Boron nitride nanotube fibers with over 1 GPa strength and 1000 °C stability for exceptional neutron shielding textiles

Abstract

Integrating boron nitride nanotubes (BNNTs) into robust macroscopic architectures for deep space radiation shielding remains a fundamental challenge. Here, we report continuous self-assembled all-nanotube (SAAN) fibers with an 80 wt% BNNT loading, combining tensile strength exceeding 1 GPa with extreme thermal stability from −196 to 1200 °C. These fibers originate from the thermodynamic co-assembly of BNNTs and carbon nanotubes (CNTs) into a lyotropic liquid crystalline mesophase. This shear-responsive dope enables cooperative reorientation and dense nanotube packing during wet-spinning. When assembled into macroscopic textiles, SAAN exhibits an exceptional thermal neutron attenuation coefficient of 2.85 mm⁻¹. Furthermore, computational simulations indicate these lightweight textiles can effectively mitigate highly penetrating secondary thermal neutrons generated by Galactic Cosmic Ray (GCR) spallation. Demonstrating their seamless integration into commercial 2D flat and 3D tubular textile manufacturing processes, this nanomanufacturing approach provides a scalable and reliable strategy for space radiation shielding in both intravehicular (IVA) and extravehicular (EVA) activities.

Researchers developed continuous, flexible boron nitride nanotube fibers achieving over 1 GPa strength and 1,000 °C stability. Interwoven into lightweight textiles, they provide thermal neutron shielding for potential manned deep space missions.

Introduction

As human space exploration extends beyond Low Earth Orbit (LEO) toward long-duration missions to the Moon and Mars, the demand for advanced materials capable of safeguarding life and equipment under extreme radiation conditions has never been greater¹. In deep space, exposure to Galactic Cosmic Rays (GCRs) and triggers spallation reactions upon interacting with surrounding matter, including spacecraft structures, extraterrestrial habitats, and planetary regolith, generating highly penetrating secondary neutrons². These uncharged neutrons bypass conventional shielding, causing severe biological damage and electronic soft errors, posing a critical threat to mission safety³. To mitigate this hazard, boron nitride nanotubes (BNNTs) have emerged as the ultimate candidate for space radiation shielding⁴. BNNTs possess a high boron content that provides a large thermal neutron absorption cross-section, while simultaneously offering thermal stability and intrinsic mechanical resilience exceeding those of conventional hydrogen-rich moderators⁵.

Despite their phenomenal nanoscale properties, translating individual BNNTs into robust, macroscopic shielding architectures has remained a formidable materials science challenge⁶. The practical application of BNNTs has been significantly hindered by their severe processing difficulties; pristine BNNTs are notoriously difficult to disperse and lack sufficient structural entanglement required to form cohesive macroscopic assemblies⁷. In the harsh environment of space, where materials must endure hypervelocity micrometeoroid impacts and extreme thermal cycling, shielding materials must inherently possess robust structural integrity⁸. Therefore, discovering a continuous, scalable nanomanufacturing strategy that preserves both the high boron density and the mechanical robustness of BNNTs is essential for next-generation space materials ⁹.

Historically, the solution spinning of lyotropic liquid crystalline (LLC) dispersions has been the foundational paradigm for mass-producing high-performance fibers¹⁰. Ranging from commercial rigid-rod polymers like aromatic polyamide and poly-p-phenylene benzobisoxazole to advanced low-dimensional nanomaterials like carbon nanotubes (CNTs) and graphene oxide, the formation of a thermodynamically stable LLC phase at critical concentrations has been instrumental in constructing highly oriented macroscopic structures^11–14. Building upon the work of Davis, Smalley, Pasquali, and co-workers, who demonstrated that chlorosulfonic acid (CA) acts as a thermodynamic solvent for one-dimensional (1D) nanomaterials, we propose a nanomanufacturing strategy that extends this approach to a hybrid system of chemically dissimilar nanotubes¹⁵. Herein, we report the scalable fabrication of self-assembled all-nanotube (SAAN) fibers via the bottom-up co-assembly and shear-driven uniaxial orientation of BNNTs with CNTs. In our approach, CA simultaneously protonates the nitrogen lone pairs of the BNNTs and the delocalized π electrons of the CNTs, inducing strong mutual electrostatic repulsion that leads to the spontaneous disentanglement of the nanotubes¹⁶. By minimizing excluded volume effects, this prevents self-agglomeration and promotes a homogeneous LLC phase ¹⁷.

The resulting thermodynamically stable LLC mesophase yields a highly shear-responsive dope, facilitating the reorientation of 1D nanomaterials during flow¹⁸. Subsequent wet-spinning facilitates the uniaxial alignment and densification of the anisotropic nanotubes, resulting in continuous macroscopic SAAN fibers with a high BNNT loading of 80 wt%. This tailored nanoscale architecture overcomes the inherent brittleness of pristine BNNTs, yielding fibers that combine a tensile strength exceeding 1 GPa with structural flexibility. When assembled into woven textiles, the SAAN architecture achieves an experimental thermal neutron attenuation coefficient (μ) of 2.85 mm⁻¹ with thermal stability across a wide temperature range from −196  to 1200 °C under an inert atmosphere. Furthermore, computational simulations indicate that these lightweight textiles could serve as a potential protective component against highly penetrating secondary thermal neutrons generated by Galactic Cosmic Ray (GCR) spallation. Such protection is of critical importance for manned space missions, encompassing both intravehicular activity (IVA) and extravehicular activity (EVA) scenarios where secondary radiation hazards are prominent. Moreover, the mechanical robustness of the continuous SAAN fibers enables continuous spinning into lengths exceeding 3.5 km. This robust structural integrity ensures direct compatibility with standard industrial textile machinery, readily yielding complex 2D flat and 3D tubular textile architectures. By successfully bridging the gap between nanoscale 1D co-assembly and macroscopic processability, this approach provides a foundational strategy for developing advanced, multifunctional radiation shielding textiles targeting next-generation space exploration.

Results

To mitigate the critical risks posed by space radiation during manned deep space missions, this study aimed to fabricate macroscopic 2D textiles specifically engineered for thermal neutron protection, mechanical robustness, and heat resistance. When considering various boron-based materials, such as 0D boron carbide nanoparticles and 2D boron nitride nanosheets, we posited that the 1D cylindrical geometry of BNNTs is intrinsically the most advantageous for constructing macroscopic 1D fibers (Fig. 1a)¹⁹. This naturally prompts a critical question: why not fabricate these textiles using 100% BNNTs? While it is theoretically anticipated that fibers with a maximized BNNT concentration would provide superior thermal neutron shielding due to the high boron content along their 1D hexagonal lattice, their practical macroscopic assembly remains fundamentally challenged by structural brittleness ²⁰.

Fig. 1. Realization of continuous macroscale BNNT fibers.

a Multiscale representation of BNNTs, showing a photograph of BNNTs (left), a SEM image (middle), and a schematic illustration of a tubular BNNT structure with a hexagonal BN lattice (right). The SEM image is representative of two independently prepared samples with similar morphologies. The schematic rendering was created using Blender v4.4 and PowerPoint v16. b Wet-spinning of a SAAN₁₀₀ solution composed solely of BNNTs. c Limited spinnability and low yield of SAAN₁₀₀ fibers. d Long and flexible SAAN₈₀ fibers with enhanced processability for textile manufacturing. e Scalable production of SAAN₈₀ fibers for space radiation shielding applications.

