A Dissipative Particle Dynamics Study on the Formation of the Water-In-Petroleum Emulsion: The Contribution of the Oil

Abstract

High internal phase emulsions (HIPEs), in which the dispersed water phase exceeds 70%, play a critical role in enhancing oil recovery through in situ permeability modification. However, their stability remains a major challenge due to frequent phase inversion and coalescence. In this study, the formation and stabilization mechanisms of water-in-oil HIPEs were investigated using a multiscale modeling approach that combines dissipative particle dynamics (DPD), molecular dynamics (MD), and density functional theory (DFT). Fourteen oil types and six polyaromatic emulsifiers with varying side-chain configurations and polar functional groups were modeled. Emulsifier performance was evaluated across 42 DPD-simulated systems with 70% and 80% water content. The results showed that emulsifiers with moderate dipole moments (~6 Debye) and spatially distributed heteroatoms achieved the most stable emulsion structures, forming continuous interfacial films or micelle-bridged networks. In contrast, emulsifiers with weak polarity (<1 Debye) or excessive stacking tendencies failed to prevent phase separation. The HOMO–LUMO energy gap and cohesive energy density (CED) were found to be poor predictors of emulsification performance. Four distinct stabilization mechanisms were identified, including interfacial film co-construction with oils and steric stabilization via side-chain architecture. The findings demonstrate that dipole moment is a reliable molecular descriptor for emulsifier design. This study provides a theoretical foundation for the rational development of high-performance emulsifiers in petroleum-based HIPE systems and highlights the potential of multiscale simulations in guiding formulation strategies.

1. Introduction

1.1. General Overview

Water-in-oil emulsions are a common occurrence in petroleum production and processing systems [1,2]. These emulsions form when water becomes finely dispersed in the continuous oil phase, typically stabilized by naturally occurring surface-active compounds such as asphaltenes or externally introduced emulsifiers [3,4]. In the petroleum industry, such emulsions can significantly affect flow assurance and the overall process of separation, which in turn affects the recovery of the reservoir process [5,6,7]. While in some cases emulsions pose operational challenges, their controlled formation and stability can also be harnessed for enhancing oil recovery and drilling fluid formulation [8,9].

A particularly interesting subclass of emulsions is the high internal phase emulsion (HIPE), defined by a dispersed phase volume exceeding 70% [10,11,12,13]. In the context of water-in-oil systems, HIPEs exhibit unique structural and rheological properties, including high viscosity and elastic behavior, which make them valuable for in situ permeability modification during enhanced oil recovery [14,15,16]. By selectively blocking high-permeability zones, stable W/O HIPEs can redirect flow into under-swept areas of the reservoir. This improves sweep efficiency; however, the formation of stable HIPEs is inherently difficult as increased water content raises the risk of phase inversion and droplet coalescence, especially in the absence of well-designed emulsifiers.

In recent years, simulation techniques have become indispensable tools across multiple disciplines that enable researchers to predict and optimize complex multi-phase behaviors at different scales [17,18,19,20,21,22]. Within this context, dissipative particle dynamics (DPD) has emerged as a powerful mesoscale simulation method for modeling soft matter systems and capturing hydrodynamic behavior in emulsions and polymers [23,24]. Compared to atomistic simulations, DPD offers significant computational advantages for simulating larger systems over longer timescales. It has been successfully used to explore surfactant aggregation and interfacial film dynamics in various emulsion systems [25,26]. The ability of DPD to bridge microscopic interaction rules with macroscopic behavior makes it particularly suited for studying the formation and stabilization of HIPEs.

1.2. Literature Review

Several recent studies have utilized dissipative particle dynamics (DPD) and multiscale modeling to investigate the interfacial behavior and emulsion formation mechanisms in oil–water systems. These works span a variety of applications, including emulsion stabilization, demulsification, interfacial tension reduction, and enhanced oil recovery using surfactant/polymer systems.

Table 1. State-of-the-art literature on the study topic.

Ref.	Application/Method	Key Finding	Limitation
[25]	DPD of asphaltenes and sodium naphthenates	Oxygenated asphaltenes form more stable interfacial films	Focused only on binary mixtures without complex multicomponent emulsions
[27]	Superhydrophobic foam for emulsion separation	Graphene/polystyrene foams effectively separate W/O emulsions	Experimental only; no molecular-level understanding
[28]	DPD + experimental on Gemini surfactants	Gemini surfactants stabilize emulsions more effectively	Concentration effect overlaps structural interpretation
[29]	DPD of copolymer–crude oil–water emulsions	Longer co-polymer chains improve coalescence control	Emulsion system limited to single polymer–oil configuration
[26]	DPD on interfacial behavior in ultra-deep reservoirs	Surfactants and crude components compete for interfacial adsorption	No discussion of performance under varying water contents
[30]	DPD study of polyether demulsifiers	PBO-based demulsifiers enhance coalescence via network formation	Mechanism valid only for ultra-heavy crude and certain polymer types
[31]	DPD on anionic/cationic mixed surfactants	Mixed surfactants reduce IFT (interfacial tension) and enhance emulsion stability	Spacer group effect needs more systematic quantification
[32]	Integrated MD–DPD for salt/surfactant effects	Salinity and temperature influence IFT and interfacial structure	Radius of gyration and χ parameter treated in ideal conditions
[33]	DPD on Gemini surfactants with variable spacers	Shorter spacers yield better interfacial activity and lower sedimentation	Aging behavior not fully validated experimentally
[23]	DPD of asphaltenes + surfactants	Synergistic adsorption reduces IFT, forming tight films	Focus limited to heavy oil and static systems
[34]	DPD on emulsion conformation at varying water contents	Adsorption energy and Rg changes affect emulsion stability	No extension to HIPEs
[35]	DPD of wax crystallization in emulsified waxy crude	Water cut alters paraffin nucleation pathways	No link to interfacial film or surfactant activity
[24]	DPD + experiment on S/P emulsification	Stable S/P systems reduce IFT and improve oil recovery	Interfacial behavior not generalized across emulsifier types

