Antigravity confined interfacial self-assembly approach for the synthesis and characterization of nanofilms

Abstract

Gravitationally induced stratification during self-assembly often leads to density-driven vertical segregation, resulting in an inherent density gradient that severely limits the synthesis of metastable nanofilms requiring inverted architectures. Here we show an antigravity confined interfacial self-assembly approach based on a liquid-liquid interface formed between hydrophilic and hydrophobic porous membranes, where capillary forces suppress gravitational effects to enable precise molecular organization. Experimental data, supported by quantum chemistry, density functional theory, and Fick’s first law, demonstrate that capillary forces enhance local concentration and interaction probability, yielding highly ordered, stable nanofilms. Our approach achieves a 17-fold increase in film area than gravity-limited methods and a 109-fold improvement over unconfined techniques. These nanofilms exhibit the stability and mechanical property, showing promise for green enhanced oil recovery and multifunctional material development. Furthermore, our strategy offers a paradigm for nanofilm mechanical characterization, paving the way for future advances in the design and application of nanomaterials.

Gravitationally induced stratification during self-assembly often leads to density driven vertical segregation limiting the synthesis of metastable nanofilms. Here, the authors address this limitation with an antigravity confined interfacial self-assembly synthesis where capillary forces dominate and enable precise molecular organization resulting in ordered and stable nanofilms.

Introduction

Confined interfacial self-assembly is a material synthesis process where molecular systems spontaneously organize into ordered structures at interfaces or surfaces under confined conditions^1–3. This approach represents a versatile and powerful approach for designing and synthesizing materials with tailored functionalities, offering broad applications in nanotechnology, environmental science, and biomedical engineering^4,5. In materials synthesis, it offers precise control over the formation of nanostructures, enabling the design of customized nanomaterials with specific properties. By utilizing confined spaces, this technique allows for the fine-tuning of the self-assembly process at the nanoscale, which is essential for creating functional materials, such as nanofilms.

However, the gravitationally induced stratification during self-assembly causes density-driven vertical segregation, typically accumulating heavier molecular components at the bottom and lighter ones at the top^6,7. This inherent density gradient critically limits the synthesis of metastable nanofilms requiring inverted architectures where dense components must occupy upper layers to ensure structural integrity. Such forced positional inversion disrupts thermodynamic equilibrium, leading to localized stress concentration and ultimately destabilizes the interfacial assembly process for compromising the synthesis stability of the material⁸.

In this work, we show an antigravity confined interfacial self-assembly (ACIS) for synthesizing self-supporting nanofilms, overcoming above limitations by employing dominant capillary forces to counteract the macroscopic effects of gravity and achieve the inverted interface. Here we strategically selected a ubiquitous and critically important cyclodextrins (CDs) to validate the feasibility and practical applicability of our methodological approach, due to CDs, a prominent class of eco-friendly compounds, have captured significant attention in the research arena^9,10. Their distinct molecular architecture bestows upon them a set of extraordinary properties, which in turn give rise to a broad spectrum of applications and exert a far-reaching influence across diverse scientific and industrial domains¹¹.

The current synthesis of CD films typically hinges on a set of interactions and techniques: host-guest interactions, hydrogen bonding interactions, interface adsorption, and electrostatic interactions, alongside methods like self-assembly, interfacial polymerization, impregnation, chemical crosslinking, casting, and various coating approaches^12–14. If the CD films are made even thinner, they will represent a critical frontier in material science due to their lightweight nature, self-healing property, fluidity, flexibility, maximize selectivity, and permeability⁷. The traditional strategy for thinning CD films primarily relies on composite synthesis using rigid supporting substrates^15,16 or liquid-liquid interface self-assembly. However, the ultrathin thickness of the films formed through such heterogeneous structures cannot be considered as the intrinsic thickness of the films themselves, as these methods result in films that cannot be transferred independently and lack self-supporting capabilities after formation. Liquid-liquid interface self-assembly uses the interface to template molecules and stabilize properties, however, the droplet interface self-assembly uses the electrodes, and two-phase interface self-assembly depends on gravity^17,18. These limitations are primarily affected by the phase density, which can hinder the selection of liquids in the self-assembly process (especially functional molecules in a phase with high phase density), large-scale synthesis of CD self-supporting nanofilms (CSNs), and the application of practical scenarios such as oil field, ultimately affecting the stability of the reaction interface, morphology, and structure of the nanofilms.

In this ACIS, the capillary force of the hydrophilic nylon and hydrophobic polytetrafluoroethylene (PTFE) membranes can effectively counteract the gravity effect to form a confined interface, which enables film formation under the same conditions where previous attempts failed (Fig. 1a). It restricts the movement direction of CD and brings them into the reduced dimensionality, enhanced proximity, and increased diffusion to interface, thus increasing the host-guest interaction and hydrogen bonding between CDs and dodecane. This interaction results in the formation of surfactant-like compounds that possess the unique capability to congregate at the water-alkane interface, thereby facilitating directional self-assembly to form the stable and highly ordered CSNs. Multi-scale modeling, based on Fick’s first law to model the material diffusion process, combined with density functional theory calculation to analyze the interaction between molecules, to reveal the driving force and influencing factors of the diffusion process from different scales (Supplementary Figs. 1–4).

