Introduction

The rapid development of science and technology leads to a strong demand for new materials that can adapt to complex and diverse working environments. Shape memory polymers (SMPs) are key candidates for solving these challenges, with the programmability of temporary shapes being their significant advantage1,2,3. In recent years, numerous SMPs have been reported, such as shape memory hydrogels4,5,6, traditional polymers7,8,9, and liquid crystal elastomers (LCEs)10,11. As a major component of shape memory materials (SMMs), SMPs offer higher flexibility, economic feasibility, and diversity, making them highly promising for expanding the application scope of shape memory technology12,13,14.

LCEs, as special smart materials, possess the stimulus-response properties of liquid crystal (LC) materials and combine the extensible properties of elastomers. The differences in molecular chain arrangement and the degree of overall orientation of LCEs lead to different phases and properties. LCEs have advantages in feedback mechanisms in response to artificial intervention or environmental changes, with promising applications in fields such as actuators and soft robots15,16,17,18,19, sensors20,21,22,23,24, and information storage and encryption25,26,27,28,29,30. Given this background, how to utilize phase changes to impart shape memory while fully exploiting the special properties of LC materials is of profound significance. To construct SMPs, common and effective strategies include introducing dynamic covalent bonds31,32, utilizing supramolecular chemistry7,33, and enhancing crystallinity34,35. The first two strategies may involve complex synthesis processes or be sensitive to environmental factors, leading to material aging and reduced durability. The third is more compatible with the phase transition of LCs while avoiding complex synthesis steps, but there exists an inherent limited crystallinity of the material (especially in chiral systems). Inspired by the metamorphosis of Buprestidae from larvae to imagoes (Fig. 1a), we note that the exoskeleton transforms from soft to rigid, preserving imago beetle’s shape. Concurrently, the hardening process retains the periodic stratified structures of chitin and cuticular proteins, stabilizing the structural coloration. Modulus determines the shape, and color acts as the medium for information, further combining the two may diversify the functionalities of SMPs. Drawing from this biological strategy, we aim to achieve memory of both three-dimensional (3D) shape and structural color through material hardening. Recently, it has been confirmed that introducing crystallizable chains in nematic LCEs induce some crystallinity36,37,38,39, and increasing the spacer chain length promotes the transition from nematic phase to smectic phase40,41. However, in chiral environments like chiral LCEs (CLCEs)—polymer-based photonic crystals with periodic structures and distinctive structural colors—the arrangement of LC units exhibits various ordering patterns, such as a helical structure (Fig. 1b, Bottom), which causes the arrangement of the crystallizable segments attached to both ends of the rod-like LC cores to deviate from the ordered arrangement of crystallization. And the crosslinked network of the elastomer restricts molecular chain motion, making the formation of crystalline (Cr) regions and the subsequent significant mechanical strengthening difficult in CLCEs, even if sufficient crystallizable components are introduced (Fig. 1b, Top). How to overcome these limitations and extend shape memory induced by phase transitions to chiral phases while fully utilizing the potential of LC materials remains great challenges.

Fig. 1: Design strategies and properties of stiffness-switchable chiral liquid crystal elastomer (SS-CLCE).
figure 1

a The inspiration derived from the metamorphosis process of the Buprestidae. b Schematics of network structures and molecular arrangements of conventional chiral nematic (N*) LCE. c Schematics of dynamic changes of network structures and molecular arrangements of SS-CLCE designed by this work. It also exhibits the local layered structure of the smectic C (SmC) phase and ordered crystalline (Cr). d Mechanical performance differences between conventional CLCE and SS-CLCE. e Physical crosslinking induced by phase transition enables the memory effect of SS-CLCEs. It should be noted that the conventional CLCE only involves the processes on the diagonal.

Full size image

In this study, we propose a practical approach to achieve intrinsic, one-way shape-color integrated memory in CLCE films through stiffness switching, which simultaneously integrates elasticity and plasticity. This material system is designated as stiffness-switchable CLCE (SS-CLCE). We achieve the formation of side chains with greater mobility by controlling the functional group relationships during the synthesis process to induce the main chains in a low-crosslink-density polymer network to participate in the transition from the chiral nematic (N*) phase to the partly-formed smectic C (SmC) phase, and the formation of crystalline phase of polymer chains, resulting in microphase separation in the CLCE (Fig. 1c). This microphase separation process is referred to as aging. The N* phase serves as the carrier of structural color, while the auxiliary SmC and Cr phases are the sources of the memory effect. Mechanical properties of the conventional CLCE remain unchanged before and after aging, whereas the SS-CLCE shows a considerable change (Fig. 1d). The initial N* phase is easily deformable, causing structural color changes, and after deformation, the material undergoes partial phase transition to the auxiliary phases at room temperature, resulting in increased stiffness and the simultaneous memory of shape and color changes. Finally, the material can be reset to its initial state by heating (Fig. 1e). Leveraging the above principle, we successfully realized shape-color integrated memory and demonstrated the application of rewritable encryption photonic paper. This SS-CLCE film avoids the issue of poor compatibility between different phases and allows the spontaneous memory process at room temperature, turning material aging into an advantage rather than a failure, enabling each phase of the material to perform its respective function in a single-layer, independent film.