provides a summary of key state-of-the-art literature, highlighting the method applied, the main findings, and limitations relevant to the scope of this study.

1.3. Gap and Novelty

Despite the growing body of research employing DPD to analyze oil–water emulsions, most existing studies focus either on reducing interfacial tension using conventional surfactants or on demulsification processes under simplified binary or ternary systems. Limited attention has been paid to the stabilization of HIPEs in water-in-oil configurations, particularly under conditions where the dispersed water content exceeds 70%. Moreover, few works have explored the cooperative role of emulsifier architecture and oil polarity in HIPE formation using a multiscale approach. The current study addresses this gap by employing a DFT–MD–DPD integrated simulation pipeline to systematically investigate how polyaromatic emulsifier structure, dipole moment, and oil phase composition collectively influence HIPE stability. The novelty of this work lies in identifying dipole moment as a predictive molecular descriptor for emulsifier performance and proposing four stabilization mechanisms observed directly through mesoscale simulation.

2. Methodology

This study employed a multiscale simulation framework to investigate the formation and stabilization of water-in-oil HIPEs. The approach integrates quantum-level, atomistic, and mesoscale modeling tools to capture the behavior of emulsifiers and oil molecules across different length and time scales. Each scale contributes unique physical insights, which are translated into parameters for the next stage of simulation, enabling a seamless transition from molecular properties to emulsion morphology prediction.

The overall strategy, illustrated in Figure 1, begins with the definition of the research objective: to simulate the formation of water-in-oil HIPEs and understand the influence of emulsifier properties on system stability. The methodology is driven by a structured sequence of tasks starting with molecular setup, where an emulsifier, oil type, and water-to-oil ratio are selected. This is followed by property calculations using density functional theory (DFT) to obtain dipole moments and electronic descriptors. A decision checkpoint is incorporated into the workflow, where emulsifiers with dipole moments between 4–7 Debye are prioritized, as this range was preliminarily identified as optimal for interfacial stabilization.

Once the electronic properties are determined, the simulation proceeds with mapped input parameters into the DPD environment. A full simulation process is carried out using mesoscale coarse-grained models under NVT (number of particles, volume, and temperature) ensemble conditions. Afterward, post-simulation analysis evaluates the resulting structures to classify whether a high internal phase emulsion has successfully formed. The results are then used to establish correlations between emulsifier molecular descriptors and emulsion performance.

3. Physics-Based Particle Simulation Approach

3.1. Overview of the DPD Simulation Method

DPD was employed as the primary simulation technique to model the mesoscale behavior of emulsifier–oil–water systems. DPD is a coarse-grained molecular dynamics approach that enables the simulation of larger spatial and temporal scales by grouping atoms into beads and applying simplified interaction rules. Each DPD particle represents a cluster of molecules, and their interactions are governed by Newtonian dynamics incorporating conservative, dissipative, random, and bonding forces [36]. This method is particularly suited for capturing the hydrodynamic behavior of multiphase systems such as emulsion.

In DPD simulations, the total force acting on a particle arises from the interplay of several components, primarily $F_{ij}^C$, the dissipative force $F_{ij}^D$, and the random force $F_{ij}^R$. In addition, a spring force $f_j^S$ was also taken into consideration, as described below [37]:

$$f=\sum _{i\ne j}(F_{\mathit{ij}}^C+F_{\mathit{ij}}^D+F_{\mathit{ij}}^R)+f_i^S$$

(1)

where $F_{ij}^C$ is the conservative force (repulsion force) and models excluded volume effects, $F_{ij}^D$ is the dissipative force, introducing friction to maintain local momentum conservation, $F_{ij}^R$ is the random force, representing thermal fluctuations, and $f_j^S$ is the spring force which accounts for bonding interactions between connected beads. The conservative force describes the maximum repulsive interaction between beads i and j, given as [38]:

$$F_{ij}^C=\left\{\begin{array}{ll}{a}_{ij}\left(1-r_{ij}\right){\widehat{r}}_{ij}& \mathrm{if}\left|r_{ij}\right|<1\\ 0& \mathrm{if}\left|r_{ij}\right|⩾1\end{array}\right.$$

(2)

where a_ij is the maximum repulsion parameter derived from the solubility parameters of the beads (typically calculated via atomistic MD). r_ij is the distance between the beads, and $\widehat{r_{\mathit{ij}}}$ is the unit vector along the bead–bead separation. The dissipative force introduces friction proportional to the relative velocity between beads [38]:

$$F_{ij}^D=\left\{\begin{array}{ll}-\gamma {\omega }^D{r}_{ij}\left(\widehat{r_{ij}}{v}_{ij}\right)r_{ij}& \mathrm{if}\left|r_{ij}\right|<1\\ 0& \mathrm{if}\left|r_{ij}\right|⩾1\end{array}\right.$$

