Regulation of Interfacial Anchoring Orientation of Anisotropic Nanodumbbells

Nanoparticles exhibiting geometrical and chemical anisotropies hold promise for environmentally responsive materials with tunable mechanical properties. However, a comprehensive understanding of their interfacial behaviors remains elusive. In this paper, we control the interfacial anchoring orientation of polystyrene nanodumbbells by adjusting interparticle forces. The film nanostructure is characterized by the orientation angle analysis of individual dumbbells from cross-sectional EM data: dumbbells undergo orientation transitions from a distinctive horizontal bilayer to an isotropic anchoring when electrostatic repulsion is suppressed by either an ionic strength increase or surface amine-modification. This anchoring orientation influences the film’s mechanical properties and foam stability, as investigated by a 2D isotherm and dark/bright-field microscopy measurements. Our findings highlight the potential for precise control of supra-colloidal structures by modulating particle alignment, paving the way for smart delivery systems.

P ickering stabilization, a phenomenon where solid nano- particles (NPs) stabilize a fluid−fluid interface, has been utilized in drug delivery, 1,2 catalysis, 3 cosmetics, 4 and food engineering. 5NPs exhibit adhesion energy significantly higher than that of molecular surfactant and thus bind to interfaces nearly irreversibly.By this, they serve as steric barriers 6,7 and make the interface highly resistant to coalescence. 6,8,9−12 Recent research has emphasized the impact of geometrical and chemical anisotropy of NPs on their interfacial behaviors. 11,13,14−18 However, most existing studies have primarily focused on individual particle level. 11Also, strategies to control the behaviors of a particle with given anisotropy profiles remain underexplored.Adding another dimension of complexity, the modulation of "supracolloidal" level characteristics (i.e., 2D film and foam formation) of ANPs by tuning the underlying interparticle forces, can provide valuable insights for a comprehensive understanding of the interfacial properties of ANPs.This knowledge can also be used to leverage the stability and programmability of ANPs in emulsification and delivery systems.
Herein, we investigate the effect of interparticle forces on the interfacial behavior and film formation of polystyrene (PS) nanodumbbells.−24 We hypothesize that changes in interparticle force would alter the interfacial orientation of nanodumbbells and subsequently modulate the mechanical properties of ANP films.
Using two-step emulsion polymerization technique, 22,23,25,26 we synthesized symmetric nanodumbbells (denoted db 1 , Figure 1A,C) from sulfonate-stabilized polystyrene spheres (sp; Figure 1A,B).A db 1 particle has two equal-sized lobes: a seed lobe (originally core−shell seed) and a budded lobe.Using an amino-silane coupling agent, we amine-modified the seed lobes of db 1 and obtained charge-anisotropic Janus dumbbells (denoted as db 2 , Figure 1A,D).Localization of amine groups on the seed lobes was confirmed by the selective adsorption of anionic gold nanoparticles (AuNPs; inset of Figure 1D and Figure S1).Using the Grahame equation and zeta potential values (Figure 1E), the charge densities (σ) of each NP and lobe were determined (Figure 1F, Table 1, and SI text). 27,28ased on the shape and charge profiles, we calculated pairwise potential energy of the three NPs anchored at the air− water interface as the sum of van der Waals attraction and electrostatic repulsion (U = V vdW + V el ; Figure 1G−I). 28,29At I = 1.7 mM, long-range electrostatic repulsion dominates, resulting in a high (∼10 4 k B T) energy barrier (U barrier ; Figure 1G).The repulsion of dumbbells is weaker than sp.For db 1 , interparticle interactions are repulsive for all approaching directions, hindering particle attachment (Figure S2).However, for db 2 , although the average energy is repulsive, attraction occurs in ∼50% of the approaching directions (U ≈ −3 × 10 3 k B T at surface-to-surface distance D = 3 nm; Figure S3), suggesting the possibility of db 2 clustering even at low salt.