Fig. 1. Synthesis mechanism of the cyclodextrin (CD) self-supporting nanofilms (CSNs) by antigravity confined interfacial self-assembly (ACIS).

a Schematic diagram of the CSNs’ synthesis, and the molecular arrangement and action mechanism of CD and alkane. Left: at the oil-water interface formed by the alkane phase and the CD solution under the influence of gravity, there are fewer CD molecules, fewer self-assembled molecules, and weak interfacial film. At the antigravity confined interface, the CD solution infiltrated in the nylon membrane and the dodecane infiltrated in the polytetrafluoroethylene (PTFE) membrane can overcome gravity, and the CD molecules can form surface-active substances through host-guest interaction with dodecane at the interface, and the surface-active molecules self-assemble into CSNs under the action of hydrogen bond. Right: snapshot of oil-water interface, snapshot, micrograph, and Scanning Electron Microscopy (SEM) image of the CSN. From the side view, the thickness of the CSN is about 90 nm. b Left: molecular configuration and intermolecular interaction at confined interfaces. The smaller the membrane pores (d) are, the stronger the capillary force (F) will be. The thicker the infiltrated liquid thickness (L) is, the greater the gravitational force (G) and buoyancy (f) will be. γ is the interfacial tension. Right: covalent and hydrogen bonds between the CD-dodecane complexes. The blue isosurface of the structure represents the presence of hydrogen bonds, and the yellow part represents covalent bonds.

At the oil-water interface formed by the alkane phase and the CD solution under the influence of gravity, the number of CD molecules in the lower layer is limited, resulting in fewer self-assembled molecules and weaker interfacial films. In contrast, at the antigravity confined interface, the CD solution infiltrated in the nylon membrane and the dodecane infiltrated in the PTFE membrane can overcome the phase gravitational constraints by the capillary force. Here, the CD molecules can form surface-active substances through host-guest interaction with dodecane at the interface, which subsequently self-assemble into the CSNs driven by hydrogen bond (Fig. 1a, Left). Characterization of the CSNs, particularly after the removal of excess liquid, reveals a remarkably uniform and smooth surface morphology, as evidenced by Scanning Electron Microscopy (SEM) imaging. Notably, the CD nanofilms exhibit thickness in the sub-100 nm range, highlighting the effectiveness of this method for synthesizing nanofilms at the antigravity confined interface.

In terms of CSNs synthesis mechanism, the integration of a β-CD infused nylon membrane with a dodecane-saturated PTFE counterpart creates an antigravity confined interface (Fig. 1b). This interface acts as a crucible where CD molecules migrate and engage in interactions with dodecane, leading to the formation of surface-active substances that mitigate interfacial tension. The diffusion rate of β-CD molecules to the interface and the amount of adsorption at the interface can be calculated by the interfacial tension (Supplementary Figs. 5, 6, and Table 1). The relationship between the transmembrane pressure difference (∆P), membrane pores (d), and the infiltrated liquid thickness (L) was calculated (Supplementary Fig. 7 and Tables 2–4). According to Laplace’s equation, Hagen-Poiseuille’s law, and Washburn equation, smaller membrane pores, characterized by low liquid surface curvature radius and high additional pressure, will produce significant resistance to the entry and passage of the liquid, eventually leading to greater liquid retention in the confined space between nylon and PTFE membranes. The congregated molecules, through a symphony of covalent and hydrogen bonding, orchestrate the synthesis of the CSNs. Quantum chemistry was employed to optimize the structure of β-CD-dodecane complexes, and the interaction region indicator (IRI) function can intuitively visualize the nature of the interactions. This approach not only realizes the synthesis of nanofilms, but also opens avenues for the development of nanofilm materials with tailored properties.

Results

Design, synthesis, and characterization of the CSNs

The synthesis of nanofilms at the liquid-liquid interface is widely recognized as a straightforward, efficient, and eco-friendly process for producing compact, robust, and easily controllable films^6,8,19. By leveraging capillary forces, liquids can be confined within micro-structured porous solid matrices to form the confined interfaces^20,21. Unlike current systems, the asymmetric hydrophilic–hydrophobic porous membranes employed here are designed to overcome gravitational limitations, significantly expanding their potential applications across diverse scientific fields²². Through the strategic arrangement of solid matrices with contrasting properties to encapsulate two immiscible liquids, a stable antigravity confined interface can be established. This interface is particularly susceptible to electrostatic interactions and van der Waals forces, providing an optimal environment for the synthesis of CD nanofilms.

