Both toluene and n-heptane (Analytical reagent grade) were purchased from Fisher Scientific. The solvents were used without further purification. Sodium chloride (NaCl), calcium chloride (CaCl₂), and magnesium chloride (MgCl₂) were purchased from Sigma Aldrich and used without purification for preparing brine solution in deionized water. Asphaltene samples were extracted from different crude oils by standard methods described elsewhere, following the standard procedure of IP143 (Table 1), and hence maltenes are obtained (Mullins, 2011; Majumdar et al., 2013). The physical properties of crude oils and their corresponding maltenes confined into engineered proxies are given below (Table 1) (Mullins, 2011; Majumdar et al., 2013). Mesoporous silica (200 nm average particle size and 4 nm average pore diameter) was purchased from Sigma-Aldrich, while mesoporous silica-2.5 nm was synthesized at the laboratory (Liu et al., 2006; Liu et al., 2013). Silica-2.5 nm was synthesized using only different carbon chain length surfactants or hydrothermal treatments (Liu et al., 2013). Then silica-2.5 nm (powder) was treated with the second hydrothermal, thus having an even stronger structure and better hydrolysis resistance. More details on the synthesis of silica-2.5 nm (powder) are given in another report by Liu et al. (2006). The silica porous monolith samples were purchased from Particle Solutions, LLC (Alachua, FL).

Measurements of BET surface areas, pore volumes, and diameters from the engineered silica rods (Table 2) were obtained with nitrogen adsorption and desorption at 77 K utilizing a Micromeritics ASAP 2020 surface area and porosity analyzer. Before the adsorption−desorption measurements, the samples were degassed at 150°C for 180 min under a vacuum pressure of 10 μm Hg.

The white powder mesoporous silica particles with a 4 nm pore diameter have a spherical morphology, and the sizes of the particles are uniform (reported by the manufacturer as 200 nm particle size). TEM imaging of this material revealed particles of approximately 4 nm diameter with an ordered array of pores, verified with BET-BJH analysis and matching the specifications provided by the manufacturer (Ok et al., 2017). Silica-4.0 nm and silica-2.5 nm (white powders) have surface areas of 597 and 1,167 m²/g, respectively. These two white powdered samples’ pore volumes are 1.04 and 0.97 cm³/g for silica-4.0 nm and silica-2.5 nm. The surface area to volume ratio for silica-4.0 nm is lower, with a value of 574, than that of silica-2.5 nm, with a value of 1,203 (Ok et al., 2020).

The nanoporous silica rods were kept in crude oils and maltenes for several weeks at 25°C for the complete filling of the nanopores. During the saturating process of the pores, the beakers or vials were sealed properly by several layers of parafilm to prevent the evaporation of lighter phases. After filling the silica rods’ nanopores, the outer surface of the silica rods was cleaned with a napkin to ensure that no crude oil or maltenes were left on the outer surface of the nanoporous silica rods.

To prepare the brine solution, 58.4 g NaCl, 111.0 g CaCl₂, and 95.0 g MgCl₂ were dissolved in 0.710 L deionized water at 25°C. The concentration of the brine is 250.000 ppm. Table 3 lists the solutions prepared. Hence, the surface area of nanoporous silica samples and the total amount of confined fluids (0.4 ml) are kept constant. This allows for comparing pore volumes’ effect on the confined fluids’ behavior. In addition, the mixtures were left at 25°C for at least 2 days for complete saturation of the nanopores before the low-field NMR measurements.

