The process begins with the reaction of the methoxy group (-CH₃O) from the silane coupling agent KH570 with water, resulting in the formation of silica hydroxyl (-Si-OH). The silica hydroxyl newly generated undergoes a dehydration condensation reaction with the silica hydroxyl groups on the SNPs, making the surface of the somewhat hydrophilic SNPs, as depicted in Figure 1. To initiate the synthesis, 1.0 g of SNPs is initially dispersed in 100 mL of anhydrous ethanol with the assistance of ultrasound for 10 min forming a dispersion, and this mixture was referred to as solution A. Different mass fractions of KH570 are added into a 50 mL ethanol solution (80 wt%) and hydrolyzed under magnetic stirring for 1 h, leading to the creation of solutions with varying concentrations of KH570, designated as solution B. Next, solution A is transferred into a 250 mL three-necked flask, followed by the adjustment of the pH to 9–10 using NaOH. While maintaining stirring at 85°C, solution B is added dropwise and continuously stirred, allowing the reaction to progress under reflux conditions for 4 h. After the completion of the reaction, the resulting product is washed with anhydrous ethanol and filtered until the filtrate becomes transparent. Subsequently, the product is dried in an oven at 60°C for 12 h and ground into a fine powder.

The surface functional groups of the SNPs were characterized using a Fourier-transform infrared spectrometer (FTIR, IRTracer-100) with a scanning resolution of 4 cm⁻¹ both before and after modification. To determine the modification degree, SiO₂ was initially mixed with a specific amount of KBr, ground repeatedly, pressed, and subjected to infrared absorption intensity measurement. By comparing the absorption peak intensity with that of the original SNPs, the reduction in surface -OH was observed through the FTIR, aiding in assessing the extent of modification. The contact angle of KH570@SiO₂ was determined using a dynamic contact angle measurement system (Data physics OCA15EC) with each measurement performed thrice for reproducibility. Subsequently, KH570@SiO₂ or original SNPs were incorporated into the AOS solution and sonicated to create KH570@SiO₂-AOS or SiO₂-AOS dispersion systems, respectively. The particle size distribution of the dispersion system was then analyzed using a laser particle size analyzer (Malvern NS 90).

Bulk foam stability testing is often used as a fundamental screening tool to assess the foamability of diverse formulations of foam, despite variations in foam size between porous media and the free state (Singh et al., 2016). To evaluate foam stability, static foam stability experiments were conducted following the Waring blender method (Guo et al., 2006). Initially, various wettability of KH570@SiO₂ nanoparticles (0.2 wt%) were dispersed in a 200 mL AOS (0.2 wt%) solution utilizing an ultrasonic disperser to establish a dispersion system. The foam was then created by agitating the system in a blender (8,000 rpm) for 3 min and subsequently transferred to a graduated cylinder. Foamability and foam stability were assessed based on the initial foam volume (V₀) and the drainage half-life (t_1/2), representing the duration for half of the liquid to drain from the foam. Furthermore, a more detailed observation of the foam texture was achieved by dropping the lower foam layer onto a slide using a dropper, followed by covering it with a cover glass to create a single-layer distribution of foam. High-resolution three-dimensional super-depth microscopy (VHX-5000, Keyence) and a high-quality camera were used to capture the bubbles’ shapes and diameters at different time points. The acquired images were analyzed using ImageJ software, which facilitated the calculation of the number and area of bubbles over time, along with determining the bubbles’ diameters through image processing and statistical analysis.

The adsorption of SNPs at the bubble surface and the mechanism of KH570@SiO₂-stabilized foam were investigated using fluorescence microscopy (Trim Scope, Germany). A fluorescent probe, Rhodamine B with a negative charge and a maximum excitation wavelength of 543 nm, was employed in this study. Initially, SNPs in the dispersion were stained with Rhodamine B, followed by centrifugation and washing with distilled water until the upper layer became clear. Subsequently, foam stabilized by the stained nanoparticles was prepared according to the standard procedure, enabling the collection of fluorescence images of the foam using a microscope.

