Designing Butterfly‐Inspired Architecture in Self‐Assembled Hybrid Composites for Synergistic Structural Thermal Management

ABSTRACT

Increasingly strict environmental regulations and sustainability goals are driving global adoption of lightweight materials in advanced transportation and defense applications. Inspired by nature's unparalleled engineering, this work employs butterfly‐hierarchical architectures to develop hybrid composites that emulate the synergy‐induced multifunctional performance of natural materials. Specifically, these composites are reinforced with hierarchical fibrous assemblies comprised of nano‐sized graphene nanoplatelets (GnPs) covalently bonded onto micro‐sized glass fibers (GFs). In detail, this work showcases a novel approach to control the in situ self‐assembly behavior of these reinforcements, achieved by tailoring the GFs' surface chemistry through functionalization, to maximize the density of covalently‐bonded GnPs. Compared to the current industrial substitute for metallic structural components, the developed hybrid composites are tailorable to achieve improvements up to 29%, 116%, and 109% in specific tensile strength, flexural strength, and impact strength, respectively, as well as 22%, 42%, and 110% in thermal conductivity, thermal management performance, and processability, respectively, with an overall 18% weight‐reduction. These advancements stem from the detailed structure–property designs, spanning across multiple length scales, encompassing crystal polymorphism, fiber alignment, and distribution, imparting a fundamental understanding of how to tune material performance to meet stringent requirements. Ultimately, these cost‐effective, industry‐ready butterfly‐inspired hybrid composites can produce lightweight, multifunctional components, demonstrating the potential of hierarchical architecture in advancing sustainable engineering solutions for a greener future.

Pioneering a butterfly‐inspired design strategy, this work engineers hybrid composites with hierarchical fibrous architectures to achieve synergistic performance. By translating biological principles into scalable composite architectures, it introduces a novel pathway to lightweight, multifunctional, and cost‐effective materials. The result, a transformative approach to structural thermal management, where biomimetic design serves not merely as inspiration, but as a blueprint for innovation.

1. Introduction

Over billions of years of evolution, nature has perfected the resource‐efficient synthesis of high‐performance, adaptive materials and structures that achieve exceptional strength, toughness, and multifunctionality while remaining lightweight. A defining characteristic of natural materials is their sophisticated hierarchical architecture, whereby structural assemblies span multiple length scales from the molecular to the macroscopic level [1, 2, 3, 4, 5, 6]. These multi‐scale assemblies enable material constituents to interact synergistically, enhancing overall performance beyond the capabilities of the individual constituents (or phases). In general, biological materials such as nacre [7, 8, 9, 10, 11], bone [12, 13, 14, 15, 16, 17, 18, 19, 20, 21], and insect cuticles and shells [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32] seamlessly integrate organic and inorganic components, leveraging intricate microstructural arrangements such as staggered brick‐and‐mortar structures, mineralized collagen fibrils, and chitin‐protein networks to surpass the mechanical performance of their constituents. These architectures enable remarkable damage tolerance through mechanisms like crack deflection, energy dissipation, and load redistribution, ensuring both durability and adaptability in dynamic environments.

Beyond mechanical performance, these biological materials exhibit inherent multifunctionality, such as self‐healing in bone [12, 13, 14, 15, 16, 17, 18, 19, 20, 21], hydrophilicity (water retention) or hydrophobicity (water repellency) in plant tissues [33, 34, 35, 36, 37, 38, 39], and advanced thermal management in natural shells and scales [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]. Many biological structures actively regulate temperature through mechanisms like heat dissipation, insulation, or infrared reflection, as seen in butterfly wings, polar bear fur, and mollusk shells. These thermal adaptations allow organisms to maintain structural integrity and function across diverse environmental conditions, further demonstrating the intricate interplay between material architecture and performance. The ability of natural materials to achieve such extraordinary properties via controlled structural hierarchy has long inspired the development of bio‐inspired composites, offering a blueprint for designing next‐generation materials with tunable mechanical performance and expanded functional capabilities.

Building on these principles, industries are increasingly embracing nature for inspiration as they seek innovative solutions to modern engineering challenges. Stringent environmental policies and growing sustainability demands are accelerating the shift toward advanced materials in the automotive, aerospace, defense, and high‐performance transportation sectors. The global pursuit of net‐zero emissions by 2050 has intensified the demand for lightweight, multifunctional materials, making them an essential component in reducing energy consumption and emissions. For instance, a 20% reduction in structural weight can lead to a 15% increase in fuel efficiency, significantly extending the operational range of both aircraft and automobiles [40, 41, 42]. Specifically, in the automotive sector, light‐weighting plays a central role in enhancing fuel economy, both reducing greenhouse gas emissions for traditional petroleum‐fueled vehicles and maximizing the driving range of electric vehicles.

However, despite significant efforts, researchers have struggled to formulate truly synergistic synthetic materials due to a limited fundamental understanding of hierarchical interactions, the absence of scalable fabrication strategies, and the challenge of precisely tuning interfacial properties. Specifically, the inadequate interaction between material constituents (or phases) in composite materials remains a significant challenge, with poor matrix‐filler adhesion leading to stress concentrations and premature failure under mechanical loads [43, 44], via mechanisms like interfacial debonding and filler pull‐out. Without carefully designed microstructural architecture and optimized phase interactions, synthetic composites struggle to replicate the robustness, adaptability, and multifunctionality that nature effortlessly achieves. Recognizing these challenges, researchers are leveraging hierarchical hybrid composites inspired by nature's intricate design strategies to develop materials with improved structural integrity and advanced thermal regulation capabilities.

For example, Aguiar et al. introduced cylindrical core–shell hybrid composites that emulate the hierarchical architecture found in osteons and bamboo, utilizing nano‐sized halloysite nanotubes (HNTs) electrostatically self‐assembled onto micro‐sized glass fibers (GFs) in a semi‐crystalline polymer matrix. Through trans‐crystallization‐encapsulation, the polymer matrix nucleates and grows around this hierarchical assembly, strengthening interfacial interactions and imparting structural integrity, which enhances mechanical performance, optimizes thermal management (insulation/stability), improves industrial processability, and reduces flame propagation speed. Sansone et al. [45] recently introduced butterfly‐inspired hybrid composites that emulate the multifunctional architecture of Ulysses butterflies by integrating hierarchical fibrous assemblies with a macro‐architectural sandwich‐structure. Specifically, nano‐sized graphene nanoplatelets (GnPs) covalently bonded to GFs mimic the structural adaptations of butterfly legs, while an alternating rigid and porous layered design replicates the thermal and mechanical benefits of butterfly wings, enhancing overall performance.

Although nature‐inspired designs have demonstrated promising multifunctional performance, many challenges persist in precisely tuning hierarchical architecture, optimizing interfacial interactions, and scaling fabrication processes to ensure consistent and reproducible material properties. Our previous efforts [46] established a novel lab‐scale approach to methodically control the self‐assembly behavior of hierarchical fibrous assemblies by tailoring the GF's surface chemistry to enhance their chemical and thermodynamic compatibility with the GnPs. However, translating these design principles into scalable, reproducible processes remains a major challenge, particularly in accommodating elevated fiber concentrations required for high‐performance applications, and in achieving reproducibility across varying reinforcement contents, where the interplay between reinforcement surface chemistry and bonding saturation concentration becomes critical. Specifically, this work formulates a fundamental understanding of how to design biomimetic architectures with maximum saturation bonding density, governed by surface chemistry, concentration, and crystal polymorphism, for optimal multifunctional performance. Additionally, expanding functionality beyond mechanical properties to enhance processability and thermal management remains an unmet need, which this work directly addresses.

For the first time, this work presents a scalable in situ approach to engineering butterfly‐inspired hierarchical hybrid composites that replicate the structural architecture of biological materials, leveraging their inherent resilience and thermal management capabilities. Through a comprehensive analysis of structure–property relationships across multiple length scales, encompassing crystal polymorphism, fiber alignment, and distribution, this study advances the understanding of synergistic mechanisms, allowing precise control over hierarchical architectures for optimized performance. As a result, our newly developed hybrid composites are tailorable to surpass the mechanical properties of the current industrial standard for automotive structural components (PPGF60, 40 wt.% PP/60 wt.% GF) [47, 48], exhibiting improvements up to 29%, 116%, and 109% in tensile strength, flexural strength, and impact strength, respectively, as well as 22%, 42%, and 110% in thermal conductivity, thermal management performance, and processability, respectively, with an overall 18% weight reduction. Ultimately, these butterfly‐inspired hybrid composites offer a cost‐effective, scalable solution for lightweight, multifunctional components, streamlining material efficiency while enhancing performance, with transformative potential across sustainable transportation, defense, and other industries.

2. Results and Discussion

2.1. Preparation and Characterization of the Self‐Assembled Butterfly‐Inspired Hybrid Composites

The butterfly‐inspired hybrid composites were fabricated using industry‐scale thermoplastic injection molding under processing conditions designed to promote the in situ self‐assembly of nano‐sized GnPs onto chemically tailored micro‐sized GFs within a polypropylene matrix. Central to this design strategy is the deliberate modification of the GF surface chemistry, which increases the density and stability of interfacial bonding and enables the formation of hierarchical fibrous assemblies that function as multifunctional reinforcements within the hybrid composites. The hierarchical self‐assembly observed in these materials arises from the coupled effects of interfacial chemistry and processing‐induced deformation. Tailored GF surface functionalization introduces non‐protonated amine groups within the fiber sizing, which react with carboxylic acid groups on the GnP surface to form amide bonds. Concurrently, high shear and extensional deformation during injection molding, together with the volume exclusion effect imposed by the micro‐sized GFs, restrict GnP mobility in the vicinity of the fiber surface. This confinement promotes preferential localization at the interface and facilitates the development of a continuous, hierarchically structured fibrous assembly. As a result, the butterfly‐inspired architecture is not only biologically motivated, but is actively engineered through designed surface chemistry.

