Enhancing Understanding of Siloxane Surface Properties and Functional Group Effects on Water Deoxygenation

Abstract

The deoxygenation process in water used in well injection operations is an important matter to eliminate corrosion in the petroleum industry. This study used molecular dynamics simulations to understand the behavior of siloxane surfaces by studying the surface properties with two functional groups attached to the end of siloxane and their effect on the deoxygenation process. The simulations were performed using LAMMPS to characterize surface properties. Jmol software version 14 was used to generate siloxane chains with (8, 20, and 35) repeat units. We evaluated properties such as total energy, surface tension, and viscosity. Then, we used siloxane as a membrane to compare the efficiency of deoxygenation for both types of functional groups. The results indicated that longer chain lengths increased the total energy and viscosity while decreasing surface tension. Replacing methyl groups with trifluoromethyl (CF3) groups increased all the above mentioned properties in varying proportions. Trifluoromethyl (CF3) groups showed better removal efficiency than methyl (CH3) groups but allowed more water to pass. Furthermore, the simulations were run using the class II potential developed by Sun, Rigby, and others within an explicit-atom (EA) model. This force field is universally applicable to the atomistic simulation of polymers, inorganic small molecules, and common organic molecules.

1. Introduction

The deoxygenation of water is a vital process in the oil and gas industry [1]. The presence of oxygen in water can cause several problems in the oil industry, including the corrosion of equipment and pipes, which can be costly and potentially dangerous. To reduce or control these problems, deoxygenation techniques are used to remove or reduce the oxygen content of water used in drilling operations [2].

Membrane deoxygenation is a valuable method for removing dissolved oxygen from water used in various industrial applications [3].

Polyargansiloxanes are a type of silicone-based polymer known as siloxane. The properties of polyarganosiloxanes can be affected by attaching different organic groups to silicon atoms. Some common organic groups include methyl, phenyl, and fluorine-containing alkyl groups. These changes in chemical structure vary the properties of polyarganosiloxanes such as flexibility, heat stability, and chemical resistance [4].

Many industries use silicon in their fields due to its unique properties. It is used in coatings, oil and gas separation processes, adhesives, lubricants, and many other products because of its resistance to heat, cold, moisture, and chemicals. For example, silicone polymers are used in coatings at low ratios to help substrates wet easily, resulting in a smooth, uniform appearance after drying [5]. In oil and gas separation processes, siloxane compounds are employed in membranes and filters to improve the efficiency of separating hydrocarbons from water; it is also used to remove oxygen from water in petroleum processes [6]. In the adhesives and sealants industry, silicon’s flexibility and chemical resistance make it ideal for applications that require long-lasting bonds, such as in construction materials [7].

The most common methods to prepare polyarganosiloxanes are condensation polymerization and addition polymerization. The properties of the synthesized siloxane can be affected by the choice of synthesis method [8]. The main type of polymer in the group of polyorganosiloxanes is polydimethylsiloxane (PDMS). PDMS is composed of repeating dimethylsiloxane units, as illustrated in Figure 1 [9].

In PDMS, each silicon atom is bonded to two oxygen atoms and two methyl groups (CH3), resulting in a flexible and relatively inert polymer chain. The methyl groups in PDMS have low polarity and contribute to a low surface tension, meaning that PDMS tends not to wet surfaces easily. This property makes PDMS suitable for applications where low adhesion to surfaces is desired, such as in mold release agents or coatings for water-repellent surfaces. Another property of PDMS is that it is permeable to gasses due to its flexible polymeric chains and the large space between methyl groups; this high permeability to gasses makes PDMS a valuable membrane in applications such as oil–gas separation [9].

The methyl group in PDMS can be replaced by other functional groups to synthesize polymers for a wide range of other exciting applications; for example, fluorine-containing alkyl groups as shown in Figure 2 lead to a high resistance to hydrocarbon solubility while maintaining low surface tension. This property is important for its application in gas–oil separation and is of direct importance for applications in the oil industry [9,10].

