Oligomerisation Drives Internal Cavity Formation and Conformational Heterogeneity in Mixed Triton X‐100/Tyloxapol Micelles Authors: Hrachya Ishkhanyan, David J Barlow, M Jayne Lawrence, Christian D Lorenz Published: 2026 ABSTRACT AbstractOligomeric surfactants exhibit physicochemical properties distinct from their monomeric counterparts, including lower critical micelle concentrations, yet how oligomerisation shapes the internal architecture of the micelles they form remains poorly understood. Here, we combine all‐atom molecular dynamics simulations with small‐angle neutron scattering (SANS) to study mixed micelles of Triton X‐100 and its oligomeric analogue Tyloxapol (trimer and heptamer) across eight compositions spanning pure Triton X‐100 to pure Tyloxapol. Tyloxapol content and oligomer length control the internal organization of the micelles: Tyloxapol‐rich micelles adopt compact, oblate structures containing internal, water‐filled cavities formed by headgroup penetration into the hydrophobic core, accompanied by reduced π–π stacking between aromatic tail groups. For pure Tyloxapol micelles, SANS corroborates the simulated size, aggregation number, oblate shape, and core hydration. Neighbor analysis shows no preferential association between Triton X‐100 and Tyloxapol, consistent with random mixing across all compositions studied, while unsupervised clustering (UMAP/HDBSCAN) of Tyloxapol conformations shows that the local micelle environment selects for distinct conformational states. Together, these findings identify surfactant oligomerisation as a route to controlling internal micelle architecture, rather than only overall micelle size and shape, with implications for the design of nanocarriers and functional soft materials with tunable internal structure.Keywords: chemistry, critical micelle concentration, micelle, molecular dynamics, neutron scattering, oligomer, pulmonary surfactant, stacking, trimer Introduction Surfactants are amphiphilic molecules that can self‐assemble in aqueous environments into a variety of structures—including spherical micelles, rod‐like aggregates, and vesicles—depending on their molecular architecture and concentration. Due to their ability to self‐assemble, lower surface tension, and wet surfaces, surfactants are widely used across industries, including cosmetics, agriculture, and pharmaceuticals, as well as in cleaning products and detergents [1, 2, 3, 4]. Surfactants are commonly classified as anionic (negatively charged), cationic (positively charged), zwitterionic (bearing both positive and negative charges), or nonionic (uncharged). Nonionic surfactants are generally favored for formulation purposes due to their lower toxicity and stability under various conditions. Among nonionic surfactants, Triton X‐100 and its oligomeric counterpart, Tyloxapol, have been the focus of extensive study. Oligomeric surfactants—typically synthesized by covalently linking monomeric surfactants via a spacer—have attracted increasing interest, as they often exhibit physicochemical properties distinct from their monomeric analogues [5, 6, 7]. These properties can be fine‐tuned by modifying the length and chemical nature of the hydrophobic tails, the structure of the linker (e.g., linear or branched), and the number of surfactant units joined together, enabling a broad range of structural variants. Several studies have reported that reducing the length of the spacer between surfactant units decreases both the critical micelle concentration (CMC) and surface tension, while also improving foaming behavior. This behavior has been observed for dimeric cationic surfactants joined via their head group (often referred to as gemini surfactants) derived from dodecyl‐ or benzyldodecyl‐trimethylammonium chloride [5]. Indeed, dimeric and trimeric surfactants generally exhibit lower CMCs and surface tensions than their monomeric counterparts [6, 7, 8, 9, 10]. For example, Bunton et al. reported a reduced CMC for a dicationic analog of the cationic surfactant cetyltrimethylammonium bromide (CTABr) [9]. Zhou et al. demonstrated enhanced antimicrobial activity in ammonium surfactants as the degree of oligomerization increased, with trimeric, tetrameric, and hexameric species showing significantly greater activity against E. coli than their monomeric or dimeric forms [10]. In structurally related surfactant series, lower