The resulting hierarchical fibrous assembly, illustrated in Figure 1, produces a hierarchical architecture analogous to that found in butterfly legs. In this natural system, nano‐sized spurs and cuticular features encapsulate micro‐sized cylindrical segments such as the coxa, trochanter, femur, tibia, and tarsus [49, 50]. In the synthetic analogue presented here, the GnPs replicate the role of these nano‐scale outer features, while the GFs serve as the primary load‐bearing skeletal elements. In butterflies, this hierarchical organization enables synergistic multi‐functionality by enhancing mechanical robustness, shock energy absorption, and thermal regulation. The nano‐textured outer layers contribute to maintaining body temperatures within an optimal range, across diverse environments, supporting stable aerodynamic performance [24, 51]. By emulating this architecture, the hybrid composites developed in this work similarly integrate structural reinforcement with enhanced thermal transport and thermal management capabilities. Accordingly, the butterfly‐inspired hybrid composites presented here demonstrate how tailoring GF surface chemistry to control in situ self‐assembly can be leveraged to engineer hierarchical architectures that deliver synergistic improvements in mechanical performance, thermal conductivity, and processability.

FIGURE 1.

Illustration of the natural architectures found in Ulysses butterflies alongside the designed synthetic hierarchical structures of the hybrid composites, adapted with permission [45].

Recognizing that the multi‐scale structure of engineered composites dictates their intrinsic properties, it is crucial to examine the effects of composite constituent interfacial interactions, crystalline microstructure, as well as morphological architecture in relation to surface chemistry (functionalization), reinforcement concentration, orientation, and distribution. XPS was conducted on the biphasic and hybrid composites, as shown in Figure 2, to understand the effects of sizing composition on the microstructural architecture of composites. For the biphasic composites with non‐functionalized GFs, the survey spectra show no discernible N1s peak, confirming the absence of amino‐silane surface functionalization. Further, this is confirmed in high‐resolution C1s spectra where only three deconvoluted peaks are observed: C─C or C═C bonds (∼284.6 eV), C─O bonds (∼286.8 eV), and C═O bonds (∼287.7 eV). In contrast, for the biphasic composites with functionalized GFs, regardless of sizing composition, a distinct N1s peak is observable in the survey spectra. This indicates the successful incorporation of amine functional groups onto the GF surface through siloxane grafting. During chemical modification, the silane terminations of the aminosilane hydrolyze to form silanol groups, which then condense onto the hydroxyl groups on the GF surface, creating stable siloxane networks [52, 53, 54].

FIGURE 2.

XPS survey spectra of biphasic and hybrid composites, high‐resolution C1s and N1s spectra, SEM micrographs, and corresponding schematics for composite with (A) Sizing 0, (B) Sizing 1, (C) Sizing 2, and (D) Sizing 3.

Additionally, in the high‐resolution C1s spectra, the same three deconvolutions are detected, with the addition of C─N bonds (∼286.1 eV), further confirming the successful silanization of the GFs.

To evaluate the efficacy of each sizing composition on introducing non‐protonated amine groups onto the GF's surface, the sizing efficiency (${\mathcal{S}}_E$) was estimated, as presented in Table 1. The ${\mathcal{S}}_E$ is determined by evaluating the percentage of non‐protonated amine bonds to the total N1s spectra (protonated and non‐protonated amine) [55] in the biphasic composites, whereby a higher ${\mathcal{S}}_E$ indicates a greater proportion of non‐protonated amine functional groups. For the GFs functionalized with sizing 1, the ${\mathcal{S}}_E$ is relatively low at 63%, as this sizing composition only contains primary amines, which are more reactive and susceptible to undergoing protonation [55]. In contrast, for the GFs functionalized with sizing 2 and sizing 3, a significantly greater ${\mathcal{S}}_E$ is determined (${\mathcal{S}}_E\cong 72\%$), due to the presence of secondary amines within a longer carbon chain (butyl in sizing 2, aminohexyl in sizing 3), which create a steric effect that inhibits amine protonation [55]. Notably, these findings are consistent with ex situ analyzed GFs [46], indicating that the ${\mathcal{S}}_E$ is not adversely affected by the presence of the matrix during composite fabrication. Thus, the proportion of available bonding sites depends on the sizing composition and reinforcement concentration, highlighting the importance of understanding their interplay to maximize synergy among constituents in hybrid composites.

TABLE 1.

Tabulated covalent bond proportions for the high‐resolution C1s and N1s XPS spectra.

Sample name	C1s bond proportions (%)				N1s bond proportions (%)
	C − C	C − N	C − O	C = O	${\mathrm{NH}}_3^{+}$	NH2	N − C(O)	${\mathcal{S}}_E$
PPGF40 – S0	91	0	2	7	0	0	0	0	0
PPGnP0.5GF40 – S0	93	0	2	6	0	0	0	0	0
PPGnP0.75GF40 – S0	93	0	2	6	0	0	0	0	0
PPGnP1GF40 – S0	92	0	2	6	0	0	0	0	0
PPGF40 – S1	87	6	2	5	38	63	0	63	0
PPGnP0.5GF40 – S1	82	13	2	4	23	17	61	—	58
PPGnP0.75GF40 – S1	81	13	2	4	23	16	60	—	57
PPGnP1GF40 – S1	82	13	2	4	23	17	61	—	58
PPGF40 – S2	84	9	2	5	28	72	0	72	0
PPGnP0.5GF40 – S2	77	18	2	4	18	18	64	—	64
PPGnP0.75GF40 – S2	68	27	2	4	14	7	79	—	79
PPGnP1GF40 – S2	69	26	2	4	14	7	78	—	79
PPGF40 – S3	83	10	2	5	29	71	0	71	0
PPGnP0.5GF40 – S3	78	17	2	3	19	22	58	—	62
PPGnP0.75GF40 – S3	73	22	2	3	16	14	70	—	70
PPGnP1GF40 – S3	73	22	2	3	16	13	71	—	71

${\mathcal{S}}_E$, sizing efficiency from XPS; , bonding density from XPS.

As such, XPS was conducted on the hybrid composites with non‐functionalized and functionalized GFs to evaluate the amount of covalently bonded GnPs at varying GnP concentrations, as shown in Figure 2 and tabulated in Table 2. For the hybrid composites with un‐sized GFs (S0‐GFs), the survey spectra, with no discernible N1s peak, imply an absence of covalent bonding between the GnPs and GFs, regardless of GnP content. In contrast, the high‐resolution N1s spectra for the hybrid composites with functionalized GFs show that protonated (∼401.3 eV) amine and non‐protonated (∼398.4 eV) amine functional groups are present from the sizing composition, along with the emergence of a new chemical bonding peak corresponding to amide bonds (∼399.5 eV). This additional amide peak confirms the successful formation of a hierarchical fibrous assembly, as the carboxylic acid groups on the GnPs surface have reacted with the non‐protonated amine groups on the GFs sizing, producing amide bonds. Thus, the amount of GnPs capable of covalently bonding onto the functionalized GFs is directly correlated to the amount of non‐protonated amine groups on the GFs sizing composition. As a result, the bonding density (), representative of the amount of covalently bonded GnPs onto the functionalized GFs surface, is determined by evaluating the percentage of amide bonds to the total N1s spectra (i.e., amide, non‐protonated amine, and protonated amine bonds), as shown in Table 1 [55].

TABLE 2.

Tabulated parameters extracted from the DSC Thermograms and XRD Diffractograms.

Sample name	T c (°C)	T m (°C)	ΔT (°C)			${\mathcal{K}}_{\alpha }$ (%)	(%)
PPGF40 – S0	122.8	163.7	40.9	51	50	94	5.9	1.2
PPGnP0.5GF40 – S0	124.5	163.6	39.1	55	54	88.9	11.1	1.2
PPGnP0.75GF40 – S0	125.6	163.6	37.9	56	55	91.2	8.9	1.1
PPGnP1GF40 – S0	125.9	163.8	37.9	54	55	92.1	7.7	1.2
PPGF40 – S1	122.9	163.6	40.7	52	53	93.5	6.2	1.2
PPGnP0.5GF40 – S1	124.7	163.7	38.9	54	54	81.4	14.9	2
PPGnP0.75GF40 – S1	125.7	163.8	38.1	56	55	88.6	13.4	1.9
PPGnP1GF40 – S1	125.9	164.1	38.2	55	56	88.5	11.5	1.8
PPGF40 – S2	123.2	163.5	40.3	51	52	93.4	6.3	1.2
PPGnP0.5GF40 – S2	124.6	163.8	39.2	56	55	91.9	9.9	2.6
PPGnP0.75GF40 – S2	125.9	164.1	38.2	57	58	80.8	21.1	3.1
PPGnP1GF40 – S2	126.2	164.3	38.1	57	57	82.7	19.3	2.9
PPGF40 – S3	123.1	163.5	40.4	51	51	93.1	6.1	1.1
PPGnP0.5GF40 – S3	124.9	163.6	38.7	55	55	90.9	9.2	2.3
PPGnP0.75GF40 – S3	126.1	163.7	37.6	57	57	83.1	16.9	2.5
PPGnP1GF40 – S3	126.4	164.3	37.9	56	57	84.4	15.6	2.6

T_c, melting temperature from DSC; T _m , melting temperature from DSC; ΔT, degree of supercooling from DSC; , percent crystallinity from DSC; , percent crystallinity from XRD; ${\mathcal{K}}_{\alpha }$, fraction of α‐crystal formed from XRD; ${\mathcal{K}}_{\beta }$, fraction of β‐crystal formed from XRD; ${\mathcal{T}}_r$, degree of trans‐crystallization from XRD.

