Mechanically-robust and humidity-tunable self-destructive polymers enabled by hydrogen-bond nanoconfinement

Abstract

Conventional materials typically maintain a stable solid state under specific application conditions. However, materials that combine high mechanical strength with reversible switching states are highly desirable for advanced technologies. Here, we present hydrogen-bond nanoconfined self-destructive polymers (HNSPs) that have robust mechanics and reversible solid-fluid switching capability at 25 °C. Upon exposure to moisture, HNSPs spontaneously transition from solid to fluid, with humidity-tunable switching rates. HNSPs with weight ratios (Rm) of 1.7 is 1.69 times higher self-destructive rate than those with Rm of 2.1, and at 90% relative humidity (RH), Rm=2.0 samples exhibit an 804.27% higher self-destructive efficiency than at 60% RH. Heating can reverse the fluid back to a solid, enabling reversible, humidity-programmable behavior. Mechanistically, large ordered hydrogen-bond clusters, reduced chain entanglement, and abundant hydrophilic groups collectively facilitate switching. This work provides a simple yet versatile strategy for designing robust, switchable self-destructive polymers, broadening their potential in next-generation devices.

Materials that combine high mechanical strength with stable and reversible switching states are highly desirable for advanced technology applications. Here, the authors reported hydrogen-bond nanoconfined self-destructive polymers that have reversible solid-fluid switching capability at 25 °C and maintain good mechanical properties.

Introduction

Achieving long-term stability, high performance, and consistent functionalization with minimal variability stands as the paramount objective in the design and development of advanced materials^1–4. However, specific demands emerge when materials must undergo self-deconstruction or physical dissolution upon mission completion, or when immediate destruction is required. These increasing needs arise for the elimination of electronic waste^5,6, the protection of sensitive information⁷, and the reduction of secondary surgeries associated with implantable medical devices^8,9. This research addresses the urgent need for these advanced materials capable of controlled self-deconstruction, which is crucial for applications ranging from electronic waste reduction and data security to minimizing invasive medical procedures.

One traditional approach to achieve self-destructive behaviors involves applying external stimuli, such as light¹⁰, heat¹¹, magnetic¹², or electricity, to induce phase transition that led to physically melting or sublimation. For instance, thermoplastic polymers undergo distinct phase transitions during crystallization melting or glass transition, enabling a reversible transformation from solid to fluid states^13,14. Despite these materials show promise, they typically require continuous external stimulation to maintain the phase change, often involving the formation of high-energy intermediates. Consequently, once the external stimulus is removed, these materials are unable to sustain their “self-destructive” state and quickly revert to their original form^15,16. Alternatively, the development of biodegradable or soluble materials offers another potential solution. Those materials such as polylactic acid (PLA)^17,18, silk fibroin (SF)¹⁹, and starch-based polymers²⁰ can degrade in response to external stimuli like light, heat and enzymes, etc. These stimuli will break the chemical bonds within the polymer chains, progressively reducing large molecules into low molecular weight fragments, ultimately resulting in physical disappearance or self-destruction^21,22. However, these molecular-level self-destructive behavior also relies on continuous stimulation (special solvents or biological fluids), leading to an uncontrollable destruction over the “Self-destructive process”, which is passively controlled by the chemical dissolution or stimulated destruction of functional materials associated with self-destructive behavior^23,24. Thus, developing transient materials or device capable of self-deconstruction or controlled dissolution at specific times and rates remains a significant challenge.

To address these limitations, researchers have increasingly focused on the materials capable of achieving reversible switching between different states while maintaining stability without continuous passively stimulation under the same environmental conditions^25–28. Worrell et al.²⁹ developed a new responsive material that can undergo a transition between a dynamic and a permanent network upon exposure to light, resulting in a stable switch between fluid and solid states at room temperature. However, this stable phase transition is unidirectional once the material transforms from a solid to a fluid state under stimulation, but it cannot revert to its solid state. Similarly, Yang et al. ³⁰introduced a polymer impregnated with a supercooled salt solution, enabling reversible changes between soft and hard states through nucleation stimulation or heating. Notably, this phase transition is both reversible and stable with the materials in the hard state being 10⁴ times harder than in the soft state. Despite these advancements, the switching capability of such materials remains limited, failing to achieve the significant transformations akin to solid-liquid phase transitions. Moreover, these materials exhibit a fixed switching rate and suffer from poor mechanical performance, with fracture strengths only reaching the kilopascal level. In addition, their self-destructive processes often need specialized stimuli, such as specific wavelength of light or seed crystals, rendering their applicability in diverse environments. Therefore, there is a pressing need to develop systems that provide materials with superior mechanical properties, utilize mild and widely applicable stimulation methods, and offer controllable switching rate. Such innovations would enable reversible solid-liquid phase change, broadening the applicability and functionality of advanced materials.

Moisture, as an abundant and environmentally benign resource, has significant advantages in triggering mechanical responses due to its low energy cost, wide range of application, easy of control, and mild operating conditions^31–34. In nature, spider silk exhibits remarkable sensitivity to moisture, a phenomenon known as super-contraction, which arises from the presence of hydrophilic 0peptide segments in spider silk proteins and surface coatings^35,36. When exposed to water or high humidity, spider silk can shrink up to 50%. This behavior occurs because water molecular disrupt hydrogen bonds, thereby increasing molecular mobility and driving the rearrangement of amorphous regions into lower energy configurations through enhanced molecular entropy^37,38. As illustrated in Fig. 1a, hydrogen bonds play a critical role in imparting the excellent mechanical properties of spider silk, as they facilitate the self-assembly of stable and ordered nanostructures (β-sheet nanocrystals), forming interlocked regions that enhance cohesive energy and enable effective load transfer along molecular chains^39–41.

Fig. 1. Schematic diagram of Hydrogen-bond nanocluster self-destructive polymers inspired by spider silk.

a Schematic diagram of spider silk with a hierarchical structure. b Schematic structure of HNSPs composed of dense hydrogen-bond nanoclusters. c Molecular dynamics (MD) simulations of the microstructure of HNSPs system with the corresponding H-bond categories diagram. d TEM images of the HNSPs (Inset: size distributions of hard segment clusters.). e Cryo-SEM images of the HNSPs. f Optical images of the HNSPs (Rm = 2.0) can easily lift a weight of 500 g.