CMC values are generally associated with larger micelles and increased solubilization capacity. For instance, Dharaiya et al. compared the solubilization of poorly water‐soluble bisphenol A in micelles formed by Tyloxapol and Triton X‐100, finding that Tyloxapol formed the largest micelle and exhibited the highest solubilization capacity [11]. Similarly, Dam et al. reported that dimeric quaternary ammonium bromides possessed both lower CMCs and greater oil solubilization capacity than the corresponding monomeric surfactant, CTABr [12], an experimental observation corroborated by coarse‐grain molecular dynamics (MD) simulations. Maiti et al. used coarse‐grained MD simulations to show that increasing the degree of polymerization in a surfactant lowers the CMC [13]. A similar trend is observed in dissipative particle dynamics simulations of SDS [14]. More recently, Wang et al. investigated the effect of spacer length on the morphology of micelles formed by dimeric dimethylcetylammonium bromide surfactants, observing a transition from micelles to vesicles as the spacer length decreased [15]. In some cases, micelles and other nanostructures are formed from mixtures of surfactants rather than from a single species [16]. Depending on the surfactants involved, mixed micellar systems may outperform their single‐component counterparts across a range of applications [16, 17, 18]. The CMCs of mixed micellar systems typically lie between those of the individual surfactants [17]. Various physicochemical properties, including micelle size, shape, interaction parameters, and CMC, have been systematically explored in mixed systems [19, 20, 21, 22]. Mixed micelles have also received attention for their ability to enhance the solubility of poorly water‐soluble drugs and to improve the therapeutic effect of the drugs they contain [23]. Experimental studies have demonstrated that two‐component micelles can solubilize more doxorubicin, an antitumor drug, and enhance its therapeutic efficacy [24, 25, 26, 27]. For example, Mehta et al. [28] showed that lecithin‐Tyloxapol micelles exhibit synergistic behavior and a high drug‐encapsulation capacity. The self‐assembly of binary surfactant mixtures has also been investigated through atomistic and coarse‐grained MD simulations [29, 30, 31], including systems based on CTAB, SDS, and glycocholate. However, to date, no computational studies have explored mixed micelles composed of a surfactant and its oligomer. This leaves several fundamental questions: (1) How does cross‐linking between surfactant molecules affect micelle formation? (2) How do micellar morphologies differ among assemblies formed by monomeric surfactants, oligomeric surfactants, and their mixtures? In this study, we investigate the structure of mixed micelles formed from Triton X‐100 and Tyloxapol (Figure 1b). Triton X‐100 consists of a hydrophilic polyethylene oxide (PEO) head group and a hydrophobic 4‐(1,1,3,3‐tetramethylbutyl)‐phenyl tail. Tyloxapol is an oligomeric analogue of Triton X‐100, composed of approximately seven monomer units covalently linked via a spacer. In this manuscript, we have studied variants of Tyloxapol that have 3 and 7 monomer units, and each have the same number, 10, of ethylene oxide groups in their hydrophilic headgroups as the Triton X‐100 molecules that have been modeled, the nearest integer to the experimentally reported average of n ∼ 9.5 (Figure 1a). We explore the structural properties of mixed micelles formed at varying ratios of Triton X‐100 and Tyloxapol (Figure 1a), and assess the effects of cross‐linking on micelle morphology and structural properties. Molecular architectures of Triton X‐100, Tyloxapol, and their mixed micelles as observed in the simulations. (a) Representative snapshots of 50:50 mixed micelles composed of Triton X‐100 and Tyloxapol oligomers of two different lengths: m=3 (TY39, top) and m=7 (TY79, bottom). (b) Chemical structures of Triton X‐100 (n≈9.5, top) and Tyloxapol (m≈7, bottom), highlighting the aromatic core and the PEG‐based headgroups characteristic of these nonionic surfactants. Methods Molecular Dynamics Simulations In this work, mixed Triton X‐100 and Tyloxapol micelles have been investigated. All‐atom MD simulations have been performed using the GROMACS package [32, 33, 34] and the CHARMM36 [35, 36, 37] force field. Triton X‐100 is modeled using CHARMM