For the hybrid composites with sizing 1 GFs (S1‐GFs), the greatest (58%) was achieved with 0.5 wt.% GnP, whereby increasing the GnP content had no effect on the creation of additional amide bonds. This suggests that the reaches saturation at 0.5 wt.% GnP, with excess (or additional) GnPs unable to participate in the hierarchical fibrous assembly, instead dispersing and potentially agglomerating within the polymer matrix. Moreover, the relative percentage of C─N bonds within the high‐resolution C1s spectra remains unchanged with increasing GnP content, supporting the hypothesis of bond saturation.

For the hybrid composites with sizing 2 GFs (S2‐GFs), the with 0.5 wt.% GnP is greater than that of the composites with S1‐GFs, aligning with the ${\mathcal{S}}_E$ results. This emphasizes the crucial role of the sizing composition in effectively introducing non‐protonated amine groups onto the GF surface to facilitate amide bond formation. Increasing the concentration to 0.75 wt.% GnP effectively increases the to 79%, as the available non‐protonated amine groups have not yet been fully utilized. However, further increasing the concentration to 1 wt.% GnP had no effect on the formation of additional amide bonds, indicating that saturation was achieved at 0.75 wt.% GnP, with the excess dispersing within the polymer matrix. Furthermore, the relative C─N bond percentage increases with GnP content up to 0.75 wt.% before plateauing, supporting the concept that for this sizing composition, amide bond formation reaches saturation at this concentration.

For the hybrid composites with sizing 3 GFs (S3‐GFs), the with 0.5 wt.% GnP is similar to that of the composites with S2‐GFs, showing direct correlation with their similar ${\mathcal{S}}_E$ results. However, increasing the concentration to 0.75 wt.% GnP results in a relatively lower (70%) than in the composites with S2‐GFs, despite initially comparable ${\mathcal{S}}_E$ values. This trend is mirrored in the C─N bond percentage, which is lower in S3‐GFs composites than in S2‐GFs composites. This suggests that steric hindrance from the hexyl group in S3 may partially restrict the bonding of non‐protonated amine groups with the GnPs. Likewise, increasing the concentration further to 1 wt.% GnP does not facilitate additional amide bond formation, indicating that saturation occurs at 0.75 wt.% GnP, with the excess dispersing within the polymer matrix. As a result, it is concluded that optimized hierarchical fibrous assemblies are achieved with 0.5 wt.% GnP for the composites with S1‐GFs, and 0.75 wt.% GnP for the composites with S2‐GFs and S3‐GFs.

The crystalline microstructure of the composites was analyzed using DSC and XRD to elucidate the influence of the hierarchical fibrous assemblies on matrix crystallinity and polymorphism. DSC crystallization and melting thermograms for the biphasic and hybrid composites are presented in Figure 3A,B, with the extracted parameters tabulated. For the biphasic composites (PPGF40), the crystallization temperature (T _c) remains nearly constant (∼123°C), regardless of surface functionality. In contrast, for the hybrid composites, regardless of GF surface functionality, the introduction of GnP leads to an increase in the T _c, signifying that the crystallization of PP was accelerated in the presence of GnP. The increase in T _c reduces the degree of supercooling (ΔT  =  T _m  −  T _c), which is directly proportional to the free energy of melting. Consequently, crystallization occurs with a lower thermodynamic driving force, as the GnPs serve as heterogeneous nucleation sites for crystallization. Regardless of GF surface functionality, it appears that increasing the GnP concentration increases the crystallization temperature of hybrid composites, supporting that GnPs can act as heterogeneous nucleation sites within the PP matrix. While the crystallization thermograms are indicative of unimodal crystallization exotherms, the melting thermograms exhibit a bimodal pattern for all composites, regardless of GF surface functionality, suggesting the existence of crystal polymorphs beyond the predominant α‐form. As a result, XRD was conducted on the fabricated composites to analyze the effect of the hierarchical fibrous assembly on the crystalline microstructure and polymorphism.

FIGURE 3.

(A) DSC crystallization thermogram and (B) DSC heating thermograms at 10°C/min, and (C) XRD diffractograms, for the fabricated composites, as well as (D) Crystal polymorphism, and (E) Degree of trans‐crystallization, ${\mathcal{T}}_{\mathit{r}}$, and Bonding density, ${\mathcal{B}}_{\mathit{d}}$, as a function of reinforcement concentration. POM images for (F) PPGnP0.75GF‐S0, (G) PPGnP0.75GF‐S1, (H) PPGnP0.75GF‐S2, and (I) PPGnP0.75GF‐S3, showcasing the variation in crystalline microstructure of the matrix, favouring trans‐crystallization with higher bonding density, ${\mathcal{B}}_{\mathit{d}}$.

Qualitatively analyzing the XRD diffractograms shown in Figure 3C, three distinct phenomena are observable: (1) the introduction of the (002) _GnP crystallographic plane intrinsic to GnPs, (2) the introduction of the (300)_β crystallographic plane, indicative of the presence of β‐crystals, and (3) the intensity trade‐off between the (040)_α and (110)_α crystallographic planes which are affiliated with the mechanisms of trans‐crystallization. The presence of the (300)_β crystallographic plane corroborates the presence of β‐crystals within all fabricated composites, directly correlating with the bimodal pattern observed in the DSC melting thermograms, as the melting temperature of β‐crystals is approximately 10.9°C lower than that of α‐crystals [37]. Since β‐crystals are recognized for their superior mechanical properties, such as toughness, tensile strength, elongation at break, and impact strength, compared to α‐crystals, it becomes imperative to quantify their formation and influence on the overall performance of the composites. Thus, the Turner‐Jones method [56, 57] was utilized to determine the total fraction of β‐phase formed (${\mathcal{K}}_{\beta }$) within the composites, as shown in Figure 3D and tabulated in Table 2.

For biphasic composites, with and without functionalization, the ${\mathcal{K}}_{\beta }$ is ∼6.1%, indicating that the presence of GF has no added effect on inducing β‐crystals, regardless of sizing composition. Additionally, the percent crystallinity for these composites (, ) remained unaltered at ∼51%. For the hybrid composite with un‐sized GFs (S0‐GFs), the ${\mathcal{K}}_{\beta }$ increased to 11.1% with the addition of 0.5 wt.% GnP, due to their heterogeneous β‐nucleation efficiency [46, 58, 59, 60, 61, 62, 63]. However, further increasing the GnP content resulted in a decreased ${\mathcal{K}}_{\beta }$, potentially due to the reduced dispersion of the GnPs, thereby forming agglomerates due to the strong π–π interactions and van der Waals forces. Thus, the β‐nucleation efficiency is sacrificed, as the aspect ratio of the GnPs is reduced, leading to fewer effective sites for heterogeneous nucleation. This behavior aligns well with findings of Aguiar et al. [55], whereby it was demonstrated that the unbonded GnP in the matrix can act as nucleation sites for β‐crystal formation, so long as agglomeration is mitigated.

For the hybrid composite with S1‐GFs, a similar behavior to the hybrid composites with S0‐GFs is observed, but with a greater maximum ${\mathcal{K}}_{\beta }$ occurring at 0.5 wt.% GnP (${\mathcal{K}}_{\beta }=$14.9%). Since a significant portion of the GnPs are covalently bonded onto the GFs, this minimizes the amount of unbonded GnPs in the matrix, thereby mitigating agglomeration. As a result, the β‐nucleation efficiency of the GnPs is promoted as the aspect ratio of the GnPs is preserved, leading to more preferred sites for heterogeneous nucleation. However, further increasing the GnP content reduces the ${\mathcal{K}}_{\beta }$, as the has surpassed saturation, forcing the excess GnPs to agglomerate in the matrix, thereby reducing their β‐nucleation efficiency. Nevertheless, the use of S1‐GFs at all GnP concentrations examined led to an obvious increase in ${\mathcal{K}}_{\beta }$ relative to hybrids with S0‐GFs.

For the hybrid composites with S2‐GFs, the ${\mathcal{K}}_{\beta }$ reaches a synergistic maximum at 0.75 wt.% GnP (${\mathcal{K}}_{\beta }=$21.1%), which directly correlates to the concentration of maximum bonding density, . Thus, the majority of the GnPs are covalently bonded onto the GFs, forming the hierarchical fibrous assembly, with the ideal proportion remaining in the matrix to effectively nucleate β‐crystals, with minimal agglomeration. Similarly, for the hybrid composites with S3‐GFs, the ${\mathcal{K}}_{\beta }$ reaches maximum formation at 0.75 wt.% GnP (${\mathcal{K}}_{\beta }=$16.9%), albeit in significantly lower proportions compared to the composites with S2‐GFs. This is attributed to the lower saturation attained in these composites, forcing the excess unbonded GnPs to agglomerate in the matrix and not participate in β‐crystal nucleation.

