Halloysite‒gold core‒shell nanosystem synergistically enhances thermal conductivity and mechanical properties to optimize the wear-resistance of a pheonlic-PBO/PTFE textile composite liner

: Polymer-textile liner composites have potential applications in aerospace applications for reducing the abrasion damage of moving parts during operation owing to their self-lubrication, light weight, and high loading capacity. Herein, Au nanoparticles (AuNPs) are successfully loaded into the lumen of halloysite nanotubes (HNTs) to construct an HNTs ‒ Au peasecod core ‒ shell nanosystem to optimize the wear resistance of phenolic resin-based poly(p-phenylene benzobisoxazole) (PBO)/polytetrafluoroethylene (PTFE) textile composites. Transmission electron microscope (TEM) characterization reveals that the AuNPs are well-dispersed inside the HNTs, with an average diameter of 6 ‒ 9 nm. The anti-wear performance of the HNTs and Au-reinforced PBO/PTFE composites is evaluated using a pin-on-disk friction tester at 100 MPa. Evidently, the addition of HNTs ‒ Au induces a 27.9% decrease in the wear rate of the composites. Possible anti-wear mechanisms are proposed based on the analyzed results of the worn surface morphology and the cross-section of the tribofilm obtained by focused ion beam transmission electron microscopy.


Introduction
Self-lubricating spherical plain bearings have been widely applied in aerospace and railway locomotive applications to reduce abrasion damage to moving parts during operation. Currently, bearings face greater challenges because of increasingly harsh service conditions. Although traditional polymer composites are lightweight and well-lubricated, they cannot meet advanced requirements because they wear out easily and have a poor loading capacity [1,2]. A woven composite liner, designed by incorporating loadcarrying fibers and a continuous polymer, can be flexibly bonded onto a steel substrate and exhibits prominent practicability in the steering systems of vehicles, fixed wings, landing gear, and operating systems of aircraft and helicopters [3]. Therefore, numerous researchers in both academia and industry have investigated extending the applications of these composite liner materials. Zhang et al. reported the tribological performances of Nomex-, polyamide-, aramid-, and PTFE-based linear composites [4,5], and Gu et al. found that PTFE/Kevlar fabric composites exhibited lower friction but poor anti-wear behavior in a vacuum [6]. Lu et al. extended the service life of a PTFE/aramid fiber composite liner to 1,000 h at a 35 Hz swing frequency by adding polysulfone/poly α-alkene oil 40# microcapsules [7]. Poly(p-phenylene www.Springer.com/journal/40544 | Friction benzobisoxazole) (PBO) fibers with strong mechanical properties, with a tensile strength and tensile modulus of approximately 5.0 GPa and 240 GPa [8] respectively, have also been reported to be suitable for linear composite materials. Nevertheless, few studies have examined the tribological behavior of PBO selflubricating liners.
Currently, clay minerals have attracted considerable attention for use in inflammation retardation [9], biomedical applications [10], catalysts [11], and polymer composites [12,13] owing to their abundance, nontoxicity, and cost-effectiveness. Clay minerals with high thermal stability and exceptional mechanical properties can constrain polymer chain motion and improve the tensile strength, tensile modulus, and hardness of the polymer matrix. Furthermore, clay, as a layered silicate mineral, can reduce friction and wear through its easy shearing and sliding between silica tetrahedra and alumina octahedra, of which the layer interactions are relatively weak van der Waals and electrostatic forces [14,15]. Driven by the potential prospects in polymer composites and tribology, Balasubramanian studied the tribological behaviors of clay-thermoset polyester nanocomposites [16]. Cheng et al. demonstrated that a nano-MoS 2 / montmorillonite hybrid exhibited good wear resistance in a DOS-based oil [15]. Zhang et al. reported that the addition of montmorillonite and kaolin formed a protective film on the counterface and improved the wear resistance of linear composites [17,18]. The halloysite possesses a similar theoretical chemical composition to kaolin but different morphologies (laminar kaolin and tubular halloysite, which resulted in different stacked layers). Therefore, the influence of halloysite on the tribological performance of linear composites must be explored.