parameters reported by Yordanova et al. [38]. Two types of Tyloxapol molecules have been constructed, a trimer (TY39) and a Triton X‐100 (TX100) heptamer (TY79) (Figure 1b), and were parameterized by the CHARMM36 general force field [39]. The water is modeled using CHARMM36m TIP3 potential. In total, eight systems, containing varying amounts of Triton X‐100 and Tyloxapol, have been constructed and simulated. In Table 1, the number and type of species in each system is presented along with their abbreviated description. The amounts of both surfactants were chosen such that the total number of Triton X‐100 monomer units, counting each Tyloxapol oligomer as its constituent TX100 subunits (3 for TY39, 7 for TY79) is approximately 150. This value was chosen to match the experimentally reported aggregation number of Triton X‐100 micelles, which falls in the range of 100–155, and to enable direct comparison with our previous simulations of pure Triton X‐100 micelles [40]. The total monomer count was held fixed across all eight systems so that composition and oligomer length are the only variables governing differences in the resulting micelle structure; composition‐dependent equilibrium aggregation numbers were not determined and are identified as a direction for future work (see Section 4). First, the surfactants in the system were preassembled into a spherical “micelle‐like” structure using the Packmol software [41]. The “micelles” were then centered in a simulation box with dimensions of 120 Å × 120 Å × 120 Å, after which water was added to fill the box. TABLE 1Composition of the eight simulation systems. Listed are the percentages and total number of each surfactant species (Tyloxapol—TY and Triton X‐100—TX100) in the systems that have been simulated. Note that Tyloxapol consisting of three (TY39) and seven monomers (TY79) and each has been studied. The number of water molecules are included for completeness. The total number of Triton X‐100 monomers (N) found in the Tyloxapol and Triton X‐100 surfactants in each simulated system is also reported.SystemTY, %TX100, %TY, N TX100, N WaterMonomer, N TY39100050051 463150TY39‐TX100 757525383851 289152TY39‐TX100 505050257551 237150TY39‐TX100 2525751311251 084151TY79100021051 373147TY79‐TX100 757525163851 158150TY79‐TX100 505050117551 242152TY79‐TX100 252575511351 041148 Composition of the eight simulation systems. Listed are the percentages and total number of each surfactant species (Tyloxapol—TY and Triton X‐100—TX100) in the systems that have been simulated. Note that Tyloxapol consisting of three (TY39) and seven monomers (TY79) and each has been studied. The number of water molecules are included for completeness. The total number of Triton X‐100 monomers (N) found in the Tyloxapol and Triton X‐100 surfactants in each simulated system is also reported. Within each preassembled structure, Triton X‐100 and Tyloxapol molecules were placed by Packmol at random positions and orientations, subject only to a minimum‐distance overlap constraint, so that the mixed micelles are randomly, not deliberately, mixed at the outset. The radius of the starting sphere was chosen to approximate the packing density of a Triton X‐100 micelle of comparable aggregation number reported in our previous work, giving a starting structure that is close to, but not fixed at, the expected equilibrium density. The subsequent energy minimization and NVT/NPT equilibration and production stages allow both the local packing and the overall micelle/box density to relax freely under the Parrinello–Rahman barostat. The size, shape, and volume of each system as a function of simulation time are shown in Figures S1–S3; all systems show substantial relaxation away from their initial Packmol values within the first ∼100–300 ns, after which each descriptor plateaus, which we take as evidence that the systems are not trapped near their initial, unrelaxed configuration. The steepest descent algorithm is applied to minimize the energy in each system. The systems were then brought to the desired temperature of 303.15 K in a canonical, NVT, ensemble using a Nosé–Hoover thermostat, after which the systems were equilibrated in an NPT ensemble to achieve a target pressure of 1 bar, coupled to a Nosé–Hoover thermostat and Parrinello–Rahman barostat. The time constants for the thermostat and barostat were set to 1 and 5 