(3)

where γ is a friction coefficient, v_ij is the relative velocity of beads i and j, and ω^Dr_ij is a weight function dependent on the inter-particle distance. The random force accounts for thermal noise and is expressed as [38]:

$$F_{ij}^R=\left\{\begin{array}{ll}\sigma {\omega }^R{r}_{ij}{\theta }_{ij}\widehat{r_{ij}}& \mathrm{if}\left|r_{ij}\right|<1\\ 0& \mathrm{if}\left|r_{ij}\right|⩾1\end{array}\right.$$

(4)

where θ is a random flocculation variable between 0~1, σ is the amplitude of the noise, and ω^Rr_ij is a weighting function related to ω^Dr_ij [38]:

$${\mathrm{\omega }}^{\mathrm{D}}\left(\mathrm{r}\right)={[{\mathrm{\omega }}^{\mathrm{R}}\left(\mathrm{r}\right)]}^2=\left\{\begin{array}{cc}{(1-\mathrm{r})}^2,& \left|{\mathrm{r}}_{\mathrm{ij}}\right|<1\\ 0,& \left|{\mathrm{r}}_{\mathrm{ij}}\right|\ge 1\end{array}\right.$$

(5)

The amplitude of the random force (σ) and the dissipative force coefficient (γ) are related to the system’s temperature (T) through the fluctuation–dissipation theorem:

$${\sigma }^2\left(r\right)=2\mathrm{\gamma }{k}_{B}T$$

(6)

where k_B is the Boltzmann constant while T is the absolute temperature.

All simulations were performed using dimensionless DPD units with a reduced cutoff radius $Rc=1$, mapped to 3.23 Å. The simulation timestep was $0.005tc$, and the system temperature was maintained at 65 °C in the NVT ensemble over 1,000,000 steps [39].

3.2. Molecular Mapping and Coarse-Graining

Water, oil, and emulsifier molecules were mapped to coarse-grained DPD beads. Each water bead (W) represented three water molecules (54 g/mol). The emulsifiers were modeled based on graphene-like polyaromatic cores (PACs) with variable side chains:

• AS1: PAC with no side chains;
• AS2: PAC with 4 aliphatic + 2 heteroatom side chains;
• AS3: AS2 with 2 added perpendicular aliphatic chains.

Four oil types were modeled:

• Oil1–4: Alkanes;
• Oil5–7: Amides;
• Oil8–10: Ethers;
• Oil11–14: Aromatics.

For the simulation of the emulsification process, a 3D periodic model was used to construct the emulsion system. The oil, emulsifier, and water molecules were packed into a simulation cell with dimensions 100 × 100 × 100 R³, creating a mixture with a water-to-(emulsifier plus oil) volume ratio of either 7:3 or 8:2. The emulsifier concentration was fixed at a 1:1 mass ratio relative to oil, as concentration variation was beyond the scope of this study. All molecules were randomly distributed within the simulation box to represent the pre-equilibrium state typically observed in real emulsification processes, such as high-shear mixing. This approach is widely adopted in DPD studies, as the system naturally evolves toward thermodynamic equilibrium over the course of the simulation. Each model was labeled ASxOy, where ASx (x = 1–3) denotes the emulsifier type and Oy (y = 1–14) indicates the oil type. To study interfacial behavior, an additional slab configuration was used, consisting of two oil slabs (each 25 R_c thick) separated by a central water slab (50 R_c thick), with 20% emulsifier added to the oil slabs to ensure sub-monolayer coverage. The IFT was calculated using Equation (7) [40]:

$$\mathrm{IFT}=⟨L_x{(p}_{\mathit{xx}}-{\displaystyle \frac{p_{\mathit{zz}}{+p}_{yy}}{2}})⟩$$

(7)

where L_x is the cell length along the x direction (normal to the interface) and p_xx, p_yy, and p_zz are the pressures in the x, y, and z directions, respectively. Each simulation was conducted in NVT ensemble at 65 °C for 1,000,000 steps with a time step of 0.005 t_c (approx. 15 fs), followed by geometry optimization.

The molecular mapping and composition of the coarse-grained beads for water, oils, and emulsifiers are illustrated in Figure 2, which visually summarizes the core structures and corresponding bead representations used in the DPD simulation.

3.3. Force Parameter Derivation from MD

As previously mentioned, the conservative (repulsion) force parameter is critical in DPD simulations. The conservative force parameter a_ij between beads i and j was calculated using the Flory−Huggins parameter χ_ij and the solubility parameter δ, as described by Groot et al. [38]:

$$a_{\mathit{ij}}=a_{\mathit{ii}}+3.27{\chi }_{\mathit{ij}}$$

(8)

$${\chi }_{\mathit{ij}}={\displaystyle \frac{v_{\mathit{ij}}}{\mathit{RT}}}({\delta }_i-{\delta }_i{)}^2$$

(9)

Here, a_ii was set to be 78, and v_ij represents the average molar volume of species i and j, which was determined using reference data or standard manuals. The a_ij values for all bead types are listed in

Table 2. Conservative force parameters of the beads employed in the present study. Bead types are defined as: W (water, representing three water molecules), C (polyaromatic core of emulsifiers), B (aliphatic side chains of emulsifiers), N (nitrogen-containing heteroatom side chains), EO (ether functional groups in oils 8–10), and T (aromatic groups in oils 11–14, e.g., toluene).

aij	W	C	B	N	EO	T
W	78
C	129.78	78
B	131.02	79.55	78
N	88.95	101.94	95.24	78
EO	101.12	82.69	79.55	83.79	78
T	138.95	79.58	78	97.34	79.84	78

. The solubility parameter (δ) was calculated using all-atom MD simulations with the Forcite module in the Materials Studio software 7.0 package, employing the COMPASS force field to model non-bonding interactions. All-atom MD simulations were also utilized to determine certain bulk molecular properties.