The influence of the hierarchical fibrous assembly to induce trans‐crystallization in the hybrid composites is related to the intensity trade‐off between the (040)_α and (110)_α crystallographic planes [46, 61, 62]. Since reinforcements that are completely encapsulated in trans‐crystals are known to expedite stress transfer at the interface, thereby reducing fiber debonding and pullout [43, 46, 61, 62, 64], it becomes crucial to quantify their formation and influence on the overall performance of the composites. As such, Figure 3E elucidates the relative fraction of the trans‐crystal polymorph formed, computed using the trans‐crystallization ratio (${\mathcal{T}}_r$) of (040)_α and (110)_α intensities, as ${\mathcal{T}}_r=I_{{(040)}_{\alpha }}/I_{{(110)}_{\alpha }}$ [65, 66], within the fabricated composites. In general, a greater ${\mathcal{T}}_r$ implies a greater amount of epitaxial growth encapsulating the reinforcing fillers.

For the biphasic composites, regardless of GF surface functionality, the ${\mathcal{T}}_r$ is approximately unity, indicating the absence of trans‐crystallization within these composites. Similarly, for the hybrid composite with S0‐GFs, trans‐crystallization remains negligible regardless of GnP concentration, as illustrated in Figure 3F. This emphasizes that the existence of unbonded GnP in the matrix is insufficient to induce trans‐crystals, as they predominantly facilitate β‐nucleation. For the hybrid composite with S1‐GFs, the ${\mathcal{T}}_r$ increases and reaches saturation at 0.5 wt.% GnP (${\mathcal{T}}_r=$2), directly correlating to the concentration of maximum and saturation for this sizing composition, as illustrated in Figure 3G. For the hybrid composite with S2‐GFs, the ${\mathcal{T}}_r$ is greater at 0.5 wt.% GnP (${\mathcal{T}}_r=$2.6), compared to the hybrid composite with S1‐GFs. This increase is attributed to the greater initial ${\mathcal{S}}_E$ and, consequently, greater . As a result, the ${\mathcal{T}}_r$ increases significantly before reaching saturation at 0.75 wt.% GnP (${\mathcal{T}}_r=$3.1), which is the concentration of maximum and saturation for this sizing composition.

As a result, these hierarchical fibrous assemblies contain a higher concentration of covalently bonded GnPs, which actively induce trans‐crystallization and are expected to achieve full reinforcement encapsulation, as illustrated in Figure 3H. Similarly, for the hybrid composite with S3‐GFs, the ${\mathcal{T}}_r$ increases and reaches saturation at 0.75 wt.% GnP (${\mathcal{T}}_r=$2.5), which is relatively lower than the composites with S2‐GFs, mirroring the behaviour observed for the ${\mathcal{S}}_E$ and . This suggests that steric hindrance from the hexyl group in S3 may partially restrict the bonding of non‐protonated amine groups with the GnPs and consequently impede the trans‐crystallization process, as illustrated in Figure 3I.

The influence of the hierarchical fibrous assembly on the rheological performance of the hybrid composites, compared to that of the biphasic composites, was evaluated under the injection molding processing temperature of 230°C, as shown in Figure 4A. For the biphasic composites, regardless of GF surface functionality, it is evident that the viscosity decreases with increasing shear rates, indicative of shear‐thinning behavior. Specifically, analyzing the viscosity under injection molding conditions with a shear rate of 10³ s⁻¹ (Figure 4B), it is evident that sizing composition has no significant effect on the composite melt's viscosity, averaging ∼28 Pa s. This is supported by the MFI measurements (Figure 4C), with all biphasic composites exhibiting average MFI values of ∼4.8 g/(10 min), regardless of surface functionality.

FIGURE 4.

Capillary rheometer, (A) Shear viscosity versus shear rate at high shear rates, and (B) Shear viscosities at 3000 s⁻¹ versus reinforcement concentration, as well as (C) Melt‐flow index (MFI) versus reinforcement concentration, showcasing the direct correlation between viscosity and fiber orientation. Micro‐CT renderings of (D) PPGnP0.75GF40‐S0, (E) PPGnP0.75GF40‐S1, (F) PPGnP0.75GF40‐S2, and (G) PPGnP0.75GF40‐S3, as well as, (H) Schematic illustrating the micro‐CT slicing direction, (I) Hermans Orientation Factor with respect to position (through thickness), and (J) Degree of orientation (DO) of the GFs with respect to reinforcement concentration, to visualize and quantify the effect of sizing composition and reinforcement concentration on fiber orientation.

For the hybrid composites with S0‐GFs, the viscosity dropped (shear‐thinning effect) with the addition of 0.5 wt.% GnP, compared to its biphasic counterpart, due to their inherent physical lubrication properties. Specifically, GnPs are known to impart a lubrication effect in polymer melts owing to layer exfoliation effects when subject to shear. Similarly, the MFI increased compared to its biphasic counterpart, demonstrating improvements in its melt processing behavior. However, further increasing the GnP concentration increases the viscosity of the composite melt, due to GnP agglomeration initiating the so‐called ‘filler effect’, which yields increased solid‐like character (i.e., increased elasticity) of polymer melts owing to the presence of high‐modulus fillers (i.e., agglomerates) [44, 67].

For the hybrid composites with S1‐GFs, a similar behavior to the hybrid composites with S0‐GFs is observed, where the maximum viscosity drop occurs at 0.5 wt.% GnP, but greater in magnitude. In this case, since a significant portion of the GnPs are covalently bonded onto the GFs, the fraction of unbonded GnPs in the matrix is minimized, thereby mitigating agglomeration of GnPs in the bulk PP. Additionally, the covalently bonded GnPs comprising the hierarchical fibrous assembly are believed to act as a lubricating coating that reduces inter‐fiber friction and promotes fiber mobility and alignment under flow. However, further increasing the GnP content significantly increases the viscosity as the has surpassed saturation, forcing the excess GnPs to agglomerate in the matrix, thereby reducing their lubrication efficiency.

For the hybrid composites with S2‐GFs, the maximum viscosity drop is achieved at 0.75 wt.% GnP, which directly correlates to the concentration of maximum and saturation . Thus, the majority of the GnPs are covalently bonded onto the GFs, forming the hierarchical fibrous assembly with a favorable proportion of GnPs remaining well‐dispersed in the matrix. Hence, the viscosity at 10³ s⁻¹ decreased by 44%, compared to its biphasic counterpart PPGF40‐S2. Similarly, for the hybrid composites with S3‐GFs, the maximum viscosity drop is achieved at 0.75 wt.% GnP, albeit with a less observable lubrication effect than with the S2‐GF hybrid composite melt. This is ascribed to the lower saturation attained in these composites, forcing the excess unbonded GnPs to agglomerate in the matrix, thereby reducing their lubrication performance. As a result, the ultimate performing hybrid composite in terms of industrial processability is PPGnP0.75GF40‐S2, lowering the viscosity at 10³ s⁻¹ and increasing the MFI of the automotive standard (PPGF60) by 78% and 110%, respectively, highlighting the role of hierarchical architecture in enhancing the processability of hybrid composite melts.

The influence of the hierarchical fibrous assembly on the microstructural architecture of the hybrid composites compared to that of the biphasic composites was evaluated as a function of viscosity using micro‐CT analysis across the thickness of the composite's cross‐sections. Here, the Hermans orientation factor was used as a metric for evaluating fiber alignment within composite materials, enabling analysis of fiber distribution and structural anisotropy within the composite, as shown in Figure 4. Notably, an S‐value of 1 denotes perfect orientation in a single direction [68, 69].

For the biphasic composites, regardless of GF surface functionality, the fiber orientation distribution exhibits complexity, as reflected in the variation of the Hermans Orientation Factor across the composite's cross‐section (Figure 4I). Specifically, the core region exhibits the lowest degree of fiber orientation (DO), resulting from the excessive shear flow field experienced by the melt during injection molding. As the composite melt is injected into the mold cavity, the reinforcements are subjected to high shear rates ranging from 10³ to 10⁴ s⁻¹, leading to preferential alignment of fibers along the flow direction. At the melt front, the fountain flow initially forces the fibers into an orthogonal orientation relative to the flow direction. However, upon reaching the mold walls, extensive shear stresses realign the fibers parallel to the flow direction, forming the so‐called ‘Skin Region’. Further behind the melt front, shear flow dominates, inducing a higher degree of fiber alignment, yielding the so‐called ‘Transition Region’. In contrast, regions of the melt exposed to minimal shear flow promote a more random fiber orientation in the Core Region. Consequently, this results in a complex distribution of fiber orientation, as previously documented by Thomason et al. [70], directly correlating with the varying Hermans Orientation Factor across the composite's cross‐section. The overall average DO is ∼61% for the biphasic composites, regardless of GF surface functionality, indicating that sizing composition has no significant effect on DO, as shown in Figure 4J.

For the hybrid composites with S0‐GFs, the average DO increases slightly with the addition of 0.5 wt.% GnP (DO = 67%), compared to its biphasic counterpart. This is attributed to the shear‐thinning effect induced by the GnPs, which act as physical lubricating agents, reducing the viscosity of the composite melt. Hence, the fibers experience reduced resistance to flow, allowing them to align more efficiently in the flow direction. However, further increasing the GnP concentration led to a decrease in the average DO of the composites, attributed to increased viscosity and the initiation of GnP agglomeration, which disrupts fiber orientation.