Key factors influencing the properties of the self-assembled nanofilms include the pore size of the solid substrate and the composition of the liquids within the ACIS. In this study, substrates with pore sizes of 1 μm, 5 μm, and 8 μm were used, while the ACIS primarily consisted of CD solutions and alkanes (Fig. 2a). CDs with biocompatible and biodegradable properties present a truncated cone shape with a central cavity, and they are widely utilized in pharmaceuticals, food industries, and biomedicine. The three natural forms (α-, β-, and γ-CD) possess hydrophobic cavity inner diameters of 4.7–5.3 Å, 6.0–6.5 Å, and 7.5–8.3 Å, respectively. Their amphiphilic nature, featuring a hydrophilic exterior and hydrophobic interior, enables host-guest interactions with hydrophobic molecules in aqueous environments^23,24. As such, CDs serve as ideal hosts for alkanes, sheltering their hydrophobic moieties. Different alkanes may form different complexes with β-CD^24,25.

Fig. 2. Synthesis and properties of the CSNs.

a Membrane pore, water phase, and oil phase options of the confined antigravity interface. There are few CD molecules in the aqueous phase, and the density is close to that of water. The oil phases are octane, decane, dodecane, and hexadecane. b Critical pressures (black) of the CSNs and carbon contents (red) in the nylon membrane of the CSNs with different CD solutions. c Critical pressures (black) of the CSNs and carbon contents (red) in the PTFE membrane of the CSNs with different alkanes. d Critical pressures for gas to flow through dodecane and β-CD solution interface (gray) and the CSNs (red) in different pore sizes. e Cyclic test of the critical pressures for the gas through the CSNs before (black) and after (purple) film synthesis. f The areas of films formed by the gravity interfacial self-assembly (GIS), the unconfined interfacial self-assembly (UIS) and the ACIS (T Triangle, S Square, and C Circle) with 26 μL of β-CD solution and 6 μL of dodecane. Error bars in (b–d, f) represent the s.d. (n = 3) and data are presented as mean values ± s.d.

Confocal microscopy (Supplementary Fig. 3) revealed that the confined space formed by infusing β-CD solution into a nylon membrane and dodecane into a PTFE membrane is approximately 4  μm wide. The aqueous phase, containing trace amounts of CD, has a density similar to water, while the oil phases, comprising octane, decane, dodecane, and hexadecane, exhibit lower densities.

Initial morphological characterization of films synthesized from water and CD solutions in dodecane was conducted via microscopy (Supplementary Fig. 8). We further investigated the transmembrane behavior for the gas through the CSNs by measuring the critical pressure required for gas to penetrate interfacial films formed between water or CD solutions and dodecane (Fig. 2b). Notably, the CSNs derived from β-CD and dodecane exhibited a higher critical pressure (~49.76 kPa) than those from other CDs, indicating greater mechanical stability and resistance to gas permeation.

We also examined β-CD behavior in alkanes with varying chain lengths (Supplementary Fig. 9). Crystal formation was observed when β-CD solution was introduced into dodecane and hexadecane, attributable to the binding of one long-chain alkane molecule to two β-CD molecules²⁵. The CSNs were fabricated by infusing β-CD solution into nylon membranes and different alkanes into PTFE membranes. Critical pressure measurements for gas transmission through these CSNs (Fig. 2c) revealed that the system incorporating dodecane required the highest critical pressure (~49.76 kPa), underscoring the superior mechanical strength and stability of nanofilms self-assembled from β-CD-dodecane complexes.

The stability of the confined interface was quantified by measuring the critical transmembrane pressure required to rupture the liquid film and induce reconfiguration of the capillary-stabilized liquids^26,27. In this system, a nylon membrane was used to confine the β-CD solution, while a PTFE membrane confined the dodecane (Supplementary Fig. 10). The selection of appropriate pore sizes for each membrane proved crucial for successful nanofilm formation. As shown in Fig. 2d, the critical pressures were evaluated for nanofilms formed using dodecane infiltrated into PTFE membranes (with 1 μm and 5 μm pores) and β-CD solution infused into nylon membranes (1 μm, 5 μm, and 8 μm pores). Following nanofilm synthesis, the critical transmembrane pressure increases across all pore sizes compared to the film-free system, indicating enhanced mechanical integrity of the interface.

The pore size of the nylon membrane had a greater impact on the critical pressure than that of the PTFE membrane, which we attribute to the diffusion of CDs through the nylon pores and the trajectory of gas pressure, which traverses the nylon membrane before the PTFE membrane. These membranes serve primarily to support the liquids in forming a stable antigravity interface. Smaller pores (e.g., 1 μm) require higher pressures to overcome capillary resistance, whereas larger pores (e.g., 8 μm) make the interface more susceptible to gravitational deformation due to the density contrast between the liquids. Among the configurations tested, the system combining a 5 μm PTFE membrane (dodecane) and a 5 μm nylon membrane (β-CD solution) proved optimal for fabricating and characterizing composite CSNs, based on the magnitude and consistency of pressure changes before and after film formation.

The ACIS establishes a liquid-liquid interface through capillary confinement of immiscible fluids within hydrophilic and hydrophobic porous matrices. The stability and antifouling properties of CSNs are critical for practical applications^26,28. A common method to evaluate thin-film stability is Laplace pressure-driven film rupture^29,30. Here, the marked increase in transmembrane pressure after CSN formation enables stability assessment via cyclic pressure loading and release (Supplementary Fig. 11). As depicted in Fig. 2e, CSNs derived from β-CD and dodecane exhibit exceptional stability: even after disruption at the critical pressure, the nanofilm reforms spontaneously upon pressure release, mediated by host-guest interactions and hydrogen bonding. This self-healing behavior ensures the long-term stability of the film as long as the β-CD/dodecane interface is maintained. Furthermore, no film residues were observed via SEM after rinsing (Supplementary Fig. 12 and Tables 5–6), confirming the antifouling characteristics of the ACIS.