Halloysite [Al 2 Si 2 O 5 (OH) 4 ·nH 2 O] occurs as a tubular clay mineral formed by rolling 1:1 aluminosilicate sheets, where the outer surface consists of tetrahedral siloxane groups (Si-O-Si), and aluminum octahedral array groups are spread over the innermost surface. This difference results in a negatively charged outer surface and positively charged inner lumen. The length of the halloysite nanotubes ranges from 0.2 to 2 μm, and their average inner and outer diameters were found to be in the range of 12−20 and 45−60 nm, respectively. Tubular halloysite has been developed as an entrapment system for loading and storing anticorrosion agents, flame retardants, drugs, and enzymes owing to its environmental compatibility, abundance, and broad lumen [19]. In particular, halloysite nanotubes (HNTs) with high length-width ratios and strong mechanical properties are promising reinforcing fillers for polymer composites such as polyamide, epoxy resin, and rubber polymeric matrices [20,21]. Motivated by the loading of drugs into the broad lumen of halloysite, we deposited Au nanoparticles (AuNPs) onto the inner surface, thereby forming a peapod structure. AuNPs with a low shear strength and characteristic thermal conductivity are usually designed as hybrids with other materials or lubricant additives in the field of tribology. For example, Zhou et al. constructed MXene@Ag hybrids to improve the tribological performance of epoxy composites [22]. Sanchez-Lopez et al. used palladium and gold nanoparticles modified with tetraalkylammonium and alkanethiolate chains as lubricant additives to form metal-containing transfer films for anti-wear applications [23]. Guo investigated the synergistic effects of ZDDP/MoDTC and Cu nanoparticles on the tribological behavior of DIOS oil [24]. Herein, we deposited dispersive AuNPs on the inner surface of HNTs by in situ reaction, avoiding the complicated modification of metal nanoparticles, which synergistically enhanced the thermal conductivity and mechanical properties to optimize the wear resistance of the pheonlic-PBO/PTFE textile composite liner.
In this study, PBO/PTFE textile composites were prepared and their heavy-loading tribological behavior was evaluated. PTFE fibers with low coefficients served as the lubricating components, PBO fibers served as the loading and bonding components, and phenolic resin was used as the continuous phase. AuNPs were loaded into the lumen of the HNTs using a rapid and effective method to form a peapod core-shell nanosystem. Subsequently, HNTs and Au nanohybrids were added to the composites as reinforcing elements. First, HNTs, as reinforcements with high specific surface area, thermostability, and strong mechanical properties, can bear and transfer loads efficiently, thus preventing severe fragmentation of 2240 Friction 11(12): 2238-2252 (2023) | https://mc03.manuscriptcentral.com/friction the resin matrix. Second, the HNTs retained in the composite matrix were gradually released onto the friction interface during rubbing. During sliding, crushed HNTs fragments were adsorbed onto the counterpart pin surface and oriented with their base planes in the rubbing direction, which was conducive to splitting the mineral lamellae along the cleavage planes under the action of shear forces. Meanwhile, the broken HNTs produced considerable amounts of reactive oxygen and Si-O dangling bonds, which resulted in tribo-oxidation and ceramic granules (SiO 2 ), accompanied by friction heat [25,26]. These ceramic phases and wear debris were compacted into a robust tribofilm that prevented direct contact between friction pairs, which resulted in effectively bearing a high load. Significantly, the design of the Au-HNTs core-shell hybrids prevented the aggregation of AuNPs in the composites. The counterface temperature increased because of frictional heat accumulation during the sliding process, thus softening the materials and degrading their mechanical properties. Thus, AuNPs with high thermal conductivity can improve heat dissipation and alleviate thermo-mechanical deformation to some extent. This work demonstrates an effective strategy for preparing HNTs-Au hybrids with excellent anti-wear behavior for various applications.

