Tri-continuous polymer templates enable scalable fabrication of hierarchical nanoparticle monoliths

Abstract

Hierarchical structures with bimodal porosity are crucial in diffusion and confinement-driven applications, such as catalysis and separation. This study introduces the first utilization of polymer blend nanocomposites as templates for isolating nanoparticle monoliths with bimodal porosity. We examined tri-continuous polymer blend nanocomposites of silica nanoparticles (SNPs) in polyethylene (PE), ethylene vinyl acetate (EVA), and polyethylene oxide (PEO) using three-channel confocal microscopy. This allowed visualization of their morphology and its evolution during quiescent annealing. The analysis extends to co-continuous polymer blend nanocomposites, with or without PEO. Our findings highlight the reinforcing effect of sequentially adding polymer phases in tri-continuous blends. This results in a refined morphology and strengthened three-dimensional particle network, as evidenced by a two-order-of-magnitude increase in the terminal modulus in frequency sweep rheometry. Conversely, co-continuous systems exhibit a significantly weaker particle network with a minimal increase in terminal storage modulus, making them prone to collapse during the polymer template removal. The interplay between domain size, nanoparticle jamming within one phase, and consequent particle network robustness enables the material to withstand deformation during polymer removal, facilitating the isolation of hierarchically structured monoliths. This novel templating method offers a scalable approach to fabricating hierarchically porous materials with potential applications in catalysis, energy storage, and gas separation.

---

New concepts

We introduce a novel concept for the scalable fabrication of hierarchically porous monoliths using tri-continuous polymer blend nanocomposites as structural templates. Unlike most existing methods that rely on multi-step, solvent-intensive, and lab-scale synthesis routes, our continuous processing strategy integrates industrially viable extrusion and sintering methods to eliminate solvent use and simplify template removal, enabling a practical path toward scale-up. Our findings introduce a new paradigm where hierarchical porosity arises from a strong particle network formed via the unique confinement effect of nanoparticles within one of the tri-continuous phases. By comparing tri-continuous and co-continuous blends, we identify key factors critical to successful monolith isolation during polymer removal. Achieving high nanoparticle dispersion within the desired polymer phase, combined with a high jamming extent within small domains, is essential for preserving structure. When these conditions are not met—as observed in co-continuous systems—network collapse occurs during template removal. This work introduces new design parameters for creating robust porous materials and advances the understanding of how blend morphology influences particle reinforcement. The concept will be of broad interest to researchers in nanomaterials, polymer science, and porous materials for mass-transport limited applications such as catalysis, energy storage, gas separation, sensing, and drug delivery.

1 Introduction

Hierarchically porous materials offer a key advantage over simple porous materials. They provide high surface area while maintaining accessibility and facilitating efficient mass transport due to contributions from all structural hierarchy levels.^1–3 This favorable balance of properties has been proven to be superior in various applications, including catalysis,^4,5 separation,^6,7 sensors,⁸ energy storage and conversion^9–11 and life sciences.^12,13 Numerous strategies have been developed to prepare hierarchically porous materials. These include dual surfactant templating,¹⁴ colloidal crystal templating,^15,16 polymer templating,^17,18 emulsion templating,^12,19 spinodal decomposition,^20,21 sol–gel process,²² freeze-drying,²³ breath figures,^24,25 phase separation,²⁶ selective leaching,²⁷ replication,²⁸ zeolitization,²⁹ and spontaneous formation.³⁰ However, multiple challenges remain including control of multimodal porosity, product stability, and most importantly, scalability and cost-effectiveness.

Co-continuous and tri-continuous polymer blend nanocomposites are multi-phase systems. In these systems, two or more polymer phases form interconnected, continuous domains, which enable unique structural and functional properties.³¹ Tri-continuous morphology refers to a co-continuous polymer system separated by a third continuous phase, offering multiple interfaces required for broader applications.³² They are typically processed using scalable melt mixing techniques, which allow for precise control over phase morphology and the selective localization of nanoparticles.^33,34 This tunable nanoparticle localization holds the potential to enable control over pore distribution and hierarchical structuring through the removal of sacrificial polymer components, making these systems highly promising templates for hierarchical porous materials. Despite this structural versatility and scalability, these systems have not yet been utilized for hierarchical porous material fabrication. Current applications primarily exploit their synergistic mechanical and functional properties in sectors such as automotive, aerospace, packaging, electronics, and biomedical devices.^35–39 By broadening the range of polymer and nanoparticle combinations, these nanocomposites could provide a commercially viable platform for designing advanced porous materials tailored to catalysis, separation, energy storage, and other applications. The scalability of melt-blending methods further strengthens their potential for large-scale production.

Here, we report the first isolation of hierarchically porous materials using tri-continuous polymer blend nanocomposite as a sacrificial template. We establish a direct correlation between the spatial distribution of SNPs in PE/EVA/PEO tri-continuous blends and the morphological and rheological properties of the tri-continuous polymer blend nanocomposite system. Multi-channel LSCM observation with two distinct dyes enabled identification of each polymer phase, while the localization of untagged SNPs is detected using the reflection mode. We show that SNPs selectively localize and percolate within PEO, thereby inducing continuity in this particular phase. Furthermore, when PE, EVA and SNPs were blended prior to the addition of PEO, the tri-continuous structure undergoes further refinement. The obtained findings imply the existence of the unique confinement effect in tri-continuous polymer blend nanocomposites, which can be further exploited to achieve a strong particle network and consequently isolate porous materials. This method is more energy-intensive than soft gel-based methods due to the high processing temperatures and polymer template removal. However, its continuous processing nature and elimination of solvents make it a scalable and environmentally favorable approach for producing hierarchically structured materials. In particular, by combining tri-continuous polymer blend morphologies with nanoparticle jamming, our approach enables the fabrication of self-supporting monoliths that can be continuously manufactured through extrusion and shaped via an industrial process such as injection molding. This approach overcomes the scale-up limitations of batch-based methods such as bijels or emulsion templating, allowing efficient monolith production with tunable morphologies.