Inspired by spider silks, we hypothesized that introducing dense hydrogen bonds between polymer chains through molecular structure design, alongside the establishment of ordered nanostructures, could impart exceptional mechanical properties to the material. Concurrently, by designing polymer chains with hydrophilic characteristics, the material can spontaneously absorb moisture from the surrounding environment. The absorbed water molecules would significantly disrupt the interactions between polymer chains while establishing new and strong interactions, resulting in the dissipation of nanostructures and enabling autonomous self-destructive behavior. Furthermore, the strong interaction between water molecules and polymer chains are not easily reversed by removing moisture stimuli, those restoration to the original state necessitates the application of strong heating and pressure.

In this study, we report the synthesis of hydrogen-bond nanoconfined self-destructive polymers (HNSPs) that exhibit controllable self-destruct behavior alongside excellent mechanical strength. The compressive yield strength of HNSPs reaches up to 14.66 MPa, with a maximum Young’s modulus of 165.21 MPa. Under the ambient conditions (25 °C), the system undergoes two reversible yet stable phase transitions. HNSPs can maintain their structural integrity for hours or even days at room temperature, contingent on the environmental humidity. In high humidity environments, the designed HNSPs can spontaneously absorb moisture, triggering a phases transition from a solid to a fluid state. Upon removal of the moisture stimulus, the system does not revert to its original state. Instead, restoration to the solid state occurs only upon the application if intense heating, facilitating the fluid-to-solid transition. Especially, this phase-switching process can be repeated multiple times. We believe that the exploration of reversible and stable HNSPs opens new avenues for innovative and practical strategies in fabricating stiffness-modulating polymers, thereby broadening their application fields in next-generation smart devices.

Results

Hydrogen-bond nanocluster assembly of HNSPs

Hydrogen-bond nanocluster self-destructive polymers (HNSPs) were synthesized via a one-step polymerization process in deionized water, those detailed synthesis process were elaborated in the Supplementary Information. As illustrated in Fig. S1 tetraethylenepentamine (TEPA) was reacted with acrylamide (AM) and N-methylolacrylamide (NMA) through the Aza-Michael reaction to obtain the final product. The detailed compositions of HNSPs with weight ratios (Rm) of NMA and Am to TEPA ranging from 1.7 to 2.1 were listed in Table S1 Rm is defined as the molar ratio of NMA (equimolar with AM) to TEPA, with Rm = 2.0 corresponding to an NMA:AM:TEPA ratio of 2:2:1.

The C = O stretching vibration in the amide group at 1652 cm^-1 was observed in Fourier transform infrared (FTIR) spectra and the stretching vibration peak of N-C can be observed at around 1407 cm^-1 and 1125 cm^-1, which belong to the amide group and the adipose group secondary amine, respectively. The results of nuclear magnetic resonance (NMR) manifested that the multi overlapping peaks at 2.80 ppm to 2.17 ppm are attributable to the -CH₂ on the main chains in the 1H-NMR spectroscopy (Fig. S3 and S4), and the characteristic peaks from at 54.2 ppm to 50.33 ppm belong to -N-CH₂- groups in the 13C-NMR spectroscopy (Fig. S5). Additionally, the results of 1H detected heteronuclear multiple bond correlation (HMBC) and heteronuclear single quantum coherence (HSQC) confirmed the correspondence between 1H-NMR and 13C-NMR (Fig. S6). The above results indicated that these HNSPs have been successfully synthesized via the Aza-Michael reaction. To further characterize the elements and groups on HNSPs, we performed X-ray photoelectron spectroscopy (XPS) experiments (Fig. S7). The XPS spectra shows that there is only C, N and O elements on the surface of HNSPs surface, including four types of binding of C in XPS spectrum of C1s corresponding to C-C, C-N, C-OH and C = O. And there are two types of binding of N in HNSPs, the characteristic peaks of 399.04 eV and 400.02 eV correspond to N-C amines and N-C = O, respectively. Meanwhile, two types of binding of O in HNSPs with characteristic peaks of 531.08 eV and 532.09 eV are correspond to O-H and O = N-C group. The XPS data further confirmed that HNSPs have abundant oxygen-containing groups and nitrogen-containing groups on the surface, which helps to provide sufficient hydrogen bonding sites. Above analysis results indicate that HNSPs containing multiple hydrogen bonds have been successfully synthesized.

As illustrated in Fig. 1b and Fig. S8 the hydrogen bond structures and supramolecular organization within hydrogen-bonded nanocluster self-destructive polymers (HNSPs) have been elucidated. The molecular segments of the target HNSPs primarily consisted of relatively flexible TEPA segments, which were covalently linked through main or side chains-aligned perpendicularly relative to TEPA segments-to more rigid components such as NMA and AM. These rigid components offered numerous hydrogen bond donors and acceptors. Furthermore, due to their shorter chain length and lower molecular weight compared with conventional polymers, HNSP molecular segments exhibited reduced entanglement, thus facilitating enhanced molecular mobility. Consequently, this distinctive structural arrangement promotes the formation of high-density hydrogen-bonded nanoclusters characterized by smaller dimensions yet densely populated hydrogen bonds, ultimately resulting in mechanically robust, humidity-responsive self-destructive polymers enabled by hydrogen-bond nanoconfinement.

Simultaneously, all-atom molecular dynamics (MD) simulations were conducted to theoretically elucidate the categories and aggregation structures of HNSPs. As shown in Fig. 1c and Fig. S9, the simulated systems comprise 30 polymer chains contains abundant hydrogen bonding sites and these hydrogen bonds are formed through the N and O groups on the molecular chains. The radial distribution function (RDF) can depict distribution with distance of other particles around a particle, which can reflect the number of hydrogen bonds⁴². The g (r) value of N groups at 0 ~ 1 Å is higher than that of O groups, which reflects the number of hydrogen bonds formed by N is greater than that of O in the HNSPs system with the fraction of H-bonded in N and O being 70% and 30%, respectively. The average bond energy of hydrogen bonds is −7.714 kJ/mol. The simulated HNSPs chain showed dense aggregation with irregular spherical shapes (nanometer scale), inducing the complexation of HNSPs molecules into aggregated nanoclusters through hydrogen-bond interactions^43–45.