Results

Material synthesis and characterization

We prepared conventional CLCE and SS-CLCEs using the well-established anisotropic deswelling method42. The polymer films were obtained through a two-step crosslinking process. However, unlike previous work, this study employed a staged approach by sequentially using different solvents to optimize both solubility and polymerization temperature—initially enhancing raw material dissolution with dichloromethane (DCM), then maximizing heating temperature with ethyl acetate (EA). RM82 was used as the primary reactive LC monomer, LC756 as the reactive chiral dopant, ODT and EDDET as chain extenders, DPA as the catalyst for the chain extension reaction, and Irgacure 819 as the visible-light initiator for the homopolymerization reaction (Fig. 2a). The 1H NMR spectra of the main raw materials are shown in Supplementary Fig. 1. Complete names of all reagents and materials are listed in the Methods section.

Fig. 2: Construction of a lightly crosslinked network and characterization of phase states.
figure 2

a The components to fabricate CLCEs. b Construction of the chain-extension crosslinked network structure based on two-step crosslinking method. c Photographs (Scale bar: 1 cm), POM images (Scale bar: 100 µm), and DSC curves of conventional CLCE and d SS-CLCE, all transition temperatures are defined in Supplementary Fig. 4. Schematic diagrams of whether phase separation exists in the 3D network are shown below the photographs of the two samples, respectively. The blue rods represent the helical arrangement of the mesogenic units, while the white rods represent the more ordered structure. e Stress-strain curves, f XRD spectra, and g FTIR spectra of SS-CLCE after heating and aging. Aging refers to placement at 30 °C for 24 h, heating refers to placement at 80 °C for 1 min.

Full size image

EDDET was used as the chain extender for conventional CLCEs, while ODT was primarily used as the chain extender for SS-CLCEs. After the completion of chain extension, the obtained oligomers were dissolved in EA, poured into molds, and the solvent evaporation and self-assembly processes were then initiated. According to our strategy, most of the acrylate and thiol groups underwent addition reactions in the first chain-extended step, as confirmed by 1H NMR spectrum. In the second step, the remaining acrylate groups underwent homopolymerization, while the residual thiol groups indicated the presence of long side chains, as verified by Fourier-transform infrared spectroscopy (FTIR)43,44. Ultimately, a lightly crosslinked network incorporating long side chains was formed (Fig. 2b and Supplementary Fig. 2).

Conventional CLCE film was soft, highly elastic, and displayed bright selective reflection colors. Polarized optical microscopy (POM) images revealed a standard grainy texture42, indicating that the film was in the N* phase. Differential scanning calorimetry (DSC) curves showed no significant enthalpy changes, but a small endothermic peak was observed at 90 °C during the heating cycle, and a small exothermic peak appeared at 83 °C during the cooling cycle. These peaks suggested phase transitions of the long side chains near the isotropic transition point (Ti) (Fig. 2c).

Due to the long linear structure and crystallizable segment-containing characteristic of SS-CLCE oligomers, gradual self-assembly occurred during the evaporation film-forming process, leading to the formation of large-sized spherulites—which was detrimental to the integrity of the polymer film (Supplementary Fig. 3). Therefore, this gave rise to the discussion on the selection of appropriate self-assembly conditions. We selected 4 h (a duration with no macroscopic defects observed) as the self-assembly time prior to polymerization, the resulting SS-CLCE presented white, and showed a controlled rigidity that disappeared upon heating. POM images revealed a polydomain texture, suggesting the presence of phase separation. DSC curves further indicated that SS-CLCE had the mentioned isotropic transition of the side chains. Additionally, during the first heating cycle, two consecutive endothermic peaks were observed at 48 °C and 60 °C. Since a pronounced cold crystallization peak was observed at 43 °C (cold crystallization temperature, Tcc) during the second heating cycle, followed by a relatively small crystal melting peak, we believed that the second endothermic peak was associated with crystal melting temperature (Tm). The appearance of dual endothermic peaks is commonly explained by the sequential melting of different crystal structures45,46,47,48,49. We propose that the two endothermic peaks correspond to the disruption of structures with different degrees of order. Notably, the absence of a distinct exothermic peak during the cooling cycle suggested that the time scale for the main chain phase transition or crystallization was relatively long (Fig. 2d).

When the self-assembly time was further extended to 8 h or more, the influence of the auxiliary phase on the primary N* phase became significant, a large number of macroscopic defects (crystalline regions) are generated, which seriously affect the uniformity of the film. To prepare uniform elastomer films, a heating (100 °C) and cooling (68 °C) process was applied to eliminate crystalline regions before polymerization. The resulted CLCE was defined as the controlled CLCE (C-CLCE), which was induced into a vertical alignment mode due to the absence of orientation effect induced by evaporation. POM images revealed a chiral coil morphology, accompanied by the complete disappearance of selective reflection. The thermodynamic and mechanical properties of C-CLCE were nearly identical to those of SS-CLCE (Supplementary Fig. 4). All the thermodynamic data were shown in Supplementary Table 1.

Since the macroscopic properties of SS-CLCE resulted from the coexistence of the N* phase with other phases, we focused on it as the primary subject for subsequent research. Based on thermodynamic analysis, POM images of SS-CLCE were observed at different temperatures in the first heating cycle (Supplementary Fig. 5). The changes in color and brightness in the images indicated variations in the optical path difference (OPD) across different regions, reflecting the changes in the periodic structure of the material and the alteration of its phase state. It is noted that the performance difference between SS-CLCE and conventional CLCE indicated that the presence of ODT affected the ability to undergo phase separation. ODT segments not only enabled polymer crystallization themselves but also promoted the formation of the to-be-discussed smectic phase. This is because the rod-like cores of nearby LC molecules acquire the characteristic of orientational order along with the regular arrangement of crystalline molecular chains.