At I = 0.17 M, the electrostatic repulsion is reduced (Figure 1H).For sp and db 1 , U barrier decreases to 125 and 25k B T, respectively, with equilibrium distances D eq (D at the secondary minimum) of 4.4 and 3.7 nm.This implies that two distant particles will be drawn together only until they reach an equilibrium spacing at D eq .In the case of db 2 , attraction is dominant and secondary minimum is absent.At higher I, the U barrier values of db 1 and sp subsequently decreased and became zero at I = 0.24 M (db 1 ) and 0.46 M (sp; Figure S4).At I = 1.0 M, repulsion is almost completely suppressed,    and attraction becomes dominant for all three particles (Figure 1I), suggesting that particles would strongly flocculate.
We examined the foaming ability of the NPs (Figure 2A,B).Typical anionic PS latex shows poor foaming due to electrostatic repulsion. 30Our sp particles also failed to form foams even at I = 1.0 M (Figure S5), while both nanodumbbells successfully stabilized foams.While db 2 formed foams even at I = 0 M (Figure 2B), the foaming ability of db 1 varied dramatically with ionic strength.While no foams were seen at I < 0.1 M, unstable (lifetime of ∼24 h) foams appeared at I = 0.17 M.This instability may be due to their equilibrium spacing, making the foams "leaky".Under higher I (≥0.25 M), db 1 stabilized foams as effectively as db 2 .These dumbbellstabilized stable foams exhibited remarkable resistance, lasting over one month.
Figure 2C quantifies defoaming due to ionic strength decrease.At 0.25 M, approximately half of db 2 foams remained, while db 1 foams disappeared to a greater extent (>80%).This ion-robustness of db 2 foams agrees with their interparticle attraction in a wider ionic regime.We examined the crosssection of dumbbell-stabilized foams with cryo-SEM (Figure 2D,E).The shell thickness was highly uniform (∼400 nm for db 1 and ∼1 μm for db 2 ).While spherical PS latex typically forms monolayer on foams, 30,31 our dumbbell-stabilized foams were bi-or multilayers, which explains their structural robustness.
We investigated the internal structure of the NP films from their cross sections at the planar interface.Figure 3 shows cross-sectional SEM images of the Langmuir−Schaefer films transferred at π c (onset pressure of the collapse phase, i.e., the highest pressure within the condensed phase; see Figure 4E for isotherm curves).In such highly compressed conditions, the film's structural characteristics are preserved even after transferring. 32,33As in foam cross sections, both db 1 and db 2 films were uniform in thickness (400−500 nm).The sp formed a well-defined monolayer at I = 0.17 M and a bilayer or trilayer at I = 1.0 M (Figure S6).
Notably, at I = 0.17 M, db 1 forms a horizontal bilayer, in which dumbbells are oriented horizontally as two separate monolayers, and those two monolayers are stacked (Figure 3A: the bottom layer is colored orange; note the gap between the two layers).To confirm the dumbbells' horizontal alignment, we trapped the particles in a polyacrylamide gel in situ and obtained top-view SEM images (Figure 3B).We could identify the majority of db 1 particles aligned horizontally (colored yellow).In contrast, visual inspection showed that dumbbells in both db 1 film at 1.0 M (Figure 3C) and db 2 film at 0 M (Figure 3D) exhibited a wider range of orientation angle.Although horizontal particles were still present, a significant portion of the particles anchored nonhorizontally.
For a precise analysis, we extracted the orientation angle (OA; angle between the film plane and a dumbbell's long axis) of individual dumbbells from cross-sectional SEM images.Briefly, the angles defining the 3D directions of the plane film's normal vector (θ p and φ p ) and a dumbbell's long axis vector (θ d and φ d ) were determined.Then we obtained the angle between the two vectors (Figure 3E−G; see SI text, Figures S7  and S8). Figure 3H−J plots the OA distributions of the dumbbells under three different conditions.For db 1 at 0.17 M, the horizontal orientation is predominant, showing the highest occurrence at the 0−5°bin.When electrostatic repulsion is reduced by either ionic strength change or surface modification (db 1 at 1.0 M and db 2 , respectively), the OA distributions widened, corroborating the aforementioned visual inspection.However, for all three conditions, horizontal anchoring (0− 30°) was preferred to vertical anchoring (60−90°).This preference is due to the high (5.5 × 10 5 k B T) rotational energy barrier (Figure S9), which makes a dumbbell particle unlikely to rotate vertically if a particle is initially anchored horizontally upon deposition.Nonhorizontally oriented particles may have originated from particle clusters formed right after deposition.Since tight clustering is energetically stable, once stably clustered, particle orientation would be preserved throughout the lateral compression.