As shown in Fig. 1d, Transmission electron microscope (TEM) images reveal densely and uniform hydrogen-bonded nanoclusters with an average radial size of 8 nm. Compared with other structures (sheet or brick), spherical nanoclusters have better fluidity and specific surface area, which provides more hydrogen bonding sites to establish the physical cross-linking structures⁴⁶. Those results are consistent with that of MD simulations, revealing the formation of nanoclusters due to the densely hydrogen bond in HNSPs. Obviously, as the weight ratio Rm increases, the hydrogen-bonded nanoclusters process more pronounced outline and larger diameters (Fig. S10). This result is also in good agreement with the all-atom MD simulations: the more AM and NMA will increase the hydrogen bond density, resulting the severe aggregation of hydrogen bonds that form bigger nanoclusters. Similarly, the cryo-scanning electron microscopy (cryo-SEM) images also show the regular and orderly arrangement of these nanoclusters, which act as nano-armor, conferring high mechanical strength and hardness to HNSPs (Fig. 1e and Fig. S11).

As depicted in Fig. 1f, a HNSP’s sample (50 × 50 × 3 mm³) is capable of easily supporting a 500 g weights without deformation or breaking, demonstrating excellent mechanical properties. Additionally, the thermal properties of HNSPs were also studied through a variety of testing methods. The results of differential scanning calorimetry (DSC) reveal that HNSPs have a glass transition temperature (T_g) around room temperatures (Fig. S12). As Rm increases, the T_g of HNSPs markedly increased, those higher T_g could be attributed to the increased cohesion energy of numerous hydrogen bonds within the HNSP. The thermogravimetric analyses (TGA) results reveal that HNSPs begin to lose their weight over 250 °C, also featuring excellent thermal stability (Fig. S13). Furthermore, X-ray diffraction (XRD) spectra display no sharp crystalline diffraction peaks, indicating that HNSPs possess typical amorphous structures regardless of different Rm values (Fig. S14). The above results indicate that the designed HNSPs have hydrogen bond nanoclusters with spider silk structure.

Mechanical and self-destructive properties of HNSPs

The mechanical properties of HNSPs are heavily depend on the characteristics of the hydrogen-bond clusters. To elucidate the superior mechanical properties of HNSPs with different Rm values, both their original structures and those under deformation were investigated in Fig. 2a. As shown in Fig. 2a, after compression testing, the thickness of the molded sample exhibited a reduction in height to 25% of its original dimension, accompanied by proportional increases in length and width, confirming the inherent ductility and plasticity of HNSPs. The stress-strain curves exhibit three distinct regions: an initial elastic deformation phase where stress significantly increases with strain, a subsequent strain-hardening phase characterized by a gradual increase in strain with a slower rise in stress, and a final densification stage where strain increases slowly while stress rises sharply (Fig. 2b). The yield strength of HNSPs, defined as the stress corresponding to 10% residual deformation, is found to be as high as 14.66 MPa. Furthermore, the Young’s modulus of HNSPs markedly increases with increasing Rm in Fig. 2c. The highest Young’s modulus of the HNSPs up to165.21 MPa is calculated in the low-strain region (<10% strain) of the stress-strain curve. Notably, as Rm increases further, the HNSPs exhibit enhanced strength and stiffness, while this is accompanied by increased brittleness (Fig. S15). This may be attributed to the fact that the strength and stiffness of the HNSPs increase because of higher hydrogen bond density forming hydrogen-bond nanocluster. The specific mechanism responsible for the excellent mechanical properties and notable ductility of HNSPs are attributable to the densely hydrogen-bond nanocluster acted as physical cross-linking points, providing robust resistance to deformation and fracture, and the dynamic nature of these hydrogen bonds enables molecular units to absorb mechanical energy and undergo plastic deformation through relative slippage and rearrangement under stress, thereby preventing brittle fracture.

Fig. 2. Mechanical performance and Self-destructive behavior of the HNSPs.

a Image of the HNSPs upon compressing (The Rm is 2.0.). b Representative stress-strain curves of the HNSPs with different Rm values. c Dependence of elastic modulus on the Rm of the HNSPs (Error bars, mean ± s. d, n = 3). d Zeta potential of the HNSPs with different Rm values (Error bars, mean ± s. d, n = 3). e Schematic illustration of the mechanically enhanced performance of the HNSPs based on effective load transfer, supported by H-bond nanoclusters (F represents the applied force.). f Images of HNSPs with macro self-destructive phenomenon upon moisture stimulation. g Self-destructive rate of the HNSPs with different Rm values (Error bars, mean ± s. d, n = 3). h Self-destructive rate of the HNSPs (The Rm is 2.0.) with different ambient humidity (Error bars, mean ± s. d, n = 3).

As confirmed in Fig. 2d, the results of zeta potential measurements reveal that HNSPs have a negative surface charge, ascribed to the abundant oxygen-containing groups on their surface. Moreover, as Rm increases, the absolute value of the zeta potential gradually decreases, suggesting that the hydrogen-bonded nanoclusters become more closely packed and the interactions between them are strengthen^47,48. The schematic diagram of the hydrogen-bond nanocluster within HNSPs bearing external stress is conceptually depicted in Fig. 2e. It shows that when Rm is low, the hydrogen-bonded clusters are relatively small and widely spaced, rendering them insufficient to bear applied loads effectively⁴⁹. The weak interactions between clusters ensures the load transfer through molecular chains, leading to the destruction of hydrogen-bonded clusters and resulting in lower mechanical performance. In contrast, at higher Rm values, the hydrogen-bonded clusters become larger and more tightly packed, adequately supporting the applied load. The strong interactions between the clusters promote the simultaneous stiffening of multiple hydrogen-bonded clusters, thereby improving excellent mechanical properties. Consequently, by adjusting the Rm value, the properties of the hydrogen-bonded clusters and, in turn, the mechanical properties of HNSPs can be easily controlled.