For the hybrid composites with S1‐GFs, similar behavior to the hybrid composites with S0‐GFs is observed, where the greatest average DO occurs with 0.5 wt.% GnP (DO = 73%), but with more discernable changes in orientation compared to hybrids with S0‐GFs. Since a significant portion of the GnPs here are covalently bonded onto the GFs, this minimizes the amount of unbonded GnPs in the bulk, thereby mitigating agglomeration. Thus, the hierarchical fibrous assembly reduces inter‐fiber friction and enhances fiber mobility. This facilitates fiber disentanglement, permitting fiber alignment along the flow direction. However, further increasing the GnP content significantly reduces the average DO as the has surpassed saturation, forcing the excess GnPs to agglomerate in the bulk, consequently reducing their lubrication efficiency.

For the hybrid composites with S2‐GFs, the greatest average DO is achieved at 0.75 wt.% GnP (DO = 88%), which directly correlates to the concentration of maximum and saturation and viscosity. Here, the optimized hierarchical fibrous assembly minimizes inter‐fiber friction and improves fiber mobility, thereby enhancing fiber alignment in the flow direction. Moreover, upon closer examination of the micro‐CT images in Figure 4, the highly oriented transition region across the composite thickness is expanded for PPGnP0.75GF40‐S2 (Figure 4F), along with a reduction of the less oriented core region, compared to PPGnP0.75GF40‐S1 (Figure 4E) and PPGnP0.75GF40‐S1 (Figure 4D). This higher‐order morphology is attributed to the combined effects of the optimized hierarchical fibrous assembly, which enhances fiber mobility and fiber disentanglement, as well as the favorable proportion of GnPs remaining in the bulk to effectively lubricate the matrix, with minimal agglomeration‐related viscosity increase. Furthermore, the DSC and XRD analyses corroborate improved GnP dispersion and distribution within the matrix, as reflected by its increased efficiency in inducing heterogeneous β‐nucleation.

Similarly, for the hybrid composites with S3‐GFs, the greatest average DO is achieved at 0.75 wt.% GnP (DO = 80%), although considerably lower compared to the composites with S2‐GFs. Upon closer examination of the composite morphology of PPGnP0.75GF40‐S3 (Figure 4G), it is apparent that the transition region across the composite thickness is slightly suppressed with a comparable expansion of the core region, compared to PPGnP0.75GF40‐S2 (Figure 4F). This is ascribed to the lower saturation achieved in these composites, leading to the agglomeration of excess unbonded GnPs within the matrix, subsequently diminishing their lubrication efficiency.

As a result, the enhanced overall DO for the hybrid composites, stemming from the expansion of the transition region and suppression of the core region, can be directly attributed to the decreased viscosity of the hybrid composites. This reduction in viscosity occurs as the addition of GnPs creates a lubricated surface on the GFs, with GnP layer exfoliation under flow fields permitting disentanglement of fibers and alignment along the flow direction. Furthermore, reducing the viscosity translates to a reduction in the shear stress needed to preferentially orient the fibers along the flow direction, causing the transition region to expand and the core region to compress, as seen for PPGnP0.75GF40‐S2. However, surpassing the specific saturation for each sizing composition results in obvious GnP agglomeration, thereby diminishing the preferential orientation of the fibers. Specifically, agglomerated GnP clusters introduce structural discontinuities that hinder fiber alignment and act as solid‐like fillers that increase melt elasticity. The subsequent sections of this work will detail the structure–property relationships associated with the hierarchical fibrous assembly in the frame of mechanical performance, thermal conductivity, and thermal management performance of the bio‐inspired hybrid composites.

2.2. Mechanical Performance

The mechanical properties of the fabricated composites were assessed through uniaxial tension, three‐point bending, and Izod impact to evaluate the degree of enhancement and synergistic effect induced by the hierarchical fibrous assembly in the hybrid composites, compared to the biphasic composites. The mechanical strengths of the composite materials are illustrated in Figure 5, while the corresponding moduli are provided in Figure S2. To validate the performance of the fabricated composites, PPGF60 was selected as a baseline for the mechanical properties and weight, as it is widely accepted as the automotive standard substitute for metallic structural components [41, 45, 47].

FIGURE 5.

(A) Tensile Strength, (B) Tensile Toughness, (C) Flexural Strength, and (D) Impact Strength, with respect to reinforcement concentration, as well as, Material Selection Charts for (E) Specific Tensile Modulus with respect to Specific Tensile Strength, and (F) Impact Strength with respect to Density, of collected literature data [46, 61, 62, 63, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91]. Note that the error bars are concealed within the data symbols due to their small and similar magnitudes.

For the biphasic composites with functionalized GFs, the tensile strength (Figure 5A), tensile toughness (Figure 5B), flexural strength (Figure 5C), and impact strength (Figure 5E) show considerable improvements, compared to the composite with un‐sized GFs (PPGF40‐S0). This behavior is attributed to the enhanced interfacial compatibility between the aminosilane sizing on the GF surface and the matrix. However, it is noteworthy that, regardless of the sizing composition, the mechanical properties exhibit no significant dependence on ${\mathcal{S}}_E$. For the hybrid composites with S0‐GFs, the mechanical strengths slightly improved with the addition of 0.5 wt.% of GnP, due to the minor crystalline microstructure and morphology enhancements. However, further increasing the GnP concentration is detrimental to the mechanical performance, as agglomerated GnPs serve as stress concentration sites that initiate composite failure.

For the hybrid composites with S1‐GFs, the mechanical strengths showed synergistic improvement with the addition of 0.5 wt.% of GnP, resulting from its tailored crystalline microstructure and morphology. Specifically, the hierarchical fibrous assembly enhances GF interfacial interactions by expanding the effective surface area with the matrix, thereby promoting trans‐crystallization, enabling more efficient load transfer, and simultaneously toughening the matrix through the optimum formation of β‐crystals (Figure 3). Additionally, the reduction in viscosity increases DO in the mechanical loading direction (Figure 4), enhancing material anisotropy and optimizing its response to applied forces along the flow direction, while the more uniform fiber orientation with reduced entanglement improves stress distribution under mechanical loading. However, a further increase in GnP concentration proved detrimental to mechanical performance as the ${\mathcal{B}}_d$ surpassed saturation, forcing the excess GnPs to agglomerate in the matrix, accelerating composite failure. These results showcase the dependence of mechanical performance on hierarchical architecture, through ${\mathcal{S}}_E$ and ${\mathcal{B}}_d$, to tailor the composite crystalline microstructure and morphology.

Similarly, for the hybrid composites with S2‐GFs, the mechanical strengths showed a synergistic improvement with the addition of 0.5 wt.% of GnP, comparable to the hybrid composites with S1‐GFs. In contrast, increasing the concentration to 0.75 wt.% GnP yielded superior synergistic mechanical performance as this concentration maximizes the ${\mathcal{B}}_d$, thereby producing an optimized hierarchical fibrous assembly that maximizes the trans‐crystal formation at the fiber‐matrix interface, without discernable GnP agglomeration in the bulk PP that would adversely affect the material mechanical properties. However, further increasing the GnP concentration deteriorates the composite mechanical performance, owing to agglomeration. Comparably, for the hybrid composites with S3‐GFs, the mechanical strengths showed optimal synergistic improvement with the addition of 0.75 wt.% of GnP, although considerably lower in magnitude compared to the composites with S2‐GFs. This is attributed to the lower saturation ${\mathcal{B}}_d$ achieved in these composites, leading to the agglomeration of excess unbonded GnPs within the matrix, which subsequently reduces the mechanical performance of the composite.

The extraordinarily enhanced mechanical performance of these hybrid composites is attributed to the synergistic effect induced by the hierarchical fibrous assembly, which was evaluated to elucidate the trends associated with the various sizing compositions, as shown in Figure S3, and statistically validated through a two‐way ANOVA analysis of the GnP–GF interaction presented in Table S2. Interestingly, the greatest and optimal effective synergistic effect ($S_E^{\%}$) for the tensile, flexural, and impact strengths is observed in PPGnP0.75GF40‐S2, achieving a maximum $S_E^{\%}$ of ∼46% (F‐statistic =  3.7  ×  10³, p‐value  =  2.4  ×  10⁻²⁰), ∼48% (F‐statistic =  1.1  ×  10⁴, p‐value =  4.4  ×  10⁻²⁴), and ∼54% (F‐statistic =  2.2  ×  10³, p‐value  =  1.3  ×  10⁻¹⁸), respectively. This behavior is directly attributed to the optimized hierarchical fibrous assembly being completely encapsulated in trans‐crystalline growth (Figure 3). Here, a gradient interphase at the fiber‐matrix interface facilitates interfacial stress transfer, with polymer chains absorbed/adsorbed along the surface of the hierarchical fibrous assembly [92]. Simultaneously, the matrix is toughened due to the maximum formation of β‐crystals, which are known to provide superior mechanical performance compared to α‐crystals, owing to strain hardening reinforcing mechanisms, as well as an increased resistance against crack propagation [59, 60, 93]. Additionally, the greater fiber alignment in the mechanical loading direction (increased DO, Figure 4), increases the material anisotropy, improving its response to applied forces along the flow direction. Moreover, a relatively homogeneous distribution in fiber orientation, with a lower degree of entanglement, can promote stress distribution when the material is subjected to mechanical loading. This, in turn, can lead to a tougher response, as stress is more evenly distributed throughout the material.