Here, we adopted CA as a thermodynamic solvent to overcome the strong van der Waals interactions between the 1D nanomaterials²¹. BNNTs and CNTs were co-dispersed in defined ratios within CA, yielding homogeneous solutions referred to as SAAN_n, where n represents the BNNT weight percentage (100, 80, 60, 40, and 20 wt%) (Fig. S1). The SAAN₁₀₀ solution forms a nematic (N) phase at 10 mg mL⁻¹, as evidenced by the thread-like Schlieren textures typical of LLCs (Fig. S2)²². As confirmed by polarized optical microscopy (POM), the SAAN₁₀₀ solution depolarized plane-polarized light and displayed characteristic opalescence at room temperature, indicating optical anisotropy and the absence of undissolved BNNT solids, which confirmed its fundamental spinnability (Fig. 1b). Since this LLC phase combines liquid-like fluidity with solid-like structural anisotropy, the mobile yet structured domains can align along the flow direction²³. However, while extrudable, the resulting SAAN₁₀₀ fibers are typically only a few centimeters long (Fig. 1c). They cannot be continuously wound onto a bobbin due to their inherently low mechanical behaviors (Fig. S3). Additionally, SAAN₁₀₀ fibers are highly brittle, resulting in limited flexibility (Fig. S4). These macroscopic shortcomings significantly limit their scalability via an industrially viable spinning process and prevent their integration into woven textiles. It is important to note that practical textile fabrication strictly requires two essential fiber attributes: macroscopic continuity and high pliability, enabling the fibers to be interlaced into the complex geometries of textiles.

By co-assembling BNNTs with 20 wt% of CNTs, we successfully addressed the aforementioned challenges of continuous spinnability, textile processability, and mass production, while preserving the high boron content essential for thermal neutron shielding. The SAAN₈₀ solution at 10 mg mL⁻¹ appeared uniformly dark with no macroscopic agglomerates visible to the naked eye, indicating the effective co-dispersion of both nanotube components in CA. Moreover, the SAAN₈₀ solutions also exhibited LLC phase behavior at this concentration. These observations confirm the successful suppression of self-agglomeration among BNNTs, CNTs, and their hybrids²⁴. Similarly, the SAAN₆₀, SAAN₄₀, and SAAN₂₀ solutions displayed strong birefringence under POM, verifying the formation of LLC phases across varying compositions (Fig. S5). The rheological measurements revealed that all SAAN_n solutions exhibited a shear-thinning behavior as a function of shear rate (Fig. S6). It indicates that the internal N domains align under shear stress, a fundamental characteristic of a homogeneous LLC mesophase rather than a random physical suspension²⁵. When a unidirectional shear force was directly applied to the SAAN₈₀, SAAN₆₀, SAAN₄₀, and SAAN₂₀ dopes, the sheared solutions on a glass substrate exhibited maximum optical birefringence when oriented at 45° relative to the crossed polarizer and analyzer, as expected for an N phase (Fig. S7)²⁶. Conversely, when the shear direction was aligned at 0°, no birefringence was detected. This confirms the anisotropic alignment of the 1D nanomaterials, demonstrating that the N director (n) is parallel to the shear direction²⁷. Crucially, the orientational feature of the LLC mixture was further verified by analyzing the polarized Raman intensity variations across the parallel (I_VV) and perpendicular (I_HH) modes, which confirmed the highly aligned state of the BNNTs and CNTs induced by shear forces (Fig. S8)²⁸. The characteristic Raman peak observed around 1,367 cm⁻¹ showed polarization-dependent intensity variations relative to the BNNT long axis²⁹. The maximum intensity of this peak occurred at I_VV, while the minimum intensity was observed at I_HH.

Leveraging the well-defined LLC behavior and the highly shear-responsive alignment of these nanomaterials, we utilized the homogeneous SAAN_n solutions as spinning dopes. By successfully translating this nanoscale orientational order into macroscopic architectures, continuous fibers were fabricated via a wet-spinning process through a multi-hole spinneret (40 circular orifices, each with a 30-gauge inner diameter) (Fig. 1d). The extruded SAAN_n fiber was extracted from the coagulant bath, collected onto a bobbin, and finally dried under reduced pressure. Notably, the resulting all-nanotube fiber is continuous and extends approximately 178 m in length when spun from merely 100 mL of the dope solution, highlighting its macroscopic scalability (Fig. 1e). Furthermore, this continuous spinning was accomplished entirely without the use of additives, such as surfactants, binders, or stabilizers, ensuring the absolute compositional purity of the fibers. This scalable outcome, which remains fundamentally inaccessible using pristine BNNTs alone, provides a highly practical pathway for developing space radiation shielding textiles while maintaining a BNNT content as high as 80 wt%.

Even high-performance nanomaterials can introduce structural defects, stress concentrations, or interfacial delamination in hybrid systems if not uniformly dispersed³⁰. To harness their intrinsic properties at the macroscopic level, precise control over the nano-to-macroscale co-assembly is therefore essential³¹. In this context, the emergence of the LLC phase ensures phase-level homogeneity of the BNNT and CNT mixtures, which subsequently enables the fabrication of robust SAAN₈₀ fibers (Fig. 2a). This LLC phase arises from a finely tuned thermodynamic balance of intermolecular interactions between the dissimilar 1D nanomaterials³². In this system, CA acts as a true thermodynamic solvent rather than a mere dispersive medium. Specifically, the BNNTs are individually dispersed via the protonation of the nitrogen atoms within their hexagonal lattice, causing the outer wall surface to carry a net positive charge³³. Simultaneously, the protonation of the CNT π-conjugated sidewalls generates analogous short-range repulsive forces that counteract inherent van der Waals attractions³⁴. As a result, both BNNTs and CNTs locally align to minimize their excluded volume, leading to the formation and growth of N domains, which eventually merge into a homogeneous anisotropic phase with randomly oriented n³⁵. Therefore, the BNNTs and CNTs were uniformly dispersed in CA by forming a homogeneous LLC phase, which served as a stable wet-spinning dope. During wet-spinning, the shear flow within the spinneret and the subsequent extensional flow during the drawing process induced the strictly uniaxial orientation of the N domains. Simultaneously, the rapid exchange of CA with acetone in the coagulation bath caused an immediate loss of fluidity, locking in the 1D long-range order.

Fig. 2. Mechanical properties of BNNT fibers.

a Fabrication of uniaxially oriented all-nanotube fibers. POM image of the homogeneous lyotropic liquid crystalline SAAN₈₀ dope in CA (left), and a schematic illustration of flow-induced alignment of N domains comprising dissimilar 1D nanomaterials during wet-spinning and continuous fiber collection on a rotating bobbin (right). The POM image is representative of two independently prepared samples showing similar optical textures. The schematic rendering was created using Blender v4.4 and PowerPoint v16. b TGA thermograms of BNNTs, CNTs, and SAAN₈₀ fibers. c High-resolution B1s (left), C1s (middle), and N1s (right) XPS spectra of the BNNTs, CNTs, and SAAN₈₀ fibers. Comparison with the BNNTs and CNTs reference spectra supports the incorporation of both nanotube components into the SAAN₈₀ fibers. The displayed binding energy ranges correspond to the narrow-scan acquisition windows for each spectrum. d Tensile strength of SAAN_n fibers as a function of BNNT content. Data are presented as the mean ± standard deviation (m = 10). e Boron content in SAAN_n fibers with varying compositions of 1D nanomaterials. f High flexibility of SAAN₈₀ fibers. The SEM image is representative of two independently prepared samples with similar morphologies. g Knot efficiency of SAAN₈₀ fibers compared with reported values for commercial high-performance fibers⁴⁴. Knot efficiency is defined as the ratio of the tensile force at break of a knotted fiber to that of an unknotted fiber, expressed as a percentage. Data for SAAN₈₀ fibers are presented as the mean ± standard deviation (m = 5). Source data are provided as a Source Data file.