Compared with prior methods for synthesizing interfacial films, a key advantage of the ACIS is its capacity to form large-area interfaces with the same volume. The dimensions of the nanofilms are determined by the membrane area over which the β-CD solution and dodecane are distributed (Supplementary Fig. 13). Using 6 μL of dodecane and 26 μL of β-CD solution infused into 25 mm-diameter membranes, the ACIS produces an interface congruent with the membrane area. In contrast, gravity interfacial self-assembly (GIS) yields an interface of only ~28 mm², and unconfined interfacial self-assembly (UIS) results merely in peripheral contact between the two liquids, leading to minimal film formation. As summarized in Fig. 2f, the nanofilm area achieved via ACIS is approximately 17 times larger than that from GIS and 109 times greater than that from UIS, highlighting the efficiency of this approach for producing large-area nanofilms under equivalent volumetric conditions.

Condition and theoretical mechanism of the synthesis of the CSNs

Based on our previous investigations into the design of aqueous and alkane phases for the properties of the CSNs, we selected β-CD solution as the aqueous phase infused in a nylon membrane and dodecane as the alkane phase within a PTFE membrane. Figure 3a illustrates the molecular mechanism at the antigravity confined interface, where β-CD molecules eventually polymerize via hydrogen bonding. Initially, β-CD molecules dispersed in the aqueous phase diffuse toward the dodecane-water interface over time and with increasing concentration. At the interface, β-CD orients itself to encapsulate dodecane via host-guest interactions, forming a complex that reduces interfacial tension and drives the interfacial assembly. Subsequently, these complexes undergo interfacial polymerization through hydrogen bonding with β-CD molecules in the bulk phase, enhancing the stability and mechanical strength of the film.

Fig. 3. Influence factors and theoretical mechanism of the CSNs.

a The molecular mechanism diagram of the β-CD and dodecane. With the increase of time and concentration of β-CD, β-CDs diffused to the interface and bonded with dodecane through host-guest interaction at first, then β-CD-dodecane complexes at the interface reduced the interfacial tension, and finally the CD molecules at the interface were polymerized by hydrogen bonding. b Critical pressures of the CSNs with different concentrations of β-CD solutions and the dodecane. Scale bar, 1 μm. c Critical pressures of the CSNs with 0.004 mol/L β-CD solutions and the dodecane at different times. Inset: molecular arrangements of the dodecane and β-CD at the antigravity interface. d Dynamic interfacial tensions of 10 μL water or 10 μL CD solution in the dodecane. e Dynamic interfacial tensions of 10 μL of β-CD solution in the alkanes. f Dynamic interfacial tensions of 10 μL β-CD solution with different concentrations in the dodecane. g The binding energy of the CD and the dodecane, and the β-CD and the alkane. h CD-alkane complexes structure with the lowest energy calculated under B3LYP/6-311 G(d) level using Gaussian 16 program. i Covalent and hydrogen bonds between adjacent β-CD-dodecane complexes. Error bars in (b–g) represent the s.d. (n = 3) and data are presented as mean values ± s.d.

The concentration of β-CD solution plays a critical role in the diffusion and interfacial binding of β-CD molecules, significantly influencing nanofilm formation. We characterized interfaces formed between dodecane and β-CD solutions at varying concentrations (Supplementary Fig. 14). As shown in Fig. 3b, we evaluated the critical transmembrane pressure required for gas to penetrate the CSNs synthesized with β-CD concentrations ranging from 0.001 to 0.005 mol/L. Compared to a pure water-dodecane interface, higher β-CD concentrations led to increased diffusion to the interface and correspondingly higher critical pressures, confirming enhanced mechanical strength due to molecular assembly. At 0.004 mol/L, a continuous nanofilm formed, effectively covering the nylon membrane pores (Supplementary Fig. 15). Below this concentration, binding between β-CD and dodecane was insufficient for nanofilm formation. Film thickness increased with concentration, as evidenced by comparisons between 0.004 and 0.005 mol/L samples (Fig. 1a and Supplementary Fig. 16).

Time-dependent diffusion and complex formation at the interface also affect film formation. β-CD molecules require time to diffuse, bind dodecane via host-guest interactions, and form hydrogen-bonded networks. As shown in Fig. 3c, the critical pressure increased over time, indicating strengthening of the film. For instance, with 0.004 mol/L β-CD, the critical pressure rose from 44.35 kPa initially to 49.76 kPa after 2 h, indicating full nanofilm formation. Beyond this period, the interfacial molecular configuration saturated, and no further increase in pressure resistance was observed.