Structural causes of stiffness-switching performance

The rigidity of SS-CLCE films decreased after heating at 80 °C, as evidenced by a substantial reduction in Young’s modulus (E) and strain energy density (U) (Fig. 2e). Meanwhile, the sharp crystalline peaks in the X-ray diffraction (XRD) patterns clearly indicated the presence of crystallinity (Xc) of 26%, reflecting the volume proportion of crystalline regions in the polymer, with the remainder being LC phases and amorphous regions. We also fabricated a wedge-shaped film, and multi-point XRD tests demonstrated the uniformity of the microstructure across the entire film as well as the wide thickness range within which aging occurred (Supplementary Fig. 6). After heating, the crystalline features disappeared (Fig. 2f). We also ruled out the influence of residual monomers through the swelling experiment of SS-CLCE and the XRD pattern of the main monomer RM82 powder. The measured gel fraction (Gf) exceeded 78%, indicating a well-formed network structure (Supplementary Figs. 7 and 8). C-CLCE films exhibited the same trend, where crystallinity corresponded to the stiffness-flexibility switchability, while conventional CLCE films remained unaffected by thermal history (Supplementary Figs. 9 and 10). This suggested that phase separation—induced by ordered structures such as polymer chain crystallization and the retained N* phase—resulted in a 1000% increase in E for SS-CLCE, from 3.3 MPa to 36.1 MPa. The fracture energy increased from 494 MJ m-³ to 1391 MJ m-³, a growth of 180%. Since the fracture elongation of this material system was far lower than that of natural rubber (which exceeds 500%), only repeatable crystallization was observed, with no obvious strain-induced crystallization phenomenon (Supplementary Fig. 8).

To correlate the dual endothermic peaks in the DSC curve with structural changes in the material, we subjected SS-CLCE to heat treatments under different conditions, followed by rapid cooling to room temperature for FTIR testing, instead of performing in-situ variable-temperature infrared spectroscopy. This approach minimized the impact of the temperature responsiveness of the LC material on our analysis (Fig. 2g). When the material was heated to 53 °C or lower, no significant spectral changes were observed. This indicated that the first endothermic peak did not correspond to the disruption of a highly oriented or highly ordered structure. However, when the heat treatment temperature reached 80 °C—covering both endothermic peaks, the spectral features exhibited several pronounced changes, including peak broadening, shifting, and weakening. Specifically, the peak at 726 cm⁻¹ slightly weakened and broadened. This peak represents the backbone vibrations of long alkyl chain structures with more than four carbon atoms50, suggesting that the spacer chains of SS-CLCE underwent a transition toward disorder, corresponding to polymer crystal melting. The peak at 1472 cm-1, corresponding to the orthorhombic crystalline form of methylene, broadened upon heating, accompanied by the prominence of the 1466 cm-1 bending vibration, which corresponds to the “amorphous” state50,51. Decreases and shifts of multiple peaks at other positions were also observed (Supplementary Fig. 11), indicating that the second melting peak corresponded to the disruption of a significantly ordered structure—namely, the melting of polymer chain crystallites52,53,54. Furthermore, crystallization was not limited to the longer alkyl chains, the thioether (650 cm−1) and benzene ring structures within the polymer network also contributed to the formation of the ordered structure.

In-situ variable-temperature 2D wide-angle X-ray scattering (2D-WAXS) experiments were conducted to demonstrate molecular-level structural changes during the heating process of aged SS-CLCE (Supplementary Fig. 12). Four temperature points were selected for testing (Supplementary Fig. 13). The 2D patterns displayed arc-shaped peaks, which originated from the sample preparation process and was minimal, having no impact on the long-period structure of the sample (Fig. 3a and Supplementary Fig. 14). The temperature-induced pattern changes were easily observed. The crystalline peaks in the wide-angle region and the long-period peaks in the small-angle region were eliminated during heating, leaving only the characteristic diffuse peak at 1.41 Å−1, typical of LC materials. The 2D images were converted into 1D curves for further analysis (Fig. 3b–d). According to Bragg’s law, at the initial temperature of 30 °C, four crystalline peaks at scattering vectors of 1.34, 1.46, 1.56, and 1.68 Å−1 corresponded to interplanar spacings of 4.65 Å, 4.30 Å, 3.97 Å, and 3.72 Å, respectively. Based on the Scherrer equation, the corresponding crystallite sizes were approximately several to a dozen nanometers. (Supplementary Fig. 15a and Supplementary Table 2). Furthermore, adjusting the ratio during the synthesis process can regulate the crystallization properties, corresponding to the control of the crystallinity (Supplementary Fig. 15b). The long-period peak at 0.141 Å−1 corresponded to a 4.4 nm layered structure of LC units and spacer chains, which is typically indicative of a smectic phase in LC materials41. However, due to the polydomain nature and localized phase separation of the material, the smectic phase characteristics were not prominent (Fig. 3e). Due to instrument limitations, data below 0.05 Å−1 were invalid. It could be inferred that the peaks below 0.1 Å−1 corresponded to larger-scale periodic structures ( > 6.2 nm), typically representing the periodic distribution of ordered and disordered regions 45.

Fig. 3: Microstructural characterization of SS-CLCE.
figure 3

a In-situ variable-temperature 2D-WAXS patterns. We set the corresponding parameters and observed changes in both the wide-angle (top) and small-angle (bottom) regions. b 1D-radial integration curves of the wide-angle region and c their locally magnified views (in the gray square), the red dashed line denotes the peak position. d 1D-radial integration curves of the small-angle region. The inset shows a magnified view of the effective data in the small-angle region. e Schematic diagram of the LC polydomain structure. f 2D-SAXS patterns before (top) and after (bottom) stretching. The arrows indicate the stretching direction. g 1D-radial integration curves (top) and 1D-azimuthal integration curves (bottom) under different strain (ε) conditions. h Schematic diagrams of the molecular alignments in the SmC phase, the red squares represent the layers before and after stretching. i Schematic diagram of the molecular arrangement and dimensions. It is noted that the omitted spacer chains enhance the clarity of the diagram, but they also participate in the formation of the ordered structure.