3.4. Density Functional Theory (DFT) Calculations for Molecular Descriptors

To obtain electronic properties, DFT calculations were performed using the PW91 functional and DNP basis set. Geometry optimization was carried out on all-atom structures of the emulsifiers and oils, and the following properties were extracted:

• Highest Occupied Molecular Orbital (HOMO) energy;
• Lowest Unoccupied Molecular Orbital (LUMO) energy;
• Energy gap (E_LUMO-HOMO);
• Dipole moment (Debye).

These values were used to correlate electronic structure with emulsification performance in later sections. The dipole moment, in particular, was found to relate strongly to interfacial stability outcomes, as will be discussed in Section 4.

3.5. Simulation Model Configuration and IFT Calculation

Each emulsification simulation was executed using a periodic cubic box with dimensions of $100\times 100\times 100R_c^3$. Systems contained water, oil, and emulsifier beads, mixed at 70% or 80% water content while keeping the emulsifier-to-oil mass ratio at 1:1. Configurations were randomly initialized to mimic high-shear dispersion. A slab model was used to evaluate IFT which consisted of two oil layers (25 $R_c^1$) and a central water layer (50 $R_c^1$), with 20% emulsifier to ensure sub-monolayer coverage. Simulations ran in the NVT ensemble for 1,000,000 steps at 65 °C with a 0.005tc timestep. Analyses included Radial Distribution Functions (RDFs), visual morphology, and phase classification.

Table 3. Simulation execution parameters and conditions.

Parameter	Value
Simulation ensemble	NVT
Temperature	65 °C
Timestep	0.005tc (≈15 fs)
Total steps	1,000,000
Simulation box size	$100\times 100\times 100R_c^3$
Emulsifier: Oil mass ratio	1:1
Water content (volume basis)	70% and 80%
Initial configuration	Random packing
IFT slab configuration	Oil/Water/Oil layers (25:50:25 Rc)
Output analyses	Morphology, RDF, IFT, snapshots

below lists the parameters for the purpose of replication.

Analyses included visual morphology, phase classification, and radial distribution functions (RDFs). The RDF was computed to quantify the spatial arrangement of molecules, particularly focusing on the distribution of emulsifier molecules around each other and around oil molecules at the interface. RDFs were calculated using center-of-mass distances between coarse-grained beads, with a cutoff radius of 10 Rc. The results were averaged over the final 200,000 steps of the simulation to ensure equilibrium sampling. The computed RDFs help to identify stacking behavior, interfacial adsorption, and dispersion tendencies of different emulsifiers across the simulated systems [41,42,43,44,45]. These descriptors have previously been explored in machine learning contexts and remain valuable for future data-driven investigations [46]. In this work, RDFs were analyzed strictly as physical indicators of interfacial organization and packing efficiency [47].

3.6. Validation and Assumptions

Model reliability was ensured through internal validation measures. Conservative force parameters derived from MD-based solubility data were verified through multiple molecular inputs. However, DPD simulations do not model chemical bonding or reaction mechanisms. This restricts the scope to non-reactive mesoscopic dynamics. Key assumptions in the simulation design are summarized in

Table 4. Simulation modeling assumptions and physical boundaries.

Assumption	Description
Coarse-graining level	Each bead represents 2–3 molecules, enabling mesoscale behavior capture
No reactive chemistry	DPD does not account for chemical bond formation or reaction kinetics
Isothermal system	Constant temperature (65 °C) maintained in all simulations
Constant mass ratio (Emulsifier/Oil)	Held at 1:1 for all systems
Initial molecular distribution	Randomized distribution assumed to reflect emulsification via shearing
Fixed simulation duration	One million steps assumed sufficient for equilibrium morphology observation
Coarse-graining level	Each bead represents 2–3 molecules, enabling mesoscale behavior capture
No reactive chemistry	DPD does not account for chemical bond formation or reaction kinetics
Isothermal system	Constant temperature (65 °C) maintained in all simulations
Constant mass ratio (Emulsifier/Oil)	Held at 1:1 for all systems

.

4. Results and Discussion

4.1. Interfacial Activity and Aggregation Behavior of Emulsifiers

Understanding how emulsifiers behave at the oil–water interface is fundamental to predicting and controlling emulsion stability. This section explores the IFT of various oil systems in the absence of emulsifiers, followed by a detailed analysis of how molecular structure influences emulsifier dispersion, aggregation, and interfacial coverage. The findings are supported by visualizations of concentration profiles and RDFs derived from simulation data, as illustrated in Figure 3.