Interfacial tension measurements provide further insight into molecular behavior. We monitored dynamic interfacial tensions between dodecane and CD solutions (Fig. 3d and Supplementary Fig. 17). While β-CD and γ-CD solutions exhibited rapid interfacial tension reduction followed by stabilization, α-CD solution showed minimal change within 60 min, indicating β-CD and γ-CD molecules bind with the dodecanes to form the surface-active substances at the interface, and ineffective binding for α-CD³¹. The lowest equilibrium interfacial tension was observed with β-CD, confirming its superior ability to form amphiphilic complexes with dodecane.

We also examined β-CD interfacial behavior with alkanes of varying chain lengths (Fig. 3e and Supplementary Fig. 18). Short- and medium-chain alkanes (octane, decane, dodecane) showed typical interfacial tension reduction curves, whereas hexadecane exhibited an initial rise followed by a decline, suggesting initial formation of non-amphiphilic complexes that later reorganize. Among these, the dodecane with medium-chain length yielded the lowest equilibrium interfacial tension with β-CD, indicating optimal complex stability.

Concentration-dependent interfacial tension profiles of β-CD solutions in dodecane are shown in Fig. 3f and Supplementary Fig. 19. Concentrations up to 0.004 mol/L of β-CD solution exhibited classical amphiphile-like interfacial tension decay. At 0.005 mol/L, however, an initial interfacial tension increases suggested formation of non-amphiphilic complexes involving multiple β-CD molecules per dodecane, leading to less efficient interfacial tension reduction and altered film morphology. After all the dynamic interfacial tensions of β-CD solutions with different concentrations reach equilibrium values, we find that the interfacial tension of 0.004 mol/L of β-CD solution is the lowest, and the interfacial tension of 0.005 mol/L of β-CD solution is the highest. In a word, one β-CD molecule in 0.004 mol/L of β-CD solution and one dodecane molecule can form a strong amphiphilic complex. Those complexes at the interface will aggregate into a nanofilm with the interaction between β-CDs and dodecanes and the interaction between β-CDs³¹.

Given the density difference between CD solutions and alkanes, antigravity confinement promotes CD diffusion toward the underlying alkane phase. To elucidate the molecular basis of these observations, we performed quantum chemical calculations to characterize host-guest interactions between CDs and alkanes³². We used the XTB program for the configuration search to find the optimal structures for each CD-alkane complex. And the semiempirical quantum mechanical methods and the hybrid density functional B3LYP together with the basis set of 6-31 G(d) are used to generate and optimize the complexes^33,34. And then, the binding energies of the complexes with the lowest energy configurations were further calculated under B3LYP/6-311 G(d) level using the Gaussian 16 program and revealed with IRI method³⁵ using the software Multiwfn³⁶. As shown in Fig. 3g, the binding energy for β-CD and dodecane was approximately −28 Kcal/Mol, significantly stronger than other CD-alkane complexes (all > −20 Kcal/Mol). Accordingly, the lowest energy configurations of CD and alkane are shown in Fig. 3h.

It is necessary to use quantum chemistry methods to optimize and search for the lowest potential energy surface to obtain the most stable conformation, and further analyze the intermolecular and intramolecular interactions. Further analysis of the lowest energy configuration revealed hydrogen bonding between adjacent β-CD molecules at the interface and with those in the bulk (Fig. 3i and Supplementary Fig. 20). The interfacial β-CD molecules are oriented with their larger rim facing downward to engage with alkane molecules. Two predominant stacking modes were identified: small-ring to large-ring and small-ring to small-ring. Both configurations exhibited similar stability, suggesting their concurrent presence at the interface.

Applications

We further demonstrate the versatility of this method across multiple liquid-liquid systems, with potential applications in materials science, emulsion processing, and the petroleum industry (Fig. 4a, Right). The morphology of the nanofilms is determined by the shape of the solid substrates used to confine the β-CD solution and dodecane. As illustrated in Fig. 4a, Left, employing nylon and PTFE membranes patterned as circles, hearts, triangles, five-pointed stars, and leaves enables the synthesis of correspondingly shaped CSNs via the ACIS.

Fig. 4. Applications of the ACIS to synthesize the CSNs.

a Applications of the ACIS. Right: application fields, such as materials engineering, food processing, emulsion preparation and oil displacement. Radar chart of the comparisons between GIS and ACIS. Left: the synthesized CSNs with different shapes, such as the circle, heart, triangle, five-pointed star, and leaf. Scale bar, 10 mm. b Stored temperature of dodecane by the ACIS. After 3 min, the dodecane by the GIS cooled to 68 °C and by the ACIS cooled to 72 °C. c Prepared dispersion system by the ACIS. Left: under the GIS, it presents a stratification phenomenon. Right: under the ACIS, it shows an obvious oil-in-water state. Scale bar, 1 cm. d Increased oil displacement effect in oil recovery by the ACIS. P₁ is the pressure of water flooding oil of the cores with the GIS, and P₂ is the pressure of water flooding oil of the cores with the ACIS. Re is the oil recovery rate. Scale bar, 1 cm. e Testing principle of mechanical strength of the CSNs. P₀ = 0 stands for uncompressed gas, P₁ is the critical pressure of the liquid clamp system without films, and P₂ is the critical pressure of the liquid clamp system with films. P₁ is enough to open the system without films but not enough to open the system with film. f The relationship between mechanical strength, interfacial tension, and transmembrane pressure of the CSNs. The three points correspond to the pressures presented in (e). ∆P is the critical pressure difference between systems with and without films. g Mechanical strength before and after film synthesis of different systems synthesized by the ACIS in materials engineering, food processing, emulsion preparation, and oil displacement. Error bars in (b, g) represent the s.d. (n = 3) and data are presented as mean values ± s.d.