The material selection charts presented in Figure 5E,F compare the specific tensile modulus with respect to specific tensile strength, and impact strength with respect to density, of collected literature data compared to the experimental data presented in this work, underscoring the effectiveness of this hybrid system in producing strong, stiff, and lightweight high‐performance materials. This performance is attributed to the favorable stiffness‐to‐weight and strength‐to‐weight ratios of the composites. It is evident that the optimum‐performing composite is PPGnP0.75GF40‐S2, whereby its stiffness‐to‐weight (5.6 MPa m³/kg) and strength‐to‐weight (94 kPa/m³/kg) ratios are maximized. Moreover, this hybrid composite outperforms the specific stiffness and specific strength of all previously reported hybrid composites despite being comprised of the same material constituents, by 15% and 6%, respectively. Additionally, PPGnP0.75GF40‐S2 outperforms the impact strength of all previously reported hybrid composites, comprised of the same material constituents, by 8%.

Overall, the butterfly‐inspired hybrid composites reinforced with hierarchical fibrous assemblies presented in this work provide synergy‐induced mechanical properties that biphasic composites are unable to achieve. This can be ascribed to multi‐scale structural characteristics in hybrid composites, including a tailored gradient interface with optimal interfacial interactions, enhanced crystalline microstructure containing β‐crystals for toughness and ductility, and improved DO for morphology homogeneity. From the perspective of material applications, PPGnP0.75GF40‐S2 has emerged as the primary candidate for replacing the current polymer‐based industrial standard (PPGF60) for metallic structural components. Specifically, PPGnP0.75GF40‐S2 exceeds the tensile strength, tensile toughness, flexural strength, and impact strength of this industrial baseline by ∼29%, ∼116%, ∼23%, and ∼109%, respectively, with an 18% weight reduction. Importantly, these performance enhancements were achieved while reducing the GF content from 60 to 40 wt.%, demonstrating the efficiency of the bio‐inspired hierarchical architecture in maximizing reinforcement effectiveness while simultaneously reducing overall component weight and material usage. Furthermore, the optimized hierarchical composites exhibited an approximately 110% improvement in processability relative to the industrial benchmark, attributed to reductions in melt viscosity arising from the tailored hierarchical architecture and enhanced melt lubrication effects. This substantial improvement in flow behavior facilitates the manufacturing of complex geometries under industrial‐scale processing conditions while simultaneously lowering processing energy requirements. Moreover, the reduced GF content is expected to significantly decrease abrasive tool wear during processing, thereby minimizing long‐term damage to industrial processing equipment, reducing maintenance requirements, and improving manufacturing sustainability and economic efficiency over extended production cycles. As a result, this formulation was employed to fabricate battery casings with complex geometries for commercial electric vehicles, as in Section S3: Prospective Application. The corresponding mechanical properties are provided in Table S1. These results demonstrate the scalability of the material and its ability to outperform the current baseline even in highly complex part geometries and under industrial‐scale processing conditions. Notably, the battery tub samples exhibited coefficients of variation below 5%, confirming excellent repeatability and processing consistency at larger scale. Collectively, these findings reinforce the material's viability for high‐performance industrial applications, with improved processability and strong economic potential.

2.3. Thermal Conductivity and Thermal Management Performance

The thermal conductivity of the fabricated composites was evaluated to highlight the degree of enhancement and synergistic effect induced by the hierarchical fibrous assembly in the hybrid composites, compared to that of their biphasic counterparts, as shown in Figure 6A. To validate the thermal management performance of the fabricated composites, PPGF60 was selected as a baseline, as it is the automotive standard for structural components, regardless of its limited inherent functionalities. For the biphasic composites, with and without functionalization, the thermal conductivity remains nearly constant at ∼28.2 × 10⁻² W/m/K, indicating that the sizing composition has no significant influence on material thermal conductivity.

FIGURE 6.

(A) Thermal Conductivity with respect to reinforcement concentration, (B) Heat Dissipation “Temperature with respect to Time”, and (C) Temperature Difference between the Front and Back Surfaces of the composites. Thermal Images for (D) PPGnP0.75GF10‐S0, (E) PPGnP0.75GF10‐S1, (F), PPGnP0.75GF10‐S2, and (G) PPGnP0.75GF10‐S3, highlighting the thermal management performance of the hybrid composites; middle images highlight T _max values for the front and back surfaces of the samples.

For the hybrid composites with S0‐GFs, the thermal conductivity slightly increased with the addition of 0.5 wt.% of GnP, due to the presence of highly conductive GnPs (29.5 × 10⁻² W/m/K) coupled with the volume exclusion effect induced by the GFs that preferentially confine the GnPs, thereby promoting the formation of thermally conductive pathways within the material [62, 63]. Interestingly, further increasing the GnP concentration yielded no further increases in thermal conductivity, indicating significant agglomeration that causes phonon scattering at the nano‐filler/matrix interphase [94]. For the hybrid composites with S1‐GFs, the thermal conductivity exhibits a synergistic improvement with the addition of 0.5 wt.% of GnP, attributed to favorable distribution of GnPs both along the matrix‐fiber interface and in the bulk PP, aligned with this composition's maximized ${\mathcal{B}}_d$. With increased GnP content of 0.75 and 1 wt.%, the thermal conductivity in fact plateaus, a likely consequence of GnP agglomeration in the bulk, which reduces the effective interfacial area for heat transfer.

Similarly, for the hybrid composites with S2‐GFs, the thermal conductivity showed synergistic improvement with the addition of 0.5 wt.% of GnP, comparable to the hybrid composites with S1‐GFs. However, in this case, increasing the concentration to 0.75 wt.% GnP imparted superior synergistic thermal conductivity as this concentration maximizes the ${\mathcal{B}}_d$, thereby producing an optimized hierarchical fibrous assembly that allows additional GnP to be incorporated in the composite and form thermally conductive pathways, with lesser agglomeration in the bulk. However, akin to S1‐GF hybrid composites, further increasing the GnP concentration beyond that of the composite's maximum ${\mathcal{B}}_d$ deteriorates the thermal conductivity. Similarly, hybrid composites containing S3‐GFs exhibited the optimal synergistic improvement in thermal conductivity with 0.75 wt.% of GnP, although considerably lower in magnitude compared to the hybrid composites with S2‐GFs. This is ascribed to the lower saturation ${\mathcal{B}}_d$ achieved in these composites, leading to the agglomeration of excess unbonded GnPs within the bulk, which subsequently increases phonon scattering.

To assess the influence of the hierarchical fibrous assembly on the thermal management performance (i.e., directional heat dissipation effectiveness) of the prepared composites, a simple heat transfer experiment was conducted. First, the front surface of the sample was heated to 100°C, and subsequently cooled to room temperature (30°C), while continuously measuring the front and back surface temperatures (Figure 6B). Finally, the maximum temperature difference (ΔT _HT, heat transfer) between the front and back surfaces was determined, as shown in Figure 6C. It should be noted that due to the anisotropic morphology of biphasic and hybrid composites, thermal conductivity is highest along the fiber orientation direction (in‐plane) and significantly lower in the perpendicular direction (through‐plane) [95]. For the biphasic composites, regardless of GF surface functionality, the thermal management performance remains unaltered (ΔT _HT  =  20.5°C), as expected, due to the lack of discernable changes in composite thermal conductivity with varying GF surface treatments. For the hybrid composites with S0‐GFs, the thermal management performance slightly improved with the addition of 0.5 wt.% of GnP (ΔT _HT  =  29.3°C), due to the improvement in thermal conductivity from the presence of the highly conductive GnPs. However, further increasing the GnP concentration showed no added effect on thermal management performance, which is in accordance with thermal conductivity observations.

For the hybrid composites with S1‐GFs, the thermal management performance shows synergistic improvement with the addition of 0.5 wt.% of GnP (ΔT _HT  =  34.1°C), attributed to the hierarchical fibrous assembly's increased anisotropy (higher DO), which optimizes phonon transport along the oriented direction, thereby increasing heat dissipation in the in‐plane direction, consequently suppressing heat transfer in the through‐plane direction. Similarly, for the hybrid composites with S2‐GFs, the thermal management performance showed synergistic improvement with the addition of 0.5 wt.% of GnP, comparable to the hybrid composites with S1‐GFs. In contrast, increasing the concentration to 0.75 wt.% GnP (ΔT _HT  =  44.6°C), imparts superior synergistic performance as this concentration maximizes the ${\mathcal{B}}_d$, thereby producing an optimized hierarchical fibrous assembly that facilitates heat dissipation anisotropically. However, further increasing the GnP concentration diminished the thermal management performance, owing to phonon scattering caused by GnP agglomeration. Comparably, for the hybrid composites with S3‐GFs, the thermal management performance showed optimal synergistic improvement with the addition of 0.75 wt.% of GnP (ΔT _HT  =  40.5°C), albeit considerably lower in magnitude compared to the composites with S2‐GFs. This is attributed to the lower saturation ${\mathcal{B}}_d$ in these composites that leads to the agglomeration of excess unbonded GnPs within the matrix, as well as a lower DO that reduces the conductivity and dissipation. The extraordinarily enhanced thermal conductivity and thermal management performance of these hybrid composites are attributed to the synergistic effect induced by the bio‐inspired hierarchical fibrous assembly, which was evaluated to elucidate the trends associated with the various sizing compositions, as shown in Figure S3, and statistically validated through a two‐way ANOVA analysis of the GnP–GF interaction presented in Table S2. The greatest and optimal $S_E^{\%}$ for the thermal conductivity and thermal management performance of fabricated hybrid composites is observed in PPGnP0.75GF40‐S2, reaching a maximum $S_E^{\%}$ of ∼34% (F‐statistic = 3.4 ×  10², p‐value = 3.2  ×  10⁻¹²). This behavior is directly attributed to the optimized hierarchical fibrous assembly, with improved fiber alignment in the in‐plane direction (increased DO, Figure 4), which increases the material anisotropy, improving phonon transport and thermal conduction along the direction of fiber alignment, as opposed to the through‐plane direction.