To quantitatively verify the exact nanotube ratio within the assembled fibers, thermogravimetric analysis (TGA) was deliberately conducted under an air atmosphere (Fig. 2b). Under these oxidizing conditions, the CNT component initiated thermal degradation at approximately 600 °C and fully decomposed by 913.3 °C, whereas the pristine BNNTs remained stable at the same temperature. Consequently, the SAAN₈₀ fibers exhibited a final residual weight of 81.48% at 913.3 °C, reflecting the initial 80 wt% BNNT loading formulated in the spinning dope. It is imperative to clarify that the mass loss observed from 600 °C is strictly an oxidation-driven event, utilized here solely for compositional verification. When operating in an inert environment where oxidation is fundamentally precluded in space, both the BNNTs and CNTs resist this degradation, thereby retaining the complete structural integrity and macroscopic stability of the SAAN₈₀ fibers (Fig. S9). The elemental composition of the SAAN₈₀ fibers was confirmed by X-ray photoelectron spectroscopy (XPS) spectra (Fig. 2c). The distinct B1s peak at 190.5 eV indicates the boron atoms from the BNNTs, while the C1s peak at 284.6 eV confirms the incorporation of CNTs³⁶. A distinct N1s peak at 397.2 eV corresponds to the nitrogen atoms within the hexagonal lattice ³⁷.

As shown in Fig. 2d, the tensile strength of the SAAN₈₀ fiber reaches 1.18 GPa, which is approximately 51 times higher than that of the SAAN₁₀₀ fiber. Progressively increasing the CNT fraction yields higher tensile strengths, reaching up to 2.96 GPa for the SAAN₂₀ fibers (Fig. S10). This enhancement in tensile strength is accompanied by simultaneous increases in both Young’s modulus and toughness (Fig. S11). The fundamental mechanism driving this dramatic phenomenological transition, from the brittle, unspinnable SAAN₁₀₀ segments to the continuously spinnable SAAN₈₀ fibers, is governed by stringent morphological constraints within the 1D co-assembly³⁸. While both nanotubes are thermodynamically co-dispersed via surface protonation, their aspect ratios differ significantly. The BNNTs had diameters of 20–150 nm and lengths of 3–30 μm, yielding an average aspect ratio of approximately 190. In contrast, the CNTs exhibit a much higher aspect ratio of nearly 6010, comprising diameters of 0.9–3.2 nm and lengths of 10–14 μm.

The incorporation of these high aspect ratio nanotubes fundamentally alters the rheological processability of the dope. While all SAAN_n solutions exhibit shear-thinning behavior, the magnitude of the viscosity reduction with increasing shear rate becomes significantly more pronounced as the CNT concentration increases (Table S1). Consequently, a decrease in the power-law index signifies a progressively stronger shear-thinning response³⁹. The longer CNTs induce a larger excluded volume minimization, enhancing cooperative interactions between the nanotubes within the N domains⁴⁰. This makes their n highly sensitive to applied shear stress, thereby facilitating a large contact area and strict uniaxial orientation during the shear flow process through the fine spinneret nozzle. Notably, the initial shear viscosity of the pure SAAN₁₀₀ solution is distinctly lower than that of the SAAN₈₀, SAAN₆₀, SAAN₄₀, and SAAN₂₀ dopes, indicating that the dispersed BNNTs possess a relatively larger free space⁴¹. The weak shear-thinning behavior of the SAAN₁₀₀ solution minimizes the probability of inter-tube collisions during shear flow. Consequently, the applied shear forces are insufficient to induce the necessary reorientation of the BNNTs parallel to the flow direction. As a result, the extruded SAAN₁₀₀ segments consist of BNNTs with a less anisotropic conformation, making them highly susceptible to inter-tube and inter-domain slipping, which ultimately leads to macroscopic fracture.

In stark contrast, the SAAN₈₀ dope, characterized by a reduced free space, exhibits robust shear-thinning behavior. During extrusion, the extensive physical contact between the CNTs and BNNTs generates sufficient shear forces to drive the reorientation of the entire nanotubes⁴². Consequently, the resulting SAAN₈₀ fiber exhibits a reduced number of nanotube junctions per unit length, thereby maximizing the interfacial contact area for efficient load transfer⁴³. This combined morphological and rheological advantage provides a highly stable pathway to increase the draw ratio (DR) during the fabrication of SAAN₈₀ fibers. We successfully spun SAAN₈₀ fibers while increasing the DR from 1.01 to 1.61 (Fig. S12). This processability proves that the introduction of CNT successfully transfers stress across the entire fiber, preventing premature breakage at BNNT-rich sites. Moreover, the ability to achieve such high DRs is a critical processing parameter that directly enhances the overall fiber production throughput. It is worth noting that for the SAAN₁₀₀ fiber, a meaningful DR cannot even be defined because the extruded segments are too brittle to be continuously collected onto a bobbin. Even if a theoretical DR is estimated based on the volumetric extrusion rate and the average length of the fractured segments, it yields a negligible value of only 0.12. Consequently, overcoming this unspinnable state translates to an 11-fold increase in manufacturing throughput for the SAAN₈₀ fibers relative to SAAN₁₀₀.

While further increasing the CNT fraction beyond 20 wt% undoubtedly yields continuous gains in mechanical properties and processability, the core objective of this study is to develop space materials optimized for thermal neutron shielding. As shown in Fig. 2e, incorporating more CNTs to maximize mechanical robustness inevitably sacrifices the total boron content in the SAAN_n fibers, which directly degrades their radiation shielding capability. Therefore, the SAAN₈₀ fiber represents the critical optimal threshold. In contrast to the fragile, unspinnable nature of SAAN₁₀₀, the SAAN₈₀ fiber exhibits flexibility alongside its high strength. SEM images reveal the uniform macroscopic 1D morphology of the SAAN₈₀ fiber with a diameter of 58 µm (Fig. 2f). The SAAN₈₀ fiber can be fully twisted and knotted without fracture, demonstrating mechanical robustness and processability for practical textile applications. To evaluate its mechanical pliability, we measured the knot efficiency, defined as the ratio of the tensile force of the knotted fiber to that of the initial unknotted straight fiber (Fig. S13). The SAAN₈₀ fibers exhibit a high knot efficiency of 87% (Fig. 2g). This level of knot efficiency is comparable to, or even exceeds, those of well-established high-performance fibers, such as Nylon, Dyneema, Zylon, Kevlar, and T300⁴⁴. Conversely, the pristine SAAN₁₀₀ segments were not only incapable of continuous spinning but also fractured when bent to an angle of merely 28° due to their brittleness. To further assess the flexibility, cyclic bending tests were performed. The SAAN₈₀ fiber demonstrated high fatigue resistance, retaining 95.9% of its initial tensile force even after 100 bending cycles at a bending angle of 50° (Fig. S14). Ultimately, the SAAN₈₀ fiber strikes an effective multifunctional balance. It delivers the high boron content essential for thermal neutron attenuation while simultaneously achieving a tensile strength exceeding 1 GPa and structural flexibility. This combined performance, enabled by the 1D co-assembly of 80 wt% BNNTs with 20 wt% CNTs, makes the SAAN₈₀ fiber a robust candidate for advanced radiation shielding textiles.