Pretreatment of halloysite
The HNTs were pretreated to remove impurities as follows: Raw HNTs were dispersed in deionized water (1 g/100 mL) and stirred for 12 h at 25 °C. After allowing the solution to stand for a certain time, it was divided into three layers. The upper supernatant was removed, the middle suspension was filtered, and the bottom sediments were discarded. This process was repeated three times to adequately purify the HNTs. Finally, the collected powder was dried at 60 °C for approximately 12 h.

Preparation of HNTs-Au nanopeapods
The preparation of the HNTs-Au nanopeapods is schematically represented in Fig. 1. HNTs (0.3 g) were dispersed in ethyl alcohol (40 mL) and sonicated for 20 min; subsequently, 40 mL toluene was added for another 20 min of sonication. Next, 10 mL oleylamine and 10 mL oleic acid were added to the solution respectively, and the mixture was stirred at 55 °C for 5 min; subsequently, 200 mg ascorbic acid was added while stirring continued for 2 min. In the process, the color of the solution gradually changed from golden yellow to black, thus indicating the formation of HNTs-Au nanocomposites. The obtained HNTs-Au solution was centrifuged and washed two times with a toluene and ethyl alcohol (1:1) mixture at 8,000 rpm for 3 min. To isolate free AuNPs, the product was sonicated twice in toluene and centrifuged at 3,000 rpm for 3 min twice. After this process, the precipitates were remained, and the supernatant was discarded every time. Finally, the precipitate was redispersed in 18 mL toluene for the next experiment.

Preparation of PBO/PTFE textile-polymer composites
In this study, PBO/PTFE textile-polymer composites were prepared by immersing the PTFE/PBO textile in a phenolic resin solution (10 g resin was dissolved in 70 mL solution consisting of ethyl alcohol, acetone, and ethyl acetate at a volume ratio of 1:1:1). The textiles were repeatedly immersed in resin solution until the proportion of resin in composites was 27 wt%. During immersion, HNTs and HNTs-Au fillers were dispersed in the resin solution to reinforce the composites. 0.05, 0.1, and 0.2 g HNTs were added to the resin solution, corresponding to S-h0 (no HNTs), S-h1 (0.5 wt% HNTs), S-h2 (1.0 wt% HNTs), and S-h3 www.Springer.com/journal/40544 | Friction (2.0 wt% HNTs), respectively. Next, 3, 6, and 9 mL of HNT-Au dispersions were added to the resin solution, corresponding to composites S-d3, S-d6, and S-d9, respectively. S-d0 is a blank sample without any filler.

Characterization
Scanning electron microscopy (SEM) images were obtained using a field-emission scanning electron microscope (FE-SEM，Thermo Scientific Apreo S). The cross section of the tribofilm was obtained using a dual-beam SEM/FIB instrument (Thermo Scientific Helios 5 UX). TEM images were obtained using a high-resolution transmission electron microscope (HR-TEM, TecnaiG2 TF20 S-TWIN, FEI). Dynamic light scattering (DLS) was used to measure the grain diameters of the HNTs using a Malvern Instruments Zetasizer Nano ZS particle size analyzer. X-ray diffraction (XRD) (Philips Corp.), X-ray photoelectron spectroscopy (XPS) (Thermo Fisher Scientificand.) and ultraviolet-visible (UV−vis) absorption spectra (Agilent Technologies Cary series UV-Vis Spectrophotometer) were utilized to systematically identify the successful loading of AuNPs in the lumen of HNTs. The Brunauer-Emmett-Teller (BET) surface areas and Barrett-Joyner-Halenda (BJH) pore volumes were determined using ASAP 2020 V4.03 (V4.03 E). Stress-strain curves were obtained using a universal material-testing machine (DY35). The heat conductivity coefficient of textile composites was measured via a TCi-3-A thermal conductivity analyzer at 25 °C. The thermal images were captured using a thermal imager (AS ONE 3-634-01). Thermogravimetric analyses were conducted in air using a PerkinElmer Pyris Diamond thermal analyzer. Wear tests were performed using a pin-on-disk tester with a 2-mm-pin diameter under 100 MPa and a sliding velocity of 0.37 m/s.