Comparisons between ACIS and GIS are showed in Fig. 4a, Right. As shown in gray, GIS in certain systems suffers from the formation of interfacial films with poor mechanical strength and limited area, and also exhibits significant limitations in practical applications. However, ACIS in these systems can effectively solve some limitations of GIS, as shown in pink. As shown in Fig. 4b and Supplementary Fig. 21a, the temperature of dodecane under ACIS confinement remains 3 °C higher than that under GIS after 3 min, underscoring the potential of CSNs in insulation materials. A finding further validated in food insulation applications (Supplementary Fig. 22).

The ACIS also facilitates unique emulsion behaviors. Unlike GIS, where alkane and CD solution rapidly separate into distinct layers, the ACIS allows the dispersion of CD solution droplets within the alkane phase (Fig. 4c). This emulsification process, stabilized by interfacial nanofilms, suggests promising applications in areas such as pesticide emulsifier formulation.

Aligned with global sustainable development goals, green petroleum extraction remains a critical endeavor. As a result, enhanced oil recovery has become an essential and indispensable part of this strategy³⁷. The tertiary oil recovery technology based on oil-water interface behavior regulation has received widespread attention all along^38,39, and the in-situ films formed by the ACIS are expected to have applications in green enhanced oil recovery (Supplementary Fig. 23 and Table 7). A key challenge in chemical flooding for the Enhanced Oil Recovery (EOR) is the preferential flow of water through high-porosity and oil-depleted channels due to viscosity contrasts and pore heterogeneity. Without interfacial films, water flooding requires only low pressure (0.26 MPa) and primarily recovers oil from large pores. Following CD film formation, however, the required transmembrane pressure rises to 8.65 MPa, enabling displacement from smaller pores (Fig. 4d). The enhanced pressure tolerance and selective permeability offered by CSNs improve adaptability to subsurface heterogeneity³⁹. Moreover, CDs are biocompatible and biodegradable, providing an ecologically superior alternative to conventional displacement agents such as synthetic surfactants and polymers³⁸. This inherent advantage not only simplifies the displacement mechanism but also synergistically amplifies the displacement efficiency, ensuring an optimal synergy between the methodological approach and the material properties for maximized outcomes.

Characterizing the mechanical strength of ultrathin films remains a significant challenge⁴⁰. There are some conventional methods to characterize the mechanical strength of ultrathin films including the SEM to directly observe the thickness of the film, the load-bearing method to place the weight above the film, the tensile method to test the stress-strain, and the Young’s modulus of the film, etc^41–44. However, using the above methods is likely to destroy the CSNs during the moving and transferring processes of the films before the test operation, and it is difficult to detect and compare the mechanical strength of the films⁴⁵. To address this, we developed a liquid clamp device based on ACIS for in-situ mechanical characterization via transmembrane pressure. In the initial state of the ACIS, opening the system requires a certain pressure (P₁) generated by the compression of gas to overcome the capillary force of the gating liquid in the membrane channel. After the ultrathin film is synthesized, the pressure that would open the initial state of the ACIS will not open the film fabricating state of the ACIS (Fig. 4e). To open the film fabricating interface, the higher pressure (P₂) is required. The critical pressure difference (∆P) required to open the interface before and after ultrathin film formation serves as a direct measure of film mechanical strength (Fig. 4f).

We applied this method across several systems (Fig. 4g and Supplementary Fig. 21b). The CSNs produced via ACIS exhibited a ~ 2 kPa higher ∆P than those from gravity confined interfaces, confirming superior mechanical integrity. For instance, millet porridge films showed greater strength than milk films.

In EOR contexts, the mechanical strength of interfacial films is critical in controlling the flow characteristics of oil-water mixtures: overly weak films permit premature water breakthrough, while excessively strong films may impede flow or elevate injection pressures. Tailoring film strength to specific reservoir conditions can promote fluid diversion and improve microscopic sweep efficiency⁴⁰. Currently, few experimental methods exist for quantifying oil-water interfacial film mechanics. Our approach enables in-situ strength measurement and controlled self-assembly of interfacial films, offering a strategy for enhancing oil recovery in challenging reservoirs. For example, we tested the in-situ CSNs synthesized by the two types of petroleum, and the North three area oil exhibited a larger transmembrane pressure difference, indicating stronger films and better displacement performance.

Discussion

In this work, we developed an ACIS for synthesizing CSNs, offering a scalable and versatile platform for molecular assembly. By leveraging confined interfaces, this method enables precise molecular organization, yielding nanofilms with high uniformity and structural integrity. We combined Fick’s first law with density functional theory calculations to model diffusion behavior and interfacial interactions, revealing how spatial confinement enhances assembly efficiency. These findings suggest that the ACIS strategy could be extended to a wide range of material systems, highlighting its potential for interdisciplinary applications.