To elucidate the nano-to-macroscale orientation of the dissimilar 1D nanomaterials, cross-sectional transmission electron microscopy (TEM) and energy-dispersive X-ray spectroscopy (EDS) were performed on the SAAN₈₀ fibers. As shown in Fig. 3a, the TEM images reveal a densely packed structure of BNNTs and CNTs. Consistent with our preceding rheological analysis, this dense packing is a direct physical manifestation of the shear-thinning behavior and cooperative reorientation during extrusion. This compacted internal morphology maximizes the interfacial contact area and inter-tube frictional forces, both of which are prerequisites for efficient stress transfer. The ability to clearly visualize distinct, circular cross-sections of the constituent nanotubes throughout the imaged area, rather than longitudinal tube walls or oblique elliptical profiles, serves as direct visual evidence of their strict uniaxial orientation parallel to the macroscopic fiber axis⁴⁵. Ultimately, both this dense spatial packing and the strict uniaxial alignment are fundamentally attributed to the incorporation of CNTs, which effectively lowers the power-law index of the homogeneous nanotube mixture⁴⁶. Furthermore, corresponding EDS mapping confirms the homogeneous co-dispersion of these structures without random aggregation; CNT regions appear red (carbon), while BNNT regions appear green and blue (boron and nitrogen).

Fig. 3. Uniaxially oriented BNNT fibers.

a Cross-sectional TEM image of the SAAN₈₀ fiber and corresponding EDS elemental maps visualizing carbon (red), boron (green), and nitrogen (blue). The TEM and EDS images are representative of two independently prepared samples with similar elemental distribution. b 2D WAXD patterns of SAAN₈₀ fibers. c 1D WAXD profiles of SAAN₈₀ fibers consisting of co-assembled BNNTs and CNTs. d Azimuthal intensity profile scanned at 2θ = 25.5°, showing the high orientation of 1D nanomaterials along the FD. e Polarized Raman spectra of SAAN₈₀ fibers. f Schematic illustration of the uniaxially oriented nanotubes within the SAAN₈₀ fibers. The schematic rendering was created using Blender v4.4 and PowerPoint v16. Source data are provided as a Source Data file.

The structural alignment is further corroborated by wide-angle X-ray diffraction (WAXD). The 2D WAXD pattern collected along the fiber direction displays anisotropic equatorial diffraction arcs, confirming that the hexagonal lattices of both carbon and BN are aligned parallel to the macroscopic fiber axis (Fig. 3b). The 1D diffraction profile extracted from the equatorial region was analyzed to investigate the intrinsic nanoscale ordered structures of the individual nanotubes, rather than the macroscopic spatial packing of nanotube bundles (Fig. 3c). The observed profile reflects a spectral convolution of the two distinct structural characteristics. Because the SAAN₈₀ fiber consists primarily of BNNTs, the profile is dominated by the intense peak at 2θ = 25.5°, which corresponds to the (002) basal plane reflection originating from the inter-wall spacing of the crystalline, concentric layers of the hexagonal BN lattice⁴⁷. Adjacent to this, the broad diffraction peak around 2θ = 24.1° originates from the structural contribution of the inter-tube spacing between the neighboring CNTs⁴⁸. The orientational order parameter (S) of the SAAN₈₀ fiber, calculated from the azimuthal intensity distribution at 2θ = 25.5° corresponding to the BNNT, was found to be as high as 0.86 (Fig. 3d), indicating a high degree of alignment.

As shown in Fig. 3e, the SAAN₈₀ fiber reveals the E_2g vibrational mode of BN bonds from BNNTs and the characteristic D- and G-bands of CNTs⁴⁹. The polarized Raman spectroscopy further supports the uniaxial alignment of SAAN₈₀ fibers. The intensity ratio of I_HH to I_vv for the SAAN₈₀ fiber was measured to be 1:13.6 at DR = 1.41 (Fig. S15). While the fibers spun at lower DRs of 1.01 and 1.21 already exhibit high orientation factors of 1:11.0 and 1:12.7, their tensile strengths remain relatively modest at 0.53 GPa and 0.77 GPa compared to that at DR = 1.41, respectively (Fig. S16). Elevating the DR beyond 1.41 yields negligible improvements in both orientation (1:13.9) and strength (1.26 GPa). Instead, it merely reduces the linear density (i.e., the mass per length) of the fiber, which inadvertently compromises the overall production throughput (Fig. S17). Building upon the established spinnability, this DR-dependent behavior indicates a direct relationship between uniaxial orientation and mechanical strength. Reaching a DR of approximately 1.41 provides the necessary mechanical stress to fully consolidate the uniaxially oriented structures, propelling the tensile strength to its peak. In essence, the introduction of the LLC mesophase fundamentally enables the homogeneous co-assembly and shear-induced alignment of CNTs and BNNTs. Ultimately, this tailored nanoscale architecture directly translates into macroscopic performance, facilitating high-throughput continuous manufacturing while maximizing the atomic density of boron for superior thermal neutron attenuation (Fig. 3f).

To explore the practical applicability of these SAAN₈₀ fibers, we next evaluated their processability into 2D woven textiles and assessed their effectiveness in thermal neutron shielding (Fig. S18). Owing to their sufficient length and flexibility, these SAAN₈₀ fibers were easily woven into plain textiles by interlacing warp and weft fibers at right angles, enabling scalable coverage over large surface areas (Fig. 4a). The fabricated all-nanotube continuous fibrous materials exhibit sufficient mechanical integrity to support mass production of textiles (Fig. S19). The high boron content incorporated in the hexagonal lattice of BNNTs, which provides a broad energy range and high absorption cross-section for thermal neutrons (Fig. 4b). Thermal neutron attenuation of 19.3% was achieved using a single-layer SAAN₈₀ textile with a thickness of 76 µm (Fig. S20). The dominant shielding mechanism involves nuclear reactions between thermal neutrons and boron nuclei, releasing lithium-7, helium-4, and gamma-ray emission⁵⁰. Femtosecond laser ablation-inductively coupled plasma mass spectrometry (ICP-MS) analysis confirmed that the BNNTs in the textiles contained 20.1% boron-10 and 79.9% boron-11 isotopes (Fig. S21). Since increasing the probability of neutron-boron interactions along the radiation path reduces radiation permeability, vertically stacking three and five layers of the textile further enhanced attenuation, achieving progressively higher shielding efficiencies of 46.2% and 59.7%, respectively (Fig. S22). The shielding efficiency was analyzed using the Beer-Lambert law, yielding a linear attenuation coefficient (μ) of 2.85 mm⁻¹ (Fig. 4c)⁵¹. On the basis of its μ, the SAAN₈₀ textile can block nearly 99% of incoming thermal neutrons at a thickness of approximately 1.62 mm (Table S2). The density (ρ) of the SAAN₈₀ textile was measured at 1.91 g cm⁻³, resulting in a mass attenuation coefficient (μ_ρ) of 14.92 cm²g⁻¹. These results highlight the advantage of SAAN₈₀ textiles in space applications, where volume and weight are critical constraints⁵². Terrestrial neutron shielding materials, such as water containers or concrete blocks used in medical accelerators and nuclear power plants, are bulky and heavy, rendering them unsuitable for use in space mission⁵³. In contrast, the SAAN₈₀ textile not only achieves superior μ_ρ but also retains the compact dimension. These advantages collectively position the SAAN₈₀ textile as a lightweight and flexible solution for future advanced protective materials in human space mission.