A microscopic etched porous media model was made for foam flooding experiments. Through a photolithographic approach, a pore network resembling the real pore structure of natural cores from the Daqing oilfield was etched onto a glass plate. Upon sintering this etched glass plate with another smooth glass plate, the microscopic model was formed. The glass micro-model was designed with a porosity of 25%, featuring an average pore depth of approximately 40 μm and a width of 100 μm. Comprising three main components of the microscopic displacement device—the foam generation device, microscopic pore model, and observation/data collection device illustrated in Figure 2–the microscopic displacement apparatus was established. Initially, the microscopic model was evacuated, following which the pores were filled with crude oil (μ = 7.69 mPa·s, 25°C). The subsequent water flood involved injecting water for 10 pore volumes (PV) until no oil production occurred, signaling the termination of the water flooding phase and the beginning of the high-water saturation stage. Foam generation involved injecting foam solution and air into the foam generator at a 1:1 volume ratio, proceeding with a 2 PV injection for AOS foam or KH570@SiO_2-AOS stabilized foam. This was followed by an additional 5 PV of water flooding, with a consistent injection rate of 1 μL/min. Throughout the displacement process, an observation microscope (VHX-5000, Keyence) and a high-resolution CCD camera with a parallel light source documented the entire process. The study focused on investigating the effect of KH570@SiO₂-AOS stabilized foam flooding on the mobilization of various residual oils during the high-water saturation period, conducted at a temperature of 25°C. Microscopic displacement images were continuously recorded during the flooding experiment. To obtain the oil saturation from the remaining oil image, the image was sharpened to improve the contrast. Water and oil in the image were then distinguished based on a gray threshold according to a program developed in Python. Finally, the oil saturation was obtained by calculating the oil hole area ratio. Upon completion of the experiments, petroleum ether and ethanol were injected into the micro-model to clean all pore throats. Subsequently, the micro-model was placed in a constant-temperature oven at 45°C for drying.

The FTIR spectra of the SNPs are illustrated in Figure 3. In the spectrum of naked SNPs, the presence of -OH on the SiO₂ surface is indicated by the -OH peak observed in the range of 3,400–3,600 cm⁻¹. Even after modification, the persistence of -OH suggests that it did not completely react with the hydrolyzed KH570. The bending vibration of the O−H bond is represented by the absorption peak at 1,635 cm⁻¹. The characteristic peaks at 1,091 cm⁻¹, 810 cm⁻¹, and 486 cm⁻¹ correspond to the asymmetric stretching vibration, symmetric stretching vibration, and bending vibration of the Si−O−Si bonds on the surface of SiO₂, indicating the presence of structural water on the K570@SiO₂ nanoparticles surface. The peak at 2,954 cm⁻¹ is attributed to the asymmetric and symmetric stretching vibrations of the C-H bonds in -CH₃. Furthermore, the stretching vibration of the C=O bond is represented by the peak at 1718.21 cm⁻¹, while the peak at 1,234.44 cm⁻¹ represents the stretching vibration of the C-O-C bond (Coates, 2000; Smith, 2018). These peaks suggest a reaction between the groups on the SiO₂ surface and KH570, leading to the successful coverage of the surface of the SNPs by -CH₃. As the KH570 content increases, the absorbance of the surface structured water at 3,200–3,600 cm⁻¹ and the surface -OH at 1,636 cm⁻¹ gradually weaken, indicating the grafting of KH570 onto the SNPs surface in the form of a Si−O−Si bond.