2 Results and discussion

To determine the optimal polymer combination for tri-continuous blends, we considered both the viscosity ratio and the compatibility with high-temperature-stable fluorescent dyes. Fig. S1a illustrates the viscosity profiles of all blend components and Fig. S1b demonstrates their corresponding storage and loss moduli. The preparation of the neat blends and the blend nanocomposites are demonstrated in Fig. 1a and the LSCM images are taken after annealing the samples for 30 min. The morphology of the neat blends are illustrated in Fig. 1b, where Sudan Blue II and Rhodamine B were used as markers for EVA and PEO polymers respectively. The EVA phase (EVA12 or EVA25, see materials section below) is displayed in green, pink regions depict the PEO phase, while PE is not tagged and is represented by white regions in the images. In both systems, (PE/EVA12/PEO and PE/EVA25/PEO), PEO phase is found in the form of droplets fully surrounded by the EVA phase.

Fig. 1 (a) Schematic representation of the tri-continuous polymer blend nanocomposite preparation using a microcompounder, quenching in liquid nitrogen and the annealing process under inert atmosphere. The top panel shows the neat polymer blend, and the middle and bottom panels depict the simultaneous and sequential blending methods, respectively. (b) LSCM images of tri-continuous polymer blends after 30 minutes of annealing: (left) PE/EVA12/PEO and (right) PE/EVA25/PEO neat blends. (c) LSCM images of simultaneous blends after 30 minutes of annealing: (left) PE/EVA12/PEO/SNP and (right) PE/EVA25/PEO/SNP. (d) LSCM images of sequential blends after 30 minutes of annealing: (left) PE/EVA12/PEO/SNP and (right) PE/EVA25/PEO/SNP. (e) Storage modulus, G′, and (f) loss modulus, G′′, as a function of frequency for PE/EVA12/PEO/SNP and PE/EVA25/PEO/SNP blend nanocomposites with different blending orders.

The wetting behavior of PE/PEO by EVA12 (or EVA25) were predicted by calculating the spreading coefficient based on Harkins theory, generalized by Torza and Mason.^40,41 For an A/B/C ternary system, a set of three spreading coefficients can be calculated as follows:

λ_BAC = γ_BC − (γ_AB + γ_AC) (1)

λ_ABC = γ_AC − (γ_AB + γ_BC) (2)

λ_ACB = γ_AB − (γ_AC + γ_BC) (3)

where γ are the interfacial tensions between components. A positive λ_ABC, indicates the tendency of component B to spontaneously spread at the interface of components A and C which is considered the complete wetting regime. In PE/EVA12/PEO system, λ_PE/EVA12/PEO, λ_EVA12/PE/PEO and λ_EVA12/PEO/PE were 4.2 mN m⁻¹, −10.2 mN m⁻¹ and −7.1 mN m⁻¹, respectively (SI, Table S1). On the other hand, λ_PE/EVA25/PEO, λ_EVA25/PE/PEO and λ_EVA25/PEO/PE were 2.5 mN m⁻¹, −5.9 mN m⁻¹ and −11.5 mN m⁻¹ in PE/EVA25/PEO system. These results indicate tendency of the EVA phase to spread at the interface of PE and PEO in both blends,⁴² and are in agreement with LSCM images shown in Fig. 1b. While the tendency of EVA to spread between the PE and PEO phases holds for both systems, the morphological differences between the blends with EVA12 and EVA25 are evident. In particular, the size of the EVA and PE domains are considerably larger in the PE/EVA25/PEO system. This variation can be attributed to the viscosity difference between EVA25 and EVA12 phases (shown in Fig. S1a), as the interfacial tension differences are marginal.⁴³ Blending higher-viscosity EVA12 system results in smaller domains and induces partial co-continuity between PE and EVA12 phase, while the system with EVA25 does not show co-continuity at this level of image magnification. The PE/EVA12/PEO and PE/EVA25/PEO blends exhibit characteristic sizes of 2.8 ± 0.1 μm and 5.6 ± 0.3 μm, respectively. Fig. 1c and d demonstrate microstructural changes in the blends when SNPs are incorporated. In both blending approaches, (simultaneous and sequential, Fig. 1a), the addition of 28 vol% SNP (30 wt%) induced continuity in the PEO phase. This occurs because nanoparticles localize in PEO, increasing its viscosity.⁴⁴ Prior studies have indicated that a high loading of fumed silica, exceeding 10 vol%, has the potential to induce phase continuity in the polymer phase that exhibits a stronger affinity towards silica. This continuity is attributed to the meniscus bonding that occurs at the interface between the SNPs and the matrix polymer.⁴⁵ The continuity of the PEO phases that are filled with SNPs is obvious in Fig. S2 which depict the superimposition of the PEO fluorescence and SNP reflectance channels. In Fig. 1c, the reduced characteristic size (3.6 ± 0.1 μm) observed in PE/EVA25/PEO/SNP system compared to PE/EVA12/PEO/SNP blend nanocomposite (4 ± 0.2 μm) when simultaneous blending approach is employed (p-value ≈0.01), can be attributed to the lower viscosity of the EVA25 phase (Fig. S1a). The lower viscosity facilitates SNP migration toward the thermodynamically preferred PEO phase, leading to a faster increase in the viscosity of the PEO phase and resulting in refinement of the morphology.