T₂ NMR measurements were performed on a Bruker Minispec mq20 NF Series instrument with a magnetic field strength of 0.47 T corresponding to a proton resonance frequency of 20 MHz at a magnet temperature of 40°C with Minispec software. The instrument was equipped with a 10 mm temperature-variable probe. Transverse magnetization relaxation (T₂) was measured using the standard Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence found in the Bruker library at 40°C. The T₂ measurements were conducted with a time delay between 90° and 180° pulses (t) of 0.2 ms for confined crude oil and maltenes, while the delay time was 0.5 ms for bulk samples. The number of data points was adjusted for each sample before each T₂ measurement, and T₂ data were acquired with 1,024 scans, and the delay time was chosen long enough, 5*T₁, to enable complete decay of the T₂ signal. The repetition time between two subsequent scans was set to 5 s. The same parameters were utilized for confined crude oil inside nanopores of silica-4.0 nm and silica-2.5 nm (white powder).

In the present study, the three-exponential fitting analysis of the acquired T₂ decays was performed in the Microcal Origin software using the following function for total NMR signal y: x = ∑ i = 1 3 A i exp − x T 2 i (3) where x stands for the signal detection time, T_2(i) is the transverse relaxation time of the i-th component with respective amplitude A_i. Fraction of each T₂ was determined as (A₁/(A₁ + A₂ + A₃)) × 100. The T₂₍₁₎, T_2(2), and T₂₍₃₎ values were later used to determine correlations between crude oil properties and NMR relaxation behavior (Canan et al., 2022). The three-exponential fitting analysis of the acquired T₁ decays was performed in the Microcal Origin software using the following function for total NMR signal y: y x = ∑ i = 1 3 A i exp x T 1 i (4) where x stands for the signal detection time, T_1(i) is the transverse relaxation time of the i-th component with respective amplitude A_i. Microcal Origin software gives T₂ results with error bars.

The NMR transverse magnetization relaxation data were analyzed: a continuous distribution of T₂ exponentials was fitted for all T₂ decays using the CONTIN algorithm (Provencher, 1982). This analysis yielded a plot for the continuous T₂ distribution.

Figure 1 exhibits the T₂ relaxation curves of the samples both in bulk and confined states. At the same time, Table 4 summarizes the T₂ relaxation values of crude oils and their corresponding maltenes in bulk and confined into nanoporous silica rods.

First, T₂ relaxation data of crude oils were treated with three component exponential decaying functions (see Eq. 3). The shortest T₂ values, T₂₍₁₎, were assigned to the heaviest SARA components (asphaltenes and resins), while the longest T₂ values, T₂₍₃₎, correspond to the saturated hydrocarbons, the lightest compounds in the sample. It was assumed (Volkov et al., 2021) that intermediate T₂₍₂₎ values characterize the aromatic compounds (heavier than the saturates but much lighter than asphaltenes and resins) in the samples. Second, in sample 1, a heavy crude oil, there is no significant deviation from bulk behavior when the sample 1 molecules are confined into nanoporous silica rod-10 nm. The confinement effect is slightly observed, as revealed by the shortening of T₂ values when sample 1 molecules are confined into nanoporous silica rod-2.5 nm. Only T₂₍₁₎ assigned to asphaltenes and resins showed a decrease both in silica-10 nm and silica-2.5 nm, while both T₂₍₂₎ and T₂₍₃₎ values decreased only inside silica-2.5 nm.

In sample 2, a medium-heavy crude oil, T₂ relaxation values decreased approximately 4-fold, confining into silica-10 nm and almost 7.5-fold inside silica-2.5 nm. In sample 3, a light crude oil, the decrease in T₂ values upon confining into either silica-10 nm or 2.5 nm is nearly 13-fold. The molecules of sample 3 feel the confinement effect independent of pore diameter. In samples 2 and 3, medium-heavy and light crude oils, aliphatics also feel the confinement as indicated by a reduction in T₂₍₃₎ values of confined crude oils.

Third, the fraction of heavy component (A₁) is not influenced by confinement in sample 1. The percentage of aliphatic fractions (A₃) with the largest T₂ relaxation values gets higher upon confinement in sample 1, a heavy crude oil, while the A₃ fraction of samples 2 and 3 decreases upon confining into silica-10 and silica-2.5 nm. The A₂ fraction of sample 1 was assigned to aromatics drops in the confined state. The A₂ percentages of samples 2 and 3 do not vary significantly in the confined state.