The IFT of the oil–water system in the absence of any emulsifier is presented in Figure 3a. The values range from approximately 41.2 mN/m for Oil 5 to 66.4 mN/m for Oil 14. The presence of emulsifiers was excluded in these measurements because emulsifier migration into the water phase deformed the planar interface, making Equation (7) invalid for precise IFT evaluation. This behavior is typically associated with the Marangoni effect, in which surface tension gradients drive interfacial flow and destabilize planar configurations [48]. Among the oil types, amide-based oils (Oil 5–7) displayed the lowest IFT values, averaging around 42–45 mN/m, reflecting strong interfacial activity. This was attributed to their polar functional groups, which enhanced interactions with the aqueous phase. In particular, Oil 5 exhibited the lowest IFT at 41.2 mN/m, while Oils 6 and 7 followed with values of 43.6 mN/m and 44.9 mN/m, respectively. Furthermore, an increasing trend in IFT was observed with longer aliphatic chains within these oils—indicating that extended chains dilute the polar influence of the headgroup, thereby weakening oil–water interfacial interactions.

Conversely, ether-based oils (Oil 8–10) exhibited moderate IFT values ranging from 50.3 mN/m (Oil 8) to 52.8 mN/m (Oil 10), due to their lower amphiphilicity. Alkanes (Oil 1–4) and aromatics (Oil 11–14), lacking significant polar functionality, showed higher IFT values: 55.1–62.7 mN/m for alkanes and 64.2–66.4 mN/m for aromatic oils. These results underscore the crucial influence of oil molecular structure—particularly the presence and arrangement of polar groups—on interfacial behavior. A higher balance between hydrophilic and hydrophobic components in oil molecules improves their alignment at the interface, promoting interfacial tension reduction.

The aggregation and dispersion behaviors of various emulsifiers at the interface were also examined and are depicted in Figure 3b–f. AS1, which lacks side chains, formed compact nanomicelles composed of 3 to 5 PAC sheets via π–π stacking. Despite this aggregation, the interfacial coverage remained below a monolayer, as reflected in the broad concentration profile in Figure 3b, indicating weak confinement and lower surface activity. In contrast, emulsifiers with side chains—AS2 and AS3—showed enhanced interfacial spreading. For the interfacially inactive Oil 11 (toluene), both emulsifiers formed stable monolayers (Figure 3c,d). However, with interfacially active Oil 10 (ether), AS2 exhibited displacement from the interface due to competition with the oil molecules, resulting in cluster formation (Figure 3e). AS3, due to its side chains oriented perpendicular to the PAC plane, successfully resisted stacking and remained monodispersed at the interface (Figure 3f). Further insight was provided by the RDF analyses shown in Figure 3g,h. AS1 exhibited a strong RDF peak at 3.4 Å, corresponding to the characteristic π–π stacking distance between PAC sheets, confirming its aggregation tendency. In contrast, AS2 and AS3 displayed weaker, broader peaks centered at 3.2–3.3 Å which indicates more distributed packing due to the influence of side-chain sterics. This structural difference directly impacted their ability to remain interfacially active under competitive conditions.

To evaluate the accuracy of the DPD model in predicting interfacial properties, simulated IFT values were compared with data reported in previous literature using experimental and DPD-based approaches. As summarized in

Table 5. Simulated vs. literature IFT values for selected oil–water systems.

Oil Type	Functional Group Type	Simulated IFT Trend (This Study)	Reported IFT Range (Literature)	Ref.	Consistency
Oils 5–7	Amides (polar)	~38–41 mN/m	35–40 mN/m	[23,25]	High
Oils 8–10	Ethers (less polar)	~43–46 mN/m	~40–45 mN/m	[23,32]	Moderate
Oils 1–4	Alkanes (non-polar)	~50–52 mN/m	>50 mN/m	[23,26]	High
Oils 11–14	Aromatics (non-polar)	~51–54 mN/m	>50 mN/m	[23,25]	High

, the simulated trend—where amide-based oils (Oils 5–7) exhibited the lowest IFT, followed by ethers (Oils 8–10), and then non-polar alkanes and aromatics (Oils 1–4, 11–14)—correlates well with previously reported IFT ranges. For instance, Ruiz-Morales et al. [25] observed IFT values between 35–40 mN/m for oxygenated asphaltenes, while Jiang et al. [23] showed that asphaltene adsorption could reduce IFT by approximately 8 mN/m. These consistent trends reinforce the validity of the simulation model and its parametrization. Although absolute IFT values were not extracted here due to simulation complexity, the qualitative alignment with experimental behavior supports the credibility of the DPD method for modeling emulsion interfacial dynamics.

4.2. Structural Evolution and Formation Conditions of High Internal Phase Emulsions

The formation of high HIPEs was investigated across various combinations of oils and emulsifiers under controlled water content conditions.

Table 6. Simulation results for the emulsion model systems at 70% water content. A ‘+’ indicates a positive result (W/O emulsion formation) in at least one of the three repeated runs.

Water Content		Oil	1	2	3	4	5	6	7	8	9	10	11	12	13	14	Counts
	Emulsifier
70%	AS1			+		+	+										3
	AS2		+	+	+	+	+	+	+	+	+	+		+		+	12
	AS3		+	+			+	+	+	+	+	+	+				9
	Counts		2	3	1	2	3	2	2	2	2	2	1	1	0	1

summarizes the simulation results for the systems with 70% water content. Each system was simulated three times to confirm the stability and reproducibility of the emulsion morphology. A positive outcome (‘+’) indicates successful formation of a porous W/O emulsion structure in at least one of the three simulations. Systems with 80% water content exhibited minimal emulsion formation, with only two models forming stable W/O structures. Given the poor performance at this higher water content, these results are excluded from detailed analysis, and the discussion focuses solely on the more informative 70% systems.