Fig. 4. Comprehensive performance of SAAN₈₀ textiles for space applications.

a Multiscale representation of SAAN₈₀ textiles for integration into an extravehicular mobility unit (EMU), showing a photograph of a 1/100 scale EMU model reproducing the multilayer spacesuit configuration and constituent materials (left), a photograph of a 10 × 10 cm² SAAN₈₀ textile (second left), a schematic illustration of aligned BNNTs and CNTs within a SAAN₈₀ fiber (second right), and a schematic illustration of the BN lattice structure and ¹⁰B mediated neutron attenuation mechanism (right). The schematic rendering was created using Blender v4.4, PowerPoint v16, and ChemDraw v25. b Thickness-dependent thermal neutron attenuation properties of the SAAN₈₀ textiles. c Thermal neutron shielding performance of SAAN₈₀ textiles. d Temperature-dependent mechanical behavior of the SAAN₈₀ textiles, with inset photographs showing the textile before and after tensile testing. e Retention of thermal neutron shielding performance across different temperatures, with inset optical microscopy images showing the structural integrity of the SAAN₈₀ textile after exposure to the indicated temperature. f SAAN₈₀ textiles exhibiting a combination of high tensile strength and thermal neutron shielding performance. g Positioning of SAAN₈₀ textiles within the domain of high thermal stability and thermal neutron shielding performance. Source data are provided as a Source Data file.

With the recent surge of interest in space exploration, including space tourism, scientific missions, resource harvesting, and space colonization, there is a growing demand for materials that can withstand the extreme conditions of space⁵⁴. Beyond thermal neutron shielding, materials deployed in these environments must endure wide temperature fluctuations and harsh external stimuli. In this context, the SAAN₈₀ textile demonstrates excellent potential. Mechanical properties were evaluated along the weft direction using textile specimens measuring 25 × 12 mm², composed of approximately 650 SAAN₈₀ fibers. The SAAN₈₀ textile demonstrated a tensile force of 1749 N, which is approximately 557 times greater than that of a single fiber (Fig. 4d). To simulate the sparse oxygen conditions and severe thermal extremes typical of extraterrestrial environments, the SAAN₈₀ textiles were exposed to temperatures of −196 °C, 400 °C, 800 °C, and 1200 °C for 1 h under a nitrogen atmosphere (Fig. S23). This exceptional tensile force remained nearly unchanged across all tested temperatures. Raman spectroscopy reveals that the characteristic E_2g peak of the BNNTs and the G-band of the CNTs remained unchanged even after these diverse thermal treatments (Fig. S24). Furthermore, XPS confirmed the complete preservation of their chemical bonding states across all temperature variations; the absence of significant shifts in binding energy proves that the intrinsic hexagonal BN and carbon lattices suffered no chemical degradation (Fig. S25). These spectroscopic results are consistent with the TGA thermogram, which revealed that the SAAN₈₀ textile maintains thermal stability up to 1200 °C under an inert atmosphere (Fig. S26). This profound thermomechanical robustness is inherently attributed to the thermal stability of the constituent nanomaterials. Furthermore, the woven structure of the textile remained intact without visible defects after exposure to −196 °C, 400 °C, 800 °C, and extended heating at 1200 °C (Fig. 4e). Because there was no structural degradation, the thermal neutron shielding efficiency of SAAN₈₀ woven textile remained stable (Table S3). Corresponding elemental mapping via SEM-EDS confirmed that the uniform distribution of boron, nitrogen, and carbon was perfectly preserved within the macroscopic dimensions, with the spectral features before and after thermal treatment being nearly identical (Fig. S27). Collectively, these results support the SAAN₈₀ textile as a high-performance shielding material that reliably maintains its chemical, structural, and functional integrity under the extreme thermal environments encountered in deep space missions.

To the best of our knowledge, the SAAN₈₀ textile exhibits the highest thermal neutron shielding efficiency reported to date among BNNT-based materials, particularly in textile structures using fibers. This enhanced performance stems from the high BNNT content of 80 wt%, achieved via the all-nanotube bottom-up self-assembly. Although uniformly dispersing nanomaterials at such concentrations typically poses a challenge, our LLC approach successfully overcomes this limitation and enables the fabrication of continuous fibers and macroscale textiles. As shown in Fig. 4f, the SAAN₈₀ fiber achieves tensile strengths exceeding 1 GPa, effectively overcoming the typical mechanical limitations of macroscopic BNNT-based fibrous materials. Additionally, the absence of organic small molecules such as dispersants, surfactants, and binders prevents the dynamic instabilities caused by molecular exchange, migration, and rearrangement, thereby preserving the physical performance of the material⁵⁵. Beyond its mechanical advantages, the SAAN₈₀ textile exhibits thermal stability (Fig. 4g). The macroscopic structure of the SAAN₈₀ textile remained intact even at −196 °C and 1200 °C for 1 h under an inert atmosphere, noting that reactive gases such as oxygen are present only in trace amounts in space, thereby maintaining a nearly constant μ of approximately 2.85 mm⁻¹. Among BNNT-based materials, the only example reported to surpass the mechanical strength and thermal stability of SAAN₈₀ textiles is a BNNT-metal composite, which reaches a tensile strength of 1.14 GPa and demonstrates resistance up to 1668 °C⁵⁶. However, due to its relatively low BNNT content, the composite’s mechanical and thermal behavior is dominated by the metallic matrix, resulting in a lower neutron shielding performance compared to the SAAN₈₀ textile⁵⁷. BNNT-polymer composites, in contrast, can incorporate a much higher BNNT content than metal-based systems, which imparts good neutron shielding capability^20,33,58. Nevertheless, most BNNT-polymer materials soften, melt, carbonize, or degrade at around 500 °C due to the intrinsic thermal limitations of the polymer matrix ^50,59–62.

Given its demonstrated thermal and mechanical stability, the SAAN₈₀ textile can be considered for use in the spacesuit to provide effective thermal neutron shielding. Current spacesuit designs comprise multiple functional layers, each tailored to serve as primary protection against micrometeoroid impacts, provide thermal insulation, maintain internal pressure, and accommodate integration with the cooling and ventilation garment⁶³. We envision incorporating the macroscopically woven SAAN₈₀ textile beneath the orthofabric outer shell, which is typically composed of para-aramid, meta-aramid, and polytetrafluoroethylene fibers, elevating thermal neutron shielding while preserving mechanical stability. Its integration is also expected to provide a synergistic effect when paired with aluminized Mylar and the Dacron restraint layer, enhancing thermal protection. It is worth noting that radiation-induced lattice rearrangement and trapped-charge formation can degrade electronic systems, leading to malfunctions that range from minor data loss to catastrophic failure⁶⁴. In addition, neutron radiation can cause atomic displacements within material structures, resulting in embrittlement and ultimately compromising the strength and integrity of the structural materials⁶⁵. Thus, the multifunctional properties of the SAAN₈₀ textile offer potential for integration not only into spacesuit architectures, but also into space exploration instruments, spacecraft exteriors, and habitat shelters, all aimed at protecting astronauts against thermal neutrons, external stimuli, and thermal fluctuations.