According to the above experiments, the organic hydrophobic groups were successfully grafted onto the surface of SNPs after modification. This grafting necessarily changes the surface hydrophobicity of the SNPs. Figure 4 illustrates the change in the contact angle of SiO₂ with the content of KH570. The initial contact angle on the surface of naked SNPs is in the range of 0°–15° (as in Figure 4A), where water dropped onto a SiO₂ thin film spreads and rapidly penetrates the substrate. The contact angle of SiO₂ first increases with the increase of KH570 content and then reaches a plateau, as in Figures 4B–H and Figure 5. This indicates that KH570 effectively alters the surface hydrophobicity of SiO₂. When the concentration of KH570 reaches 7.7%, the contact angle of SiO₂ reaches 91.2° as in Figure 4H. However, as the content of KH570 continues to increase, When the contact angle ranges from 30° to 80° does not continue to increase. This is because a self-condensation side reaction occurs between the hydrolysis products of KH570 and Si−OH. With an increase in KH570 content, both the grafting of silane alcohol with SNPs and the extent of the self-condensation reaction intensifies. However, when KH570 content is excessive, it tends to favor self-polymerization rather than grafting onto the surface of SNPs.

In an AOS solution, a good foam stabilizer must be dispersed well. Figure 6 presents the median diameter (D₅₀) of the aggregation of SNPs (0.2wt%) in an AOS (0.2wt%) solution. The naked SNPs tend to cluster together, forming aggregates with a D₅₀ of around 200 nm, which lose their nanoparticle properties. Conversely, the KH570@SiO₂ nanoparticles exhibit mitigated aggregation when the contact angle ranges from 30° to 80°, leading to an improvement in dispersion within the AOS solution as evidenced by a decrease in D₅₀ to below 100 nm as the contact angle reaches 50°–60°. However, when the contact angle surpasses 80°, severe aggregation reoccurs, with the D₅₀ exceeding 200 nm. This phenomenon is attributed to the heightened hydrophobicity that arises from the introduction of hydrophobic groups, thereby intensifying van der Waals forces between particles. The enhanced intermolecular interactions cause the particles to repel the aqueous phase, which ultimately results in their increased proximity and aggregation.

Foaming was made using the Waring blending with dispersions of SiO₂ or KH570@SiO₂ nanoparticles (0.2wt%) with different contact angles and AOS (0.2wt%). The foam volume and t_1/2 are presented in Figure 7. The foaming volume and t_1/2 of the dispersion with naked SNPs were 658 mL and 448 s, respectively. The foaming volume decreased slightly with increasing contact angle, possibly attributed to KH570@SiO₂ being better dispersed and increasing the viscosity of the foaming dispersion. In these experiments, the rate of the blender is 8,000 rpm, which means the energy supplied for foaming is constant. Therefore, the foam volume decreases. The t_1/2 initially increased first with increasing contact angle and then decreased. When the contact angle of KH570@SiO₂ nanoparticles was around 60°, the t_1/2 (580 s) at maximum increased by 29.4% which is better than the foam system stabilized by (CH₃)₂SiCl₂-modified silica particles (Zhu et al., 2017) compared to the naked SNPs stabilized foam, indicating that the stability of the foam is strongest when the contact angle of KH570@SiO₂ nanoparticles is near 60°. Therefore, when the hydrophilicity and hydrophobicity are too strong, the synergistic effect with the surfactant cannot achieve a moderate hydrophobic effect on the nanoparticle surface. Consequently, the SNPs cannot firmly adsorb at the gas-liquid interface, resulting in ineffective enhancement of foam stability. The properties of the foam are a combination of foaming volume and foam stability, and the effect is optimal when the contact angle of KH570@SiO₂ nanoparticles is around 60°.