Based on these results, treatments for crude oil production need to be decided appropriately before production. Similar treatments could be applied to produce light and medium-heavy crude oils. The ^oAPI gravity borderline value above which the treatment should be switched to heavy crude oil is approximately 30 (Speight, 2002; Riazi, 2005). The results suggest that the SARA analysis of crude oils in bulk needs to be revisited more systematically for the confined crude oils.

T₂ distribution data obtained by Inverse Laplace Transformation (ILT) of T₂ decaying curves show a systematic trend for all three samples (Figure 2). As the crude oils were confined into silica-10 nm and 2.5 nm, the T₂ peaks became narrower and shifted towards lower values. Comparison of T₂ distributions indicates broader T₂ peaks for sample 3 in bulk compared to samples 1 and 3. This clearly shows the influence of confinement on the dynamics of crude oil molecules. Considering SARA fractions as saturates, aromatics, resins, and asphaltenes, it is possible to say that heavy crude oil has lower intensity aliphatic T₂ peak. In contrast, the T₂ peak assigned to aliphatics becomes broader for medium-heavy and light crude oils.

T₁ is a more sensitive experiment. The T₁ values are obtained by the exponential growth function shown above (see Eq. 4). Table 5 summarizes the T₁ results, and Figure 3 compares the T₁ curves of crude oils in bulk and confined states.

As expected, T₁ values in bulk and confined states are longer than corresponding T₂ values. T₁ values reflect some physical properties of crude oils, such as ^oAPI gravity in bulk. For instance, the T₁ values of sample 1 with lower ^oAPI gravity are much shorter than those of samples 2 and 3 with higher ^oAPI gravity. Similar to T₂ results, T₁ results in a confined state also deviate from bulk attitude. As in T₂, the deviation degree of sample 1 (heavy crude oil) in a confined state compared to bulk is lower than that of samples 2 and 3 in a confined state compared to their bulk dynamics. T₁₍₃₎ of sample 1 becomes only 2.2 times shorter inside silica-2.5 nm nanopores, while T₁₍₃₎ values of samples 2 and 3 are shortened 4.4 and 13.9 times, respectively, inside silica-2.5 nm nanopores. Surprisingly, the T₁(1) and T₁(2) values of sample 1 become longer in silica-10 nm nanopores, then reduced inside silica-2.5 nm nanopores. When the molecules are confined into nanoporous systems in heavy crude oil, the confinement effect suggests thinking beyond the conventional concept of SARA (saturates, aromatics, resins, and asphaltenes). A remarkable portion of saturates and aromatics-resins feel the confinement which in turn causes the molecules of aromatics and resins to act similar to asphaltenes. This is also reflected in A₁ and A₂ percentages in sample 1 inside silica-10 nm. A different situation occurs in samples 1 and 2 inside silica-2.5 nm nanopores. Both fractions of saturates-like and resins/aromatics-like fractions increase while the percentage of asphaltenes decreases. In the light crude oil (sample 3) confined into silica-2.5 nm nanopores, the percentage of saturates-like increases, but fractions of resins-aromatics and asphaltenes decrease. These fluctuations in the percentages of different sub-components of crude oils indicate that the conventional SARA approach needs further analysis and better definition for especially confined states of matter.

Light crude oils contain approximately 1% or less than 1% asphaltenes, and medium-heavy crude oils might have a maximum of 3%–4% asphaltenes. On the contrary, the fraction of heavy crude oil asphaltenes might exceed 8%–9% (Volkov et al., 2021). Asphaltenes in the confined state might influence the dynamics of the other components of crude oils (Espinat et al., 2017; Korb et al., 2013). For instance, according to the study by Espinat et al. (2017) on the dynamics of confined crude oil in the porous alumina catalysts by LF-NMR, asphaltene nanoaggregates and assemblies can act as large substances collected in the pores with mobility that affects toluene mobility. When the porous catalyst has macropores, the asphaltene collection with a larger population becomes less influential, permitting faster dynamics of asphaltenes and toluene. Referring to the work by Espinat et al. (2017), the current results suggest a similar explanation.