Characterization and analysis of HNTs-Au hybrids
As observed in SEM micrographs of the original HNTs shown in Figs. 2(a) and 2(b), the length of the HNTs ranged from 0.2 to 2 μm, and the average particle size was analyzed by DLS to be 1,070 nm ( Fig. 3(a)). The TEM images revealed that the pristine HNTs possessed a unique tubular structure with an inner and outer diameter of 20 and 60 nm, respectively (Figs. 2 (c), 2(d), and 2(g)), whereas the inset in the top     Fig. 3(b). UV−vis absorption spectra were obtained for different samples, as shown in Fig. 3(c), minimal absorption was observed in the visible region, and free AuNPs exhibited acharacteristic peak at 535 nm. The incorporation of HNTs and AuNPs shifted the absorption to higher wavelengths, thus suggesting the formation of HNTs-Au nanocomposites. The loading of AuNPs into the halloysite lumen was further confirmed by XPS, measurement as shown in Figs. 3(d)-3(f). The presence of Si 2p and Al 2p signals was attributed to HNTs, and Au 4f signals were detected in the XPS spectrum of HNTs-Au, which were associated with the loading of AuNPs. The high-resolution XPS spectrum of Au 4f shows two peaks at 83.7 eV and 87.5 eV binding energy (Fig. 3(e)). In addition, the O content in the HNTs was significantly higher than that in HNTs-Au, as shown in Fig. 3(f), which suggests that the OAc and OAm were not completely removed during washing. The typically BET surface area and pore structure of HNTs and HNTs-Au were analyzed by nitrogen adsorption-desorption isotherms. As shown in Fig. 4(a), the isotherm plots of HNTs and HNTs-Au were both type IV, thus indicating their micro-and mesoporous characteristics. The corresponding BET specific surface areas and pore volumes of HNTs and HNTs-Au were 60.24 m 2 /g and 30.50 m 2 /g, and 0.31 m 3 /g and 0.18 m 3 /g, respectively. As is well known, HNTs with a high aspect ratio can provide reinforcement to the polymer matrix (elastic modulus of HNTs is 140 GPa) [27]. Additionally, the interaction between the HNTs and the matrix was intensified because the polymer chains could enter the lumen of the HNTs to some extent. However, the specific surface area and pore volume declined after loading AuNPs, which is attributed to the fact that the lumens of the nanotubes were filled with metal nanoparticles. Figure 4(b) shows the BJH pore-size distribution plots of HNTs and HNTs-Au. The narrower distribution of the diameter at approximately 20 nm was attributed to the lumen of the nanotubes, and another peak of the pore volume versus pore width at 100 nm belonged to the pores among the tubes [28].

Tribological behaviors of composites
To gain insight into the dependence of the tribological behavior on the HNT content, every wear test was performed at least three times for linear composites with different HNT contents. As shown in Fig. 5(a), the addition of HNTs reduced the wear rate of pure composites from 0.93×10 -14 to 0.78×10 -14 m 3 /(N·m) when the HNTs content was up to 1.0 wt%. This data  indicates that HNTs with strong mechanical properties can provide reinforcement to polymer composites. However, the weak interaction (electrostatic interaction and van der Waals force) between the phyllosilicate interlayer favors the construction of a robust tribofilm, which prevents direct contact between the composites and steel pin, thereby resulting in mild to moderate abrasion and low wear loss. Presumably, poor HNTs (0.05 wt%) cannot effectively construct a continuous tribofilm to separate friction pairs. Additionally, agglomeration is likely when a superfluous filler (2.0 wt% HNTs) is present in the composites, which deteriorates the composites during rubbing. Figure  5(b) shows the evolution of the friction coefficient of the composites as a function of sliding time under the same conditions. Initially, the friction coefficients increased for all samples, and then leveled off and reached a steady state within a running time of approximately 5-6 min. Nevertheless, the friction coefficients of the pure composites (S-d0) were lower than those of the composites containing HNTs, which can be interpreted as follows. The pure composites suffered from severe abrasion compared with the HNT-reinforced composites, thereby resulting in more exposed PBO and PTFE fibers, particularly PTFE fibers with good lubrication, which favored friction reduction of the composites.