During the initial stage of sequential blending approach, a co-continuous structure is established between PE and EVA, wherein SNPs are selectively localized within the EVA phase (Table S1). The disparity in viscosity between EVA12 and EVA25 (Fig. S1a) results in the generation of smaller domains in the PE/EVA12 system (1 ± 0.1 μm) compared to the PE/EVA25 system (1.7 ± 0.1 μm) where p-value ≪ 0.01. This, coupled with the migration of particles from the EVA phase to the PEO phase, results in the kinetic arrest of a finer microstructure in the PE/EVA12/PEO/SNP blend nanocomposite. Therefore, the average characteristic length became significantly smaller during sequential blending approach when PEO is added in the second step of blending. The comparison of characteristic size of the blends is illustrated in a histogram in Fig. S3. In the EVA12 system, the tri-continuous blend exhibits an increase in characteristic domain size compared to the neat blend, primarily due to the formation of large, interconnected PEO domains. In contrast, the sequentially processed sample yields the smallest domains. For EVA25, the domain size decreases gradually; however, it never reaches the reduced size observed in EVA12, since the system inherently begins with much larger PE and EVA domains, as shown in Fig. 1c.

Rheological investigation of the neat blends and PE/EVA/PEO/SNP composites as shown in Fig. 1e and f provides additional insights into the morphology and SNP packing within the PEO phase. The addition of SNP results in formation of the SNP particle network but its effect is highly dependent on the system as well as type of blending. The appearance of a plateau in viscoelastic moduli at low frequencies is a signature of the particle network.⁴⁶ Simultaneous blending of all components in the PE/EVA25/PEO system yields a higher plateau modulus compared to the PE/EVA12/PEO system. This observation aligns with the morphological signature exhibited by these blends. The presence of smaller PEO domains in the PE/EVA25/PEO system contributes to higher jamming extent of SNP within these domains. This, in turn, enhances the solid-like behavior of the blends. In contrast, sequential blending leads to a significantly more pronounced plateau and an increase in G′ (storage modulus) compared to the simultaneous and pure blends. In jammed systems, the material exhibits solid-like behavior due to the constraints on particle movement. This is reflected in the high storage modulus (G′), which indicates the material's ability to store elastic energy. Despite the solid-like state, jammed systems exhibit a viscoelastic behavior characterized by energy dissipation (evidenced by a high loss modulus, G′′) resulting from particle rearrangement under applied shear stress. Under oscillatory shear, this viscoelasticity is more evident, as the material shows both solid-like and fluid-like responses depending on frequency and strain amplitude. The closeness of G′ and G′′ values (in Fig. 1e and f) in a frequency sweep suggests that the material's elastic and viscous properties are comparable. This balance is typical in viscoelastic systems near the jamming transition, where a material exhibits properties of both states.⁴⁷

Fig. 2a and b illustrate the polymer template removal process with the specific thermal profile using a muffle furnace. Fig. 2c–e show the confocal images of the quaternary polymer blends after 30 min annealing and the SEM images after removal of polymer components through the heating process. Highly porous self-standing SNP structures were isolated only in the sequentially blended samples (Fig. 2d and e), while the larger PEO domains filled with SNPs were not preserved during the polymer removal process (Fig. 2c). The visual inspection of samples before and after the template removal process indicates retention of sample shape as a result of interconnected SNPs network during the polymer removal process of the sequentially blended sample. To confirm the jamming extent of nanoparticles, a water droplet was deposited on the PE/EVA12/PEO/SNP (sequential) surface for 30 s and then removed, selectively dissolving surface PEO and revealing the underlying particle network (Fig. 2f).

Fig. 2 (a) Schematic representation of the polymer removal process of the annealed samples using a muffle furnace. (b) Thermal profile used in the polymer removal process of annealed samples for SEM imaging of the hierarchically porous silica structure. LSCM images of annealed blends (top row) and SEM images of the residues of quaternary polymer blends after polymer components removal (bottom row) through the heating process for blends prepared by (c) simultaneous blending of PE/EVA12/PEO/SNP with 30 wt% SNP, (d) sequential blending of PE/EVA12/PEO/SNP with 30 wt% SNP, and (e) sequential blending of PE/EVA25/PEO/SNP with 30 wt% SNP. (f) Jamming study procedure and SEM images of the jammed SNP within PEO phase.