Due to the abundance of hydrophilic groups in their molecular structure, HNSPs can spontaneously absorb water molecules from the environment. Surprisingly, HNSPs can achieve reversibly transition between two states of matter, shifting from a sturdy and tough solid to a viscoelastic liquid and back. As shown in Fig. 2f a cured HNSP sample was placed on a metal grid and maintained shape stability in a dry environment, demonstrating the system integrity. Upon exposure to a high-humidity environment, the HNSPs will spontaneously enter a self-destructive state, forming viscoelastic liquids that gradually drip and migrate beneath the metal grid over time. To investigate the dynamic mechanical behavior of HNSPs during their self-destructive, we analyzed their rheological properties of HNSPs (Fig. S16). Upon exposure to high humidity, HNSPs exhibited spontaneous self-destructive behavior, reflected in the progressive decrease of their storage modulus (G′) and loss modulus (G″) relative to the original state. This gradual reduction in modulus indicates an increased fluidity of the polymer matrix, attributed directly to moisture-induced weakening of intermolecular hydrogen bonds, which are critical for maintaining the mechanical integrity and stiffness of HNSPs. To obtain deeper insights into the molecular and microstructural evolution occurring during the self-destructive process, wide-angle X-ray scattering (WAXS) was conducted (Fig. S17). Upon exposure to moisture-induced self-destructive, the scattering intensity became significantly weaker and broader, indicating moisture infiltration gradually disrupts intermolecular hydrogen bonded nanoclusters, leading to a loosened and disordered microstructure. The UV-vis spectra showed a characteristic peak at ~370 nm, corresponding to n-π* transitions of C = N and C = O groups (Fig. S18). As self-destructive time increased, the absorption peak intensity progressively decreased, indicating dispersion of HNSP molecular fragments into the aqueous phase, leading to a reduction in the concentration of chromophoric groups (C = N and C = O). Furthermore, to achieve secondary recycling under special applications, the self-destructive system can recover its initial state through high-temperature drying, as evidenced by macroscopic shape transformation and the restoration to a sturdy solid (Fig. S19).

Additionally, further experiments were conducted to assess the self-destructive rate of HNSPs, defined as the mass decreases speed of HNSPs (passing through a specified aperture sieve) over time and can be expressed to the amount of mass loss per unit time (g/h). As shown in Fig. 2g, with increasing Rm at a relative humidity of 80%, the self-destructive rate gradually declined and ultimately stabilized within a narrow range. The difference in self-destructive rates between the fastest (Rm = 1.7) and slowest (Rm = 2.1) systems is 1.69 times. When exposed to varying external humidity levels (Fig. 2h), HNSPs (Rm = 2.0) exhibits more significant changes, achieving a self-destructive efficiency of 804.27% at 90% RH compared to 60% RH. Therefore, by adjusting both humidity and Rm, we can effectively control the self-destructive rate of HNSPs over a broad range.

Based on the controlled self-destructive rate of HNSPs under moisture, the experimental model of temporary buildings, self-destructive Quick Response (QR) code and transience devices were designed to elaborate the significance and practical benefits (Fig. S20). A model of temporary bridge can sustain a standard weight of 100 g yet leads to the structural destruction of itself induced by moistures under 75% RH at 25 °C for 180 min (Fig. S20a). Intriguingly, a self-destructive QR code consisted of the black and white squares that arranged according to specific coding rules gradually become unrecognizable liquid mixtures due to the self-destructive behavior of HNSPs under moisture, realizing the self-erasure of information and the protection of information leakage (Fig. S20b). The moisture-triggered self-destructive transience devices that were constructed by mixing HNSPs with Ni powder, can be easily connected in series to the circuit and light up the small light bulbs, but ultimately became a diluted liquid and Ni powder after absorbing moisture, effectively reduce the pollution of electronic waste (Fig. S20c).

Hygroscopicity kinetics of HNSPs

To elucidate the mechanism of reversible transformation between the drying stability and high humidity-induced self-destructive, the hygroscopicity kinetics of HNSPs were further investigated^50–52. as shown in Fig. 3a. The adsorption equilibrium points of HNSPs with different Rm values at 25 °C is presented. At lower humidity (RH < 40%), the weight of HNSPs shows minimal change, indicating limited water adsorption. As humidity increases, the water absorption capability of HNSPs progressively enhanced, with a maximum weight increase of 68.65% obsessed at 90% RH. Moreover, as Rm increases, the water absorption capability of HNSPs decreases gradually, with a 12.34% difference in the highest and lowest water adsorption values. During the desorption process, HNSPs gradually release water molecules, resulting in a decrease in total mass as humidity decreases. The residual mass of HNSPs with different Rm values shows minor variations, generally ranging between 16% and 19%. Fig. 3c illustrates the mass change of HNSPs (Rm = 2.0) throughout the adsorption-desorption cycle. The absorption rate significantly exceeds the desorption rate, and the mass cannot revert to the initial state of HNSPs. Table S2 details the mass changes of HNSPs in the adsorption-desorption cycle (The RH range from 0% to 90%), showing the phenomenon of desorption hysteresis. This hysteresis phenomenon may result from the formation of strong intermolecular force between water molecules and HNSP molecules, hindering water desorption. HNSPs exhibit a sensitive response to humidity changes, with notable differences in the equilibrium times of the two states (Fig. 3d). The equilibrium time of adsorption is longer than that of desorption, likely attributed to the process of breaking the existing hydrogen bonds between HNSPs molecular chains and then forming new hydrogen bonds during adsorption.

Fig. 3. Influence of hygroscopicity kinetics on self-destructive performance for HNSPs.

a Adsorption isotherms and b Desorption isotherms of HNSPs with different Rm. c Adsorption and desorption cycle isotherms of HNSPs (The Rm is 2.0.). d The mass change of HNSPs during adsorption and desorption cycle. e LF-NMR spectra during adsorption and desorption process. f The compositions of water in the adsorption process. g The composition of water in the desorption process. h, i In-situ FT-IR spectra of HNSPs under same RH conditions.