The in-situ the CSNs exhibit adaptive reconfigurability under complex conditions, such as in oil reservoirs, where they sustain intrusion pressures 4.7 times higher than traditional surfactant systems, enabling efficient oil recovery from micropores. The methodology also allows customization geometries and accurate mechanical characterization of the CSNs, making they promising for applications in materials science, food processing, and chemical engineering.

Our future work will pursue an integrated strategy to advance this method for complex field applications. This entails scaling up the synthesis and developing a comprehensive materials database to discover ACIS-compatible functionalities and optimize chemistries. In parallel, we will advance diagnostic methods for characterizing interfaces under realistic temperature and pressure. The culmination of these efforts will be the validation of our method’s performance in specific industrial applications, such as enhanced oil recovery in ultra-low permeability formations.

Methods

Synthesis of the antigravity confined interfacial self-assembly

The water, 0.004 mol/L of α-CD, 0.001 mol/L of β-CD, 0.002 mol/L of β-CD, 0.003 mol/L of β-CD, 0.004 mol/L of β-CD, 0.005 mol/L of β-CD, and 0.004 mol/L of γ-CD solutions were used as the functional liquids infused to the nylon membrane used as the porous material in the upper layer of the ACIS. And in the lower layer of the ACIS, the octane, the decane, the dodecane, and the hexadecane were used as the functional liquids infused to the PTFE membrane used as the porous material. The ACIS was the overlap of the nylon membrane infused with the CD solution and the PTFE membrane infused with the alkanes to form the confined antigravity interface.

Characterizations of membranes

Microstructures and energy spectrum diagrams of the nylon membrane, the nylon membrane infused with the CD solution, the PTFE membrane, and the PTFE membrane infused with the alkanes in the ACIS were measured by using Field emission SEM (Zeiss GeminiSEM 500).

Transmembrane critical pressure measurements

The CD solutions were infiltrated into the nylon membrane, and the alkanes were infiltrated into the PTFE membrane. And then, the gas was injected to pass through the overlapping the nylon membrane with CD solutions and PTFE membrane with the alkanes. Transmembrane critical pressures for gas passing through the ACIS system were measured by using the liquid gating intelligent tester and analyzer (GIFT-E1 Pro) from Gating Inspired Future Technology Co., Ltd. (Xiamen, China). In all experiments, the gas flow pressures were produced by a syringe pump with a pressurization rate of 2 mL/min. Unless otherwise specified, the pore sizes of the nylon membranes and the PTFE membranes with diameter of 25 mm are 5 μm. All experiments were repeated at least three times.

Stability of film

The gas was injected to pass through overlapping the nylon membrane infiltrated with β-CD solution and PTFE membrane infiltrated with the dodecane. When gas passes through the system, the film is destroyed. After the gas input stops, the film continues to synthesize. After several times of gas injection and gas removal, the transmembrane pressure of the film is stable, indicating that if the liquid-liquid interface exists, the film will stably synthesize. Transmembrane critical pressures produced by the compression of the gas to pass through the CD self-supporting nanofilms (CSNs) were measured by using the liquid gating intelligent tester and analyzer (GIFT-E1 A). In all experiments, the pressures were produced by the 2 mL/min of gas flow by a syringe pump. The pore sizes of the nylon membranes and the PTFE membranes with diameter of 25 mm are 5 μm. All experiments were repeated at least three times.

Characterization of film area

In this part, three methods are used to synthesize the films. By the ACIS, because the 26 μL of β-CD solution will infiltrate in the nylon membrane, and 6 μL of dodecane will infiltrate in the PTFE membrane, the area of film formed will be consistent with the area of the membranes. By the gravity interfacial self-assembly (GIS), 6 μL of the dodecane floats on 26 μL of the β-CD solution. By the unconfined interfacial self-assembly (UIS), when the 26 μL of β-CD solution touches the 6 μL of dodecane, it pushes the spread 6 μL of dodecane toward the edge, so that only the edge of the β-CD solution forms an interface with the dodecane. The areas of films are calculated using the radius measured by the ruler.

Interfacial tension measurements

The water, 0.004 mol/L of α-CD, 0.001 mol/L of β-CD, 0.002 mol/L of β-CD, 0.003 mol/L of β-CD, 0.004 mol/L of β-CD, 0.005 mol/L of β-CD, and 0.004 mol/L of γ-CD solutions were prepared by the deionized water. 10 μL of these solutions were injected into the alkanes to directly measure via the pendant drop method on the OCA100 system (Dataphysics, Germany) at room temperature. All experiments were repeated at least three times.