Many fundamental questions related the relationship between the backbone structure and the functional groups attached to the backbones can be addressed using molecular simulations, with carefully chosen interaction potentials and target boundary conditions with correct sampling procedures.

The beneficial properties of PDMS motivated us to study and better understand the surface behavior of siloxane, aiming to enhance oxygen removal from water in petroleum industry applications to reduce corrosion.

In this study, MD simulations by LAMMPS were performed to evaluate the surface properties of siloxane with two different functional groups attached to the end of the chain, which are the methyl group and fluorinated alkyl, represented by trifluoromethyl. Three types of PDMS chain lengths were chosen, namely (8, 20, 35) monomers per chain. By modifying the chain length, the surface tension, wettability, and adhesion characteristics of PDMS can be adjusted to meet particular requirements. For example, the accurate surface tension estimation of oil and gas systems is important in gas well stimulation where minimizing surface tension is required to minimize capillary forces that trap the aqueous phase in the formation [11]. Surface tension, viscosity, and total energy were examined for both cases. Subsequently, we replicated the 20 monomer per chain for both end functional groups and used this to evaluate the efficiency of the deoxygenation of water.

2. Results and Discussion

2.1. Surface Tension

The study of surface tension is of great importance in understanding fluid behavior in petroleum gas–oil separation processes [12]. As mentioned in the introduction, trifluoromethyl groups (CF3) have the unique property of providing high resistance to the solubility of hydrocarbons while maintaining low surface tension. This property enhances the efficiency of separation processes.

Table 1. Chain length and surface tension for methyl and fluorinated alkyl function group.

Type of Chain	Chain Length	Surface Tension (m N/m)
CH3	8	21.44
CH3	20	19.97
CH3	35	17.8
Average		19.74
CF3	8	24.76
CF3	20	20.183
CF3	35	19.81
Average		21.58

and Figure 3 show that surface tension increased slightly when changing the end functional group from methyl to trifluoromethyl. In addition, surface tension decreased as the chain length increased for both types of functional end groups. These results can be used to check the accuracy of the results as the surface tension remained at low value. The strong electronegativity and spatial barriers of trifluoromethyl (CF3) groups on siloxane molecules lead to higher surface tension compared to siloxane molecules with methyl (CH3) groups [13]. However, surface tension decreases with increasing chain lengths, supporting the hypothesis that longer chains enhance van der Waals interactions and surface area, which reduces surface tension [14,15,16]. This result is observed for both types of end groups, indicating that chain length is a critical factor in determining surface tension. The results show that, although the surface tension of siloxane is slightly higher in the presence of a trifluoromethyl group compared to a methyl group, it still maintains a relatively low value, making it suitable for applications in gas–oil separation processes.

2.2. Viscosity

This section presents the simulation results for viscosity at different chain lengths with two functional end groups. Higher viscosity reduces water permeability, which is vital for applications where controlled filtration is necessary, such as in water purification or gas separation processes [17]. The results in

Table 2. Chain length and viscosity for methyl and fluorinated alkyl function group.

Type of Chain	Chain Length	Viscosity (cp)
CH3	8	8.83
CH3	20	18.4
CH3	35	20
Average		15.74
CF3	8	11.9
CF3	20	17.8
CF3	35	29.9
Average		19.86

and Figure 4 indicate that as the chain length increases, viscosity also increases, in agreement with the simulation results of JIC et al. [18]. Furthermore, the results showed that the viscosity for the trifluoromethyl (CF3) functional group is generally higher than for the methyl (CH3) group, with the exception of a chain length of 20, where the CH3 group exhibits slightly higher viscosity. Longer chain lengths and higher viscosity decrease the water permeability through the membrane [18]. The increase in viscosity with increasing chains lengths is due to increased molecular interactions and cross-linking, which impedes the flow of molecules [18]. The observation of viscosity has significant effects for membrane technology. Understanding this can be essential for balancing membrane performance. In addition, understanding the effects of viscosity helps in selecting appropriate manufacturing techniques and conditions, ensuring that the membrane is produced with consistent quality and performance [19].

