1a).
Wood sample with a thickness of 1 mm was obtained by cutting the natural bulk balsa wood. NaOH/Na2SO3/anthraquinone reaction system was utilized to remove most of the lignin and hemicellulose, yielding an oriented cellulose fibril skeleton. Firstly, delignification solution rapidly throughout wood lumens and vessels, allowing the sulfite ions and anthrahydroquinone dianions under alkaline conditions to attack the β-aromatic ether bonds of lignin, which caused the lignin macromolecular chains quickly decompose into soluble small molecular fragments38. The addition of methanol with low surface energy and low viscosity reduced the Laplace resistance at the interface between wood and water, which greatly enhanced the penetration flux of delignification solution to the wood, thus achieving a uniform delignification process39. Furthermore, the recessive reducing terminal group on C1 of cellulose chains could be selectively oxidized by anthraquinone to form carboxyl groups. The above process largely prevented the peeling reactions of cellulose under alkaline conditions. Fourier transform infrared (FT-IR) spectroscopy was used to characterize the removal of lignin and hemicellulose. The peak at 1735 cm− 1 was assigned to the carboxyl groups in hemicellulose. The absorption band at 1245 cm− 1 corresponded to the ester linkage of the carboxyl groups in hemicellulose and lignin. FT-IR analysis has shown that the functional groups of SSWF assigned to hemicellulose and lignin significantly disappeared compared to natural wood (Supplementary Fig. S1)40. After NaOH/Na2SO3/anthraquinone treatment, the SSWF consisted mainly of cellulose with little hemicellulose and lignin (Supplementary Fig. S2).
Owing to the continuous synthesis and deposition of lignin in the intercellular layer of parenchyma cells, natural wood has developed buckling resistance to compressive loads. Therefore, wood shrinkage behavior did not occur during the evaporation of moisture. As for the delignified wood, the dense cellular structure of rigid wood became soft and loose, especially in the fully swollen state. The tightly connected interface disappeared and numerous nanoscale pores appeared densely on the wood cell wall after DMSO swelling and moisture replacement (Supplementary Fig. S3). Noticeably, the above process also resulted in a small amount of cellulose dissolution and regeneration behavior within the wood lumens. The (200) diffraction peak reflections slightly shifted to small diffraction angles, revealing an increase in the lattice spacing due to the weakening of cohesion caused by DMSO swelling and moisture replacement (Supplementary Fig. S4)41. According to the classical shell theory, the driving force for SSWF originated from the capillarity generated by the evaporation of water flow in delignified wood. The effective distance of capillarity force on solid deformation was affected by the solution surface tension and Young's modulus, which could be evaluated by Seq. (2) (Supporting Information)42–44. The wood sample exhibited an elastic modulus of about 1.18 ± 0.46 KPa due to the removal of lignin and sufficient swelling of DMSO (Supplementary Fig. S5). It can be seen that the effective distance of capillary tension in delignified wood was about 61.02 µm, which satisfied the requirements of cell wall shrinkage. Therefore, the wood cell wall tend to shrink along the direction of capillary force (Supplementary Fig. S6).
In addition, partial removal of lignin and hemicellulose effectively exposed more hydrophilic hydroxyl groups on the surface of cellulose nanofibers, which facilitated re-formation of H-bonds between adjacent nanofibres. The resulting denser fiber structure promoted efficient stress transfer and mitigated structural defects (Fig. 1b and 1c). After air-drying, the SSWF exhibited a more densified structure in the thickness direction (the thickness was reduced by 96%, only 40 µm) compared to natural wood (Supplementary Fig. S7a). In addition, the SSWF film could be curled in the transverse direction and fully folded in the radial direction, showing excellent flexibility (Supplementary Fig. S7b, 7c).
Optical transparency is one of the critical properties of CNF materials, which has attracted extensive attention in the development of optoelectronic devices. The densification process induced by moisture evaporation endowed the refractive index of the wood film uniformly stable at 1.53, greatly reducing the probability of light scattering45. The SSWF film has a total transmittance of around 75.7% (Supplementary Fig. S8). Unlike materials assembled from nanostructured pulp fibers, the natural multiscale arrangement of cellulose microfibrils in wood cell wall was well preserved, opening up the possibility to develop woody materials with superior properties (Supplementary Fig. S7d-f). A similar “top-down” method was reported previously for the transparent wood film from delignified wood and TEMPO-oxidized delignified wood26, 27. Noteworthily, the fabrication method in our work only required harvesting energy from natural physical phenomena that occur at any time and everywhere, exhibiting lowest energy consumption.