T₁ and T₂ relaxation dynamics of heavy crude oils with a large amount of asphaltenes do not change in the confined state in silica-10 nm (see Tables 4, 5; Figures 1–3). T₁ and T_2, relaxation dynamics of medium-heavy and light crude oils with less fraction of asphaltenes, changed remarkably in silica-10 and silica-2.5 nm rods. Asphaltene nanoaggregates are formed in the sub-regions closer to the walls of the nanopores of silica-10 nm in sample 1 (a heavy crude oil). Nanoaggregates of asphaltenes of sample 1 in silica-10 nm rod result in faster dynamics of the rest of the molecules, such as maltenes, as suggested by Espinat et al. (2017). Nanoaggregate formation and assemblies of asphaltenes of sample 1 “confine” maltene molecules in the pores of the silica-2.5 nm rod. The stronger confinement of maltene molecules in silica-2.5 nm rod is reflected as shorter T₁ and T₂ relaxation values than those of crude oil in bulk. For samples 2 and 3, medium-heavy and light crude oils, respectively, reductions in T₁ and T₂ values are observed in silica-10 and silica-2.5 nm rods. Since samples 2 and 3 do not contain a high fraction of asphaltenes, the relaxation dynamics of maltene molecules in the nanopores of silica rods observe a severe constraint confinement effect by direct interaction with the pore walls. The above explanation could be summarized as follows: a relaxation model of irregular surface dynamics of maltene molecules near nanoaggregates of asphaltenes and bulk dynamics between clusters of these nanoaggregates where the nanoaggregates are closer to the pore walls in heavy crude oils (Korb et al., 2013). These clusters based on the nanoaggregates are less influential for medium-heavy and light crude oils because of their lower fraction.

The present results also show that a relatively higher fraction of asphaltenes, as in heavy crude oils, influence wettability alteration, defined as making the reservoir rock more water-wet in the petroleum industry (Mohammed and Babadagli, 2015). According to the current results, wettability alteration is an internal factor in heavy crude oil. In other words, nanoaggregate clusters of asphaltenes of heavy crude oil formed in silica-10 nm rod wet the pore walls. On the other hand, in medium-heavy and light crude oils with a lower fraction of asphaltenes, the fraction of asphaltenes is insufficient to form the nanoclusters that might influence the wettability alteration.

The wettability alteration in reservoirs is an essential issue because wettability in the pores of oil reservoirs could be either homogeneous or heterogeneous. In homogeneous wetting, the whole rock surface is wetted uniformly by either water or oil. On the contrary, in heterogeneous wettability, different surface zones show another tendency of wetting regimes for oil or water. Altering the wettability, when there is complete water wetting or heterogeneous wetting by distinct regions, has a direct relationship with enhanced oil recovery processes (Yefei et al., 2011). The method of treating oil reservoirs by different approaches, such as nanofluids and surfactants (Yefei et al., 2011; Eltoum et al., 2021), aims at improving oil recovery and reducing the trapped crude oil in reservoirs. The present results show that in confined crude oils, oil-trapped “isolated globules” could be treated by, at first, understanding the nature of crude oil, especially in a confined state. To provide which conditions could be better, the correlations between the identification of crude oil (heavy, medium-heavy, and light), asphaltenes (the heaviest fraction of crude oils), wettability alteration treatment (nanofluids, the salinity of water, surfactants, etc.), and potential enhancements in oil recovery need to be mapped.