2.3. Total Energy

Evaluating the total energy of the molecular interactions of siloxane is essential to understanding its behavior in various applications because it provides understandings of the stability and overall energetics of the system [20]. Accurately evaluating the total energy helps predict how siloxane molecules will interact with other materials, which can improve their performance in practical applications such as coatings, adhesives, and separation processes [20]. This section presents the simulated results of the total energy of siloxane chains with (CF3) and (CH3) functional end groups. As illustrated in

Table 3. Total energy for siloxane as a function of chain length and types.

Type of Chain	Chain Length	Total Energy (Kcal)
CH3	8	171.83
CH3	20	422.95
CH3	35	723.16
Average		439.32
CF3	8	177.44
CF3	20	424.47
CF3	35	722.9
Average		441.6

and Figure 5, the total energy increases as the chain length becomes longer. This trend is consistent across both CF3- and CH3-functionalized siloxane, indicating that the longer the chain, the greater the number of atomic bonds and the more substantial the interactions within the siloxane structure. These results are expected since the addition of monomer units increases the number of atoms within the molecule, thereby enhancing the overall interaction energy.

The results showed that, while the functional group does influence surface properties such as viscosity and surface tension, it does not significantly affect the total energy of the siloxane chains. This could be due to the fact that the contribution of the functional group to the overall molecular energy is relatively small compared to the interactions along the siloxane backbone. The implication of this finding is that the choice of functional group (whether CF3 or CH3) may not drastically alter the energetic stability of the siloxane chains, which could be advantageous in applications where consistent energy profiles are required.

2.4. Evaluation Dissolved Oxygen Removal

This section highlights the effect of functional end groups on the removal of dissolved oxygen by using different concentrations of dissolved oxygen in water, as mentioned in the Materials and Methods section.

Table 4. Dissolved oxygen removal at 300 K.

Chains Type	Force in z-Direction (kcal/mol)/Angstrom	Od Concentration with Water	Od Reomoval %	Water Pass %	(Od Removal %)/(Water Moleules Pass %)
CH3	0.25	20%	0.00	0.00	0.00
		50%	1.98	40.11	4.95
		80%	3.93	62.87	6.25
		100%	157.56	95.68	164.67
		average	40.87	49.67	82.29
CF3	0.25	20%	1.99	43.69	4.56
		50%	63.70	91.62	69.53
		80%	139.73	85.90	162.66
		100%	185.15	82.36	224.80
		average	97.65	75.89	128.66
CH3	1	20%	2.29	37.60	6.09
		50%	94.66	67.51	140.21
		80%	141.68	65.67	215.73
		100%	188.82	69.53	271.55
		average	106.86	60.08	177.86
CF3	1	20%	15.96	58.32	27.37
		50%	91.90	72.52	126.72
		80%	139.73	85.90	162.66
		100%	185.15	82.36	224.80
		average	108.19	74.78	144.68
CH3	2	20%	11.95	38.92	30.70
		50%	75.62	38.28	197.55
		80%	119.92	58.79	203.98
		100%	172.35	52.41	328.86
		average	94.96	47.10	201.62
CF3	2	20%	26.90	64.72	41.57
		50%	63.70	91.62	69.53
		80%	78.84	65.55	120.28
		100%	170.20	71.53	237.95
		average	84.91	73.35	115.76
CH3	4	20%	13.93	50.56	27.55
		50%	53.72	35.75	150.24
		80%	117.24	53.66	218.49
		100%	133.07	47.50	280.13
		average	79.49	46.87	169.60
CF3	4	20%	25.92	73.48	35.27
		50%	69.80	60.19	115.96
		80%	137.03	65.34	209.71
		100%	157.75	62.91	250.75
		average	97.62	65.48	149.09