Full size image

As the temperature increased to 46 °C, the diffuse peak at 1.41 Å−1 continuously intensified, while the crystalline peaks showed a slight decrease. A minor shift was observed at 1.69 Å−1, which resulted from the overlapping intensity variations of multiple peaks rather than the actual shift of an independent peak. This indicated the melting of a small number of imperfect crystallites. Meanwhile, no decrease was observed in the small-angle peaks, suggesting that both types of long-period structures remained largely intact. At 53 °C (the final stage of the first endothermic peak and the early-to-mid stage of the second endothermic peak in the DSC curve), further crystal melting occurred. However, the intensity of peaks below 0.1 Å−1 did not decrease, indicating that the melting of a small number of crystallites did not disrupt the large-scale periodic distribution of ordered and disordered regions. Meanwhile, a 7 °C temperature increase led to a 50% reduction in the intensity of the smectic layering peak, suggesting that the latter stage of the first endothermic peak corresponded to the breakdown of the smectic phase. Finally, when the temperature reached 80 °C, all crystalline peaks disappeared, the periodic ordered-disordered structure was completely disrupted, and the smectic phase degraded into the N* phase55. Based on these findings, we inferred that the two independent endothermic peaks spanning 40 °C in the DSC curve corresponded to the phase transitions from the smectic phase to the N* phase and from polymer crystallization to the LC state, respectively. The enthalpy-change relationship corresponding to these two peaks implied the relative quantitative relationship between the smectic phases and the crystalline phases.

To confirm the existence of the smectic phase, we further calculated the energy-minimized structure of the LC monomer and chain extender using the MM2 force field (Supplementary Fig. 16). The molecular length of the LC unit was found to be approximately 3.8 nm, the chain extender’s molecular length was 1.2 nm, so the length of the extended unit was 5 nm, which was larger than the 4.4 nm interlayer spacing shown by WAXS. This suggests that the phase is SmC33,56. 2D small-angle X-ray scattering (2D-SAXS) tests were conducted on both the un-stretched aged sample and the aged sample at ε = 1.5, showing significant orientation differences (Fig. 3f and Supplementary Fig. 17). 1D data were obtained through integration, revealing only a slight change in the interlayer spacing, from 4.55 nm to 4.39 nm (Fig. 3g, top), which was the result of both chain extension and the directional change of the interlayer spacing. Additionally, a weak peak was observed at 1.7 nm, which was not visible in the wide-angle experiment due to its low intensity. This peak was unaffected by stretching orientation and may represent short-range ordering of LC units at covalent crosslinking sites. Azimuthal integration (Fig. 3g, bottom) showed that the orientation of the layered structure changed from the standard ring-like unoriented state to a double-dumbbell state, reflecting that stretching caused the layered structure to misalign, forming a V-shaped structure with an angle of 120°, indicative of the chevron SmC (cSmC) phase (Fig. 3h)57. Based on the periodic arrangement of semicrystalline polymer crystalline grains, intermediate phases, and amorphous regions58,59, we provided a schematic diagram of the molecular arrangement of the aged SS-CLCE in the un-stretched state, showing the molecular spacing data in different phase regions (Fig. 3i). In summary, SS-CLCE underwent diverse phase transition processes, where microphase separation occurred after aging, involving the primary N* phase, the auxiliary SmC phase, and chain crystallization. The coexistence of SmC and Cr in N* phase enabled physical crosslinking. While retaining partial LC properties, the material exhibits a significant increase in stiffness and can be softened by heating.

Mechanochromism and self-reinforcement effect of SS-CLCE

Due to the presence of the helical structure, CLCEs exhibit selective reflection of light, and the variation in reflected color is correlated with changes in the helical pitch (p) (Supplementary Fig. 18)60,61,62. In this study, microphase separation was achieved within the primary N* phase to obtain switchable mechanical properties while maintaining effective reflection color after stretching. A uniaxial stretching experiment was performed on SS-CLCE films after heating at 80 °C for 1 min (Fig. 4a). The initial film appeared white with a faint orange-red reflection, corresponding to a broad reflection peak around 630 nm. This is because the helical structures in the sample had uneven orientations and variations in p at different positions, influenced by the relatively weak orientation force of solvent evaporation during preparation. Meanwhile, the difference in refractive index for light among different microdomains leads to intense light scattering, exhibiting non-selective unpolarized light. Considering comprehensively, the white color under natural light is the combined result of unpolarized light scattering and broad-peak reflection of right-handed circularly polarized light. In the early stage of stretching, the surface alignment of the film improved, reducing diffuse reflection and enhancing the intensity of reflection in the perpendicular direction. When the strain increased to 60%, the alignment of the LC units was further optimized, resulting in a well-ordered perpendicular arrangement of the helical structure. Consequently, the broad-reflection feature diminished, and a single reflection peak at 560 nm (green) emerged. With further stretching, the peak became narrower as the film approached a monodomain state; consequently, the blue color became more vivid while the white background gradually faded. However, the overall reflectance decreased, indicating partial disruption of the helical structure. Overall, the color change in the film was uniform, demonstrating its soft-elastic behavior. It should also be noted that the structural color exhibited by SS-CLCEs has certain limitations compared with existing CLCEs, which is attributed to the molar ratio of acrylate to thiol (used for oligomer preparation) being close to 1:1, as well as the selection of the microphase separation strategy and the solvent evaporation method. The trade-off involved causes SS-CLCEs to sacrifice the quality of structural color in exchange for switchable mechanical properties63.