In addition to mapping the correlations mentioned above, a potential issue that might affect the wettability in the current results might arise from the aging of the crude oil samples, as suggested by Medina-Rodriguez et al. (2020). The work of Medina-Rodriguez et al. analyzes glass beads solely with oil and demonstrates the effect aging has on surface wettability by changes in T₂ distribution and diffusion. The effect of aging is more pronounced in untreated glass beads, with a shift from a peak height of 40.0 ± 0.1 ms at the end of 1 day, attributed to a water-wet situation, to a peak height of 35.7 ± 0.1 ms at the end of 3 weeks, explained by an oil-wet situation. However, in treated glass beads, there was a slight shift from 35.1 ± 0.2 ms to 34.7 ± 0.1 ms as a function of aging (1 day versus 3 weeks). The slight shift was considered to further enhance oleophilicity. Similar to the oil-wetted treated glass beads with 3 weeks of saturation, silica rods were kept long enough in the crude oil samples to reach complete saturation of the tiny nanopores. Therefore, the behaviors of crude oils confined in the nanoporous silica rods are studied in an aging regime closer to the treated glass beads. As a result, oil-wetted nanoporous silica rods were subjected to analysis in the present study.

It is not possible to measure the dynamics of asphaltenes in solid-state by low-field NMR relaxometry. However, asphaltenes’ indirect influence on maltenes’ dynamics is NMR measurable. Therefore, the next focus is the analysis of the dynamics of maltenes in bulk and confined states (Figures 4, 5; Tables 6, 7). The first significant result is a reduction in both T₁ and T₂ maltenes’ values compared to the corresponding crude oils in bulk. There is also a systematic decrease in both T₁ and T₂ values of maltenes in a confined state compared to those of bulk.

The confinement effect is more pronounced in maltenes (sample 6) of light crude oil (sample 3). The fractional analysis by the A₁, A₂, and A₃ components of T₂ does not reflect significant changes. In samples 4 and 5 inside nanopores of silica-2.5 nm rod, the percentages of resins decrease, and resin molecules act similarly to the aromatics. In sample 6, maltenes of light crude oil, while the fraction of resins decreases, the saturates’ percentage increases. This was also observed in the light crude oil (sample 3) inside nanopores of the silica-2.5 nm rod. These results clearly show that resin molecules might indicate similar attitudes to saturates in natural confinement with pores in different scales ranging from nano to micro and even macro, while aromatics might show resin-like behaviors. That means in addition to saturates, aromatics, resins, and asphaltenes. There could be aromatics-to-resins and resins-to-saturates as sub-groups in crude oil. Such sub-groups might affect wettability alterations of crude oils in rock cores. Therefore, these issues must be considered before deciding on proper treatments for crude oil production.

The dynamical behaviors of sample 3-a light crude oil whose physical properties are shown in Table 1-inside the nanopores of white powder samples of silica-4.0 nm and silica-2.5 nm are studied. The primary goal, in this case, is to compare the effect of pore diameter and S/V ratio on the dynamics of light crude oil (see Table 2 of Ok et al. (2020) of listing the physical properties of the two silica materials). Table 8 lists the T₁ and T₂ values of sample 3 in bulk and a confined state in white powder nanoporous silica samples, while Figures 6A, B displays the comparisons of T₂ and T₁ relaxation.

Chemically speaking, the two white powdered silica samples are provided as SiO₂, but it was already shown that they are H₂O absorbing systems, and their pore walls are decorated with −OH groups (Ok et al., 2017). Thus, nanoporous SiO₂(−OH) matrixes are ideal enough to resemble geologically relevant sub-surfaces. As shown in Table 2 of Ok et al. (2020), the silica-2.5 nm material has a relatively low pore volume but has the highest pore surface area due to its nearly micropore feature and mesoporous structures than the silica-4.0 nm probe. The surface-to-volume ratio (S/V) demonstrates that a 2.5 nm pore diameter silica has a higher S/V value. This ratio will be used as a parameter in the following discussion. As listed in Table 8, the T₂ and T₁ values of confined crude oil inside the nanoporous systems are lower than those in bulk. However, the T₂ and T₁ values of crude oil in silica-4.0 nm are lower than those in silica-2.5 nm. This shows that the dynamics of crude oil in the confined state do not change as a function of pore diameter and instead by S/V ratio. Timur claimed that in a three-component NMR model, the pore spaces of a porous medium were divided into three groups based on their S/V ratio distribution (Timur, 1969). He explained that the longer T₁ times correspond to the smaller S/V ratios and the larger pores. T₁ times become longer when sample 3 molecules are confined into silica-2.5 nm with the highest S/V ratio, while the shortest T₁ times are observed upon confining H₂O molecules into silica-4.0 nm with the lowest S/V ratio.