To evaluate the practical effectiveness of the SAAN₈₀ textile against space radiation environments, computational simulations using the Particle and Heavy Ion Transport code System (PHITS) based on Monte Carlo method were conducted for two critical human exploration scenarios⁶⁶. First, an IVA scenario was modeled with a GCR proton source incident on a silicon carbide (SiC), reinforced carbon-carbon (RCC), and aluminum (Al) alloy plate followed by an onboard computer, representing a spacecraft hull and an electronics-dense cabin, respectively (Fig. S28 and Table S4). Second, an EVA scenario was simulated, wherein GCR protons strike a deep layer of lunar regolith (Fig. S29 and Table S5). In both cases, initial spallation generates a broad energy spectrum of secondary neutrons (Fig. S30). These secondary neutrons either propagate forward, transmitting through the targeted materials along the incident GCR trajectory, or backscatter into the open environment⁶⁷. Therefore, assuming the astronaut is equipped with a spacesuit integrating the SAAN₈₀ textile, the simulations were strategically designed to evaluate two distinct directional threats. For the IVA spacecraft model, as the astronaut is situated inside the cabin, the primary objective is to assess the shielding efficacy against the transmitted thermal neutrons penetrating the hull (Fig. S31). These energetic particles undergo continuous multiple scattering collisions with internal structures, eventually thermalizing and becoming trapped within the confined cabin environment (Fig. S32)⁶⁸. Conversely, for the lunar EVA model, where the astronaut operates externally on the surface, the focus shifts to evaluating protection against the albedo thermal neutrons reflecting off the regolith (Fig. S33). Inherently acting as a massive moderator, the regolith generates a substantial flux of these thermal neutrons (Fig. S34)⁶⁹. A standard spacesuit alone offered negligible protection against this specific threat (Tables S6 and S7), producing a transmitted thermal neutron spectrum nearly identical to the unshielded environment (Fig. S35), as clearly visualized in the spatial distribution maps (Fig. 5a, 5b) and secondary neutron energy spectra. In contrast, integrating the SAAN₈₀ textile attenuated the neutron flux within the thermal energy region, which otherwise penetrates standard spacesuit architectures. (Fig. 5c, 5d). The broad-energy neutron transmittance spectra for both the IVA (Fig. 5e and Table S8) and EVA (Fig. 5f and Table S9) scenarios further demonstrate that neutrons with energies below 0.6 eV are effectively blocked, resulting in a transmittance approaching nearly 0% upon the integration of the SAAN₈₀ textile (Fig. S36). This computational result aligns with the theoretical prediction derived from our experimental thermal neutron shielding coefficient, which indicates that a 99% thermal neutron attenuation is achievable at a textile thickness of 1.62 mm.

Fig. 5. Computational evaluation in deep space environments and scalable manufacturing of SAAN₈₀ textiles.

a Simulated spatial distribution maps of thermal neutrons penetrating a standard spacesuit architecture in IVA, and b lunar EVA scenarios. c Spatial distribution maps demonstrating the attenuation of thermal neutron flux by the SAAN₈₀-integrated spacesuit in corresponding IVA, and d EVA environments. e Broad energy neutron transmittance spectra for IVA, and f EVA models, highlighting the absorption of neutrons below 0.6 eV upon SAAN₈₀ integration, with inset images corresponding to the IVA and EVA scenario models. The scenario models were created using Blender v4.4 and PowerPoint v16. g Demonstration of scalable manufacturing showing 21 independent batches of SAAN₈₀ fibers. h Macroscopic SAAN₈₀ textiles processed via commercial textile machines, producing a flat 2D structure, and i, a tubular 3D structure. Source data are provided as a Source Data file.

Building on these multifunctional properties, we evaluated the mass production feasibility of SAAN₈₀ fibers to demonstrate their practical applicability for next-generation spacesuit textiles. We demonstrated process reliability and reproducibility by producing 21 separate batches of SAAN₈₀ fibers, each approximately 170 m in length (Fig. 5g). These independent fiber batches exhibited consistent tensile strength, modulus, and toughness values, indicating the stability of the spinning dope and the extrusion process (Fig. S37). Furthermore, this batch-to-batch consistency supports the efficacy of acetone as the coagulation bath. The fundamental purpose of the coagulant is to rapidly extract the CA to solidify the fiber without disrupting the uniaxially aligned N domains⁷⁰. While ether exhibits excellent non-reactivity with the CA solvent, its extreme volatility severely destabilizes the spinning line, hindering continuous mass production. Conversely, water is typically the preferred coagulant for industrial scale-up owing to its abundance; however, it triggers a violent exothermic hydrolysis reaction with CA, generating hydrogen chloride gas and severely degrading the assembled nanostructures⁷¹. Consequently, acetone induces rapid and uniform CA extraction, instantaneously freezing the highly aligned 1D co-assembled nanotubes established in the spinneret. This process shows sufficient stability of the coagulation bath required for continuous, scalable spinning. Demonstrating the scalability of our process, we manufactured a continuous SAAN₈₀ fiber exceeding 3.5 km in length. Utilizing the structural reliability and flexibility of these long fibers, we processed them into macroscopic textiles using commercial equipment. Specifically, we fabricated a flat 2D structure (Fig. 5h) and a tubular 3D structure (Fig. 5i) using weft and circular textile processing machines, respectively. Processing such architectures under dynamic multi-axial stresses requires batch-to-batch reproducibility and consistent mechanical reliability of the constituent fibers. These results indicate that SAAN₈₀ fibers can be scaled up for practical applications in deep space exploration garments.

Discussion

The performance of the SAAN₈₀ textile is enabled by a nanomaterials-driven nanomanufacturing strategy incorporating three hierarchical design principles: the exclusive use of 1D nanomaterials at the primary level, guided homogeneous co-assembly at the secondary level, and continuous uniaxial alignment of fibers at the tertiary level. Fabricated via a high-throughput wet-spinning process, the structural integrity of these fibers is driven by the thermodynamic co-assembly of the dissimilar nanotubes into the LLC mesophase. This nanoscale architecture yields a shear-responsive dope, cooperatively facilitating the disentanglement, reorientation, and dense packing of the nanotubes under applied drawing forces. Ultimately, this structural synergy overcomes the inherent brittleness of unspinnable SAAN₁₀₀ segments, enabling the continuous spinning of SAAN₈₀ fibers with a tensile strength exceeding 1 GPa.

Further enhancement of the μ of the SAAN₈₀ textile can be achieved by enriching the boron-10 isotope content in the BNNTs. Since thermal neutrons are primarily absorbed by boron-10 nuclei, undergoing nuclear reactions that convert them into charged alpha particles accompanied by gamma emission, the neutron shielding performance of BNNTs is directly influenced by their boron-10 composition⁷². Accordingly, increasing the boron-10 fraction in the SAAN₈₀ fibers would proportionally improve the attenuation efficiency. If the BNNTs in the SAAN₈₀ textile, which already contains a high BNNT content of 80 wt%, were isotopically enriched to consist entirely of boron-10, the μ is projected to increase to 14.2 mm⁻¹, approximately 4.98 times higher than that of the current system.