The binding energy calculations

Quantum Chemical (QC) methods were used to investigate noncovalent interactions, such as hydrogen bonds and van der Waals forces, which are essential for the stability of CD complexes. To investigate the host-guest interactions between CDs and alkane molecules, we adopted a multi-level computational strategy. Given the substantial size of CD complexes, the semiempirical quantum mechanical methods GFN1-xTB³³ and GFN2-xTB³⁴ were first applied for efficient conformational sampling. Then, the single point energies and binding energies based on the structure with the lowest energy were further calculated under B3LYP/6-311 G(d) level using Gaussian 16 program, with Grimme’s D3BJ dispersion correction applied throughout to enhance accuracy. Solvation effects were incorporated via the Solvation Model based on Density (SMD) as an implicit solvent model. Finally, the binding natures in the studied complexes were revealed with IRI method³⁵ using the software Multiwfn³⁶, and binding energies were obtained based on the following equation:

$$E_{binding}=E_{cplx\_opt}-\left(E_{CD\_opt}+E_{guest\_opt}\right)$$ 1

All the visualization of molecular structures and contour surfaces were performed by VMD⁴⁶. For the conformer search, the temperature is 298.15 K, and the energy window is 6 kcal mol⁻¹.

Visual analysis

The types of interactions that stabilize such complexes, including hydrogen bonding, van der Waals forces, and both hydrophobic and dipole-dipole interactions, typically cannot be modeled with the necessary accuracy using molecular mechanics methods. Graphical analysis of interactions is a useful way to examine chemical problems and visualize the type of interaction at a particular location in the system. The interaction area indicator function (IRI)³⁵ is one of them, through its isolines, all types of interactions can be clearly grouped in a picture. IRI methods are excellent at revealing chemical bonds and weak interaction regions. Unlike Bader’s atoms in molecules (AIM) method, which is a method of calculating atomic charge based on real space segmentation, the IRI method more intuitively demonstrates the nature of the interaction. It is defined as follows:

$$IRI\left(r\right)=\frac{\mid \nabla \rho \left(r\right)\mid }{{\left[\rho \left(r\right)\right]}^a}$$ 2

Where ρ(r) is the electron density. If the iso-surface of IRI is blue, it indicates the van der Waals force. If the iso-surface is purple, it indicates that there is some steric hindrance. The yellow ones are covalent bonds. In addition to the intermolecular interactions, various intramolecular weak interactions can also be observed. And in this system of work, the weak interaction refers to hydrogen bonding.

Different shapes material synthesis

The nylon and PTFE membranes are cut into different shapes, including circles, hearts, triangles, pentagrams, and leaf shapes. The β-CD solution will infiltrate in the nylon membrane, and the dodecane will infiltrate in the PTFE membrane. After overlapping the two membrane systems for two hours, the corresponding shape of the films are synthesized. Dyes are soluble in alkanes and insoluble in water, and we found yellow dye on the hydrophilic nylon membrane, which further confirmed the synthesis of the film.

Insulation experiment

25 mL of the β-CD solution and the dodecane samples are heated to 85 °C and transferred to a beaker to observe the temperature change of the samples. In the GIS system, where the dodecane is on top, the dodecane dissipates heat faster. In the ACIS system, if the PTFE membrane and nylon membrane are fixed, the upper layer of the β-CD solution will not flow into the lower layer, and the system will slow heat dissipation. The sample without films cooled faster within 3 min.

Dispersion system preparation

There are two methods for dispersion system preparation. The GIS uses the dense β-CD solution phase as the dispersing phase and the dodecane as the dispersing substance. The ACIS uses the dense β-CD solution phase as the dispersing substance and the dodecane as the dispersing phase. The dispersing substance is placed into the syringe and dripped through the corresponding nylon and PTFE membranes with confined space to the dispersing phase.

Oil displacement experiment

A uniform Berea core with a gas permeability of 30 × 10⁻³ μm² was vacuumed for 4 h at room temperature and saturated with deionized water to calculate pore volume and porosity. The core was then heated for 12 h at a constant temperature of 25 °C. The original oil saturation of crude oil was calculated by using the oil to flush water. Next, the deionized water or β-CD solution was injected into the core. When the cumulative injection volume reached 2 PV, the water or β-CD solution was stopped. In this work, it is worth noting that we first injected 0.5 PV of the deionized water or β-CD solution, and then suspended for 2 h before further injected remaining 1.5 PV of liquid. Dynamic pressure and recovery were recorded during displacement. During the experiment, the injection rate was constant at 0.35 mL/min. A similar core is used when the injection liquid is changed. Each experiment was repeated three times.

Author contributions

X.H. conceived the project. X.H., J.M.L., Z.H.Z., and W.F.L. designed the research. J.M.L., Z.Y.Z., G.L., and S.J.Y. performed the experiments. J.M.L., Y.Y.C., and X.H. designed and analyzed the theoretical model. J.M.L., X.Z., Y.F., and X.H. designed the devices for the applications. Z.H.Z., J.M.L., G.L., Q.Z., L.H., N.H.J., B.Z., and W.F.L. designed and analyzed the applications. J.M.L., X.H, and Z.H.Z. wrote the manuscript. All authors edited, read, and approved the manuscript.

Peer review

Peer review information

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

Data availability

The data supporting the findings of this study are available within this article and its Supplementary Information. Source data are provided with this paper. All data are available from the corresponding author upon request. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-68447-8.