Structural And Mechanical Properties Of Sswf Film
As a highly hierarchical porous material, wood cell walls are mainly divided into S1, S2, and S3 layers, where S2 layer with the largest proportion of wall thickness and the lowest microfibril orientation angle are the main contributor to the longitudinal mechanical properties16. It is noteworthy that the nanocellulose in SSWF film can still inherit the high orientation and present a denser arrangement. To prove the orientation structure, we observed the fiber arrangement morphology of SSWF film. Under polarized light, the microfibrils of natural wood appeared dark gray due to the existence of amorphous hemicellulose and light-absorbing lignin (Supplementary Fig. S9). For comparison, the microfibrils of SSWF film exhibited bright birefringence along wood growth direction, indicating excellent alignment and crystallization behavior. The alignment was confirmed by using the 2D-WAXS patterns. Four bright spots, appearing in equatorial direction, correspond to overlapping (1\(\stackrel{-}{1}\)0), (110) crystal planes and (200) structure of the anisotropic \({I}_{{\beta }}\) cellulose crystallites, which were aligned parallel to the longitudinal axis, suggesting their orientation along the microfibrils (Fig. 2a, b)46. The orientation index was calculated to be 0.88 from the diffraction intensity of the crystal reflection along the equatorial direction of (200). The XRD profiles of the natural wood and SSWF film both exhibited two broad diffraction peaks at approximately 2θ = 18.3° and 22.2° arising from the scattering of the (110) and (200) planes of cellulose crystals, respectively47, 48. XRD analysis confirmed that rapid delignification and self-densifying process did not change the crystal structure of cellulose molecules. Meanwhile, the (200) diffraction peak reflections slightly shifted to larger diffraction angles by approximately 0.18° compared with natural wood, revealing a decrease in the lattice spacing of the SSWF film due to the re-formation of H-bonds between cellulose molecules (Fig. 2c)49.
The longitudinal tensile strength of the SSWF film could reach as high as 596.24 ± 57.01 MPa, and the toughness was up to 10.43 ± 2.07 MJ m−3. In addition, the specific strength of the SSWF film (density: 1.35 ± 0.08 g cm−3) could reach 441.66 MPa cm3 g−1 (Fig. 2d, e). To the best of our knowledge, the tensile strength and specific strength of the SSWF film developed in this study surpass those reported in the current literature, e.g., CNFs film (tensile strength: 289 MPa, specific strength: 192.67 MPa cm3 g−1)50 and transparent wood film (tensile strength: 469.9 MPa, specific strength: 391.60 MPa cm3 g−1)27.
To shed light on the strengthening and toughening mechanism of SSWF, we first simulated the dynamic deformation and failure process during interlayer shear deformations. From the distributions of interlayer H-bond networks and corresponding local coarse-grained shear strain, we can see that the interlayer shear deformation will lead to a large interlayer shear strain, thereby destroying the interlayer H-bond networks (Fig. 2f, g). Moreover, under shear deformations, the stick-slip behaviors could be observed (Fig. 2h), which was attributed to the dynamically destroyed and re-formed H-bond networks51. This process dissipated a lot of strain energy and increase the toughness of SSWF consequently.
To investigate the overall mechanical behaviors of SSWF, we simplified the microstructure to the regular staggering nacre-like structure (Supplementary Fig. S11). The atomic configuration could be homogeneous and continuous as a continuum model, i.e., the deformable tension-shear (DTS) model52–54. In DTS model, the n-layer stacked cellulose blocks in SSWF were considered hard layers bearing only in-plane loading, while interlayer H-bond networks were simplified as soft layers bearing only shear loading. Note that, the hard layer defined here was not only a single cellulose chain, but also a cellulose block with n layers of cellulose stacked together (Supplementary Fig. S11b)52.