The morphology of the worn surface is essential for analyzing the tribological behaviors and wear mechanisms of the composites. Figure 6(a) shows the  www.Springer.com/journal/40544 | Friction SEM image of the original PBO/PTFE textile before immersion. After immersion in the resin solution approximately 14-16 times and high-temperature curing under 0.2 MPa, the composites were obtained and their surface morphology is shown in Fig. 6(b). The gaps and surfaces of the fibers were filled and covered with a resin layer with an obsolete cicatrix, as indicated by the square. When the pure composites experienced abrasion under a high load of 100 MPa, the superficial resin layer was squashed and worn off. The movement of the polymer exposed the fibers, which gradually fractured and were even pulled out during rubbing, as shown in Figs. 6(c)-6(f). In addition to numerous fractured fibers, considerable wear debris, derived from the friable resin and crushed fibers, was spread over the worn surface, as shown in the virtual ellipse marked in Fig. 6(e). In some areas, the bulky resin was removed from the composites, thus leaving large pits on their surfaces (Fig. 6(f)). At the edge of the wear scar, not only fractured fibers but also apparent phase separation between the resin and fibers were observed (Figs. 6(g)-6(i)). The sophisticated factors, including poor adhesion and different moduli between the resin and fibers, can induce phase separation. In addition, the brittleness and poor toughness of the cured resin matrix may be responsible for the observed phase separation. Thus, in conclusion, abrasive wear is the dominant form of wear failure for pure composites, given the large amount of wear debris and fractured fibers that act as grinding materials during sliding.
To explore the synergistic effect of HNTs-Au hybrid on wear resistance for composites, wear tests were conducted under 100 MPa loading at 0.37 m/s and the results are shown in Fig. 7. In particular, the wear rate of the S-d6 sample decreased by 27.9% compared with pure composites (from 0.93×10 -14 to 0.67×10 -14 m 3 /(N·m)). These results indicate that the combination of HNTs and Au should improve the anti-wear behavior of the composites. Typical friction curves of the composites with respect to time are shown in Fig. 7(b). Although all curves exhibited similar running stages and fluctuations, the friction coefficients of the HNTs-Au-reinforced composites were always higher than those of the pure composites. This is because the S-d6 specimen with mild to moderate abrasion has an intact surface topography, and few fibers participate in the lubrication.
In agreement with the mild-to-moderate wear of the composites containing the HNTs-Au hybrids, the worn surface remained intact (Figs. 8(a)-8(f)). First, the wear scar depth of the S-d6 sample was significantly lower than that of S-d0 ( Fig. 8(a)), which was consistent with its low wear loss. This result is also indicated by the three-dimensional profiles of the composite worn surface shown in Figs. 8(g) and 8(h). The average scar depth was approximately 49 μm for S-d0, whereas it decreased to 35 μm for S-d6, which indicates that S-d0 suffered much more damage during sliding. Second, the worn surface of S-d6 was smoother than that of S-d0, and only a few fractured fibers were observed. Instead of widespread severely fractured fiber bundles, grooves, delamination of resin, and a few extruded fibers (dotted arrows in Figs. 8(b) and 8(d)) were prevalent on the worn surfaces. In addition, the fuzzy boundary of the wear scar coincided well with the superior wear resistance of the S-d6 composites (Figs. 8(e) and 8(f)), thus suggesting that the addition of HNTs-Au improves the loading capacity and mechanical properties of the polymer matrix. The stress-strain curves of the composites before and after reinforcement prove that the HNT-Au hybrids  indeed improved the tensile strength and failure strain ( Fig. 10(a)). The moduli of PBO/PTFE textile, S-d0, and S-d6 were 0.94, 1.94, and 2.19 GPa, respectively. Notably, high specific surface areas and the penetration of polymer chains into the lumen can prevent severe fragmentation of the resin matrix. Thus, almost no severe crack propagation was observed in the S-d6 composites except for some delamination of the resin. In some areas, individual local-regional crack growth resulted in resin crumbling and even debonding between fibers and the resin matrix, namely phaseseparation, as shown in Fig. 8(c). In this area, the exposed fibers were visualized with integration rather than fracturing, and adhesive wear and fatigue wear were suggested to be the main wear mechanisms.