To assess the achievement of isolated hierarchical monoliths through the combination of high particle loading and small phase domains, co-continuous polymer blend nanocomposites of PE/EVA12/SNP and PP/EVA12/SNP were examined. PE/PEO/SNP and EVA/PEO/SNP blend nanocomposites were also prepared as shown in Fig. S4 and S5, but were not further investigated because of the lack of co-continuity in their morphology. Fig. 3 illustrates the co-continuous polymer blend nanocomposites that were prepared using both sequential and simultaneous blending approaches. The co-continuous polymer nanocomposite systems exhibited comparable domain sizes to the tri-continuous polymer nanocomposites, along with an SNP loading equivalent to a single polymer phase (30 wt% relative to EVA). While a porous structure of SNP monoliths was achieved, the post template removal samples exhibited structural collapse in each case. The calculated wetting parameters for PP/EVA12/SNP and PE/EVA12/SNP are −3.2 and −4.1, respectively (SI, Table S1), indicating that EVA is the thermodynamically preferred phase. However, high magnification LSCM images shown in the insets of 2D images of Fig. 3 indicate the presence of SNP aggregates at the blend interfaces. In the sequential blending process, pre-blending of the SNP particles with PP or PE phase, results in the formation of SNP aggregates. The subsequent addition of the thermodynamically favorable EVA12 phase induced a co-continuous morphology, creating Laplacian pressure at the highly curved blend interface. This pressure prevents the migration of the primary particles and small aggregates towards the EVA12 phase, while it is unfelt by the larger agglomerates,⁴⁶ which cross the interface and localize within EVA12 phase. Although primary particles and small aggregates at the interface help suppress coarsening,⁴³ micron-sized aggregates create localized stress concentrations and weak points within the interfacial network. Conversely, larger aggregates localized within EVA12 reduce dispersion extent and result in a less jammed SNP network, compromising mechanical integrity. During polymer removal, the capillary forces exerted on the rigid walls of these structures can cause deformation, making the interface susceptible to collapse. In contrast, in tri-continuous polymer blend nanocomposites, SNPs are well-dispersed within the PEO phase, and no particle aggregates are observed at the interface between the PEO and EVA phases (Fig. 1 and Fig. S2). The morphology of tri-continuous and co-continuous polymer nanocomposites is directly reflected in the inner pore structure of the remaining monoliths. This correlation is evident in the comparison between domain sizes of the annealed polymer nanocomposites and the pore sizes of the obtained monoliths. This dispersion extent and complete localization within the PEO phase enable interfacial rearrangement during polymer phase removal, thereby preserving structural integrity during the polymer removal process. Considering the interfacial tensions, SNPs exhibit the highest affinity for the PEO phase, with a low interfacial tension of 3.2 mN m⁻¹. In contrast, the interfacial tension of SNP/EVA12 and SNP/EVA25 are 8.7 mN m⁻¹ and 12.6 mN m⁻¹, respectively, indicating a weaker interaction with EVA compared to PEO. This indicates that SNPs are more effectively distributed within the PEO phase in tri-continuous polymer blend nanocomposites, while in co-continuous blends, particles are more likely to aggregate at the blend interfaces. Consequently, the main difference between the co-continuous and tri-continuous systems examined here is the extent of SNP dispersion in the thermodynamically unfavorable phase during the first blending step and the differing affinities of SNPs for the thermodynamically preferred phase during the second blending step. Additionally, a key disadvantage of co-continuous systems is that processing at higher nanoparticle loadings (30 wt% relative to EVA12 phase) presents scalability challenges for our templating method compared to tri-continuous systems.

Fig. 3 Confocal images of (a) PE/EVA12/SNP 30 wt% simultaneous, (b) PE/EVA12/SNP 30 wt% sequential, (c) PP/EVA12/SNP 30 wt% simultaneous and (d) PP/EVA12/SNP 30 wt% sequential blends before and after 30 min annealing followed by the SEM images of hierarchical structures after the template removal process. High magnification images show the aggregation of the nanoparticles at the interface of the blends. (e) Storage modulus, G′, and (f) loss modulus, G′′, as a function of frequency for PP/EVA12/SNP and PE/EVA12/SNP blend nanocomposites with different blending orders.

The rheological behaviour of the PP/EVA12/SNP and PE/EVA12/SNP blend nanocomposites confirm the significant difference in the SNP network formation between co-continuous and tri-continuous blend nanocomposites (Fig. 3e and f). The neat PP/EVA12 and PE/EVA12 blends exhibit slightly lower modulus than the PE/EVA12/PEO neat blend (Fig. 1e and f), indicating the addition of a third polymer phase and creation of an additional interface increases the modulus of pure blends. However, a significant difference is observed when comparing the co-continuous and tri-continuous blends that contain SNP within one confined phase. The storage and loss moduli of the tri-continuous blend nanocomposites are two orders of magnitude higher than those of the co-continuous blend nanocomposites (Fig. 1e, f and 3e, f). This increase in moduli demonstrates a higher extent of structural refinement in the tri-continuous blend nanocomposite, allowing isolation of SNP monoliths after the polymer template removal. In contrast, the co-continuous blend nanocomposites tend to collapse and deform after the template removal process. Fig. S6 also highlights the difference between co-continuous and tri-continuous polymer blend nanocomposites templates, as shown in the strain sweep data. The linear viscoelastic region is evident in all samples prior to the yield point. Beyond the yield strain, the tri-continuous blend nanocomposites exhibit a sharp drop in shear stress, indicating the de-jamming of the nanoparticles, where the applied shear breaks down the previously jammed nanoparticle network. Following this drop, the system stabilizes into a plateau, signifying the onset of a new steady flow regime. In contrast, this sharp drop is absent in both neat polymers and co-continuous blend nanocomposites, suggesting that the nanoparticles in the co-continuous system are not sufficiently jammed which confirms the structure collapse after polymer removal in co-continuous polymer blend nanocomposites.^12,48

To demonstrate the complete removal of the polymer template, TGA was conducted on monoliths derived from co-continuous and tri-continuous blend nanocomposites and compared with the thermograms of pure polymers. The pure polymers exhibited a significant weight loss of up to 100% due to their degradation at temperatures above 400 °C (Fig. 4a). However, this characteristic weight loss due to polymer degradation was not observed in the monoliths, indicating the absence of polymer residue after template removal (Fig. 4b).