Overall, the butterfly‐inspired hybrid composites reinforced with hierarchical fibrous assemblies presented in this work provide synergy‐induced thermal management performance that biphasic composites are unable to achieve, owing to preferential phonon transport in the in‐plane direction of hybrid composites. In fact, PPGnP0.75GF40‐S2 has emerged as the primary candidate for replacing the current polymer‐based industrial standard (PPGF60) for metallic structural components. Specifically, PPGnP0.75GF40‐S2 exceeds the thermal conductivity and thermal management performance by ∼22% and ∼42%, respectively, with an 18% weight reduction. Importantly, these enhanced thermal management capabilities were achieved simultaneously with reduced GF content relative to PPGF60, further demonstrating the efficiency of the bio‐inspired hierarchical architecture in promoting multifunctional performance while reducing overall material usage and component weight. Additionally, the improved thermal management characteristics are particularly advantageous for electric vehicle battery enclosure applications, where efficient heat distribution and thermal regulation are critical for improving battery safety, reliability, and long‐term operational performance. As a result, these composites were employed to fabricate battery casings with complex geometries for commercial electric vehicles, as detailed in Section S3: Prospective Application, demonstrating their viability for use in high‐performance industrial applications with enhanced thermal management characteristics.

3. Materials and Methods

3.1. Materials and Sample Preparation

A commercially available PP homopolymer (HIVAL 2435, Nexeo Plastics, Texas, United States), with a melt flow rate of 35 g/(10 min) (230°C/2.16 kg) and a specific gravity of 0.9 g/cm³ serves as the polymer matrix. The GnPs (GrapheneBlackTM 3X, NanoXplore, Quebec, Canada) were used as‐received, having an average flake diameter of 38 µm, with approximately 6–10 layers, and a bulk density of 0.18 g/cm³. The non‐functionalized (or un‐sized) GFs (E‐Glass, Allendale Fibertech Corporation, South Carolina, United States), were obtained by placing the as‐received GFs (l  =  6.45 mm, d  =  14 µm) in a high‐temperature oven at 600°C for 2 h to remove the commercial surface modification. The functionalized (or sized) GFs, with aminosilane surface modification, were prepared using an aqueous 95 wt.% ethanol/water solution with a controlled pH of 4.5–5, using acetic acid [96]. The various aminosilanes, as detailed in Table 3, were added dropwise to achieve a weight fraction of 3%, and then gently stirred for 20 min. Next, the GFs were removed, drained, and dried in an oven at 80°C under vacuum for 12 h.

TABLE 3.

Aminosilane sizing chemical composition and structure.

Sizing	Chemical name	Chemical structure
1	3‐AMINOPROPYLTRIMETHOXYSILANE
2	N‐BUTYLAMINOPROPYLTRIMETHOXYSILANE
3	N‐(6‐AMINOHEXYL)‐AMINOMETHYLTRIETHOXYSILANE

Biphasic PP/GF composite masterbatches with 40 wt.% GF were prepared by melt‐compounding the non‐functionalized and functionalized GFs and the as‐received Neat PP in a twin‐screw extruder (MIC27/6L‐40D, Leistritz, Nürnberg, Germany). The extrusion process was conducted at 40 RPM with a linear temperature profile across the 10 heating zones, from 140°C (feeding) to 190°C (die). The extruded filaments were pelletized using a tabletop pelletizer (BT25, Scheer Bay Co., Michigan, United States) with an end‐pellet size of ∼5 mm in length and ∼3 mm in diameter. The biphasic PP/GnP composite masterbatch with 30 wt.% GnP was produced by melt‐compounding the as‐received GnP powder and the as‐received Neat PP, under the same temperature conditions at 45 RPM, to ensure thorough mixing of the nanomaterial. The hybrid PP/GnP/GF composite blends were prepared by mixing the biphasic composites into the desired concentrations through dry blending.

A 50‐ton injection molding machine (Allrounder 270/320C, Arburg, Lossburg, Germany) with a 30 mm diameter screw equipped with MuCell Technology (Trexel, Woburn, United States) was used to fabricate the biphasic and hybrid composite materials, utilizing the following conditions: melt temperature of 230°C, mold temperature of 80°C, injection speed of 100 cm³/s, back pressure of 16 MPa, packing pressure of 25 MPa, and a cooling time of 60 s. A custom‐designed mold comprised of dual tensile (ASTM D638 – Type IV) and flexural (ASTM D790) mold cavities was utilized to standardize the sample geometry. To simplify, Figure S1 provides a high‐level illustration of the composite preparation process. The composite materials were designated by indicating the matrix, the amount of GnP, the amount of GF, and the sizing composition, according to this format: PPGnPζGF40–S‡, where the ‘ζ’ corresponds to the amount of GnP by weight, and ‡ corresponds to the sizing composition.

3.2. Interfacial Interactions, Crystalline Microstructure, and Morphological Architecture

X‐Ray Photoelectron Spectroscopy (XPS; K‐Alpha, Thermo Fisher Scientific, Massachusetts, United States), was conducted on the biphasic and hybrid composites to elucidate the surface chemical composition of the GFs, with and without functionalization, and highlight the density of covalently bonded GnPs onto the sized GFs in the hierarchical fibrous assemblies. Additionally, scanning electron microscopy (SEM; Quanta 250 FEG, FEI Company, Oregon, United States) was conducted on the hierarchical fibrous assemblies to visualize their hierarchical architecture, extracted from the hybrid composites, with different sizing compositions and GnP concentrations, by chemically etching the composites in boiling Xylene for 1 h to remove the matrix.

The crystalline microstructure of the composites was elucidated using non‐isothermal DSC (DSC 250, TA Instruments, Delaware, United States) under an inert N₂ atmosphere at atmospheric pressure. The fabricated composites (samples of 5–10 mg) were equilibrated at −50°C and then heated at a rate of 10°C/min to 250°C, to surpass the injection molding melt temperature. The crystallinity of the composites was calculated from the DSC thermograms as:

$${({\mathcal{X}}_c)}_{DSC}=\frac{\mathrm{\Delta }{H}_f}{\varphi \mathrm{\Delta }{H}_0}\times 100\%$$ (1)

where ΔH_f is the measured enthalpy of fusion of the sample, ΔH ₀ is the enthalpy of fusion for perfectly (100%) crystalline PP (ΔH ₀ = 209 J/g) [71], and ϕ is the weight fraction of PP. Additionally, POM was conducted to visually illustrate the in situ crystallization of the fabricated composites, following the same heating profiles as the DSC, using an optical microscope (BX53, Olympus, Tokyo, Japan) coupled with a heating stage (CSS450, Linkam Scientific Instruments Ltd., Salfords, United Kingdom), and a gypsum waveplate. XRD (Ultima IV Diffractometer, Rigaku, Tokyo, Japan) was conducted to investigate the polymorphism variations within the fabricated composite, with a Cu K‐α source (λ = 1.54184 Å) and a scanning range (2θ) of 2° to 42° at a step size of 0.05°, in reflection mode. The crystallinity of the composites was calculated from the XRD spectra as:

$${\left({\mathcal{X}}_c\right)}_{XRD}=\frac{I_c}{I_c+I_a}\times 100\%$$ (2)

where I_c and I_a are the intensities of the crystalline and amorphous peaks, respectively.

The Turner‐Jones et al. [56, 57] method was used to determine the fraction of β‐phase formed within the fabricated composites, as:

$${\mathcal{K}}_{\beta }=\frac{I_{\beta \left(300\right)}}{I_{\beta \left(300\right)}+I_{\alpha \left(110\right)}+I_{\alpha \left(040\right)}+I_{\alpha \left(130\right)}}\times 100\%$$ (3)

where I _α(110), I _α(040), and I _α(130) are the intensities of the (110), (040), and (130) diffraction peaks of the α‐form, respectively, and I _β(300) is the intensity of the (300) peak of the β‐form.

The overall morphology of the fabricated composites was investigated using micro‐CT (EasyTom 230, RX Solutions, Chavanod, France) with an accelerating voltage of 40 kV to preserve the internal microstructure. The degree of fiber orientation was investigated across the sample width (skin, transition, and core regions), prior to mechanical deformation, using eleven projections parallel to the melt flow direction, as shown in Figure 4H. From these projections, 200 GFs were randomly selected, and their tilt angles with respect to the melt flow direction were quantified using ImageJ (National Institutes of Health, Maryland, USA).

Then, the various Hermans Order Parameters across the sample width were determined based on these measurements, as [68, 69]:

$$S=\frac{3〈{\cos }^2\theta 〉-1}{2}$$ (4)

where,

$$〈{\cos }^2\theta 〉=\frac{{\sum }_{i=1}^{200}{\cos }^2{\theta }_i\sin {\theta }_i}{{\sum }_{i=1}^{200}\sin {\theta }_i}$$ (5)

and S = 1 for perfectly oriented fibers.

The rheological performance of the composites was evaluated using a Capillary Rheometer (RH2200, Rosand, Cirencester, United Kingdom) in twin‐bore (19 mm diameter) configuration, at 230°C, an internal pressure of 0.6 MPa, and a 1.5 min melt saturation time. Rheological measurements were taken at shear rates of 600, 1000, 3000, 6000, and 10 000 s^{− 1}. Additionally, the Melt‐Flow Index (MFI; MFI2 Melt‐Flow Indexer, Dynisco, Massachusetts, United States) of the composites was measured following the ASTM D1238 standard.