Additionally, the moisture diffusion behavior of HNSPs during the self-destructive process was conducted through Two-dimensional concentration diffusion simulation based on lattice Boltzmann method (LBM) (Fig. S21). At low humidity, the internal concentration of HNSPs remained nearly constant, with slightly lower concentration at boundaries. Increasing humidity elevated boundary concentrations, which then diffused inward, reaching equilibrium. Under constant high humidity (90% RH), initially, only boundaries exhibited high concentration, with low internal concentration. Over time, moisture diffused inward, leading most regions to equilibrium except a small central area. Simulation results closely matched experimental data, confirming Fickian diffusion behavior.

Moreover, to further elucidate above mechanism, low-field nuclear magnetic resonance (LF-NMR) was used to characterize the state changes of water during the adsorption-desorption process, distinguishing free water, immobilized water, and bound water by measuring proton relaxation times (Fig. 3e). Initially, HNSPs contained only trace amounts of immobilized water, with no detectable bound or free water, due to spontaneous moisture absorption during testing and sampling process. Upon moisture adsorption, the total water absorption mass of HNSPs increases, and the internal structural composition of the moisture changes. As depicted in Fig. 3f, when the proportion of bound water reaches 52%, immobilized water accounts for 43%, and free water constitutes a minor fraction of 5%. Over time, as the water absorption mass of HNSPs continues to increase (Fig. 3f), the proportion of bound water content further rises to 63%, while immobilized water decreases to 31%, and the proportion of free water remains nearly unchanged. This phenomenon suggests that immobilized water transforms into bound water during the adsorption process, indicating that HNSPs initially forms weak interactions with water molecules (immobilized water). As time progresses, more water molecules infiltrate between HNSP molecular chains, disrupting existing interactions and forming stronger hydrogen bonding (bound water). Subsequently, samples subjected to room temperature vacuum treatment (to more rigorously simulate desorption at low humidity) exhibit a significant reduction in total moisture absorption mass, accompanied by a marked decrease in water molecules across all three states (Fig. 3g). Additionally, the composition of water molecules in HNSPs undergoes significant changes, with an increase in the proportion of bound water and a sharp decrease in the proportion of immobilized and free water. This indicates that HNSPs preferentially remove free water and weaker immobilized water during the desorption process, due to the weak hydrogen bonding forces between some water molecules and HNSPs molecules. In contrast, bound water is difficult to remove, attributed to the presence of more abundant and stronger hydrogen bonds between certain water molecules and HNSP molecules, preventing water detachment during the desorption process⁵³.

As shown in Figs. 3h, i, the interactions between water molecules and the hydrogen-bond clusters of HNSPs are also investigated. In-situ Fourier-transform infrared (FT-IR) spectroscopy under consistent RH conditions is used to examine changes in hydrogen bonds during water adsorption. In high-humidity environments (RH = 95%), the infrared absorption peak corresponding to the N-H stretching vibration shifts from 3370 cm^-1 to 3430 cm^-1 over time, indicating that the pre-existing hydrogen-bonded state of the amino groups gradually convert to free amino groups. Additionally, as shown in Fig. 3i, the shift of the 1410 cm^-1 peaks to a lower wavenumber further indicates the gradual dissociation of the stretching vibration peak of N-C, which belong to the amide group. These results reveal that exposure of HNSPs to high humidity leads to the breaking of hydrogen bonds within HNSPs by water molecules.

Self-destructive mechanism of HNSPs

The hydrogen-bond nanoclusters in HNSPs are the primary contributors to their exceptional mechanical properties, formed through dense hydrogen bonds. To better reveal elucidate the inherent driving forces and mechanisms behind the self-destructive behavior of HNSPs, molecular dynamics (MD) stimulations using the all-atom OPLS force field is conducted to model the kinetic behavior of hydrogen-bond nanocluster structures under varying environmental conditions^54,55. As shown in Fig. 4a, molecular segments autonomously aggregated under dry conditions, resulting in the formation of highly aggregated and overlapping hydrogen bond nanocluster structures. Surprisingly, when HNSPs are exposed to prolonged high humidity, water molecules will naturally infiltrate between the molecular chains, causing the hydrogen-bond clusters to relax. This infiltration can even lead to the detachment of individual molecular chains detaching from the cluster aggregates, causing them to disperse into the surrounding aqueous environment.

Fig. 4. The self-destructive mechanism of HNSPs.

a The molecular structural changes of HNSPs in their initial and self-destructive states. b Total radius of gyration of HNSPs molecules in initial and self-destructive state from MD simulations. c Hydrogen number of HNSPs molecules in initial and self-destructive state from MD simulations. d Hydrogen number of HNSPs of water molecules. e The lifetime of hydrogen bond in HNSPs molecules in initial and self-destructive state. f Self-destructive mechanism of HNSPs.

To further investigate the self-destructive mechanisms at molecular scale, the dynamics of individual molecular segments were analyzed during the transition from the initial state to the self-destructive state. The mean-square radius of gyration (R_g) and end-to-end distance (Å) are used to quantify the degree of molecular curling and the rigidity or flexibility of the molecular chains. As depicted in Fig. 4b, the R_g and the end-to-end distance of HNSPs increase by 123.53% and 170.82%, respectively, in the self-destruct state compared to in the initial state. As shown in Fig. S22 and Fig. S23, The longest end-to-end distance in the initial state is 15.46 Å, while it increases to 24.10 Å in the self-destructive state, which is 56% higher than that in the initial state. This suggests that the molecular segments are becoming more and more loose, indicating a major alteration in molecular configuration, with a stretching behavior in the self-destructive state and a curling behavior in the initial state. Notably, the change in hydrogen bonds is a key factor responsible for the pronounced differences between these two states. In dry environment, the hydrogen bonds between HNSP chains are tightly bound, maintaining a high number of hydrogen bonds. However, under high humidity conditions, the number of hydrogen bonds within the clusters sharply decreases, leading to a loose cluster structure (Fig. 4c). These hydroxyl and amide groups in HNSPs, released from the clusters, will further interact with the water molecules to form new hydrogen-bond configurations, significantly increasing the number of hydrogen bonds in the water molecules (Fig. 4d). Concurrently, the lifetime of hydrogen bonds decreases when transitioning from the initial state to the self-destructive state (Fig. 4e).