For this reason, the trend in T₁ times of sample 3 confined into silica materials is attributed to the S/V ratios rather than pore diameter. These measurements were conducted by keeping the pore volume of each nanoporous silica constant at around 120 cm³ (see Table 2 of Ok et al. (2020)). When the pore volume is kept constant at approximately 120 cm³, water has the longest T₁ value in the mixture with silica-2.5 nm. Similarly, the highest T₂ was obtained for the same crude oil and silica-2.5 nm mixture.

When the pore diameter is enormous and the S/V ratio is small, fluid molecules first wet the surface of the pore walls. Further filling happens smoothly from the pore wall toward the center of the pore. Then complete filling of the pores is achieved (Grünberg et al., 2004). In the case of nanoporous silica with a small pore diameter and large S/V ratio, the filling mechanism occurs differently: first, the center is filled; hence there is a coexistence of filled pore segments with wetted pores. Further filling of the pores grows axially in the direction of the pore axis. The pore volume is also related to pore length when cylindrical pores are assumed. The nanoporous silica with a small pore diameter and long pore axis leads to larger pore volumes. The long pore length (large pore volume) gives enough freedom for confined crude oil molecules. Relatively free confined crude oil molecules prefer motion in the axial direction, resulting in weaker interactions with pore walls. This is a possible explanation of larger T₁ values for pores with smaller pore diameters and larger pore volumes.

The last set of experiments conducted included the confinement of crude oil with ^oAPI gravity of 35.92 and density of 0.83707 g/cm³ at 25°C and brine into powdered silica-4.0 nm and silica-2.5 nm matrixes. The results are shown in Table 9. The shorter relaxation values of light crude oil arise from the heavy components such as asphaltene and resins. In contrast, the longer relaxation values are assigned to the lighter fractions, including saturates and aromatics (Ok et al., 2021).

The effect of confinement is more pronounced for brine, according to the T₂ results. More importantly, both fluids’ relaxation values, both T₂ and T₁, are higher when the fluids are confined into silica-2.5 nm. Three-component fittings evaluate the relaxation values when both fluids are confined together in the nanoporous proxies. T₂(3) and T₁(3), the shortest relaxation values, are assigned to brine, while T₂(1) and T₁(1), the longest relaxation values, are attributed to the mixture of brine and crude oil, and T₂(2) and T₁(2) to mainly crude oil. A portion of brine forms an interphase layer with pore walls; this might help enhance crude oil recovery if, for example, brine is enriched with inorganic nanoparticles such as ZnO (Alomair et al., 2022). Brine fills the nanopores faster than crude oil molecules due to its lower viscosity and forms a layered structure via interaction with −OH decorating the pore walls. Hence, the free space in the pore volume is decreased, and less crude oil is confined into the nanoporous silica proxies. This is reflected in longer relaxation values, T₂(1), T₂(2), T₁(1), and T₁(2), when both fluids were confined than when either of the fluids was confined separately. The results of the last set of samples can be related to wettability alteration. The intense interaction between water molecules and the pore walls shows the possibility of wettability alteration, referring to the process of making the reservoir rock more water-wet (Mohammed and Babadagli, 2015) by a facile approach of utilizing brine.