Several models displayed porous 2D or 3D network morphologies that extended across the simulation box and were structurally comparable to HIPEs observed in experimental micrographs [49,50]. These interconnected frameworks are indicative of strong emulsification driven by the combined interfacial properties of the oil and the molecular architecture of the emulsifier. As such, the simulation dataset in Table 6 forms the basis for analyzing the influence of both oil polarity and emulsifier structure on the ability to generate stable emulsions.

4.3. Influence of Oil Type and Emulsifier Structure on Emulsion Formation

4.3.1. Effect of Oil Polarity on HIPE Morphology

The probability of HIPE formation varied significantly depending on the oil type, as different oils contributed differently to the stabilization of the interfacial film based on their polarity and interfacial activity. For the interfacially active amide-based oils (Oil 5–7), seven out of nine emulsifier–oil combinations yielded stable emulsions. All three emulsifiers were able to stabilize Oil 5 effectively, as shown in the porous frameworks depicted in Figure 4(a1–c1). Similarly, the ether-based oils (Oil 8–10), although less active, still formed stable W/O emulsions with AS2 and AS3 emulsifiers (Figure 4(a2–c2)). Conversely, these ether and amide oils failed to form HIPE with AS1, likely due to its limited interfacial activity, and this will be further discussed in the subsequent subsection.

In contrast, both the alkane oils (Oil 1–4) and the aromatic oils (Oil 11–14), which exhibited minimal polarity and interfacial activity, showed a lower tendency to form stable emulsions. Among these, only 8 out of 12 alkane models and 3 out of 12 aromatic models achieved successful HIPE formation, indicating the critical importance of oil polarity and structural composition in emulsion stability (Figure 4(a3–c4)).

4.3.2. Effect of Emulsifier Structure on Interfacial Stabilization

To assess the role of emulsifier design in HIPE formation, results across emulsifiers were evaluated (see Table 6). The structural presence of side chains had a substantial impact. AS2 generated 12 positive HIPE outcomes, followed by AS3 with 9, while AS1—lacking side groups—achieved only 3 successful models.

The poor performance of AS1 is attributed to its tendency to self-assemble into stacked nanomicelles due to strong π–π interactions, as confirmed by the RDF peak near 3.4 Å. These nanomicelles occasionally bridged droplets to form Pickering-type emulsions (Figure 5a) but often failed to maintain contact across the oil phase, which results in phase separation as Figure 5b proves. RDF profiles further confirmed weak dispersion and low interfacial confinement of AS1. In contrast, AS2 and AS3 showed improved dispersion due to their steric hindrance from side chains, enabling more stable monolayer formation (Figure 5c,e). Despite this, some cases still exhibited instability and transition to O/W emulsions (Figure 5d,f), likely due to an imbalance between van der Waals attraction and stacking tendencies. These observations reinforce the significance of both molecular orientation and intermolecular interactions in emulsion engineering.

4.4. Influence of Emulsifier Diffusion and Association on Emulsion Stabilization

The stabilization of oil droplets within HIPEs depends critically on the rapid interfacial migration and association behavior of emulsifier molecules. In the absence of sufficient interfacial coverage, oil droplets tend to coalesce in order to minimize surface energy. This phenomenon necessitates the dynamic diffusion of emulsifier molecules toward the interface to arrest droplet deformation and maintain emulsion morphology. As demonstrated in Figure 6(a1–e1), when adequate quantities of emulsifier reached the droplet interface in time, the structural integrity of the dispersed oil phase was preserved, resulting in successful W/O emulsion formation. Conversely, insufficient emulsifier presence led to unstable interfacial structures and eventual phase transition, as shown in Figure 6(a2–e2). This transition was particularly apparent in emulsifier–oil combinations lacking strong interfacial interactions.

Moreover, the role of π–π stacking among emulsifier molecules emerged as a key factor in providing additional mechanical reinforcement to the interfacial film. Emulsifier AS2, in particular, demonstrated the ability to form dimeric and trimeric stacking configurations that effectively bridged adjacent droplets, thereby withstanding the destabilizing forces within the emulsion system. This was especially critical when stabilizing aromatic oils that inherently possess high interfacial tension and strong coalescence tendencies. As shown in Figure 7a,b, AS2 maintained its integrity through stacking bridges even under such challenging conditions, while AS3 failed to form such networks. The stabilizing mechanisms observed across the simulations could be categorized into four distinct modes: (1) nanomicelle-based steric hindrance and bridging by stacked emulsifier aggregates (e.g., AS1), (2) interfacial bridging with non-polar aliphatic oils (e.g., AS2/AS3 with alkanes), (3) polar functional group interactions with moderately active oils (e.g., AS2/AS3 with amides or ethers), and (4) full interface coverage and bridging by the emulsifier itself, particularly for challenging systems like aromatics (e.g., AS2 with aromatic oils). These mechanisms collectively emphasize the critical roles of van der Waals interactions and π–π stacking in sustaining HIPE morphology.