summarizes the results obtained from the second part of the investigation. Ting Li et al. demonstrated that crosslinked PDMS membranes effectively removed dissolved oxygen, and their findings corroborate our observations of longer siloxane chains, which enhance oxygen removal efficiency [21]. In addition, Yinhua Yang et al. reported that high-hydrophobicity membranes, such as those with trifluoropropyl groups, showed improved separation performance. Their results validate our findings on the impact of the CF3 functional group on membrane efficiency [22]. We studied four oxygen concentrations, represented by (20, 50, 80, 100)% of dissolved oxygen in water. The system for these concentrations consists of three types of chain length of PDMS, which are (8, 20, 35) monomers per chain and each thus contain (87, 207, 357) atoms, respectively. To this, we then add 100 molecules of water and (20, 50, 80, 100) molecules of oxygen to represent the four concentrations mentioned. To apply a force in the z-direction for both water and oxygen molecules, we also applied forces of (0.5, 1, 2, 4) (Kcal/mol.Å) to enable the proposed separation within the timescale limitations.

We calculated the percentage of water molecules permeating through the membrane and then determined the ratio of dissolved oxygen to water flux or the passage of water through the membrane. Figure 6 presents a simulation snapshot of water and dissolved oxygen on the siloxane surface during the simulation period.

Figure 7 shows that siloxane with the CF3 functional group is more effective than the methyl group in removing dissolved oxygen at all concentrations. This is because CF3 groups are more polar and electronegative than CH3 groups, leading to stronger interactions with oxygen molecules and, consequently, increased removal efficiency [23]. The high polarity and electronegativity of the CF3 group increase the dipole moment within the siloxane structure, which promotes strong dipole interactions with dissolved oxygen molecules. This may be due to the more electron-attracting nature of the fluorine atoms in the CF3 group, which enhances the electrostatic attraction between the functional group and the oxygen molecules.

On the other hand, although the CF3 functional group is more efficient than the CH3 functional group, the percentage of water passing through the membrane is higher than in CH3 functional group, as can be seen in Figure 8. Despite the strong electronegativity of CF3, simulations predict that strongly hydrophobic pores could produce ultrahigh water permeability due to the low friction between the water molecules and the membrane; however, in practice, achieving high water permeability in these hydrophobic pores is difficult because of the high threshold pressure drop [24]. In addition, the position of the CF3 group could significantly influence how the modified PDMS interacts with water. The increased polarity and potential hydrophilicity of certain configurations could either enhance or reduce water permeability, depending on how these groups are positioned within the polymer structure [25]. Understanding this balance can guide the development of membranes tailored for specific applications where either high selectivity or high permeability is prioritized.

In both types of membranes, we evaluated the ratio of oxygen removal percentage to water passage percentage, as shown in Table 4 and Figure 9. Overall, we can conclude that the ratio of oxygen removal to water passage is better for the CH3 functional group compared to CF3. While CF3 groups are more effective at capturing and removing dissolved oxygen due to their stronger polar interactions, the strongly hydrophobic pores and specific positioning of CF3 groups within the PDMS structure might also reduce the membrane’s selectivity, allowing more water to pass through [24,25]. In contrast, CH3 groups, despite being less effective in removing dissolved oxygen, provide a better balance by allowing less water passage, thereby achieving a higher dissolved oxygen removal efficiency ratio. The key takeaway from this finding is that membrane design needs to account for the specific application and desired balance between oxygen removal and water passage. For applications where minimal water passage is critical, CH3-functionalized membranes might be more advantageous, despite their lower dissolved oxygen removal efficiency. Conversely, for scenarios where higher oxygen removal is prioritized and water passage is less of a concern, CF3-functionalized membranes may be more suitable. This nuanced understanding reinforces the need to tailor membrane functionality based on the specific requirements of the intended application, whether the priority is maximizing oxygen removal, minimizing water loss, or achieving a particular balance between the two.

3. Materials and Methods

3.1. Simulation Model

Molecular dynamics simulations through LAMMPS were performed for this study [26]. Jmol software version 14 was used to generate chains of siloxane structures with lengths of (8, 20, and 35) repeat units, as illustrated in

Table 5. Number and molecular weight of PDMS structures.