Figure 8 illustrates the dynamic changes in the foams stabilized by AOS (0.2wt%)-SiO₂ (0.2wt%) or AOS (0.2wt%)-KH570@SiO₂ (0.2wt%)with different contact angles at 0 min, 10th min, 20th min, and 30th min. Compared to the foam stabilized by naked SNPs (Figure 8A), it is evident that the foam stabilized by KH570@SiO₂ nanoparticles exhibits a slower increase in size (Figures 8B–D). Specifically, after 20 min, the foam stabilized by naked SNPs displays bubbles arranged in a polygonal pattern. These bubbles have larger diameters, resulting in the foam skeleton. In contrast, the foam stabilized by KH570@SiO₂ particles maintains fine textures, particularly for particles with a contact angle of approximately 60° (Figure 8C). The presence of KH570@SiO₂ particles leads to a higher proportion of smaller bubbles and less variability in bubble sizes, suggesting a significant deceleration in foam coalescence. However, as the contact angle of KH570@SiO₂ increases to 90.3°, the coalescence of foam becomes more obvious (Figure 8D, evidenced by a reduction in the number of bubbles and a decrease in roundness at the 20-min mark. These observations align with the findings derived from t_1/2 measurements conducted on the foam. Figure 9 shows changes in the D₅₀ of the foam at the beginning and 20th minute. Initially, the D₅₀ of the foam stabilized by naked SNPs is 246 μm. However, D₅₀ decreases to 214 μm when the contact angle reaches 60°. This decrease can be explained by the nanoparticles present in the liquid phase aggregating on the surface of the bubbles. Notably, the variation in D₅₀ at 20 min aligns with the trends observed in the images.

After grafting -CH₃ onto the surfaces of SNPs, the surfaces become partially hydrophobic, allowing SNPs to adsorb at the gas-liquid interface. To investigate the adsorption of KH570@SiO₂ particles on the gas-liquid surface, the optical microscope experiments were performed first and the microscopic images from generated foams by KH570@SiO₂ nanoparticles (0.2wt%)-AOS (0.2wt%) with a contact angle of 61.5°dispersions were demonstrated in Figure 10. Fluorescence microscopy analysis, as shown in Figure 10B, reveals a distinct visualization of the foam structure stabilized by KH570@SiO₂ particles, compared to foam only stabilized by AOS as illustrated in Figure 10A. The adsorption of KH570@SiO₂ particles onto the gas bubble surfaces, with some particles forming layers separated by platform boundaries, facilitates the formation of a three-dimensional network structure, thereby significantly enhancing foam stability. To visualize the particle distribution more clearly, the liquid from the foam in Figure 10B was drained to form dry foam skeletons, represented in Figures 10C. Notably, some particles are embedded within the dry foam structure, confirming the adhesion of particles to gas bubble surfaces. The retention of particles within the dry foam framework provides evidence that if the particles were in the surrounding continuous phase and not adhering to gas bubble surfaces, they would have been expelled during liquid drainage, rather than being retained within the dry foam skeleton.

The stability of the foam in the static condition of the microscopic model is given by Guo et al. (2006), and the flow state of the foam stabilized by the modified nanoparticles is described in this experiment. Figure 11 demonstrates the flow of foam in microscopic pores. A comparison is presented between foam generated by KH570@SiO₂ (0.2wt%) with a contact angle of 61.5°and only AOS (0.2wt%) dispersion, and foam generated only by AOS (0.2wt%). Foam stabilized by KH570@SiO₂-AOS, in comparison to foam stabilized by AOS, can keep smaller sizes over a longer period. This is achieved through the formation of bridging blockages with small bubbles and single-bubble blocking up, as Figures 11A, C. The small bubbles aggregate, with liquid films facilitating interactions between neighboring bubbles, leading to enhanced resistance to flow. Consequently, subsequent foam is compelled into narrower channels. Conversely, Figure 11B illustrates the flow of AOS foam, where bubbles amalgamate into larger trains that can go through the pores by deformation without effective blockage. Notably, the slightly larger bubbles formed by KH570@SiO₂-AOS lead to less deformable liquid films due to particle adsorption on the bubble surface. Consequently, it is prone to blockage when passing through a throat, as illustrated in Figure 11C. Contrastingly, in AOS foam, bubbles without soil-absorbed surface are easier to deform and are more likely to flow through a throat, resulting in ineffective blockages, as evidenced in Figures 11D–F, indicating a weaker blockage.