Fig. 4: The selective reflection and mechanically switchable properties of SS-CLCE.
figure 4

Changes in the reflection color with stretching after a heating and b aging. The red dashed circle indicates the data collection positions. c Stress-strain curves for strains ranging from 0 to 100% at different aging times at 20 °C. The right half shows the calculated values of Young’s modulus (E) (the inset is a magnified view of the data for shorter aging times) and strain energy density (U). d Cross-section SEM image of the SS-CLCE, with yellow lines indicating the pitch length and direction, while the dashed lines mark the regions where the pitch is absent. e DSC curves after aging at different temperatures. f Stress-strain curves for the 0–100% strain range after aging at different temperatures at a fixed time of 24 h. The dashed arrows reflect the effect of increasing aging temperature on the self-reinforcement effect. g Schematic diagrams of temperature-time dependent phase transitions.

Full size image

A different phenomenon was observed when a uniaxial stretching experiment was conducted on aged films (Fig. 4b). The initial color remained as a weak red reflection on a white background due to the polydomain structure. Due to the higher crystallinity, the aged film exhibited increased rigidity. In the early stage of stretching, stress concentration led to significant deformation at the film’s edges first, causing a sudden shift in reflection color to blue at the edges. As the applied stress increased, deformation gradually extended to the central part, reducing the polydomain regions. However, due to physical cross-linking, the stretching-induced alignment of LC units was restricted, leading to minimal changes in reflectance. Since the regions connecting the edges and the center displayed intermediate colors, we also collected the reflection data at the edge positions (Supplementary Fig. 19), these data reflected the differences between the center and the edge positions. In summary, SS-CLCE still exhibited controllable reflection colors within the visible spectrum. The reflection spectra of other samples are shown in Supplementary Fig. 20.

Heating eliminated phase separation, leading to material softening. Conversely, the softened material could undergo aging under certain conditions, resulting in an increase in stiffness. To investigate the time-dependent self-reinforcement effect, the tensile properties were tested for aging times ranging from 0 to 72 h, with the treatment temperature fixed at 20 °C (the second variable affecting the aging degree) to ensure a uniform aging rate. The stretching rate was fixed at 25 mm min−1. Since stiffness changes are primarily reflected in the elastic deformation region, the study focused on the strain range of 0 to 100% (Fig. 4c), although local preferential strain occurs during the process (local necking first occurs at a specific point), this is also a reflection of the overall performance. The material was tested after removing thermal history and cooling to room temperature, serving as the baseline data before aging. When the aging time was less than 8 h, the stress at 100% strain remained below 1 MPa, with a low modulus and a limited elastic deformation region. As the aging time increased, a distinct elastic deformation region emerged, and the modulus gradually increased. After 48 h, the stress at 100% strain reached 3.5 MPa, marking a 1066% increase compared to the no-aging value of 0.3 MPa. The E followed an S-shaped curve, indicating a slow-fast-slow trend in the physical cross-linking rate, peaking at 16 h before gradually approaching zero. The early-stage data points were densely distributed, clearly illustrating the mid-stage acceleration in cross-linking. After 48 h of aging, E increased from 2.2 MPa to 43 MPa, a rise of 1854%. The U exhibited a similar trend, increasing from 197 kJ m-³ to 3218 kJ m-³, an improvement of 1533%. These results demonstrate that the material’s stiffness and toughness significantly improve over time, highlighting its time-dependent self-reinforcement effect. Scanning electron microscopy (SEM) revealed the absence of helices in some regions, reflecting the cause of the self-reinforcement effect—namely, the presence of phase separation. SEM further demonstrated the distribution of p (Fig. 4d)64,65,66. To examine the mechanical performance potential of SS-CLCE, we set the aging time to two weeks and even five months for testing (Supplementary Fig. 21). The stress-strain curves indicated that after 2 weeks of aging, the Young’s modulus increased from 43.1 MPa (at 48 h) to 63.4 MPa. However, no significant increase in Young’s modulus was observed for the material aged for 5 months, with only a partial improvement in toughness. This phenomenon demonstrated that the SS-CLCE film still underwent a self-reinforcement process after 2 weeks. Nevertheless, after 48 h, the material required much more time (tenfold to hundredfold) to achieve a much smaller enhancement. The result was consistent with the S-shaped enhancement trend. Guided by the principle of efficiency, we selected 24 h or 48 h as the main research timeframes. The results showed that the material exhibited a self-enhancement effect over a long-time scale. As time progressed, the mechanical properties continuously improved, making aging an advantage rather than a failure process.