To elucidate the 2D interparticle interactions, we performed surface affinity and force measurements.Figure 4A shows the interfacial NP adsorption affinity (quantified from subphase turbidity after particle deposition) as a function of ionic strength.The db 2 demonstrated robustly high surface affinity under all ion conditions.However, the adsorption affinity of db 1 and sp showed dramatic increases from ∼0% to >60%, likely due to suppressed interparticle repulsion.The db 1 surface affinity was higher than that of sp, consistent with stronger interparticle repulsion of sp.
Dark-field microscopy (DFM) revealed self-assembled cluster structures of nanodumbbells at zero surface pressure.The db 1 cluster morphology was dependent on the ionic strength.At 0.17 M, db 1 forms uniformly distributed 10 μm scaled clusters (Figure 4B).At 1.0 M, db 1 clusters grew to submillimeter sizes and became polydisperse in thickness (greater variance in pixel brightness; Figure 4C).The db 2 clusters were even larger (>1 mm) and also showed polydisperse thickness (Figure 4D).The sp clusters were not visible at 0.17 M but were observed at 1.0 M (Figure S10).
Figure 4E demonstrates the surface pressure−area per particle (π−A p ) isotherm curves of the NPs.The x-axis was rescaled with interfacial adsorption affinity (accounting for particle submersion upon deposition) and close-packing area (A CP ; Table 1).For each curve, the collapse area (A c ) and the pressure (π c ) were quantified.By calculating A CP /A c , we estimated the effective number of layers (Figure 4F).See Figure S11 for the entire isotherm curves.
The db 1 isotherm curves transformed dramatically with changing I.At very low salt (≤2 mM), π increased monotonically up to <10 mN m −1 , without sign of collapse.This suggests that db 1 particles adsorbs only weakly and submerge upon compression. 30At 0.05 M, collapse occurs near A CP , suggesting the formation of monolayer.In the range of 0.1−0.2M, the film thickness increases to bilayer (Figure 4F), which supports the occurrece of horizontal bilayer seen in Figure 3.The db 2 isotherm did not change with I with a thickness corresponding to a bilayer under all ionic conditions.The sp film's thickness showed a gradual increase from monolayer at 0 M to ∼2.5 layers at 1.0 M.This is in line with the cross-sectional SEM images in Figure 3.
Figure 4G illustrates the I-dependent changes in π c .The db 2 displayed a highly constant (25 mN m −1 ) π c .In contrast, the π c of db 1 abruptly decreased from 35 to 21 mN m −1 at I = 0.3 M. The sp also showed a sudden decrease from 40 to 23 mN m −1 at I = 0.5 M.These thresholds align well with the energy barrier decrease (Figures 5A and S4).Under the presence of equilibrium distancing, the repulsive force might effectively dissipate the lateral compressive stress throughout the film.Therefore, the film might withstand stronger lateral compression, leading to higher π c .If attraction is dominant (db 1 and sp at high salt or db 2 ), particle clusters assembled right after deposition are jammed upon compression.Cluster−cluster boundary regions might be structurally weaker, making the entire film more heterogeneous in thickness and mechanical strength. 32When stress is concentrated at those weaker regions, the film may collapse at a lower π c .This π c decrease may also explain the thinning of the db 1 film from 2 to 1.5 layers at I ≥ 0.3 M. Because collapse occurs at a lower π c , the film thickness at A c may also have decreased.
Before reaching π c , NP films undergo a wrinkling phase transition, as revealed by bright-field microscopy (BFM; Figure 4H−J, Movie S1).For db 1 at 0.17 M, three local minima are observed in the 1D autocorrelation curve (i.e., higher spatial coherence; Figure 4K).For the other two conditions, only the first minimum is identified, indicating more heterogeneous wrinkle pitch (i.e., lower spatial coherence).Consequently, the DFM, BFM, and π c data consistently indicate that ANP clustering becomes heterogeneous when repulsion is reduced.