Fig. 4 TGA of (a) polymer components and (b) the hierarchical monoliths prepared from co-continuous and tri-continuous polymer blend nanocomposites after polymer removal. (c) N₂ adsorption–desorption isotherms at −196 °C for tri-continuous PE/EVA12/PEO/SNP monoliths prepared using the sequential method. (d) Differential pore volume distribution as a function of pore size, determined by DFT for the microporous region (left inset) and by BJH analysis for the mesoporous/macroporous region (right inset), both calculated from the adsorption branch.

The thermogram for the pure silica nanoparticles used in this study shows the expected trend of two distinct weight loss stages. The first stage, up to 150 °C, is attributed to the release of adsorbed water on the hydrophilic silica nanoparticles. The second stage, occurring at temperatures above 250 °C, results from the removal of water produced during the non-catalyzed self-condensation reactions of silanol groups on the silica nanoparticle surfaces. For the monoliths, traces of adsorbed water are visible at temperatures below 150 °C, with a weight loss ranging from 1 to 3 wt%. However, the condensed water loss at temperatures above 250 °C is not observed in the monoliths, as the jammed silica nanoparticles have already undergone self-condensation reactions during the polymer removal process.^49–52 The lower amount of absorbed water in the monoliths isolated from PE/EVA12/PEO/SNP and PE/EVA25/PEO/SNP samples is attributed to the highly jammed structure in these samples, reducing the penetration of water molecules into the jammed structure. however, in all cases, the weight loss difference remains within 2 wt%, indicating complete removal of the polymer phases.

The porosity of the PE/EVA12/PEO/SNP monolith prepared by the sequential method, which exhibited the most refined structure, was further evaluated by N₂ adsorption–desorption analysis. The rapid uptake of N at p/p₀ < 0.1 indicates the presence of micropores, while the Type IV isotherm with an H3 hysteresis loop is characteristic of mesopores larger than 4 nm (see Fig. 4c).⁵³ The hysteresis loop, which does not exhibit limiting adsorption at high relative pressures, reflects the presence of meso/macropores formed as a result of particle jamming during fabrication.⁵⁴ The nearly parallel adsorption and desorption branches further suggest that the mesoporous network is fully accessible and extends to the outer surface of the monolith, which is desirable for applications where enhanced mass transfer is required by shortening diffusion lengths in the micropore.^55,56 The hierarchical porous structure was confirmed by combining density functional theory (DFT) and Barrett–Joyner–Halenda (BJH) analyses, used for the microporous and meso/macroporous regions, respectively. DFT analysis revealed micropores in the range of 1.2–1.7 nm, while BJH analysis identified mesopores with peaks at 21 nm and 40 nm, along with larger macropores centered around 93 nm (see Fig. 4d). It should be noted that our SNPs are ∼100 nm in diameter, and larger monolith pores present in the micron range are not detected by gas sorption analysis; however, their existence contributes to the relatively modest measured surface area (∼102 m² g⁻¹) and total pore volume (0.20 cm³ g⁻¹), despite the highly porous nature of the monolith. This contrasts with silica monoliths produced by sol–gel processing,⁵⁷ which typically have ∼15 nm silica nanoparticles,⁵⁸ possess uniform small pores that yield very high surface areas, but lack hierarchical connectivity, thereby limiting pore accessibility. In comparison, the co-continuous and accessible hierarchical porosity of the PE/EVA12/PEO/SNP monoliths makes them promising candidates for catalysis, storage, and separation applications.

3 Conclusion

This study marks the first successful isolation of hierarchically structured monoliths from polymer blend nanocomposites. Hierarchically porous monoliths are highly versatile and find applications in energy storage, catalysis, filtration, sensors, drug delivery, biomaterials, and environmental remediation due to their high surface area, interconnected pore networks, and tunable porosity. The most intriguing aspect of this approach is the potential for continuous processing of polymer blend nanocomposite templates and versatility of the monoliths that can be developed. Typically, polymer blend nanocomposites are produced continuously through high-throughput extrusion, followed by injection molding into the desired shapes. A critical addition to this process would be the polymer removal step to achieve the final monolithic structure. Our results demonstrate that achieving a high dispersion extent of nanoparticles in the preferred polymer phase, along with significant confinement induced by the small size of polymer domains and high jamming extent of nanoparticles, is crucial for the successful isolation of monoliths during the polymer removal phase. The presence of small aggregates or particles at interfaces was found to create rigid structures prone to collapse, underscoring the need for precise control over interfacial interactions, viscosity ratios, and nanoparticle compatibility with the localized polymer phase. In our study, this balance was struck in tri-continuous polymer blend nanocomposites, with sequential blending approach where the high dispersion extent of SNPs in PEO, coupled with their high affinity for it, was instrumental in achieving a high jamming extent of particles. This was evidenced by 3D confocal images, quantified characteristic size, and a two-order-of-magnitude increase in terminal moduli in the frequency sweep measurements. Additionally, TGA analysis demonstrated the complete removal of polymer phases, confirming the achievement of binder-free monoliths.

Since this is the first successful isolation of monoliths from the polymer blend nanocomposite systems, there is a significant potential for future research in this field. Further work should focus on modifying nanoparticles to achieve balanced dispersion in the thermodynamically unfavorable phase and strong affinity for the thermodynamically preferred phase across a range of practical polymer blend systems. While stable monoliths were not realized in the co-continuous systems studied, we do not rule out this possibility, recognizing that two-component polymer systems offer simplicity but face processing challenges with higher effective particle loadings. Additionally, refining monolith properties for targeted applications, such as cleantech and catalysis, will broaden their impact in high-performance materials. This approach not only tackles existing scalability challenges but also unlocks new opportunities in advanced material design, enabling the creation of multifunctional, high-performance monoliths with customized morphologies and enhanced properties for modern applications.