3.3. Mechanical Performance

The tensile and flexural mechanical properties of the fabricated composites were measured using a uniaxial universal testing machine (5965 Series Premier, Instron, Massachusetts, United States) in accordance with the ASTM D638 standard and ASTM D790 standard, respectively, with a load cell of 5 kN. For the tensile properties, a crosshead speed of 5 mm/min and a gauge length of 25 mm were used, at ambient conditions. For the flexural properties, a crosshead speed of 1.3 mm/min and a span length of 48 mm were used, at ambient conditions. The impact mechanical properties of the fabricated composites were measured using a pendulum‐based impact testing machine (IT504, Tinius Olsen, Redhill, United Kingdom) in accordance with the ASTM D4812 standard. Prior to all mechanical testing, the samples were maintained at atmospheric conditions for at least 48 h. The average values corresponding to five independent replicates were obtained and reported together with their standard deviations, and the corresponding coefficients of variation have been included in Table S1.

The synergistic effect (or percent synergy) was quantified using the effective percent synergy method [97, 98], as:

$$S_E^{\%}=\frac{k-\left(p+q\right)}{\left(p+q\right)}\times 100$$ (6)

where k represents the magnitude of the enhancement of the hybrid composite relative to the matrix material, and p and q represent the magnitude of the enhancements of the individual reinforcing materials alone, relative to the matrix material. A positive $S_E^{\%}$ is indicative of a synergistic enhancement generated within the hybrid composite, while a negative $S_E^{\%}$ suggests a discord within the hybrid system resulting in an undesirable decrease in its properties.

Statistical validation of the experimental data was conducted using a Two‐Way ANOVA with replication to assess the individual effects of the reinforcements and their synergistic interaction. The experimental design followed a factorial framework, evaluating the concentrations of the reinforcements as independent factors across five distinct replicates for each material composition to ensure statistical robustness. This analysis was applied to all properties, specifically tensile stiffness and strength, flexural stiffness and strength, impact strength, and thermal conductivity, and the complete statistical results are presented in Table S2.

The significance of the main effects and the interaction term, the latter serving as the mathematical indicator for synergy, was determined at a 95% confidence level (α = 0.05). Verification of synergistic behavior was established by evaluating the resulting P‐values and F‐statistics from the interaction term, ensuring that observed performance enhancements were statistically significant and not attributable to experimental variability. Furthermore, 95% confidence intervals were calculated for each mean value to provide a verifiable range for the performance of the composites.

3.4. Thermal Conductivity and Thermal Management Performance

The thermal conductivity of the fabricated composites was measured using a multipurpose thermal conductivity analyzer (TPS2500S, Hot Disk Instruments, Göteborg, Sweden) with a Kapton‐insulated sensor (5465 F1, Hot Disk Instruments, Göteborg, Sweden), and in accordance with the ISO/DIS 22007‐2.2 standard. The obtained thermal conductivity represents the isotropic overall average conductivity of the composites, considering the dissipated heat in all directions (i.e., both in‐plane and through‐plane). The thermal management performance or heat dissipation effectiveness of the fabricated composites was determined through a conventional heat transfer experiment. Simply, the front surface of room temperature equilibrated samples was heated to 100°C through illumination using a heating light source (V‐WLP‐1000, Newport, California, United States), and then subsequently cooled back to room temperature. The temperature of the front and back surfaces of the sample was simultaneously measured as a function of time, during heating and cooling, using an infrared thermal camera (TC002, TOPDON, Shenzhen, China).

4. Conclusion

Inspired by nature's unparalleled engineering, this work employs hierarchical architectures to develop hybrid composites that emulate the synergy‐induced multifunctional performance of butterflies. Specifically, these composites are reinforced with hierarchical fibrous assemblies comprised of nano‐sized GnPs covalently bonded onto micro‐sized GFs. In detail, this work showcases a novel approach to control the in situ self‐assembly behavior of these reinforcements, achieved by tailoring the surface chemistry of the GFs through surface functionalization, to maximize the density of covalently bonded GnPs. Specifically, distinct aminosilane surface functionalizations were selected to systematically decouple the effect of sizing efficiency (amount of available non‐protonated amine groups) and reinforcement concentration on the bonding density of the hierarchical fibrous assembly. In fact, our newly developed hybrid composites are tailorable to surpass the mechanical properties of the current industrial standard for automotive structural components (PPGF60, 40 wt.% PP/60 wt.% GF), exhibiting improvements up to 29%, 116%, and 109% in tensile strength, flexural strength, and impact strength, respectively, as well as 22%, 42%, and 110% in thermal conductivity, thermal management performance, and processability, respectively, with an overall 18% weight reduction.

These significant material property enhancements were made possible due to the in‐depth analysis of structure–property relationships (spanning multiple length scales) presented in this work, creating a fundamental understanding of how to tailor material performance to achieve desirable properties that meet stringent safety standards and regulations. At the atomic scale, the chemical structure of the sizing governs the density of covalently bonded GnPs that bond onto the surface functionalized GFs; this bonding density is demonstrated to be both a function of sizing efficacy and GnP concentration. Specifically, the maximum bonding density of 79% was achieved with n‐butylaminopropyltrimethoxysilane, Sizing 2, and 0.75 wt.% GnP. Above this GnP concentration, supersaturation of the GF surface transpires, which adversely affects GnP distribution in the bulk and leads to severe agglomeration.

At the micron scale, for the biphasic composites, with and without functionalization, the GFs show minimal influence on the matrix's crystalline microstructure, while the hybrid composites exhibit vastly different crystallization tendencies. Specifically, the unbonded GnPs in the matrix bulk preferentially nucleate β‐crystals, while GnPs covalently bonded to the GF surface preferentially nucleate trans‐crystals at the fiber–matrix interface. Moreover, excessive unbonded GnPs appear to agglomerate and diminish the GnPs’ β‐nucleation efficiency. At GnP concentrations where optimal bonding densities are achieved, complete trans‐crystallization encapsulation transpires, with simultaneous β‐crystal nucleation occurring in the bulk, which positively affects composite bulk properties. Additionally, for the biphasic composites, with and without functionalization, there is minimal influence on the matrix's overall morphology, while the hybrid composites drastically alter the micro‐morphology. Specifically, for the hybrid composites with optimal bonding density, maximum melt lubrication is achieved, translating to apparent changes in the size and DO of both the core (decreases in size and increases in orientation) and transition region (expands in size and increases in orientation). However, further increasing the GnP concentration beyond saturation leads to a decrease in the average DO of the composites, as GnP agglomeration increases melt viscosity, suppressing alignment of GFs along the flow direction.

At the macro‐scale, the mechanical performance was found to be governed by the synergistic enhancements resulting from the intricate interrelationship between crystalline microstructure and overall morphological architecture, influenced by GF surface functionality and reinforcement concentration. Specifically, this can be ascribed to multi‐scale structural characteristics in hybrid composites, including a tailored gradient interface with optimal interfacial interactions, enhanced crystalline microstructure containing β‐crystals for toughness and ductility, and improved DO for morphology homogeneity. Moreover, the higher DO in the hybrid composites, compared to the GFs in biphasic composites, results in preferential anisotropic heat transport, thereby facilitating heat dissipation.

Ultimately, the efficacy of this Avant‐garde hybrid composite approach in an industrial‐scale setting was demonstrated by the fabrication of a car battery casing, illustrating the viability of this technology in improving material processability in a cost‐effective and scalable manner. The development of such lightweight hybrid composites not only imparts superior mechanical performance but also offers other functionalities, including improved thermal management performance. Such multifunctionality is crucial in developing materials and structures that overcome redundancies in greater systems, presenting a greener, value‐added material alternative to conventional structural components. Importantly, these multifunctional enhancements were achieved while simultaneously reducing GF content from 60 to 40 wt.% relative to the current industrial benchmark, demonstrating the exceptional reinforcement efficiency of the butterfly‐inspired hierarchical architecture. Beyond lightweighting advantages, this reduction in abrasive GF content is also expected to significantly decrease long‐term tool wear during manufacturing, thereby reducing maintenance requirements, improving processing sustainability, and extending industrial equipment lifespan during high‐volume production. Furthermore, unlike many high‐performance composite systems requiring specialized processing routes or complex manufacturing modifications, the developed hybrid composites remained fully compatible with conventional industrial‐scale injection molding infrastructure, highlighting their strong potential for rapid industrial implementation and commercialization.

Collectively, these findings establish a scalable pathway for the development of next‐generation multifunctional structural composites capable of simultaneously enhancing mechanical performance, thermal management, lightweighting efficiency, and manufacturability. More broadly, this work demonstrates how bio‐inspired hierarchical architectures can bridge fundamental structure–property relationships across multiple length scales to engineer industrially relevant composite systems with transformative multifunctional capabilities.

Author Contributions

Nello D. Sansone: conceptualization, data curation, formal analysis, investigation, methodology, writing – original draft, and review & editing; Rafaela Aguiar: conceptualization, data curation, formal analysis, investigation, methodology, writing – original draft, and review & editing; Nichole Cheung: data curation, investigation, methodology, and writing – review & editing; Anthony V. Tuccitto: data curation, investigation, methodology, and writing – review & editing; Zirui Huang: data curation, investigation, methodology, and writing – review & editing; Matthew Leroux: funding acquisition, project administration, resources, and writing – review & editing; Patrick C. Lee: conceptualization, funding acquisition, resources, supervision, and writing – review & editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll74948‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.