From 1D autocorrelation curves, we quantified the wrinkles' spatial wavelengths (λ = 2 × L 1 min [L 1 min : lag at first minimum]) as λ = 5.2 μm (db 1 at 0.17 M), 6.5 μm (db 1 at 1.0 M), and 7.8 μm (db 2 ).From these values, the bending rigidity of the films was estimated.The total energy of a film on a fluid substrate is minimized when the bending energy and the substrate deformation energy are balanced. 34Thus, the relationship between the equilibrium wrinkle wavelength (λ) and the bending rigidity (B) is given as B = ρg(λ/2π) 4 , where ρ is fluid density and g is the gravitational acceleration. 34The calculated B values are 1.12k B T (db 1 at 0.17 M), 2.73 k B T (db 1 at 1.0 M), and 5.66 k B T (db 2 ).The horizontal bilayer structure of db 1 at 0.17 M has relatively weak layer−layer interaction because most particles lie horizontally and therefore do not engage in interlayer anchoring (note the interlayer gap in Figure 3A).This may make the film more susceptible to bending.Moreover, as shown in energy calculations, db 1 could exhibit equilibrium distancing instead of tight packing, which also accounts for its low rigidity.At higher salt, db 1 particles become more isotropically (nonhorizontally) oriented, increasing the number of interlayer anchoring points, which can explain increased rigidity.The db 2 film exhibits an even wider OA distribution, indicating greater rigidity.These demonstrate that interparticle energy profile and subsequent anchoring behavior of the ANPs can impact the film's mechanical properties.
From the energy calculation and experimental results, we could specify three interparticle force conditions of repulsive, weakly attractive, and strongly attractive, each corresponding to a specific ionic regime for db 1 (Figure 5A).These force conditions subsequently govern the film forming mechanism of our nanodumbbells.The boundary between regimes I and II is 0.05 M, corresponding to the lowest ionic strength condition that allowed stable layer formation (Figure S11B).The boundary between regimes II and III is 0.24 M, the condition where U barrier becomes zero and equilibrium distancing disappears.This is also where π c abruptly decreases in the 2D isotherm and the foam formation ability of db 1 becomes similar to that of db 2 .
At repulsive conditions (db 1 in regime I), dumbbells exhibit low surface affinity and do not form foams or films (Figure 5B).Tight packing does not occur and particles submerge by overcoming the attachment energy barrier of 7.3 × 10 5 k B T (Figure S9).At weakly attractive conditions (db 1 in regime II), dumbbells spontaneously approach up to the equilibrium distance (D eq ), forming thin and weak clusters.Because vertical rotation is energetically unfavorable (Figure S9C), particles transition to a bilayer arrangement while maintaining horizontal anchoring, which we have termed the horizontal bilayer.This structure may be "leaky", i.e., gaps may be present between individual particles, as evidenced by low foam stability.At strongly attractive conditions (db 1 in regime III and db 2 ), repulsion is greatly reduced, further enhancing spontaneous cluster assembly, which leads to a broader distribution of orientation angles.Upon lateral compression, intercluster jamming occurs, resulting in lower collapse pressure and higher bending rigidity.
We have demonstrated the ionic regulation of the interfacial 2D anchoring of dumbbell-shaped anisotropic nanoparticles, noting a horizontal-to-isotropic transition with reduced interparticle repulsion.This shift also changed the mechanical properties of the films.The systemic investigation and control of supra-colloidal level behaviors (i.e., interfacial orientation and film formation) of anisotropic nanodumbbells are unique to our work. 11,16,35We did so by extracting the relative orientation between the film and individual dumbbells, enabling a precise analysis of angle distribution.
We have also highlighted the distinct characteristics of db 1 and db 2 : db 1 is ion-responsive, which can switch its film structure upon an ionic change.On the other hand, db 2 exhibits ion-robustness, which maintains stable film structures under ionic changes.This can be implicated in the delivery system and emulsion stabilizers.
Future perspectives include exploring more geometrical and chemical anisotropies, such as size ratio, charge density, and polymer grafting, to manipulate a wider variety of interparticle interactions.We anticipate that this will enable innovative strategies for optimizing nanoparticle performance, increasing their applicability across industries.