End Functional Group	PDMS Structure per MONOMER	Total Atoms of PDMS in Simulation Box	Molecular Weight (g/mol)
CH3	8	87	607.310
	20	207	1497.169
	35	357	2609.493
CF3	8	87	931.1384
	20	207	1532.729
	35	357	2284.717

[27]. Methyl (CH3) and trifluoromethyl (CF3) functional groups were independently incorporated into each sample. The simulations were run using the class II potential developed by Sun, Rigby, and others, as shown in

Table 6. Potential parameters for bond interactions [28,29].

bond	Kt	r0	Angle	Kθ	θ0	Dehidral	KØ	n
C-F	496	1.363	C-Si-C	44.4	113.5	F-C-Si-C	0.9006	3
C-H	340.6175	1.105	C-Si-O	42.3	113.1	F-C-Si-O	0.9006	3
C-Si	238	1.809	F-C-F	60	109.5	H-C-Si-C	0.9006	3
O-Si	392.8	1.665	F-C-Si	95	107.8	H-C-Si-O	0.9006	3
			H-C-H	60	109.5	Si-O-Si-C	0.9006	3
			H-C-Si	34.6	112.3	Si-O-Si-O	0.9006	3
			O-Si-O	44.4	113.5
			Si-O-Si	60	109.5

[28,29]. This class II potential is applied within an explicit-atom (EA) model, treating Si, O, C, and H atoms separately. A timestep of 0.01 fs and rc = 11 Å was used for all of the simulations. The temperature T was controlled with a Nose–Hoover thermostat at 300 K.

The methodology consists of two parts. Firstly, we use both the NVT ensemble and NPT ensemble with a constant density of 0.95 g/cm³ to evaluate the surface properties of siloxane with both functional groups, such as the total energy, surface tension, and viscosity.

In the second part of the investigation, we chose a chain length of (20) repeat units and replicated this four times for both types of functional group. The main aim of this part was to study the efficiency of removing dissolved oxygen in water. In this part, PDMS was equilibrated using the NVT ensemble with a density of 0.98 g/cm³ for the methyl functional group and 1.2 g/cm³ for the trifluoromethyl functional groups. Packmol [30] was used to configure the simulation box containing equilibrated PDMS in both functional groups as a membrane, including water with TIP3P force fields and dissolved oxygen randomly placed above the membrane, as illustrated in Figure 10. We choose four different concentrations (20, 50, 80, 100) % of dissolved oxygen with four forces in the z-direction (0.25, 1, 2, 4) (Kcal/mol. Å). The configured simulation box was run with the NVE ensemble and Langevin to control the temperature.

We used the common form of class I intermolecular potentials for bonds, angles, and torsional bonds. Bond lengths, bond angles, and torsional bond angles are preferred in Potential Harmony, as mentioned in the equations below, respectively [28,29]:

E = k_b (r − r₀)²

(1)

E = k_θ (θ − θ₀)²

(2)

E = k_t [1 + cos(n_Ø)]

(3)

where r₀ is the equilibrium bond distance, k_b is (energy/distance²), θ₀ is the equilibrium value of the angle, k_θ and k_t are energy, n is an integer > = 3, and Ø is the torsion bond angle.

All potential parameters for bond interactions and non-bond interactions used in the simulations are summarized in Table 6 and

Table 7. Potential parameters for non-bond interactions [28,29].

Nonbonded	∈ [kcal/mol]	ơ [Å]	q	Mass
C	0.054	4.01	−0.294000	12.0115
F	0.0598	3.2	−0.344300	18.9984
H	0.02	2.995	0.053000	1.0079
O	0.24	3.535	−0.445000	15.9994
Si	0.07	4.284	0.715000	28.085501

, respectively.

The van der Waals and Coulomb potential interactions occur through electrical interactions between two or more atoms or molecules in close proximity. The parameters for these interactions are listed in

Table 8. Van der Waals interaction parameters for the class II force field.