The DSC curves of SS-CLCE after aging for 48 h at different temperatures were compared (Fig. 4e). It was observed that as the aging temperature increased, both endothermic peaks shifted to higher temperatures. The SmC-N* phase transition peak maintained its shape, while the melting peak became sharper. Combined with XRD and WAXS data, these results indicated that within a certain range, higher temperatures lead to a more complete crystalline structure, increased grain size, reduced crystal plane spacing, and a narrower melting range, which aligns with general polymer crystallization behavior. Additionally, the enthalpy of fusion decreased from 19.43 J g−1 to 12.22 J g−1, corresponding to a reduction in phase separation, which could potentially lead to a decrease in mechanical performance. To further investigate the temperature-dependent self-reinforcement effect, the tensile properties of SS-CLCE films were tested at different temperatures with a fixed aging duration of 24 h to exploring the impact of temperature on the rate of aging. The stretching rate was fixed at 25 mm min-1 (Fig. 4f). The glass transition temperature (Tg) of the material is −12.2 °C. When the aging temperature was close to Tg, molecular chain motion was restricted, resulting in only a slight enhancement of the mechanical properties. As the temperature approached room temperature (20 °C), a significant improvement in mechanical performance was observed. When the temperature reached 30 to 40 °C, near the melting range, the mechanical performance peaked. When the temperature increased beyond the early stage of the melting range (50 °C and 80 °C), the mechanical reinforcement weakened. This phenomenon reflected that high temperature would also inhibit the occurrence of phase separation, further underscoring the importance of selecting an appropriate temperature for phase separation. (Supplementary Fig. 22). Despite this, the slope of the stress-strain curve at 50 °C increased significantly, suggesting the formation of more complete and larger crystalline structures.

In summary, we emphasized that the phase transition of LC polymer materials exhibits a dual dependence on both temperature and time. Unlike traditional studies on LC materials that focus solely on temperature as the determinant of phase state (Fig. 4g, top), we introduced time as a second dimension. At an optimal temperature, self-enhancement of mechanical properties can be achieved within a short time, whereas excessively low or high temperatures prolong or even prevent phase separation from occurring (Fig. 4g, bottom). Heated-aged cycle tests were conducted to demonstrate the stability of the self-reinforcement effect and the durability of the SS-CLCE. The mechanical properties after aging at different temperatures and the structural color after cycling both indicated the durability and fatigue resistance (Supplementary Fig. 23). This finding provides valuable insights for the controlled tuning of mechanical properties.

Shape-color integrated memory based on stiffness switching

The stiffness-switching capability of the material endows it with programmability. Just like the metamorphosis of Buprestidae (Fig. 1a), a soft, white larva evolves into a colored beetle with a hard exoskeleton after long-time sustained growth. Specifically, SS-CLCE can fully soften within 10 seconds when heated to 80 °C and remain flexible for the following 8 h at room temperature, attributed to the slow reinforcement rate in the initial stage of the S-shaped reinforcement process. The modulus remains relatively low within 8 h, allowing for large and complex deformations with minimal force. Certainly, the 8-h duration is merely a reasonable choice based on the reinforcement rate, it is not fixed and can be flexibly adjusted according to specific requirements. This period is defined as the programming region. Near room temperature, the material undergoes spontaneous phase transitions, forming physical crosslinks that enhance stiffness. The phase with a rapid increase in stiffness is defined as the crosslinking or setting region. When the crosslinking time exceeds 24 h, the material’s stiffness and mechanical properties have significantly improved, and the crosslinking rate slows down. Although the material has not yet reached its maximum mechanical strength, this duration balances time efficiency with mechanical enhancement, making it suitable for practical applications. Therefore, the region beyond 24 h is designated as the application region after programming (Fig. 5a). If the requirements change, reheating the material can return it to its initial state, enabling reprogramming. This defines the application logic of SS-CLCE materials.

Fig. 5: Demonstration of stiffness switching and shape-color integrated programming.
figure 5

a Explanation of the programming strategy. The two dashed lines correspond to 8 h and 24 h, respectively. b Loading experiment of stripe-shaped SS-CLCE. Heating disrupted the physical crosslinking, returning it to the unaged state. c One-way programming of SS-CLCE film demonstrating the memory process and degree of two temporary shapes. d Feasibility verification of large-size and complex-shaped programming. All scale bars in Fig. 5 are 1 cm.

Full size image

First, a force test was conducted on strip-shaped film samples. In the aged state, the sample could withstand 1667 times its own weight without significant deformation, maintaining a white, multi-domain, rigid state. Upon heating, the sample instantly softened, stretched into a single-domain transparent state accompanied by a blue shift in the reflection color, and remained unbroken (Fig. 5b). When extensively heated, the film exhibited the same deformation behavior (Supplementary Movie 1). This clearly illustrated the difference in the mechanical properties of the material before and after aging. Next, we demonstrated the shape-color integrated programming process of SS-CLCE. The strip-shaped film served as the permanent shape, which could be molded into a shelf-shaped (temporary shape) after softening, displaying structural color changes due to deformation. After aging, the shelf regained stiffness and retained its structural color, capable of supporting a certain weight. Under excessive load, the shelf collapsed but could stand upright again once the weight was removed (Supplementary Movie 2). Throughout these processes, bending and tilting due to applied pressure were visibly reflected in the selective reflection colors. Upon reheating, the temporary shape was erased, restoring the permanent shape, which could then be reprogrammed into a spiral shape (temporary shape, Methods Section). This spiral shape was entirely composed of curved surfaces, exhibiting a rich structural color (Fig. 5c). It demonstrated a certain load-bearing capacity and could recover its spiral shape after being flattened and unloaded (Supplementary Movie 3). These demonstrations highlighted the stability and reusability of the shape-color integrated memory effect in SS-CLCE materials. Finally, a large-sized film (90 × 50 mm) was fabricated and programmed into the shape of the Monkey King’s facial mask (62 × 58 × 30 mm). The mask accurately retained the shape of facial features with fine pattern details (Fig. 5d). Upon heating, it was also capable of reverting to its original state, thereby enabling the next usage cycle. It is worth noting that the aforementioned programming time requires more than 24 h to achieve better memory effect. By introducing more crystallizable segments and increasing the aging temperature to accelerate chain rearrangement, the duration can be further shortened to 8 h (Supplementary Fig. 24).