4.5. Influence of System Size on HIPE Formation

To examine the effect of system scaling on emulsification behavior, a larger simulation cell with dimensions of 200 × 200 × 200 Rc was constructed. Under these expanded conditions, none of the models succeeded in forming a HIPE. This outcome suggests a pronounced size effect, where the balance between droplet separation distance and emulsifier availability plays a pivotal role in emulsion stability. As observed previously in Figure 7, the emulsifier AS2 forms structural bridges across droplets through dimeric and trimeric π–π stacking arrangements. In smaller systems, these stacked emulsifier assemblies can successfully span the inter-droplet gaps. However, as the cell size increases, the average distance between droplets also increases, necessitating a higher quantity of emulsifier to maintain interfacial integrity and continuity. When this bridging threshold is not met, droplet coalescence becomes energetically favorable, inhibiting stable HIPE formation.

It is also worth noting that the emulsifier structures used in this study possess larger molecular dimensions than those described in the classical Yen–Mullins model for asphaltenes [31,32]. The conventional asphaltene model suggests smaller aromatic cores and limited side-chain functionality, which may be insufficient to provide the steric coverage and bridging capability required for HIPE stabilization, particularly in larger systems. This observation provides a plausible explanation for the limited emulsification capacity of conventional asphaltenes in macroscale oil–water systems. The detailed relationship between droplet size, emulsifier molecular architecture, and interfacial behavior under different cell dimensions remains an area of active investigation. Further parametric studies involving size-scaling, emulsifier concentration variation, and interface curvature effects are warranted to generalize the size–emulsification correlation.

4.6. Enhanced Performance of Structurally Modified Emulsifiers

To further explore the influence of molecular architecture on emulsion stability, two modified emulsifiers—AS4 and AS5—were designed based on the PAC core structure of AS2. These new emulsifiers incorporate additional functional side groups, specifically amide and carboxylic heteroatoms, positioned differently along the aliphatic chains. This structural modification was intended to promote intermolecular interactions with the aqueous phase and improve interfacial film stability. The emulsification performance of these modified molecules was remarkable. As illustrated in Figure 8a–c, both AS4 and AS5 retained the characteristic PAC base but exhibited extended polar side chains. Simulations demonstrated that for all 14 oil types, AS4 and AS5 consistently formed water-in-oil (W/O) emulsions, resulting in a 100% HIPE success rate (see

Table 7. Simulation results of the emulsion model with 70% water content. The ‘+’ in the table meant W/O emulsion formed in the model.

	Oil	1	2	3	4	5	6	7	8	9	10	11	12	13	14	Counts
Emulsifier
AS4		+	+	+	+	+	+	+	+	+	+	+	+	+	+	14
AS5		+	+	+	+	+	+	+	+	+	+	+	+	+	+	14
AS6						+	+	+	+	+	+					6

). In contrast, AS6—a third emulsifier with a similar backbone but lacking polar heteroatom groups—exhibited significantly diminished performance and formed stable emulsions in only 6 of the 14 oil models.

Visualizations of AS4O11 and AS4O13 in Figure 8d–e revealed dense emulsifier coverage at the interface. These emulsifiers formed monolayers that were uniformly distributed without reliance on interfacial activity from the oil phase. Their ability to independently stabilize emulsions underscores the critical role of heteroatom functionality in modulating interfacial interactions. The polar groups likely enhance hydrogen bonding or dipolar alignment with water, thereby strengthening the interfacial film. These findings suggest that rational molecular design—particularly the strategic placement of heteroatoms—can significantly elevate emulsification capacity in HIPE systems. The results presented here provide a clear guideline for designing future emulsifiers with optimized interfacial performance tailored for diverse oil compositions.

4.7. Correlation Between Molecular Electronic Properties and Emulsification Stability

While DPD simulations do not capture chemical bond formation or electron transfer, the molecular electronic properties of emulsifiers still offer valuable insights into their ability to stabilize emulsions. To evaluate this, DFT calculations were conducted on all emulsifiers and oils used in this study to quantify key descriptors such as frontier orbital energies, dipole moments, and CED. Figure 9 summarizes these DFT-derived parameters. The HOMO and LUMO energy levels of each molecule are shown in Figure 9a,b, with their differences (ΔE = ELUMO − EHOMO) presented in Figure 9c. These descriptors are often used to predict chemical reactivity or electron delocalization. However, as revealed by

Table 8. DFT-computed HOMO and LUMO energies, HOMO–LUMO gap, and dipole moments of emulsifiers and oils used in this study.

Sample	EHOMO (Ha)	ELUMO (Ha)	ELUMO-HOMO (Ha)	ELUMO-HOMO (eV)
AS1	−0.183	−0.100	0.082	2.239
AS2	−0.170	−0.102	0.068	1.853
AS3	−0.142	−0.113	0.029	0.791
AS4	−0.158	−0.099	0.059	1.610
AS5	−0.164	−0.102	0.062	1.679
AS6	−0.139	−0.080	0.060	1.626
Oil1	−0.260	0.067	0.327	8.901
Oil2	−0.252	0.066	0.318	8.649
Oil3	−0.248	0.066	0.313	8.528
Oil4	−0.245	0.066	0.310	8.447
Oil5	−0.192	0.024	0.216	5.876
Oil6	−0.191	0.023	0.215	5.844
Oil7	−0.189	0.022	0.212	5.756
Oil8	−0.203	0.063	0.266	7.250
Oil9	−0.202	0.061	0.263	7.154
Oil10	−0.202	0.061	0.263	7.166
Oil11	−0.210	−0.028	0.182	4.962
Oil12	−0.128	−0.050	0.077	2.107
Oil13	−0.184	−0.058	0.126	3.426
Oil14	−0.135	−0.112	0.023	0.623