4 Experimental section

4.1 Materials and methods

Polypropylene (PP) with a melt index of 8 g/10 min (230 °C/2.16 kg, grade 1508) and low-density polyethylene (LLDPE) with a melt index of 2 g/10 min (190 °C/2.16 kg, grade GA502119) were provided by the Ingenia Polymers. Polyethylene-co-vinyl acetate (EVA) containing 12 wt% vinyl acetate with melt index of 8 g/10 min (190 °C/2.16 kg) and EVA containing 25 wt% vinyl acetate with melt index of 19 g/10 min (190 °C/2.16 kg) were purchased from Sigma-Aldrich. PEO (Polyox WSR N10) with molecular weight of 100 000 g mol⁻¹ was purchased from the Dow Chemical Company. Rheological properties of polymers at compounding and annealing conditions (discussed later) are shown in Fig. S1a. Sudan Blue II (98%, Sigma-Aldrich) and Rhodamine B (Sigma-Aldrich) were used to dye the EVA phase and PEO phase of the blend respectively. SNP (SEAHOSTAR KE-P10, d = 0.1–0.2 μm) was provided by Nippon Shokubai. The refractive index (RI) of PP, PE, EVA, PEO, and SNP is 1.49, 1.51, 1.50, 1.45 and 1.43 respectively.⁵⁹ The minimal RI mismatch resulted in transparent samples and allowed for deeper laser penetration and high-quality image acquisition during LSCM.

4.2 Melt-compounding and annealing

All samples were prepared using a conical twin-screw micro-compounder (MC15, Xplore) at 200 rpm and 190 °C under a nitrogen atmosphere. Extrudate from microcompounder was instantly quenched in liquid nitrogen to preserve the morphology of the blends. The loading of SNPs used in this study was calculated as wt% relative to the thermodynamically preferred phase. In case of composites based on PE/EVA/PEO blends (either with EVA12 or EVA25), the blending ratio of each polymer was fixed at 33 wt%/33 wt%/33 wt% and the amount of each fluorescent dye (Sudan Blue II, Rhodamine B) was fixed at 0.05 wt% of the total blend weight. In the simultaneous blending approach, all components (PE, EVA, PEO, SNPs and dyes) were loaded simultaneously and blended for 5 min at 190 °C. In case of sequential blending, PE, EVA, SNPs and dye (Sudan Blue II) were pre-blended for 5 min after which PEO and Rhodamine B were added and the mixing was continued for an additional 30 s. It took 20 s to load materials into the compounder at each step, and the mixing time measurement was initiated after that. Extruded samples were cut into 100–200 μm thick slices using a razor blade, sandwiched between two cover glasses, and put onto 190 °C hotplate under a nitrogen atmosphere. After pre-determined annealing time, they were air cooled, and directly used for LSCM observations. In case of composites based on PE/EVA and PP/EVA blends, the blending ratio was fixed at 50/50 wt% and the amount of fluorescent dye (Sudan Blue II) was fixed 0.05 wt% to the total blend weight. The SNPs were pre-blended with PP or PE for 5 min at 190 °C after which EVA and dye were added and the mixing was continued for an additional 5 min. In our previous work, we showed that the addition of Sudan Blue II does not alter the rheological properties of the EVA phase.⁴³

4.3 Shear rheology

Samples were molded into 25 mm discs at 190 °C for rheological measurements. Rheological tests were performed at 190 °C in a nitrogen atmosphere using a 25 mm parallel plate, rotational rheometer (MCR302 by Anton Paar). The frequency sweep tests were performed at a strain of 5%. All the samples were kept at the desired temperatures for 3 min in the rheometer before performing the tests to ensure that samples were completely molten, and any shear history was removed.

4.4 Morphology analysis

LSCM imaging was conducted using a Leica SP8 inverted confocal microscope, equipped with an 8 kHz resonant scanner. 488 nm and 638 nm lasers were used to visualize Rhodamine B-dyed PEO, and Sudan Blue II-dyed EVA respectively via fluorescent observation. 552 nm laser was used to visualize SNPs via the reflection mode. Although the refractive index mismatch between the phases is minimal, the presence of jammed silica within the PEO phase causes distinct light scattering compared to the polymer phases. This results in variations in the detected intensity of reflected light by the microscope, making it possible to detect silica nanoparticles using the reflection mode. Fig. 5 demonstrates the separation of the channels within the ternary blend following Avizo software processing, wherein the untagged PE phase is visually highlighted by assigning a distinct color.

Fig. 5 Visualization of separated channels of polymer blend images by LSCM, followed by the binarization and reconstruction using Avizo software.

20× or 63× oil-immersion objective lens was used depending on the characteristic size of the derived blend microstructure. Every image was captured with 15 μm depth (z-step size 0.25 μm). Since the signal attenuation of SNP and Sudan Blue II channels decreased with increasing imaging depth, we compensated it by the laser gain adjustment to normalize the signal. The stack of 2D images from LSCM was binarized, reconstructed and analyzed in 3D using Thermo Scientific Avizo (CMC Microsystems) to calculate the characteristic length of PP/EVA blend interface. The software used a generalized marching cubes algorithm to generate a triangular mesh to the 3D interface and calculate the interfacial area, by adding up the area of the triangles. The average characteristic length, λ, was then determined by eqn (4),

where V is the volume of the probed sample and is the sum of the area of all triangles. Further details of this calculation method have been reported elsewhere.⁶⁰ The reported values are the mean value and standard deviation of the characteristic size measurements of four samples. The representative 3D image of a tri-continuous blend of PE/EVA12/PEO is shown in Fig. 1. The PE phase, initially untagged, is visually represented as a dark channel under LSCM. To achieve better clarity, we have color-coded this channel in blue, as illustrated in Fig. 5.