Rewritable photonic paper for encrypted writing

The previously described whole-film processing of the film enabled the retention of both its 3D shape and structural color. In contrast, localized processing allowed for the recording of encrypted textual information. Previous work reported stiffness control via crosslinking degree: (1) using photomasks on single films to form deformation patterns via differential photopolymerization; (2) making matrices with small CLCE pieces of varying crosslinking degrees for regional color differences26,67,68. Similarly, differences in thickness will convert the same external force into different stresses, thereby resulting in color distinction during the stretching process69,70. These measures fixed film modulus, precluding erasure/rewriting (crosslinking control) or single-film color variation (dot matrix). While prior studies are valuable, our work enabled post-fabrication non-directional patterning via localized heating to reduce modulus. Heating the entire film and re-aging allowed rewriting, achieving write-erase on a single film without matrixization or custom photomasks.

As shown in Fig. 6a–b, the initially white film exhibited uniform color and modulus. By locally heating the film to 80 °C using a heating pen, the desired information was written, making the information section soft and more susceptible to stretching within a few hours. Although the modulus of the heated regions decreased, the apparent color remained white. At this moment, the information was in an encrypted state and cannot be observed. Only slight marks caused by excessive writing pressure may remain on the film surface. Due to the polydomain and phase-separated nature of the film, its non-selective broadband reflection hindered the readability of the recorded information. In this case, a right-handed circular polarizer (R-CP) is the first key to decrypt the information, the intensity of unpolarized reflection and scattering is reduced by half (left-handed polarization is blocked), and the intensity of right-handed polarized light centered at 630 nm, dominated by the N* phase, remains unaffected60,67. Thus, a distinct orange-red background was observed through a R-CP. Biaxial stretching as the second key induced uneven deformation in soft and hard regions, resulting in structural color variation that decrypted visual information. Subsequently, the self-reinforcement effect helped preserve the displayed information by maintaining external stress at 30 °C for 1 day. The film can recover after globally heating the film at 80 °C and leaving at 30 °C for one day, allowing for a new cycle of information encrypting. As an example, we successively wrote in the year “2025” and key material’s abbreviation “CLCE” (Fig. 6c). The modulus of the information-bearing regions was lower, resulting in greater deformation and a more pronounced blue shift in structural color, thereby displaying high-contrast yellow-green patterns under R-CP. Supplementary Movie 4 illustrated the process of information decryption.

Fig. 6: Demonstration of encrypted information writing.
figure 6

a Schematic diagram of the thermal writing-decrypting-erasing process, a right-handed circular polarizer (R-CP) and stretching serve as the two keys for decrypting. b Photograph of the thermal writing process. c Photographs of the results of the information writing, decrypting and erasing. All scale bars in Fig. 6 are 1 cm.

Full size image

Discussion

In this work, we report the bioinspired SS-CLCE designed via a two-step crosslinking strategy, where linear long side chains were introduced to enhance phase transition tendencies. Within the temperature range of 10–40 °C, the originally pure N* phase of the crosslinked network spontaneously transformed into SmC and Cr regions under the influence of the mobile long side chains, while overcoming the constraints of helical molecular alignment. This transformation occurred in localized domains, leading to microphase separation and subsequent physical crosslinking that hardened the material. Meanwhile, the residual N* matrix preserved its selective reflection of light, resulting in vivid structural colors. The resulting mechanical performance was highly dependent on both temperature and time. At 30 °C, the material exhibited a 1000% increase in E after 24 h of aging, meeting the requirements for shape-color integrated memory effects. Moreover, the stiffness of the material continued to improve over time. In the programming process of SS-CLCE, the stiffness switching enabled significant energy savings. Brief heating (80 °C for 10 s) dramatically reduced the stiffness, allowing large deformations with minimal energy input. Self-strengthening could proceed spontaneously at ambient temperatures, with optimal memory efficiency observed between 20–40 °C. The programmed shapes demonstrated sufficient rigidity to bear weight and displayed structural colors. When new demands arise, the devices can be reprogrammed through simple reheating and reshaping. The material’s special properties further endowed it with the potential for encrypted writing, enabling thermal encrypting, mechanical decrypting and erasure of information. The reversible stiffness switching and dynamic structural coloration of SS-CLCE offer a practical design paradigm for shape memory polymers, breaking the long-standing limitations between mechanical performance and color in polymeric systems.

Methods

Materials

In this study, 1,4-bis[4-(6-acryloyloxyhexyloxy)benzoyloxy]-2-methylbenzene (RM82, 95%) and 1,4:3,6-Dianhydro-D-glucitolbis[4-[[4-[[[4-[[1-oxo-2-propenyl]oxy]butoxy]carbonyl]oxy]benzoyl]oxy]benzoate] (LC756, 90%) were purchased from Kindchem (Nanjing) Co., Ltd. 1,8-octanedithiol (ODT, 98%) was purchased from Macklin. 3,6-Dioxa-1,8-octanedithiol (EDDET, 98%) was purchased from Energy Chemical. Dipropylamine (DPA, 99%) was purchased from Aladdin. Bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide (Irgacure 819, 99%) was purchased from Heowns. Dichloromethane (DCM, >99.5%) and ethyl acetate (EA, >99.5%) were purchased from FUYU Chemical. All chemicals and reagents were used as received.