The self-destructive mechanism for HNSPs was proposed to reveal how microscopic changes strongly influence their macroscopic physical properties (Fig. 4f). In the initial state, polymer chains are firmly interlocked through high-density hydrogen bonds, forming large and ordered hydrogen bond clusters that can endow HNSPs with excellent mechanical properties. The molecular chains are curled, possessing high Gibbs free energy, which compromises the system’s stability⁵⁶. However, environmental changes that trigger the self-destructive mode allow water molecules to infiltrate the cluster aggregates, breaking the internal high-density hydrogen bonds. As water molecules peel molecular segments away, new hydrogen bonds form, leading to the dissociation of hydrogen-bond clusters. The dissociative molecular chain stretch, resulting in a lower Gibbs free energy (ΔG = −41.837 KJ mol^-1 compared to the initial state) and an increase in system stability. According to the trend of decreasing Gibbs free energy and increasing entropy, HNSPs will spontaneously transition from an initial state to the self-destructive state upon exposure to water molecules. This transition is manifested as a macroscopic self-destructive phenomenon, characterized by a reduction in mechanical properties and a change in the state of matter.

Discussion

In summary, we have developed an efficient strategy for fabricating hydrogen-bond nanocluster self-destructive polymers (HNSPs) that seamlessly integrate high mechanical performance with controllable self-destructive behavior. The devised HNSPs can overcome the limitations of existing self-destructive material technologies while maintain exceptional mechanical properties. Impressively, these HNSPs have a high compress strength of 14.66 MPa and an elastic modulus of 165.21 MPa. The key to this strategy lies in three key factors: First, the formation of high-density hydrogen bonds facilitates the creation of ordered hydrogen-bonded nanoclusters, which can confer superior mechanical properties. Second, the incorporation of short molecular segments effectively reduces entanglement between molecular chains, increasing molecular chains and enhancing the dominant role of hydrogen bonds within the system. Third, the abundance of enriched hydrophilic groups enables efficient moisture absorption from the environment, allowing the formation of new hydrogen bonds with water molecules, thereby disrupting the original equilibrium and leading to the dissociation of hydrogen-bond clusters. Our work provides an efficient strategy for building self-destructive materials with broadened practical application, positioning the devised HNSPs as ideal candidates for temporary structures, self-destructive drones, and other applications that require both load-bearing capacity and stability.

Methods

Materials

Tetraethylenepentamine (TEPA, C₈H₂₃N₅, Mw = 189.30, technical grade purity, CAS No. 112-57-2), Acrylamide (AM, C₃H₅NO, Mw = 71.08, 99.0% purity, CAS No. 79-06-1), N-Methylolacrylamide (NMA, C₄H₇NO_2, Mw = 101.1, 98% purity, CAS No. 924-42-5), Deuterated dimethyl sulfoxide (DMSO-d_6, 99.8% purity) and Deuterium oxide (D₂O, ≥99 atom% D) were purchased from Aladdin (Shanghai) Chemistry Co., Ltd. Deionized water (conductivity <0.5 us/cm) was bought from the Chengdu Kelong chemical plant (China). Ni powder purchased from Brofos Nanotechnology Co. Ltd. All reagents were used as received without further purification unless otherwise noted.

Synthesis of HNSPs

Taking the molar weight ratio Rm = 2.0 as an example, the addition amount of NMA and AM were 20.22 g (0.2 mol) and 14.22 g (0.2 mol), respectively. The addition amount of TEPA was 18.93 g (0.1 mol). Deionized water was used as the solvent with an addition amount of 22.5 g, and the concentration of all synthesized samples were controlled at 70%. NMA and AM were added to the three-neck round bottom flask and dissolved in Deionized water. Consequently, TEPA was evenly dripped into flask and further kept at 25 °C for 2 h under stirring. The obtained mixture was heated at 40 °C and stirring for 3 h and then heated to 65 °C for 4 h. The as-obtained viscous polymer solution was next poured into homemade silicone mold and then heated on a hot plate at 80 °C and finally placed in a vacuum oven at 100 °C for12 h to remove the deionized water to obtain the final materials.

Characterization

General Characterization

Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy (Nicolet is50, Thermo fisher Scientific) with the scanning range from 4000−400 cm^-1.

X-ray photoelectron spectroscopy (XPS PHI 5000 Versa Probe III) can effectively detect the elemental composition and chemical valence state.

Proton-1 and carbon-13 nuclear magnetic resonance (1H-NMR) (AV III HD 400 MHz) were used to determine the chemical composition of the obtained materials The ideal molecular structure HNSPs measured in DMSO-d6 and D₂O.HMBC (1H detected heteronuclear multiple bond correlation) and HSQC (Heteronuclear Single Quantum Coherence) were measured in DMSO-d6.

X-ray diffraction measurements were carried out on an Ultima IV X-Ray diffractometer (Rigaku) with Cu Kα radiation (λ = 0.15418 nm) at the accelerating voltage of 35 KV and emission current of 30 mA at a scanning range of 5°−55° and a scanning rate of 5 ° min^-1.

Differential scanning calorimetry (DSC, Netsch DSC 204 F) was used to measure the glass transition temperature. The whole test procedure is that the temperature increased from 25 °C to 80 °C, followed by isothermal at 120 °C for 3 min, and then cooled down to −50 °C while maintaining at −50 °C for 3 min. Finally, the sample was heated to 120 °C again. Both the heating rate and cooling rate were 10 °C·min⁻¹. The nitrogen flow rate is 50 mL·min⁻¹.

Thermogravimetric analyses (TGA) were performed on the SDT Q600 instrument in the temperature range from 25 °C to 600 °C at the heating rate of 10 °C min^-1 in a nitrogen atmosphere.