In general, the interchain H-bonds formed between cellulose molecules are thought to be the weak points in SSWF, then the fracture of SSWF can be attributed to the break of interlayer H-bonds. In this work, we adopt the strain-failure criterion, i.e., the failure strain of interlayer H-bond networks was assumed as \({\gamma }_{\text{c}\text{r}}\). In this regard, the maximum interlayer shear strain reach \({\gamma }_{\text{c}\text{r}}\), we can obtain the strength as53
$${\sigma }_{\text{S}}^{\text{S}}=\frac{D{\gamma }_{\text{c}\text{r}}}{2{l}_{0}}\frac{\text{s}\text{i}\text{n}\text{h}{k}_{1}}{1+\text{c}\text{o}\text{s}\text{h}{k}_{1}}$$
1
where we define the characteristic length \({l}_{0}=\sqrt{{Dh}_{\text{s}}/4G}\), \({k}_{1}=L/{l}_{0}\) represents the dimensional size of SSWF, \(D\) represents the in-plane stiffness of equivalent hard layer, G and \({h}_{\text{s}}\) are interlayer shear modulus and distance, respectively. The longitudinal tensile strength will increase with the increase of length and converge rapidly (i.e., \(L/{l}_{0}\approx 4.0\)) to the maximum value \(D{\gamma }_{\text{c}\text{r}}/2{l}_{0}\) (Fig. 2i). Of course, if the failure of the n-layer stacked cellulose block is also considered, two failure modes can be distinguished, i.e., the failure of the n-layer stacked cellulose block (mode G) and the failure of interlayer H-bond networks (mode I) (Fig. 2j). In this regard, the tensile strength will increase with the increase of length and converge to \({\sigma }_{\text{s}}^{\text{H}}=D{\epsilon }_{\text{c}\text{r}}/2{h}_{s}\) with \({\epsilon }_{\text{c}\text{r}}\) the failure strain of the n-layer stacked cellulose block. Note that, the densifying patterns in the experiment cannot be regular staggering, thus the convergence speed will be much lower for more random staggering patterns, such as stair-wise staggering55.
Here, we only consider the yz slip plane (Supplementary Fig. S10a) where the interlayer H-bonds induce the strongest interlayer shear modulus34. Thus, all the interlayer shear properties are fixed in this work, i.e., \(G=16.37\) GPa, \({h}_{s}=8.7 \text{\AA }\), \({\gamma }_{\text{c}\text{r}}=0.1\). Whereas, the in-plane stiffness of the hard layer is strongly related to the thickness of the n-layer stacked cellulose block. The in-plane stiffness of the n-layer cellulose block increase with the increase of layer number n, which can be well fitted by
D = nD0 (2)
where D0 = 54.96 N/m represents the stiffness of monolayer cellulose. Note that, the effect of layer number n on the in-plane failure strain \({\epsilon }_{\text{c}\text{r}}\) is negligible (Supplementary Fig. S12). Then the characteristic length becomes \({l}_{0}=\sqrt{n{{D}_{0}h}_{s}/4G}\), which increases with the increase of layer number n following the scaling law \(\sqrt{n}\).
Combining the above analyses (Fig. 2i, j), we can conclude that to increase the longitudinal tensile strength of SSWF, we should theoretically increase the dimensional length \(L/{l}_{0}\). In the experiment, we directly slow down the depolymerization of cellulose during alkali treatment to maintain the length L of cellulose, thereby improving the overall mechanical properties. In addition, upon increasing the density of the interlayer H-bond networks and thus the interlayer shear modulus G,55 we can also reduce \({l}_{0}\) and increase the dimensional length \(L/{l}_{0}\) indirectly. Therefore, the proposed DTS model can well capture the strengthening and toughening mechanism of SSWF film.
The degree of polymerization of cellulose reached 1972.6 ± 103.1, 1574.8 ± 89.5, 1465.98 ± 85.3 after being treated for 1 h, 1.5 h, and 2 h (Supplementary Fig. S13). The tensile strength of SSWF film reached the best mechanical properties when the treatment time was 1.5 h, which was mainly attributed to the optimal balance between the degree of cellulose polymerization and H-bond density. In addition, we compared the dynamic thermal stability of SSWF film with some typical plastic films. The storage modulus of the SSWF film could be maintained almost at about 20 GPa in the variable temperature range from −10°C to 100°C, which was higher and more stable than plastics (polypropylene, polyethylene), thus guaranteeing the reliability in practical application (Supplementary Fig. S14).