Figure 1 .
Figure 1.Nanoparticle synthesis and characterization.(A) Schematic illustration of ANP synthesis by two-step seeded emulsion polymerization and selective amine-modification.The numbers of charges are in scale.(B−D) Representative SEM images of (B) sp, (C) db 1 , and (D) db 2 (inset of D: negatively charged AuNPs adsorbed on amine-modified seed lobes of db 2 ).Scale bars: 200 nm.(E) Zeta potential and (F) effective charge density of three NPs vs ionic strength at pH ≈ 7.5.Solid lines: logarithmic fits.(G−I) Pair interaction potential energy (U) of three NPs vs surfaceto-surface distance D at I = (G) 1.7 mM, (H) 0.17 M, and (I) 1.0 M. The curves of db 1 and db 2 are the average of all approaching directions and relative orientations.In part I, the curves of db 1 and db 2 are nearly overlapped.

aR:
radius of the host sphere; L: length of the dumbbell.Errors are standard deviations (n = 30).A CP : close packing area.The A CP values of dumbbells correspond to a horizontal orientation.BL: budded lobe.SL: seed lobe.

Figure 2 .
Figure 2. Ion-dependent foaming and defoaming by ANPs.(A) Dumbbell-stabilized air−water foams at I = 0.5 M 1 h after foaming.(B) Foam height measured in glass tubes (inset) 10 min after foaming as a function of ionic strength.(C) Defoaming of db 1 and db 2 by salt dilution (initial I = 1.0 M).Corresponding ionic strength is denoted at the top axis.Cryo-SEM images of foam cross sections of (D) db 1 and (E) db 2 at I = 1.0 M. Scale bars: 2 μm.

Figure 3 .
Figure 3. Anchoring structures of ANPs in the films.(A) Cross-section of db 1 film at I = 0.17 M. Horizontally oriented db 1 particles form two layers.Bottom layer is colored orange.(B) Cryo-SEM top-view image of a gel-trapped db 1 film at 0.17 M. db 1 with horizontal anchoring colored yellow, others green.(C) Cross-section of the db 1 film at 1.0 M. (D) Cross-section of db 2 film (I = 0 M).Scale bars: 1 μm.(E) Schematics of orientation angle (OA) of an ANP relative to the film plane.(F) Zoom-in of (C).Film plane and the normal vector angles (θ p and φ p ) are depicted.Four representative dumbbell particles are highlighted in red.(G) Schematics of the dumbbell particles in (F) embedded in the film plane.Dumbbell vector angles (θ d and φ d ) and orientation angles (OA) are depicted.OA distribution plots (n = 74) are (H) db 1 at 0.17 M, (I) db 1 at 1.0 M, and (J) db 2 at 0 M. OA = 0°(horizontal) and 90°(vertical anchoring) depicted with schematics.

Figure 4 .
Figure 4. 2D phase behaviors of the NP films.(A) Ion-regulated interfacial adsorption of NPs.NP adsorption was quantified by measuring subphase turbidity after deposition.Dark-field microscopy (DFM) images were taken for (B) db 1 at 0.17 M, (C) db 1 at 1.0 M, and (D) db 2 at 0 M. Images were obtained before lateral compression (π = 0 mN m −1 ).ANPs in (C) and (D) are highly clustered.Scale bars: 200 μm.(E) Surface pressure−area per particle (A p ) isotherm curves.A p (x-axis) is normalized by the close-packing area (A CP ).Collapse points are indicated by asterisks.(F) Ratio of A CP and collapse area A c showing the effective number of layers.(G) Collapse pressure π c as a function of subphase ionic strength.Bright-field microscopy (BFM) images of (H) db 1 at 0.17 M, (I) db 1 at 1.0 M, and (J) db 2 , showing wrinkle phase of ANP films.Scale bars: 100 μm.(K) 1-D autocorrelation of (H), (I), and (J).First minima are indicated with arrows.

Figure 5 .
Figure 5. Schematic illustration of nanodumbbell film formation mechanisms.(A) Three ionic regimes of db 1 and U barrier (maximal energy in pair interaction potential).U barrier of db 1 reaches zero at I = 0.24 M. (B−E) Interfacial behaviors of nanodumbbells upon lateral compression at various interparticle force conditions.(B) At repulsive condition (db 1 in regime I; I ≤ 0.02 M), dumbbells submerge upon compression and particle film is not formed due to strong repulsion.(C) At weakly attractive condition (db 1 in regime II; I = 0.05−0.24M), dumbbells form a monolayer with equilibrium spacing (D eq ).When compressed below A CP , out-of-plane slipping occurs and a horizontal bilayer is formed.(D, E) At strongly attractive conditions ((D): db 1 in regime III; I ≥ 0.24 M and (E) db 2 ), dumbbells spontaneously form tight clusters.Upon compression, intercluster jamming occurs, resulting in an isotropic film.

Table 1 .
Size Parameters and Charge Characteristics of NPs a Extended results and discussion about pH-dependent zeta potential, surface charge density, and attachment energy calculation; Experimental section; Additional data for sp; Full isotherm curves (PDF) Movie S1: Wrinkle formation of nanoparticle films (MP4) Joon Heon Kim − Advanced Photonics Research Institute, Gwangju Institute of Science and Technology, Gwangju 61005, South Korea; Email: joonhkim@gist.ac.krJin Nam − AMOREPACIFIC R&I Center, Yongin 17074, South Korea; Email: apjnam@amorepacific.com Myung Chul Choi − Department of Bio and Brain Engineering, Korea Advanced Institute of Science and Technology, Daejeon 34141, South Korea; Email: mcchoi@ kaist.ac.kr