The compress tests were implemented at room temperature using a universal testing machine (Instron 5966, 10 KN) with 5 mm min^-1 rate. The mechanical tests were performed at a strain rate of 5 mm min^-1 at room temperature with a humidity of 40−50%. The specimens were model into dimensions of 25 mm × 10 mm × 10 mm. The result was obtained by averaging five specimens.

Cryogenic scanning electron microscopy (Cryo-SEM) experiments were performed on a ThermoFisherScientific (FEI)-Apreo S HiVoc scanning microscope (FEI Company, US) equipped with a Quorum cryo-stage PP3010T. The samples for observation were prepared by a freeze-drying method: samples were first frozen in liquid nitrogen (−185 °C) for 30 s, and then were transferred to a chamber to sublimate the moisture under vacuum at −90 °C for 10 min. Subsequently, the samples do not need to be coated (like Au) and then observed at-140 °C at an acceleration voltage of 5 kV. The reason why the samples do not need to be coated is that the size of the gold nanoparticles is much larger than the observed microstructure, which can affect the observation of the sample. Due to the lack of gold coating, there may be a decrease in image contrast and an increase in signal noise during sample observation.

Zeta potential measurements were conducted using a Malvern Zetasizer Nano ZSE instrument by detecting back-scattered light at an angle of 173°, equipped with a 4 mW He-Ne laser operating at 633 nm. Aqueous dispersions of the copolymer nanoparticles were diluted to 0.10 % w/v using deionized water. Aqueous electrophoresis measurements were conducted using disposable folded capillary cells supplied by Malvern (DTS1070) using the same instrument.

Transmission Electron Microscope (TEM) images were obtained by depositing droplets of 0.5% w/v aqueous dispersion of HNSPs onto carbon-coated cooper grids using a JEOL/JEM-1400Flash instrument at an operating voltage of 120 kV.

Rheological measurement was carried out using a rheometer (Thermo Fisher, Mars 60) with a parallel plate geometry (20 mm diameter rotating top plate). storage modulus (G′) and loss modulus (G″) and temperature (80°C) with an increase in frequency (from 0.1−100 rad s^-1) were obtained.

The wide-angle X-ray scattering (WAXS) measurements were performed on a synchronous diffraction laboratory instrument, Nano-inXider (Xenocs, France), equipped with a micro-focus source generating X-rays of a wavelength λ = 1.542 Å (Genix3D), operating at 50 kV and 0.6 mA.

The UV-vis spectra of HNSPs were observed on a UV-3600 Spectrophotometer (Shimadzu).

The water vapor sorption is tested with a dynamic vapor sorption (DVS) system under constant humidified airflow. The sorption-desorption performance was measured by a DVS system (Surface Measurement Systems LTD, DVS Adventure). The airflow of DVS is fixed at 200 mL min^-1 for all samples as the standard setting. The sorption-desorption performance was measured by a DVS system (Surface Measurement Systems LTD, DVS Adventure). The test samples are preheated at 105 °C, 0% RH for 5 hours in the testing chamber for completed drying, and then placed at 25 °C for 2 hours stabilization. The mass of test HNSPs sample was about 30 mg, and the adsorption-desorption performance was measured at 25 °C with the humidity range from 0% to 90% and then to 0%, conducted at a humidity rate of 10% as a step. When the weight of the sample is finally stable at a certain humidity (the change in the mass of the sample within 10 min is less than 0.01% of the absolute dry mass), the sample is considered to reach a moisture content equilibrium state at that humidity, then the testing program adjusts RH adjusted to the next level for repeated testing.

Low-field ¹H nuclear magnetic resonance (NMR) was used to reflect the existence and distribution of water molecules. By measuring the relaxation time to elucidating the interaction between reflect structural heterogeneity and mobility of molecules. In general, the relaxation time of free water is long, while that of bound water is short. LF-NMR spectra were measured on a VTMR20-010V-I NMR analyzer (Su-zhou Niumag Analytical Instrument Corporation, Suzhou, China).

In-situ FTIR: The Fourier transform infrared spectra were collected on a Thermo Fisher Nicolet iS20 spectrometer equipped with a glow bar source, KBr beamsplitter and a liquid-nitrogen-cooled mercury cadmium telluride detector. Furthermore, all the spectra were collected in the ATR mode on a self-supported wafer of the sample using a custom-built infrared cell. The sample atmosphere was controlled by a mass flow controller and a pressure gauge.

Self-destructive efficiency test

The parameters of HNSPs used for self-destructive tests are as follows: diameter ≈ 20 mm, thickness ≈ 1 mm, and mass ≈ 0.4–0.8 g. The samples were placed on a stainless-steel mesh (aperture ≈ 3.8 mm, wire diameter ≈ 0.5 mm) and placed in a Temperature-Humidity Test Chamber (DHT-500, computer-controlled). By adjusting the humidity of the testing environment through Temperature-Humidity Test Chamber, the shape and weight decay of the sample with humidity changes were record. And we observe the remaining HNSPs samples on the stainless-steel sieve, and determine the self-destructive efficiency of HNSPs (with different Rm) at different humidity levels.

Here, we have defined that the self-destructive rate of HNSPs as follows: The speed that the decreases mass of the test sample (passing through a specified aperture sieve) over time, in which the standard sample perform the programmed self- destruction behavior triggered by moistures under the same temperature and humidity conditions. It can demonstrate the self-destructive speed of materials under the environmental conditions, which can express in terms of the amount of mass loss per unit time (g/h) or the percentage of mass loss (%/h). The calculation formula is$X=\frac{M_I-M_c}{T}$ (X: Self-destructive speed; M_I: Initial mass (g); M_c: Current mass (g); T: Time (hours)). Considering the hygroscopicity of HNSPs, we have deducted the additional hygroscopic mass generated by the hygroscopic process to obtain the current mass of the material (M_c) (based on the DVS experimental data or the percentage of overall device hygroscopicity mass). Notably, the above results are the results of testing the relative self-destructive rates of different Rm, which may be affected by the size of the aperture line diameter.