ps, respectively, with isotropic pressure coupling applied to the barostat. The durations for the equilibration simulations were set to 75 and 325 ps for the NVT and NPT simulations, respectively. Then each system is simulated for 600 ns in an NPT ensemble using a timestep of 2 fs. The TY39‐TX100 75 system did not fully stabilize within the standard run time and was therefore extended to 800 ns. The LINCS algorithm [42] is applied to constrain hydrogen‐containing bonds. A cutoff of 1.2 nm is used to truncate short‐range interactions. Long‐range interactions were calculated using the Particle‐Mesh Ewald algorithm. Analysis of MD Simulations The analysis of the systems is conducted using tools provided by the GROMACS software package as well as in‐house Python scripts. The calculated values for each parameter were averaged over the portion of the trajectory where the system has reached its stability determined by the root‐mean‐square deviation (RMSD) of the coordinates of the atoms within the molecules constituting the micelle. Uncertainties reported throughout (Table 2, Figure 2, and elsewhere) represent the standard deviation. TABLE 2Global structural and morphological descriptors obtained from the molecular dynamics simulations of the Triton X‐100 and tyloxapol micelles. The table includes the average solvent‐accessible surface area (SASA), ellipticity, volume, and radius of gyration for each system, together with the corresponding nanoparticle shape classifications. These metrics provide a comparative view of overall micelle size, compactness, and anisotropy, enabling direct assessment of how tyloxapol composition and oligomer length influence micelle morphology and structural stability.SystemSASA (Å2)EllipticityVolume, nm3 R g, ÅShapeTY39 47887±1335 1.370±0.04 75.8±1.6 28.03±0.2 OblateTY39‐TX100 75 51581±1658 2.1±0.1 84.8±2.8 30.5±0.4 TriaxialTY39‐TX100 50 53353±1462 1.902±0.2 79.5±5.4 31.50±0.6 ProlateTY39‐TX100 25 55087±1400 2.1099±0.1 91.0±5.1 32.02±0.4 ProlateTY79 50409±1359 1.4438±0.07 90.3±1.7 29.3±0.2 OblateTY79‐TX100 75 54235±1260 1.682±0.08 74.9±2.7 30.64±0.3 ProlateTY79‐TX100 50 54884±1478 1.453±0.05 74.6±2.7 29.437±0.2 OblateTY79‐TX100 25 54590±1512 1.5496±0.1 88.3±2.5 31.877±0.4 ProlateTX100 60100±1279 1.48±0.08 80.2±4.1 29.215±0.3 Prolate Global structural and morphological descriptors obtained from the molecular dynamics simulations of the Triton X‐100 and tyloxapol micelles. The table includes the average solvent‐accessible surface area (SASA), ellipticity, volume, and radius of gyration for each system, together with the corresponding nanoparticle shape classifications. These metrics provide a comparative view of overall micelle size, compactness, and anisotropy, enabling direct assessment of how tyloxapol composition and oligomer length influence micelle morphology and structural stability. Key global structural descriptors of Tyloxapol‐containing micelles obtained from our molecular dynamics simulations as a function of Tyloxapol composition. Plotted are the average solvent‐accessible surface area (SASA), ellipticity, volume, and radius of gyration versus the percentage of Tyloxapol present in the micelle. Gray circles indicate the corresponding average values for 100% Triton X‐100 micelles from our previous work [40], providing a reference point for assessing how increasing Tyloxapol content influences micelle size, anisotropy, and overall morphology. Size and Shape of Micelles To determine the stable micelle structure, the RMSD and the radius of gyration have been calculated. Furthermore, the solvent‐accessible surface area (SASA) is calculated to assess the surface area of each aggregate. To determine the shape of the micelle, the inertia tensor is calculated and averaged over the last 200 frames of the trajectory, saved every 100 ps, corresponding to the final 20 ns. This window was chosen because it lies within the plateau region reached by all global size/shape descriptors (Figures S1–S4), typically after the first 100–300 ns of simulation. The micelle shape can be classified as spherical, oblate, prolate, or triaxial. The criterion for a micelle to be classified as a certain shape is described in terms of the principal moments of inertia (I1, I2, and I3) as follows: 1. I1≃I2≃I3, micelle is spherical. I1≃I2≃I3, micelle is spherical. 2. I1≃I2>I3, micelle is oblate. I1≃I2>I3, micelle is oblate. 3. I1≃I2