Looking ahead, the mechanical properties of the SAAN₈₀ fibers may be further enhanced by tuning intrinsic BNNT parameters such as diameter and length, which govern inter-tube load transfer. While the BNNTs utilized in this study possessed diameters of 20-150 nm and lengths of 3–30 μm, incorporating BNNTs with higher aspect ratios may elevate the tensile strength beyond 3 GPa, a benchmark often associated with super fibers, by increasing lateral friction and interfacial surface energy between the nanotubes⁷³. Furthermore, considering that the localized and partial wall dissociation or unzipping of BNNTs at elevated temperatures can result in the loss of their intrinsic nanostructure, ultimately leading to the collapse of uniaxially oriented macroscopic 1D architectures, the use of highly crystalline BNNTs with fewer walls is expected to further enhance thermal stability ⁷⁴.

Since the development of BNNTs is analogous to the early stages of CNT research, future breakthroughs in BNNT synthesis will intrinsically elevate their absolute physical properties⁷⁵. However, until such milestones are fully achieved, our current 1D co-assembly strategy provides an immediate and effective structural workaround. By bridging the gap between nanoscale 1D co-assembly and macroscopic processability, the SAAN₈₀ textile provides a practical, scalable, and reliable structural solution for multifunctional radiation shielding. Beyond its neutron attenuation capability, the textile’s combination of low weight, mechanical robustness, and thermal durability makes it suitable for both internal and external components in future long-duration human deep space missions.

Methods

Fabrication of of SAAN fibers

BNNT (Naieel, NanoBorNT) and CNT (Meijo Nano Carbon, DX-2) powders were dried for 24 h at 60 °C under reduced pressure to eliminate moisture in the samples. The BNNTs and CNTs were mixed in a glass vial filled with CA, followed by the preparation of a series of solutions with different weight ratios (100/0, 80/20, 60/40, 40/60, and 20/80) in a golve box. To ensure complete dissolution and homogeneity, these mixtures were magnetically stirred for 48 h at room temperature and then planetary mixing to avoid agglomeration of nanotubes. The resulting solution was carefully loaded into a glass syringe, avoiding air bubbles, and then extruded through a spinneret (inner diameter 159 μm, 40 circular holes) into coagulation bath filled with an acetone at the dew point of −20 °C. During this wet-spinning process, the emerging SAAN_n fibers were wound onto a bobbin. The extrusion rate was precisely controlled at 0.5 mL m⁻¹ in with winding speeds of 0.64 m min⁻¹, 0.76 m min⁻¹, 0.89 m min⁻¹, and 1.02 m min⁻¹, resulting in DR of 1.01, 1.21, 1.41, and 1.61, respectively. The collected SAAN_n fibers were rinsed in a water bath to remove residual solvents. Then, the SAAN_n fibers were dried in a vacuum oven at 80 °C for 24 h. A macroscopic woven SAAN₈₀ textile was crafted by interlacing these continuous fibers as warp and weft yarns. The continuous SAAN₈₀ fibers successfully withstood the dynamic tensions and multi-axial stresses inherent in commercial textile processing. Consequently, macroscopic flat 2D and tubular 3D structures were fabricated using commercial weft and circular knitting machines, respevtively, with technical support (Carvri Future Materials), ultimately demonstrating their exceptional processability under standard industrial conditions.

Characterizations

The optical textures were observed using POM (Nikon Eclipse E600) with an analyzer and polarizer. The processing of POM data was performed by NIS Elements BR v5.1 software. The packing structures were charaterized by 1D WAXD patterns using a X-ray diffractometer (Rigaku SmartLab) with Cu Kα radiation operated at 40 kV and 200 mA with a step size of 0.02°. The processing of 1D WAXD data was performed by SmartLab v1.5 software. The 2D WAXD data of SAAN₈₀ fibers were collected using a custom-built laboratory X-ray diffraction instrument (Rigaku NANOPIX). The setup includes a micro-focus Cu rotating anode X-ray generator (1.2 kW, Cu Kα = 1.5418 Å) and a 2D hybrid pixel array detector (pixel size: 100 × 100 μm). The sample-to-detector distance was set to 63 mm, corresponding to a q-range of 0.050–35.11 nm⁻¹. X-rays were incident normal to the fiber axis of the SAAN₈₀ sample, resulting in a 2D diffraction pattern. This pattern was processed using X-ray diffraction analysis software to obtain a 1D diffraction profile. To evaluate the S, the azimuthal intensity distribution was scanned from 0° to 360° at 2θ = 25.5°. Based on this profile, the S was calculated using Hermans’ orientation parameter. The processing of 2D WAXD data was performed by Nanopix Guidance v1.3 software. The characteristic Raman shifts were evaluated using a Raman spectrometer (Renishaw inVia Reflex). The samples were placed on a glass substrate and analyzed within a Raman shift range of 700 to 1800 cm⁻¹. An excitation laser with a wavelength of 785 nm was focused onto the sample surface using an optical microscope equipped with a ×50 objective lens, yielding a spatial resolution of approximately 1 µm. To determine the degree of alignment, we monitored the intensity variations of the characteristic Raman peaks as the polarization angle was varied between the parallel (vertical-vertical, VV) and perpendicular (horizontal-horizontal, HH) modes relative to the applied shear force direction. The scattered light from the samples was dispersed by a monochromator equipped with a 1200 grooves mm⁻¹ grating. The processing of Raman data was performed by WIRE v3.4 software. High-resolution TEM images and elemental mapping were performed using an FEI Titan Cubed 60 operated at an accelerating voltage of 80 kV with a Cs-corrector and EDS. The processing of TEM data was performed by GATAN v3.4 and Espirit v1.9 software. The cross-section of the SAAN fiber was prepared using a focused ion beam technique (FEI Helios SEM) prior to analysis. The surface morphology and elemental composition were investigated by SEM (FEI Helios NanoLab 650). The analysis was conducted at an accelerating volatge of 15 kV, and elemental mapping was performed using an EDS detector. The processing of SEM data was performed by xTM v5.5 and TEAM v4.5 software. The chemical state was measured by X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha). The processing of XPS data was performed by Thermo Advantage v5.9 software. The isotope of the boron element in SAAN fiber and textile was analyzed using a femtosecond laser ablation inductively coupled plasma mass spectrometry (ICP-MS, Applied Spectra J200). The thermal degradation temperature was investigated by TGA under nitrogen (Netzsch, STA449F5) and air (PerkinElmer, TGA 4000) atmospheres at a heating rate of 10 °C min⁻¹. The tensile strength of 10 independent SAAN_n fibers was evaluated by fiber tester (Textechno FAVIMAT) and texture analyzer (Brookfield AMETEK CTX) equipped with a 100 N load cell. These measurements utilized a fixed gauge length of 25 mm and a strain rate of 2 mm min⁻¹. To evaluate the cyclic bending durability, the initial gauge length of 25 mm was compressed to 3 mm to achieve a bending angle of 50°. Returning to the initial gauge length constituted one complete cycle, with a single cycle duration of 50 s. This procedure was repeated for 100 consecutive cycles, with the resultant tensile force monitored continuously throughout the testing period. Additionally, the flexibility and mechanical integrity of the fibers under severe deformation were assessed using the knot efficiency (E), calculated as E (%) = (F_knot × F_unknot⁻¹) × 100, where F_knot and F_unknot represent the tensile forces at break for knotted and unknotted fibers, respectively. The tensile force of SAAN textiles was measured using a universal testing machine (SALT ST-1002) equipped with a 30 kN load cell. For the textile samples, testing was performed with a gauge length of 12 mm and a crosshead displacement rate of 1 mm min⁻¹ at room temperature. The modulus was calculated from the initial linear elastic region of the stress-strain curves, and the toughness was quantified by integrating the area under the curve up to the strain at failure. The thickness of the SAAN₈₀ textiles was measured at nine distinct locations using a digital thickness gauge with a resolution of 0.001 mm (Mitutoyo K-PG-01). The thickness is reported as the mean ± standard deviation of these measurements. An ultramicrotome (RMC Boeckeler PTXL) equipped with a 3.0 mm diamond knife (SYNTEK SYM3045HD) was utilized to cut 100 nm thick ultra-thin sections at a cutting speed of 1.5 mm s⁻¹. Prior to sectioning, the sample were prepared by embedding the SAAN textiles in an resin (QATM KEM90), followed by overnight polymerization at ambient temperature. The section thickness was subsequently monitored using OM. The linear density (mg m⁻¹) of the SAAN fibers was determined by measuring the mass of a 100 m long fiber segment using a precision balance (AND GH300). The density (ρ, g cm⁻³) was obtained using a gradient density column (Ray-Ran POLYTEST). Macroscopic optical photographs were obtained using a digital camera (Sony HX90V).