Bonds	∈ [kcal/mol]	ơ [Å]
C-C	4.0100	0.0540
C-F	3.7119	0.0459
C-H	3.6691	0.0233
C-O	3.8091	0.1062
C-Si	4.1583	0.0603
F-F	3.2000	0.0598
F-H	3.1059	0.0339
F-O	3.3881	0.1146
F-Si	3.9199	0.0459
H-H	2.9950	0.0200
H-O	3.3189	0.0615
H-Si	3.8875	0.0229
O-O	3.5350	0.2400
O-Si	3.9952	0.1107
Si-Si	4.2840	0.0700

. Non-bonded potential is defined as follows [28,29]:

$$\begin{array}{cc}{\mathrm{U}}_{\mathsf{\alpha }\mathsf{\gamma }}^{\mathrm{nonbond}}\left(\mathrm{r}\right)=& \{\begin{array}{}\mathrm{(4)}& {\mathrm{U}}_{\mathsf{\alpha }\mathsf{\gamma }}^{\mathrm{vW}}\left(\mathrm{r}\right)+{\mathrm{k}}_{\mathrm{q}}\frac{{\mathrm{q}}_{\mathsf{\alpha }}{\mathrm{q}}_{\mathsf{\gamma }}}{\mathrm{r}}& \mathrm{r}<{\mathrm{r}}_{\mathrm{c}}\mathrm{(5)}& {\mathrm{k}}_{\mathrm{q}}\frac{{\mathrm{q}}_{\mathsf{\alpha }}{\mathrm{q}}_{\mathsf{\gamma }}}{\mathrm{r}}& \mathrm{r}>{\mathrm{r}}_{\mathrm{c}}\end{array}\end{array}$$

where ${\mathrm{U}}_{\mathsf{\alpha }\mathsf{\gamma }}^{\mathrm{vW}}\left(\mathrm{r}\right)$ are the van der Waals interactions, r is the distance between two atoms of type α and γ, q is the electric charge, and r_c is the cutoff distance for the van der Waals interactions. To determine the interaction parameters for non-bonded interactions, we used the mix sixthpower rule as shown below [28]:

$${\in }_{\mathsf{\alpha }\mathsf{\gamma }}={\displaystyle \frac{2{\mathsf{\sigma }}_{\mathsf{\alpha }}^3{\mathsf{\sigma }}_{\mathsf{\gamma }}^3\sqrt{{\in }_{\mathsf{\alpha }}{\in }_{\mathsf{\gamma }}}}{{\mathsf{\sigma }}_{\mathsf{\alpha }}^6+{\mathsf{\sigma }}_{\mathsf{\gamma }}^6}}$$

(6)

$${\mathsf{\sigma }}_{\mathsf{\alpha }\mathsf{\gamma }}={\left({\displaystyle \frac{{\mathsf{\sigma }}_{\mathsf{\alpha }}^6+{\mathsf{\sigma }}_{\mathsf{\gamma }}^6}{2}}\right)}^{{\displaystyle \frac{1}{6}}}$$

(7)

3.2. Calculation of Properties

3.2.1. Surface Tension

Surface tension was calculated using the Irving–Kirkwood definition (see Equation (8)), which involves computing the difference between the normal and tangential components of the pressure tensor. The pressure tensor components P_xx, P_yy, and P_zz were calculated during the simulation using the compute pressure command in LAMMPS. These components were averaged over a sufficient simulation time to ensure statistical accuracy [31].

$$\mathsf{\gamma }={\displaystyle \frac{{\mathrm{L}}_{\mathrm{z}}}{2}}\left\{{\mathrm{P}}_{\mathrm{z}\mathrm{z}}-{\displaystyle \frac{1}{2}}({\mathrm{P}}_{\mathrm{x}\mathrm{x}}+{\mathrm{P}}_{\mathrm{y}\mathrm{y}}\right\}$$

(8)

where L_z is the length in the z-direction of the simulation box.