4.5 Polymer removal process

Removal of polymer components was performed using a muffle furnace (FB1415M, Thermo Scientific). The extruded samples were cut into 500–600 μm thick slices using a razor blade, placed onto a crucible, and then placed inside the muffle furnace. The temperature was increased gradually at a rate of 40 °C h⁻¹ from room temperature to 500 °C, where it was maintained for 20 min. The surface of the residue was observed using table-top SEM (Phenom ProX by Phenom World) without metal-coating.

4.6 TGA

To ensure complete polymer removal, TGA was conducted on the samples after polymers were removed. TGA tests were performed using a TGA/DSC 3+ by Mettler Toledo. Heating was applied to 5 to 10 grams of the samples, increasing the temperature at a rate of 10 °C min⁻¹, starting from room temperature and reaching 800 °C. Analysis was conducted under a nitrogen atmosphere with a flow rate of 40 mL min⁻¹.

Author contributions

Aylin Habibiyan: writing – original draft, conceptualization, data curation, formal analysis, validation, investigation, methodology, project administration, visualization, writing – review and editing. Shohei Yoshida: writing – original draft, data curation, formal analysis, validation, investigation, methodology, project administration, visualization, writing – review and editing. Rajas Sudhir Shah: supporting formal analysis, methodology, validation, writing – review and editing. Milana Trifkovic: conceptualization, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, writing – review and editing.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. “There are no conflicts to declare”.

Data availability

The data supporting this article have been included as part of the SI. See DOI: https://doi.org/10.1039/d5mh01159k.