Oil displacement experiments were conducted using AOS (0.2wt%) and KH570@SiO₂ (0.2wt%)-AOS (0.2wt%) foams. Dead-end pores, very common in rocks, are characterized by only one end connected to the pore-throat network. Upon water flooding, residual oil is distributed in dead-end pores, with the driving force of water proving insufficient to penetrate the higher viscosity dead-end residual oil. This results in residual oil in dead-end pores as depicted in Figure 12A. It is still a great challenge to recover dead-end residual oil. The subsequent AOS foam displacement process, shown in Figures 12B, C, reveals the residual oil in dead-end pores cannot be displaced completely using the foam only stabilized by AOS. The foam stability generated only by AOS is inadequate, with a limited number of small-sized bubbles maintained, and most of the bubble sizes are larger than the throat size, so the larger bubbles are difficult to squeeze into the dead-end pores. Even if smaller AOS bubbles manage to touch the top oil in dead-end pores, they quickly coalesce and enlarge due to poor oil tolerance, preventing further penetration, and leaving some residual oil still in dead-end pores, as illustrated in Figure 12C. In contrast, KH570@SiO₂-AOS foam more effectively displaces residual oil from dead-end pores compared to only AOS foam. Figures 12D–F demonstrate the displacement process by the KH570@SiO₂-AOS foam. The adsorption of particles on the bubble surface enhances interfacial viscoelasticity, thus increasing the foam’s blockage capability and facilitating efficient flow channel blockage. KH570@SiO₂-AOS foam stability is superior, maintaining relatively small bubbles for an extended period, reducing flow resistance when entering dead-end pores, and enabling deeper penetration into oil-containing dead-end pores, as illustrated in Figure 12D. Within the dead-end pores, bubbles deform under the influence of driving forces. While AOS bubbles are compressed against the surface of the residual oil due to the high viscosity of oil, KH570@SiO₂-AOS bubbles possess higher interfacial viscoelasticity. Deformed KH570@SiO₂-AOS bubbles tend to restore their original shape, gradually displacing residual oil into a film or droplets and allowing for efficient displacement from the dead-end, as observed in Figure 12E. Ultimately, the majority of oil in the dead-end pores is displaced by the KH570@SiO₂-AOS foam, as shown in Figure 12F. The result indicates that the foam stabilized by the KH570@SiO₂ particles displacement system drives the residual oil in dead-end pores by way of the lots of small bubbles blocking the water channel, enhancing the driving force of the bubble into dead-end pores, then squeezing, deformating, and restorating.

ImageJ software was used to calculate the residual oil in the full microscopic model after different fluid displacements. Figures 13A, C were raw images of initial oil distribution before flooding. The pores of microscopic model is full of oil. In the process of microscopic displacement, water flooding was carried out firstly and then a slug (0.3 PV) of the foam was injected at the residual oil saturation, followed by an extended water flooding. During water flooding, there was a significant “viscous fingering” phenomenon due to the low resistance of the large pores. The injected water preferentially entered into the large pores, and only a small amount of water was driven into the small pores. There is a large amount of residual oil after water flooding, Figures 13B, D. The displacement effects of water flooding, AOS foam flooding, and KH570@SiO₂-AOS were compared (Figure 13). After AOS foam flooding and extend water flooding, there's less residual oil, as shown in Figures 13E, G. However, after KH570@SiO₂-AOS foam flooding and extend water flooding, the distribution of residual oil is further reduced, as shown in Figures 13F, H. The recovery of water flooding, AOS foam flooding, and KH570@SiO₂-AOS was 59.5%, 70.4%, and 79.1% respectively. Less oil was trapped during the KH570@SiO2-AOS and extended water displacement, indicating that KH570@SiO₂-AOS and extended water displacement could achieve the best oil recovery. This system was more effective in mobilizing and displacing residual oil to the production outlet than water flooding and AOS foam flooding followed by water flooding. The displacement behavior is consistent with the pore-scale visualization observations in Figure 12.