3.2.2. Viscosity

The Green–Kubo relation was used to calculate the viscosity of siloxane, as shown in Equation (9), which relates the viscosity to the integration of the autocorrelation function of the off-diagonal stress tensor components. The stress tensor components and their autocorrelation functions were calculated during the simulation using the fix ave/correlate command. The viscosity was then obtained by integrating these functions according to the Green–Kubo relation as follows [26]:

$$\mathsf{\eta }={\displaystyle \frac{\mathrm{V}}{{\mathrm{K}}_{\mathrm{B}}\mathrm{T}}}\underset{0}{\overset{\mathrm{\infty }}{\int }}〈{\mathrm{P}}_{\propto \mathsf{\beta }}\left(0\right){\mathrm{P}}_{\propto \mathsf{\beta }}\left(\mathrm{t}\right)〉\mathrm{d}\mathrm{t}$$

(9)

where V is the system volume, k_B is the Boltzmann constant, T is the temperature, and ${\mathrm{P}}_{\propto \mathsf{\beta }}$ are the off-diagonal components of the stress tensor. The average viscosity was computed as the mean of the integrated autocorrelation functions for all relevant stress tensor components.

3.2.3. Total Energy

The total energy of the system was computed and monitored throughout the simulation using the LAMMPS molecular dynamics software. The total energy includes both kinetic and potential energy contributions and is given by [26]

$${\mathrm{E}}_{\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}={\mathrm{E}}_{\mathrm{K}\mathrm{i}\mathrm{n}}+{\mathrm{E}}_{\mathrm{P}\mathrm{o}\mathrm{t}}$$

(10)

The kinetic energy is calculated based on the atomic velocities and masses, while the potential energy includes contributions from bonded interactions (bonds, angles, dihedrals) and non-bonded interactions (van der Waals and Coulombic interactions). The LAMMPS input script was configured to output these energy components at regular intervals using the thermo_style command, ensuring the continuous monitoring of the system’s total energy. The thermodynamic properties, including total energy, potential energy, and kinetic energy, were printed every (t) timesteps during both the equilibration and production runs.

3.2.4. Oxygen Removal Calculation

The concentration of dissolved oxygen was monitored using the count command to calculate the number of oxygen molecules in two regions (above and under siloxane membrane) of the simulation box. The initial and final concentrations of oxygen in these regions were compared to determine the amount of oxygen removal.

4. Conclusions

In this paper, the effect of functional groups and chain length on the properties and performance of siloxane-based membranes in separation processes was investigated. Our study revealed that CF3 functional groups, due to their high electronegativity and polarity, slightly increase the surface tension compared to CH3 groups. However, the overall surface tension remains low, making these membranes suitable for gas–oil separation applications. As well as this, we concluded that increasing the chain length reduces the surface tension of both functional groups; understanding this behavior is important for optimizing the chain length to enhance membrane performance.

In terms of studying the effect of chain length on the total energy, we concluded that the increasing chain length in both functional groups will increase the total energy within the siloxane structure, while replacing the CH3 functional group with the CF3 functional group has a relatively small effect on the stability of the total energy. This result suggests that siloxanes containing CF3 and CH3 functional groups can provide proportionate energy profiles, which is an advantage for industrial applications that require stable membrane performance under variable conditions.

Additionally, we concluded that the efficiency of removing dissolved oxygen from water for membranes with the CF3 functional group is higher than for the CH3 functional group due to the stronger dipole–dipole interactions between CF3 groups and oxygen molecules. Despite this, a higher percentage of water is allowed to pass through. Conversely, membranes with CH3 functional groups are less efficient in removing dissolved oxygen than membranes with CF3 functional groups but with less water permeability, making them more suitable for applications where water retention is required.

These results emphasize the importance of tailoring membrane design to the specific needs of the application. For processes that prioritize maximum oxygen removal, membranes containing the CF3 group are more effective. Conversely, in cases where minimizing water passage is essential, membranes containing the CH3 group offer a more balanced solution. This precise understanding of the interaction between functional groups and membrane performance provides valuable insights for optimizing siloxane membranes in various industrial applications, including gas–oil separation and water treatment.