As is well known, the tribofilm and friction heat play vital roles in the wear of polymer composites. First, the superficial resin layer of the S-d6 composite was broken and removed. Simultaneously, the HNTs-Au hybrids reserved in the matrix were released onto the frictional interface and formed a protective coating on the steel pin surface with dynamic balancing. Figure 9 shows a comparison of the morphologies of the counterpart pin before and after sliding at 100 MPa for 120 min. The steel pin was polished using abrasive paper (800 mesh) for 2 min at 12.5 MPa, and then scrubbed with ethanol to reduce system error before every wear test. As illustrated in Figs. 9(a) and 9(b), plow grooves were spread all over the polished pin surface. Figures 9(j)-9(l) show that a protective film covered the pin surface after rubbing with the S-d6 composites. Furthermore, EDS elemental mapping analyses demonstrated the existence of the tribofilm in Figs. 9(c)-9(i). Evidently, Al, Si, Au, and O pervaded the entire pin surface, thus indicating a homogeneous coverage originating from the HNT-Au hybrids. The distribution of C was derived primarily from the packed resin debris.
Generally, most frictional power is converted into thermal energy during wear, thus resulting in an increase in temperature of the contact area. The generated frictional heat typically softens the polymer www.Springer.com/journal/40544 | Friction matrix and intensifies wear loss owing to its sensitivity to high temperatures. Clay minerals possess high thermal stability as depicted in Fig. 10(c), and the comparatively low loss of quality for HNTs is attributed primarily to the removal of adsorbed water, zeolitic water, and even hydroxyl groups in the structure [29] when the temperature was raised from 25 to 700 °C at a heating rate of 10 °C/min. Thus, the incorporation of HNT-Au hybrids restricts the movement of molecular chains to a certain extent and positively affects the thermal stability of the composites. In addition, establishing a rational thermally conductive network is an effective means for evacuating concentrated heat to mitigate softening of the polymer matrix. The uniform loading of AuNPs into the lumen of HNTs improved the thermal conductivity of the composites, as depicted in Fig. 10(e). In particular, the addition of 6 mL HNTs-Au hybrids solution led to a 37%   increase in the thermal conductivity coefficient, from 0.075 to 0.103 W/(m·K), compared with raw composites. Although more incorporation of HNT-Au hybrids results in a higher thermal conductivity, superfluous reinforcement deteriorates the wear resistance of the composites. Figure 10(f) shows the thermal images of the S-d0 and S-d6 composites at the same intervals; evidently, the thermal transmission speed of S-d6 was higher than that of S-d0, thus indicating that S-d6 can effectively evacuate the concentrated heat.