Notes and references

1. X.-Y. Yang, L.-H. Chen, Y. Li, J. C. Rooke, C. Sanchez and B.-L. Su, Chem. Soc. Rev., 2017, 46, 481–558.
2. M.-H. Sun, S.-Z. Huang, L.-H. Chen, Y. Li, X.-Y. Yang, Z.-Y. Yuan and B.-L. Su, Chem. Soc. Rev., 2016, 45, 3479–3563.
3. H. Lv, N. Li, J. Zhang, Y. Hou, X. Fan, X. Liu and F. Dang, Small, 2025, 21, 2411525.
4. X. Li, X.-T. Wu, Q. Xu and Q.-L. Zhu, Adv. Mater., 2024, 36, 2401926.
5. J. Wang, J. Tang, B. Ding, V. Malgras, Z. Chang, X. Hao, Y. Wang, H. Dou, X. Zhang and Y. Yamauchi, Nat. Commun., 2017, 8, 15717.
6. J. Wang, Y. Liu, Q. Cai, A. Dong, D. Yang and D. Zhao, Adv. Mater., 2022, 34, 2107957.
7. K. Tuo, J. Li, Y. Li, C. Liang, C. Shao, W. Hou, Z. Li, S. Pu and C. Deng, Mater. Horiz., 2024, 11, 3364–3374.
8. M. Gao, J. Zeng, K. Liang, D. Zhao and B. Kong, Adv. Funct. Mater., 2020, 30, 1906950.
9. L. Wu, Y. Li, Z. Fu and B.-L. Su, Natl. Sci. Rev., 2020, 7, 1667–1701.
10. Y. Zhong, M. Yang, X. Zhou and Z. Zhou, Mater. Horiz., 2015, 2, 553–566.
11. T. Qiu, Z. Liang, W. Guo, H. Tabassum, S. Gao and R. Zou, ACS Energy Lett., 2020, 5, 520–532.
12. H. Liu, F.-Y. Wu, G.-J. Zhong and Z.-M. Li, Mater. Des., 2023, 227, 111773.
13. R. C. Gonçalves, M. B. Oliveira and J. F. Mano, Mater. Horiz., 2024, 11, 4573–4599.
14. Ö. Sel and B. M. Smarsly, Hierarchically Struct. Porous Mater., 2012, 41–53.
15. O. D. Velev and A. M. Lenhoff, Curr. Opin. Colloid Interface Sci., 2000, 5, 56–63.
16. S.-W. Bian, Z. Ma, L.-S. Zhang, F. Niu and W.-G. Song, Chem. Commun., 2009, 1261–1263.
17. J. Xue, D. Liu, D. Li, T. Hong, C. Li, Z. Zhu, Y. Sun, X. Gao, L. Guo, X. Shen, P. Ma and Q. Zheng, Adv. Mater., 2024, 37, 2312596.
18. Q. Liu, J. Xiong, W. Lin, J. Liu, Y. Wan, C. F. Guo, Q. Wang and Z. Liu, Mater. Horiz., 2025, 12, 2436–2466.
19. C. Stubenrauch, R. Tessendorf, A. Salvati, D. Topgaard, T. Sottmann, R. Strey and I. Lynch, Langmuir, 2008, 24, 8473–8482.
20. R. Arabjamaloei, R. S. Shah, S. Bryant and M. Trifkovic, Phys. Fluids, 2021, 33, 032020.
21. Y. Lu, M. Kamkar, S. Guo, X. Niu, Z. Wan, J. Xu, X. Su, Y. Fan, L. Bai and O. J. Rojas, Small, 2023, n/a, 2300686.
22. Y. Wan and D. Zhao, Chem. Rev., 2007, 107, 2821–2860.
23. S. Deville, E. Saiz and A. P. Tomsia, Acta Mater., 2007, 55, 1965–1974.
24. U. H. F. Bunz, Adv. Mater., 2006, 18, 973–989.
25. K. H. Wong, M. Hernández-Guerrero, A. M. Granville, T. P. Davis, C. Barner-Kowollik and M. H. Stenzel, J. Porous Mater., 2006, 13, 213–223.
26. G. Hasegawa, K. Kanamori and K. Nakanishi, Mater. Lett., 2012, 76, 1–4.
27. L. Zhang, H. B. Wu, B. Liu and X. W. Lou, Energy Environ. Sci., 2014, 7, 1013–1017.
28. X. Zhang, K. Tu, Y. Xie, C. Tung and S. Xu, Adv. Mater., 2006, 18, 1905–1909.
29. J. Zheng, Q. Zeng, J. Ma, X. Zhang, W. Sun and R. Li, Chem. Lett., 2010, 39, 330–331.
30. M. Sun, C. Chen, L. Chen and B. Su, Front. Chem. Sci. Eng., 2016, 10, 301–347.
31. S. Ravati and B. D. Favis, Polymer, 2010, 51, 4547–4561.
32. Y. Shao, Z.-X. Yang, B.-W. Deng, B. Yin and M.-B. Yang, Polymer, 2018, 140, 188–197.
33. M. Nofar, R. Salehiyan and S. S. Ray, Composites, Part B, 2021, 215, 108845.
34. R. Zong, X. Hu, M. Shang, C. Wu and B. Shentu, Ind. Eng. Chem. Res., 2023, 62, 8289–8296.
35. M. J. Mochane, J. S. Sefadi, T. S. Motsoeneng, T. E. Mokoena, T. G. Mofokeng and T. C. Mokhena, Polym. Compos., 2020, 41, 2958–2979.
36. A. K. Naskar, J. K. Keum and R. G. Boeman, Nat. Nanotechnol., 2016, 11, 1026–1030.
37. S. A. Attaran, A. Hassan and M. U. Wahit, J. Thermoplast. Compos. Mater., 2017, 30, 143–173.
38. J.-W. Lee, T. H.-Q. Nguyen, E. S. Oh, S. Lee, J. Choi, H. S. Kwon, C. Wang, S. Lee, J.-Y. Lee and T.-S. Kim, et al., Adv. Energy Mater., 2024, 14, 2401191.
39. L. Ma, C. Liu, R. Dou and B. Yin, Composites, Part B, 2021, 224, 109158.
40. W. D. Harkins, J. Chem. Phys., 1941, 9, 552–568.
41. S. Torza, R. G. Cox and S. G. Mason, J. Colloid Interface Sci., 1972, 38, 395–411.
42. H. Zhang, Adv. Mater., 2020, 32, 1806328.
43. R. S. Shah, S. Bryant and M. Trifkovic, Macromolecules, 2020, 53, 10918–10926.
44. D. S. Sarul, D. Arslan, E. Vatansever, Y. Kahraman, A. Durmus, R. Salehiyan and M. Nofar, Colloid Polym. Sci., 2021, 299, 987–998.
45. D. Amoabeng, A. Tempalski, B. A. Young, B. P. Binks and S. S. Velankar, Polymer, 2020, 186, 121831.
46. R. S. Shah, S. Bryant and M. Trifkovic, J. Rheol., 2023, 67, 863.
47. T. G. Mason and D. A. Weitz, Phys. Rev. Lett., 1995, 75, 2770–2773.
48. D. I. Verrelli and A. R. Kilcullen, Adv. Condens. Matter Phys., 2016, 2016, 1716598.
49. Y. Jiang, S. Wang and Y. Zhang, Polym. Bull., 2023, 80, 12373–12392.
50. J. Zheng, D. Han, X. Ye, X. Wu, Y. Wu, Y. Wang and L. Zhang, Polymer, 2018, 135, 200–210.
51. C. Zhang, Z. Tang, B. Guo and L. Zhang, Compos. Sci. Technol., 2019, 169, 217–223.
52. B. Charmas, K. Kucio, V. Sydorchuk, S. Khalameida, M. Zie zio and A. Nowicka, Colloids Interfaces, 2019, 3, 6.
53. C. Schlumberger and M. Thommes, Adv. Mater. Interfaces, 2021, 8, 2002181.
54. K. S. Sing, Pure Appl. Chem., 1985, 57, 603–619.
55. J. C. Groen, L. A. Peffer, J. A. Moulijn and J. Pérez-Ramrez, Chem. – Eur. J., 2005, 11, 4983–4994.
56. S. Palvai, D. Kpeglo, G. Newham, S. A. Peyman, S. D. Evans and Z. Y. Ong, ACS Appl. Mater. Interfaces, 2024, 16, 5568–5581.
57. A. Feinle, M. S. Elsaesser and N. Huesing, Chem. Soc. Rev., 2016, 45, 3377–3399.
58. M. Di Luigi, Z. Guo, L. An, J. N. Armstrong, C. Zhou and S. Ren, RSC Adv., 2022, 12, 21213–21222.
59. T. Kohoutek, M. Parchine, M. Bardosova and M. E. Pemble, Colloids Surf., A, 2020, 593, 124625.
60. M. Trifkovic, A. T. Hedegaard, M. Sheikhzadeh, S. Huang and C. W. Macosko, Macromolecules, 2015, 48, 4631–4644.

---

Footnote

† These authors contributed equally to this work.

---