, emulsifiers such as AS1 and AS6, which possess moderate gaps (2.239 eV and 1.626 eV, respectively), failed to generate stable high internal phase emulsions (HIPEs). In contrast, AS4 and AS5, with slightly lower gaps (1.610–1.679 eV), consistently produced stable emulsions. Thus, energy gap alone did not correlate with emulsifier effectiveness, indicating it is not a decisive predictor of HIPE stability.

Dipole moment, however, showed a more consistent trend. As illustrated in Figure 9d, AS4 and AS5, both with dipole moments close to 6 Debye, achieved the highest HIPE success rates. In contrast, AS1 and AS6, with dipole moments below 1 Debye, were among the least effective. Interestingly, AS2 (10.04 D) and AS3 (4.61 D) exhibited intermediate performance, suggesting that moderate polarity—rather than extremes—is optimal for achieving emulsifier stability in oil–water systems. These findings are reinforced by the quantitative data in Table 8 that support the utility of dipole moment as a reliable predictor. The role of CED, presented in Figure 9e, was also examined. CED quantifies internal molecular cohesion and is sometimes used to infer interaction potential with other phases. However, no clear correlation was observed between CED and HIPE formation. For example, Oil 11 and Oil 12 had high CED values but did not yield stable emulsions, indicating that bulk cohesive strength may play a secondary role compared to interfacial forces such as van der Waals interactions and dipole alignment.

For structural context, Figure 10 depicts the all-atom molecular structures used in the DFT study, highlighting the polyaromatic cores and various side chains. These architectures influence both molecular polarity and spatial coverage at the interface. Moreover, Figure 11 visualizes the HOMO and LUMO distributions of representative emulsifiers. These orbitals were localized within the PAC region, suggesting a lack of orbital overlap with surrounding molecules. This supports the conclusion that emulsification in these systems is governed by non-bonding interactions—rather than chemical bonding or charge transfer.

In summary, while the HOMO–LUMO energy gap and cohesive energy density provide some insights, it is the dipole moment that shows the strongest correlation with emulsifier performance. These findings suggest that appropriate polarity enables the formation of stable interfacial films through non-covalent interactions, particularly van der Waals forces. Although dipole moment has been previously linked to electro-coalescence phenomena in petroleum systems [51], this study presents the first application of dipole correlation to HIPE stabilization, offering valuable guidance for designing future emulsifiers with targeted electronic properties.

5. Conclusions

A multiscale simulation study was conducted to investigate the stabilization mechanisms of water-in-oil HIPEs, with up to 70–80% water content, using DPD, MD, and DFT. A total of 42 emulsion models were simulated using three core emulsifiers (AS1–AS3) and fourteen oil types which revealed key relationships between molecular structure and emulsion stability. Among the tested emulsifiers, AS2 and AS3 exhibited significantly higher performance, achieving HIPE formation in 12 and 9 cases, respectively, compared to only 3 cases for AS1. The superior performance of AS2 and AS3 was attributed to enhanced interfacial dispersion, steric effects, and molecular polarity. Modified emulsifiers AS4 and AS5 further improved performance as they successfully stabilized HIPEs in all 14 oil cases for the benefit of optimized side-chain architecture and heteroatom inclusion.

DFT results showed that AS4 and AS5 possessed dipole moments of approximately 6.1 Debye, whereas poorly performing emulsifiers AS1 and AS6 had values below 1.0 Debye. Emulsifiers with excessively high dipole moments, such as AS2 (10.04 Debye), exhibited moderate stability, indicating that an optimal polarity range enhances interfacial film formation. In contrast, the HOMO–LUMO energy gap (ranging from 0.029 to 0.082 Ha) and CED showed weak correlation with emulsification performance, suggesting that these descriptors are less reliable in predicting HIPE stability. Visualizations of orbital distributions confirmed that HOMO and LUMO regions were localized within the PAC as it limited electron delocalization and supported the dominance of non-bonding interactions in stabilization. Four distinct mechanisms of HIPE stabilization were identified: nanomicelle bridging, emulsifier–oil co-construction of interfacial films, side-group anchoring in interfacially inactive oils, and full interface coverage by emulsifier monolayers. This study demonstrates that dipole moment is the most predictive molecular property for HIPE stabilization, introducing a new descriptor for emulsifier design. These findings provide actionable criteria for the development of next-generation surfactants in petroleum recovery, chemical flooding, and complex emulsion technologies.

Although these simulations provide valuable mechanistic insights into HIPE stabilization, the absence of direct experimental validation remains a notable limitation. To enhance the credibility and applicability of the findings, future research should incorporate experimental comparisons, particularly through interfacial tension measurements, microscopic characterization of emulsion microstructures, and rheological analyses. Additionally, the current DPD framework does not account for potential chemical adsorption or reactivity at the interface. Addressing this limitation will require the development and integration of reactive force fields capable of capturing chemical bonding phenomena relevant to complex emulsion systems.