Tribofilm and wear mechanisms
Tribofilms are significant in anti-wear and frictionreducing mechanisms. The abovemotioned SEM and optical analyses shown in Fig. 9 do not effectively illustrate the nanostructures of the tribofilm. Therefore, FIB-TEM characteristics were employed to reveal the cross-section of the transfer film on the pin counterface after sliding with S-d6 composites. Figures 11(a) and 11(b) show an adaptive silicate mineral layer of approximately 100 nm, which is porous and resembles the ceramics [30] deposited on the pin surface. The EDS elemental maps and line scanning analyses confirmed the fine silicate coating, as shown in Figs. 11(c) and 11(d), respectively. The distributions of O, Si, and Al were derived from the HNTs, whereas Fe was distributed primarily in the steel substrate. During the rubbing process, the mineral particles reserved in the matrix were released onto the friction interface and their base planes gradually became oriented with the rubbing direction. The weak interaction between the layers of HNTs facilitated splitting along the laminar structure under the action of shear forces, and this constructed an effective wear-resistant protective layer on the counterface. In addition, the destruction of the crystal structure of HNTs produced numerous reactive oxygen atoms and dangling Si-O bonds [31]. Consequently, a loose iron oxide layer adjacent to the pin was constructed when these active oxygen atoms and Si-O dangling bonds encountered activated iron atoms caused by friction heat and extraordinary pressure, as shown in Figs. 11(e) and 11(h). Figure 11(g) shows a high magnification image of iron oxide, where the lattice fringes of 0.33 and 0.24 nm correspond to the crystal faces (022) and (021) of Fe 3 O 4 and FeO, respectively. A high flash temperature and pressure may induce the tribo-oxidation of minerals and the appearance of SiO 2 ceramic particles, as shown in Fig. 11(f) [25,26]. A protective mineral layer with a higher load-carrying capability that improved the abrasion resistance of the composites was obtained and continuously replenished during sliding. Figure 12 illustrates the friction-reduction and anti-wear mechanisms of the HNT-Au hybridreinforced PBO/PTFE textile composites. Polymer molecular chains entangled with HNTs-Au physically infiltrated the lumen of the HNTs, thereby increasing the strength of the composites. When the composites were subjected to a high applied load, most of the stress was transferred from the resin matrix to HNT-Au, and the elastic modulus and tensile strength were higher than those of the resin phase. During sliding, the accumulated plastic deformation broke the critical deformation capacity of the composites and generated cracks. When the growing cracks met the HNTs-Au, the hybrids acted as crack bridges and consumed increasing energy until the filler broke. With increasing applied stress, partial HNTs-Au were pulled from the matrix, and consequently, its strengthening effect disappeared consequently. In addition, partial HNT-Au acting as a barrier induced microcrack deflection and prevented crack propagation, as shown in Fig. 12.
During the rubbing process, the hard steel pin removed the superficial resin of the composites and the HNT-Au reinforcement was gradually exposed to the friction interface. Moreover, they experienced a period of distribution with their base planes in the rubbing orientation, essentially favoring the splitting of silicate minerals along the cleavage planes under severe tangential forces [31]. The HNTs were exfoliated and adsorbed onto the counterface, thereby forming an adaptive self-organization tribofilm that was porous and resembled a ceramic silicate mineral [30]. The design of a halloysite-gold core-shell nanosystem fully utilized the synergistic effects, thus avoiding severe agglomeration of AuNPs in the polymer matrix. The HNT-Au hybrid was gradually exposed on the sliding surface as friction and wear proceeded. In addition, soft metallic nanoparticles with low shear strength smeared onto the contact area, thereby forming a ductile layer that contributed to the achievement of moderate abrasion in composites.

Conclusions
Overall, AuNPs were deposited onto the lumen of the HNTs, essentially forming a halloysite-gold core-shell nanosystem by in situ reduction, with uniformly distributed AuNPs with a narrow size range of 6-9 nm. Reportedly, the incorporation of HNTs-Au hybrids reduced the wear rate of PBO/PTFE textile composites from 0.93×10 -14 to 0.67×10 -14 m 3 /(N·m) under a high load of 100 MPa and linear velocity of 0.37 m/s. The entanglement between the molecular chains and HNTs enhanced the mechanical properties and load-bearing capacity of the polymer matrix. Significantly, the HNTs-Au hybrids eliminated the conjugation of AuNPs and synergistically improved the thermal conductivity coefficients of the composites. Most importantly, the weak interaction between the layers of HNTs facilitated splitting along the laminar structure under the action of shear forces, constructed an effective protective layer of approximately 100 nm on counterface-filling asperities, and effectively prevented direct contact between frictional pair surfaces. Meanwhile, the destruction of the crystal structure of the HNTs released abundant reactive oxygen and Si-O dangling bonds, thus leading to the formation of SiO 2 , FeO, and Fe 3 O 4 . This investigation sheds light on reducing the wear loss of textile composites for potential industrial applications.