Flexural Strengths And Processability Of Sswf
We simultaneously realized the self-densifying process of 1 cm thick wood blocks, in which balsa wood and pine wood were successfully transformed, demonstrating a universal strengthening strategy (Fig. 3a-h). The shrinkage rates of balsa and pine could reach 90% and 80% along the thickness direction, and 30.29% and 28.57% along the width direction, respectively. The flexural strength of the self-densified balsa wood and pine wood reached as high as 418.51 ± 23.86 MPa and 296.52 ± 10.41 MPa, which were 42.35 times and 4.89 times higher than that of natural wood, respectively. In addition, the swollen wood has excellent formability, enabling it capable to form complex machined shapes similar to metals and plastics. We only need to place the surface-dried swollen wood in the mold of specific shapes to achieve self-densifying and shaping process via moisture evaporation process, such as M-shape, S-shape, and spiral-shape (Fig. 3i-k). Here, based on the hydroplasticity of swollen wood, we fabricated a self-densified wood nail similar to the commercial steel nail. We hammered self-densified wood nail and steel nail together into three layers of natural poplar wood sections (each layer: length ⋅ width ⋅ thickness = 50.0 ⋅ 20.0 ⋅ 5.0 mm), and neither the self-densified wood nail nor steel nail were damaged during the process of penetrating natural poplar layers. Although the initial stiffness of the self-densified wood nail was lower than that of the steel nail, the load-carrying capacity shown roughly the same trend (both reached the maximum peak at around 238 N), indicating that the failure mode of the samples was caused by the rupture of the poplar layer (Fig. 3l).
The Flexibility Of Sswf Film
Benefiting from the excellent toughness of the SSWF film, various 3d structures, such as cubic structure, corrugated structure, wood box, curved structure, star structure, and hexagonal structure could be all successfully displayed (Fig. 4a-e). The bending cycle stability test was carried out using a laboratory-made electric linear translation stage. After1000 tensile-bending cycle tests, the fabricated SSWF film (via 7 folds) exhibited no bending fracture even when fully bent to 180° (Supplementary Fig. S15). The above result was attributed to the covalent bond bridges between residual lignin and polysaccharide acting as a colloidal matrix at the interface between load-bearing nanofibers, which enhanced the interface interaction and improved the toughness60. To demonstrate the above phenomena, we have used the sodium chlorite method (95 ℃, 20 h) to remove the lignin (1.3%) and evaporate the wood under the same process conditions. The fully delignified wood film has shown apparent fracture after bending, indicating that the retained part of lignin and hemicellulose was positive for the improvement of toughness (Supplementary Fig. S16 and Video S1). Benefiting from these characteristics, we demonstrated the possibility of using SSWF film as a substitute for practical applications in flexible electronics. Interconnect lines drawn with a 9B pencil on SSWF film could act as a conductive wire to be integrated into 2D or 3D circuits with commercial electronic components such as LEDs (Supplementary Fig. S17a). Similar structural designs could also be used as flexible Joule heating elements for healthcare applications. We systematically investigated the relationship between the pencil writing cycles and the electrical conductivity of graphite. The sheet resistance of graphite was reduced from 61.16 ± 2.25 kΩ sq− 1 in the first time to 153.26 ± 20.57 Ω sq− 1 in the 20 times (Supplementary Fig. S17b). Owing to the high thermal conductivity of graphite, the temperature of the SSWF-based Joule heating elements increased sharply from room temperature to 119.4°C within 20 s at a voltage of 12 V (Supplementary Fig. S17c). Under the applied voltages of 5 V, 10 V, and 15 V, the temperature of the SSWF-based Joule heating elements reached 36.8°C, 78.4°C, and > 150°C in a very short time, respectively (Supplementary Fig. S17d-f), indicating the rapid electrical heating responsiveness.
In addition, the excellent mechanical strength and formability of the SSWF film provide new ideas for designing and manufacturing lightweight, load-bearing materials such as honeycomb cores. The hexagon structure of woody honeycomb panel was fabricated by origami forming along the direction of wood fibers. The woody honeycomb panel exhibited an ultra-low density of 0.025 g cm− 3, which was about the half density of the current commercial honeycomb panel. This porous structure has impressive support properties, which can carry an adult (65 Kg, more than 23550 times its own weight) without structural collapse (Supplementary Fig. S18 and Video S2).