Rheological behaviors

The shear rate-dependent viscosity of the prepared samples was measured using a viscometer (Brookfield DV2 Plus-LV SC4-34). A Couette geometry of a spindle (Brookfield SC434) was employed to optimize shear stress sensitivity across both high- and low-viscosity regimes. The measurement system consisted of an inner rotating spindle (radius = 5 mm) with a fixed gap of 3 mm to a glass cylinder. Prior to measurement, all samples were thoroughly degassed using a planetary centrifugal mixer (THINKY ARM-310) to eliminate entrapped air bubbles and ensure high data reproducibility. All rheological measurements were performed in a temperature-controlled laboratory at 21 °C. Shear viscosity was measured over a shear rate range from 1.0 to 14.0 s⁻¹. To ensure data reliability, the viscosity at each shear rate was monitored for a minimum of 1 min until the torque reached a stable steady state, after which the corresponding average value was recorded. All experiments were conducted in triplicate, and the relationship between shear viscosity and shear rate was analyzed using the power-law model, expressed as η = y × γ^x−1 where η, γ, x, and y is the shear viscosity, shear rate, power-law index, and power-law coefficient, respectively⁷⁶. The value of x was determined by performing a linear regression on the log-transformed viscosity data within the power-law region.

Theraml neutron shielding performances

The thermal neutron shielding effectiveness of the SAAN₈₀ textiles was evaluated within a thermal neutron calibration field established at the Korea Research Institute of Standards and Science, utilizing an SP9 ³He proportional counter (Centronic) coupled with an ²⁴¹Am-Be radionuclide neutron source (Amersham) exhibiting an emission rate of 1.23 × 10⁷s⁻¹⁷⁷. To ensure the reliability and stability of the results, the environmental temperature (20 ± 3 °C) and relative humidity (50 ± 20 %) were continuously monitored. To isolate the thermal neutron flux, a standard cadmium (Cd) difference method was employed. First, the total count rate (R_t) for all incident neutrons was measured for 1800 s using a 0.6 mm-thick Cd aperture (overall diameter: 7 cm) featuring a 2 cm-diameter opening. Subsequently, to exclusively count neutrons with energies above 0.6 eV (R_e), the aperture was completely blocked with a Cd cover. To ensure sufficient counting statistics for this attenuated flux, the measurement time was extended to 14,400 s. Finally, the incident thermal neutron count rate (I₀) was determined by subtracting R_e from R_t (I₀ = R_t - R_e). Next, the Cd cover was removed from the aperture, and a 2.5 × 2.5 cm² specimen of the SAAN₈₀ textile was loaded centrally over the 2-cm-diameter measurement hole. The detector’s count rate for all neutrons passing through the specimen was measured for 1800 s and defined as S_t. Following this, the Cd cover was reinserted to measure the transmitted neutrons with energies above 0.6 eV during 14,400 s (S_e). Ultimately, the transmitted thermal neutron count rate (I) penetrating the SAAN₈₀ textile was calculated by subtracting S_e from S_t (I = S_t - S_e). For all experimental conditions, the total accumulated thermal neutron count was directly determined as the product of the respective count rate and its designated measurement time. The atteunation of thermal neutrons passing through the sample was determined using the following equation of µ = ln(I₀ × I⁻¹) d⁻¹, where µ is the linear attenuation coefficient, d is the representative thickness of the textile. The thermal neutron shielding efficiency (%) was calculated as (1 - I × I₀⁻¹) × 100.

Computational modeling of IVA and EVA space radiation environments

All radiation transport simulations were performed using the Particle and Heavy-Ion Transport code System (PHITS v3.3), a general-purpose Monte Carlo code, to evaluate the radiation shielding performance of the SAAN₈₀ textiles. To evaluate the neutron shielding effectiveness, the 2D spatial distributions of the particle flux and the transmitted thermal neutron spectra penetrating the target surfaces were quantified. The primary radiation source was defined to simulate a specific GCR environment based on energy spectrum data derived from the Space Environment Information System (SPENVIS) operated by the European Space Agency (ESA)⁷⁸. The incident GCR proton spectrum was generated utilizing the standard model (ISO-15390). The primary radiation source in this study was defined exclusively as protons, as they constitute the dominant component (~85%) of the GCR spectrum compared to helium nuclei (~14%) and heavier ions (~1%)⁷⁹. The spectrum was specifically extracted for solar minimum conditions, during which the weakened heliospheric magnetic field allows for the maximum penetration of GCRs into the inner solar system, representing the most severe radiological threat to astronauts⁸⁰. The GCR was geometrically defined as a 100 × 100 cm² rectangular planar source, providing a parallel proton beam incident upon the geometry of the spacecraft and the lunar surface. Upon striking these targets, the high-energy protons generate secondary neutrons through complex spallation interactions. The simulations were designed to track these secondary neutrons, ultimately quantifying the final flux passing through the spacesuit equipped with or without the SAAN₈₀ textiles. For the IVA scenario, the spacecraft hull was composed of SiC, RCC, and an Al alloy (Table S10)^81–84. Additionally, the internal electronics located in the cabin were incorporated into the geometry, with the spacesuit positioned behind a 1 cm air gap. For the EVA scenario, a lunar regolith layer comprising a complex multi-elemental composition was modeled at the bottom boundary to account for the generation of secondary albedo neutrons (Table S11)⁸⁵. To evaluate the shielding effectiveness of the SAAN₈₀ textiles against these albedo thermal neutrons, the spacesuit was positioned above the lunar surface, separated by a 1 cm vacuum gap (Table S12)⁸⁶. Data were analyzed using Excel v16 and OriginPro v9.6.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

D.Y.K. conceived the idea and directed the research. K.H.R. and S.I.L. fabricated the materials and evaluated their physical properties. S.G.J. and S.A. conducted the radiation shielding analyses. Y.J.C., K.U.J., S.G.K., and J.Y.H. characterized the structural evolution. K.H.R. and D.Y.K. co-wrote the manuscript, and all authors discussed the results and commented on the manuscript.

Peer review

Peer review information

Nature Communications thanks Hyeong Min Jin, Fengmei Su, Zhiyuan Xiong, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was financially supported by the Korea Institute of Science and Technology institutional program (26E0430, 26E0353) and the Ministry of Trade, Industry & Energy (2300258591, 2410000736).

Data availability

All data supporting the findings of this study are available within the Article and its Supplementary Information. 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 https://doi.org/10.1038/s41467-026-77427-x.