Molecular dynamics (MD) simulations

Models and computational details

In this work, a standard molecular mechanics’ potential model was used with the following functional form:

$$u\left(r^N\right)={\sum }_{bonds}\frac{k_i}{2}{\left(l_i-l_{i,0}\right)}^2+{\sum }_{angles}\frac{k_i}{2}{\left({\theta }_i-{\theta }_{i,0}\right)}^2+{\sum }_{torsions}\frac{V_n}{2}\left(1+\cos \left(n\omega -\gamma \right)\right)$$

$$+{\sum }_{i=1}^N{\sum }_{j=i+1}^N\left(4{\epsilon }_{ij}\left[{\left(\frac{{\sigma }_{ij}}{r_{ij}}\right)}^{12}-{\left(\frac{{\sigma }_{ij}}{r_{ij}}\right)}^6\right]+\frac{q_i{q}_j}{r_{ij}}\right)$$

where the first three terms are the bonded interactions, including bond, angle, and torsion interactions, and the second terms are nonbonded interactions, including van der Waals and Coulombic interactions. For different kinds of atoms, the Lorentz-Berthelot mix rules were adopted for van der Waals interactions, which is following the equation:

$${\sigma }_{ij}=\frac{1}{2}\left({\sigma }_{ii}+{\sigma }_{jj}\right){;\epsilon }_{ij}={\left({\epsilon }_{ii}*{\epsilon }_{jj}\right)}^{1/2}$$

According to the experimental conditions, two systems were constructed. In the system, 30 molecules were placed randomly in a cubic box with dimensions of 6 × 6 × 6 nm³. The simulations were performed using GROMACS package (version 2023.3)^57–60 with the all-atom OPLS force field⁶¹. The steep descent method was used to minimize the energy of the system. Then, molecular dynamics simulations under NVT ensemble at 298 K were performed for 20 ns for the first system. After the simulation, all 30 molecules have aggregated into a cluster. Water molecules were filled into the box to build the initial configuration of the second system. Then, molecular dynamics simulations under NPT ensemble were performed for 20 ns for the second system. LINCS algorithm⁶² was applied to constrain the bond lengths of other components. Periodic boundary conditions were applied in all three directions. The temperature was maintained using the V-rescale thermostat algorithm⁶³. The cut-off distance for the Lennard-Jones and electrostatic interactions was 1.2 nm. Particle mesh Ewald method was used to calculate the long-range electrostatic interactions⁶⁴. Configurations were visualized using Visual Molecular Dynamics software⁶⁵.

Two-dimensional concentration diffusion simulation based on lattice Boltzmann method (LBM)

Background

Diffusion phenomena are widely observed in both natural and engineering contexts, such as groundwater solute migration, geotechnical contamination dispersion, and mass transfer in chemical reactions. Fundamentally, concentration diffusion is a mass transport process driven by the random thermal motion of molecules, typically described by Fick’s second law:

$$\partial \mathrm{C}/\partial \mathrm{t}={\mathrm{D}\nabla }^2\mathrm{C}$$

where C (x, y, t) represents the solute concentration at a given time and position, D is the diffusion coefficient, and ∇² is the Laplacian operator in two dimensions.

Traditional numerical methods such as the Finite Difference Method (FDM) and Finite Element Method (FEM) can encounter limitations when dealing with complex boundaries or coupled physical processes. The Lattice Boltzmann Method (LBM), a mesoscopic numerical approach based on kinetic theory, simulates the propagation and collision of particles on a discrete lattice to solve the macroscopic diffusion equation indirectly. It offers several advantages:

• High numerical stability;

• Well-suited for parallel computing;

• Efficient handling of complex boundaries and multi-physics coupling;

• Naturally accommodates anisotropy and spatial heterogeneity due to its inherent structure.

Modeling principle of lattice Boltzmann method

The D2Q9 model utilizes nine discrete velocity directions: one rest particle, four along the Cartesian axes, and four diagonals. The evolution equation is given by:

$${\mathrm{f}}_{\mathrm{i}}\left(\mathrm{x}+\mathrm{e},i\Delta t,t+\Delta t\right)={\mathrm{f}}_{\mathrm{i}}\left(x,t\right)-\left(1/\mathrm{\tau }\right)\left({\mathrm{f}}_{\mathrm{i}}-{\mathrm{f}}_{\mathrm{i}}\mathrm{eq}\right)$$

Where, f_ieq = w_i $*$ C. The macroscopic concentration C = ∑f_i

The relationship between the diffusion coefficient D and the relaxation time τ is as follows⁶⁶:

$$\mathrm{D}={\mathrm{c}}_{\mathrm{s}}^2\left(\mathrm{\tau }-0.5\right)\mathrm{\Delta }\mathrm{t},{\mathrm{c}}_{\mathrm{s}}=1/\sqrt{3}$$

Numerical implementation process

The model was implemented in MATLAB. The main computational steps are:

1. Initialization of the concentration field and distribution functions

2. Collision computation

3. Streaming (propagation) step

4. Application of boundary conditions

5. Calculation of macroscopic concentration

6. Iteration over time steps until the simulation ends

Boundary conditions and initial conditions

A 150 × 150 square domain was established, with each grid cell representing 40 micrometers. The simulated scenario is an absorption process under 90% relative humidity. Initially, all boundary nodes are set to a high concentration corresponding to equilibrium at 90% RH (157%), while the internal nodes are initialized to a low concentration corresponding to 80% RH (137%).

Author contributions

H.X. performed the most materials synthesis and characterizations; T. Y. helped in hygroscopicity kinetics characterizations and data analysis; Y. L. and X. F. helped in mechanical performance characterizations and data analysis; Z. L. helped in molecular dynamics stimulations and data analysis; S. L. helped in microscopic morphology and thermodynamic performance tests; L. J. and J. L. developed the concepts behind this research and supervised this work. Manuscript was prepared by all the authors.

Peer review

Peer review information

Nature Communications thanks Masanobu Naito and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

Additional data are available from the corresponding author upon request. The data generated in this study are provided in the Supplementary Information/Source Data file. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-66044-9.