Synthesis of CLC linear oligomers

All the studied CLCEs were prepared based on a two-step crosslinking method, with the first step being the synthesis of oligomers. In the SS-CLCE and C-CLCE systems, the molar ratio of acrylate to thiol was 1:0.99. The molar ratio of RM82 (240 mg, 0.357 mmol) to LC756 (11 mg, 0.011 mmol), which provided acrylate functional groups, was 1:0.032, while the molar ratio of ODT (52.7 mg, 0.296 mmol) to EDDET (13.5 mg, 0.074), which provided thiol functional groups, was 1:0.25. For the conventional CLCE system, only EDDET (64.2 mg, 0.353 mmol) was used to provide thiol functional groups. All monomers were dissolved in DCM, 1.5 wt% DPA (excluding solvent mass, 5 mg, 0.049 mmol) was added, and 1.5 mL of DCM was used per 300 mg of solute for thorough mixing. The mixture solution was then placed in a glass dish at 33 °C, allowing the solvent to evaporate within 2 h while carrying out the initial Michael addition reaction, which served as the chain-extension step. Finally, the oligomer was dissolved in EA (half the volume of DCM), followed by the preparation of the polymer film.

Preparation of CLCE films

0.3 wt% Irgacure 819 (excluding solvent mass, 1 mg, 0.002 mmol) was added to the oligomer solution, and then poured into a glass mold at 68 °C. The solvent was evaporated under dark conditions, and a certain self-assembly time was maintained. When the mass of RM82 was 240 mg, a glass mold with dimensions of 50 mm * 30 mm * 10 mm was used. The self-assembly times for SS-CLCE, C-CLCE, and conventional CLCE were 4 h, 8 h, and 4 h, respectively. And C-CLCE underwent an annealing process (heated to 100 °C and then cooled to 68 °C). Finally, irradiate with a 450 nm blue LED (100–260 V, 50 W) for three minutes to carry out the photopolymerization process.

Characterization

1H NMR spectra were recorded on a Bruker AVANCE III HD 400 MHz spectrometer using CDCl3 as solvent. FTIR spectra were measured on a Nicolet 6700 FTIR spectrometer. POM images of CLCEs were characterized by Leica DM2500P polarized optical microscopy operating in transmission mode, coupled with a temperature-controlled hot stage (Linkam THMS-600) calibrated to ±0.1 °C. The temperature-dependent POM images were taken after heating at a rate of 2 °C min−1, followed by a 1-min hold. DSC measurements were performed using a DSC 3500 Sirius (Netzsch) between −30 and 140 °C with heating and cooling rates of 10 °C min−1 under nitrogen. The tension measurement was carried out using a stretching machine F105-EM (Mark-10). XRD measurement was conducted using a Shimadzu XRD-7000 diffractometer with Cu Kα radiation. The reflection spectra are recorded with an Avantes AvaSpec-2048 spectrophotometer in the dark at room temperature. SEM images of the SS-CLCE was obtained using a TESCAN MIRA LMS scanning electron microscope at an acceleration voltage of 3 kV. Film cross sections were prepared by fracturing after freezing in liquid nitrogen. 2D-WAXS measurements were performed using an Anton Paar SAXSpoint 2.0 system equipped with a Cu Kα radiation source (λ = 0.154184 nm, 40 kV) and a 2D hybrid photon-counting detector (EIGER R 1 M, pixel size = 75 μm). The sample-to-detector distance was 0.54 m. The in-situ temperature-dependent test was conducted by holding at 30 °C, 46 °C, 53 °C, and 80 °C while collecting data. 2D-SAXS measurements were performed using an Xenocs Xeuss 2.0 system equipped with a Cu Kα radiation source and a 2D hybrid photon-counting detector (platus3R, pixel size = 172 μm). The sample-to-detector distance was 0.38 m.

One-way shape-color integrated memory

The shape of the mold during polymerization determined the permanent shape of the SS-CLCE. We prepared strip-shaped film using a simple quadrilateral mold. The film was used after being stored at room temperature for 24 h following preparation. Then, it was fully softened by heating on an 80 °C hot stage for 10 s, causing a sharp decrease in stiffness (modulus) to facilitate strain generation. The softened film was then molded into a shelf shape and placed in a 30 °C oven for 24 h to enable efficient physical crosslinking, enhancing stiffness to fix the shelf shape and the structural color induced by deformation. Finally, the SS-CLCE in the shelf shape was demolded.

The shelf-shaped SS-CLCE was heated on an 80 °C hot stage to restore its permanent shape. It was then stretched and coiled around a metal rod to form a spiral structure. After fixing both ends, the film underwent a crosslinking process, and the spiral-shaped SS-CLCE was obtained upon demolding. Notably, benefiting by the intrinsic nature of spontaneous physical crosslinking, the material could be reused multiple times.

A large-area SS-CLCE film was fabricated for more intricate programming. After heating, the film was stretched and spread over a plaster mask mold. A commercial plastic film was first applied to the surface, followed by commercial clay to ensure firm contact. The plastic film was used to prevent clay residue from adhering to the SS-CLCE surface. After placing the setup in a 30 °C oven for 24 h, the demolded SS-CLCE mask retained its shape and structural color.

Encrypted information writing

The film was used after being stored at room temperature for 24 h following preparation. A heating pen set to 80 °C was gently applied to the surface of the SS-CLCE film to write target information (e.g., “2025” or “CLCE”). The written information was more distinguishable under right-handed circular polarizer. The strain was maintained for 24 h to fix the color pattern. Erasure was performed as described in the One-way shape-color integrated memory section. The film was then stored at room temperature for 24 h before the next writing cycle.