Mineral composition and pore structure on spontaneous imbibition in tight sandstone reservoirs

Abstract

The pore structure of tight oil reservoirs is extremely complex, with a wide range of pore sizes, diverse pore types, and well-developed nanopores and throats. Although researchers have conducted in-depth studies on the mechanism of spontaneous imbibition, issues such as low recovery and short imbibition distance have not been well explained. This study characterized the relationship between mineral composition and pore types through XRD and SEM, and combined NMR-C, LF-NMR, and micro-CT to characterize the pore size distribution of tight sandstones. The effects of pore structure characteristics and displaced phase properties on spontaneous imbibition were investigated. Using NMR-C and low-temperature evaporation methods, the proportions of free water and bound water in pores with diameters ranging from 3.33 to 600 nm were quantitatively characterized. The results indicated that the maximum pore size and its proportion were the primary factors influencing permeability, whereas the content of small pores and mesopores played a crucial role in accessing the quality of connectivity. The content of micron-scale pores in the core exhibited a positive correlation with quartz content, whereas the amounts of feldspar and clay minerals were closely associated with the development of nanoscale pores. The proportion of large pores and the viscosity of the displaced phase had the most significant effect on imbibition efficiency, followed by the homogeneity of pore distribution and connectivity. Moreover, the bound water content in nanopores exceeded 50% of the pore volume, which reduced pore size and connectivity during the imbibition process, resulting in a shorter imbibition distance. This study investigated the effect of pore structure and mineral composition on spontaneous imbibition and examined the factors contributing to the limited imbibition distance in tight oil reservoir development. The findings offer new insights and strategies for enhancing recovery in tight oil reservoirs.

Introduction

In recent years, the development of unconventional oil and gas resources, represented by tight oil, has rapidly increased globally¹. According to data from the U.S. Energy Information Administration, tight oil production in the United States accounted for 64% of its total crude oil production in 2022, playing a crucial role in improving the U.S. energy structure². China has enormous potential for tight oil resource development, with preliminarily proven reserves of approximately 17.82 billion tons, and an annual production of 1.05 million tons in 2018³. The effective development of tight oil resources is of great significance for ensuring China’s energy security and improving its energy structure. However, the average recovery factor of tight oil reservoirs worldwide is currently very low, only 5%~10%^4,5. How to increase the recovery of tight oil is an important issue that urgently needs to be addressed.

The development of nanoscale and microscale pores and throats, poor pore connectivity, complex pore-throat shapes, small pore radii, and strong heterogeneity in tight oil reservoirs severely limit the improvement of oil recovery^6,7. An in-depth understanding of the pore structure of tight oil reservoirs is crucial for optimizing their spontaneous imbibition efficiency and increasing the recovery of tight oil reservoirs. As shown in Table 1, the pore structure information obtained by a single characterization method is limited, and the problems and focuses addressed by each technique are different. Therefore, in actual research, two or more techniques are usually combined to characterize the pore structure of tight oil reservoirs, thereby obtaining more accurate, reliable, and comprehensive pore structure characterization results.

Table 1 Characterization methods for pore structure in tight oil reservoirs.

Different pore-throat types have significant differences in their impact on the permeability of tight oil reservoirs. Primary intergranular pores connected by larger pores and throats contribute significantly to permeability, while clay-dominated micropores connected by small pores and throats contribute less to permeability⁸. Pore types in tight reservoirs can be classified into four categories: intercrystalline pores, intragranular pores, micropores, and microfractures⁹. These pores range in size from 50 nm to 400 μm, and the variations in pore size lead to significant heterogeneity in the microscopic pore structure of tight oil reservoirs. This results in the formation of distinct pore networks, which greatly influence development of various tight sandstone reservoirs. In addition, mineral composition has a significant influence on imbibition. Clay minerals contain a large number of nanopores and exhibit water-wetness, which has a positive effect on spontaneous imbibition oil recovery¹⁰.However, clay minerals can fill the pore-throat system and swell when encountering water, blocking fluid transport channels^11,12.

Numerous studies have shown that pore structure properties such as pore size, pore-throat shape, and pore connectivity significantly affect spontaneous imbibition. According to the Young-Laplace equation, the smaller the pore size, the greater the capillary force, and the easier it is for fluids to enter the pores. Therefore, in theory, when the fluid properties are the same, small pores will be filled with fluid first, followed by large pores. However, Zheng et al.¹³ discovered that the proportion of small pores exhibited a negative correlation with oil recovery, whereas medium pores showed a positive correlation with tight oil recovery, and large pores had almost no effect on recovery. Considering that the pore size of tight sandstone reservoirs changes continuously in actual situations, Gong et al.¹⁴ simulated the imbibition process in variable-diameter capillaries using a two-phase lattice Boltzmann model. The results showed that the degree of imbibition increased significantly with the increase in pore size reduction, and even the occurrence of imbibition interruption, resulting in a very low degree of non-wetting phase recovery. Pore size is also related to connectivity. Studies have shown that micropores have the strongest connectivity, followed by mesopores, and macropores have the poorest connectivity^15,16. Dong et al.¹⁷ established a more realistic imbibition model for porous media based on average geometry and fractal geometry, revealing the influence mechanism of pore connectivity on initial and equilibrium imbibition behavior. The study showed that for pore structures with weak connectivity, the initial imbibition ability was weak, the corresponding imbibition time was prolonged, and the imbibition recovery was low. Therefore, it is crucial to study the influence of pore structure, such as nanopore size and connectivity, on imbibition.

This study characterized the correlation between mineral composition and pore types through XRD and SEM, and combined NMR- C, LF-NMR, and micro-CT to characterize the pore size distribution of tight sandstones. Four groups of water/gas imbibition and water/oil imbibition schemes were compared to investigate the effects of pore structure and displaced phase on imbibition. Considering the influence of bound water on the pore size and connectivity of nanopores, NMR-C and low-temperature evaporation methods were used to quantitatively characterize the proportions of free water and bound water in pores with diameters ranging from 3.33 to 600 nm.

Methods

The tight sandstone samples used in this study were obtained from the Changqing, Daqing, Xinjiang, and Jilin regions. The characteristics of tight reservoirs in the Changqing, Daqing, Xinjiang, and Jilin regions were shown in Table 2. Core samples similar to the reservoir characteristics were selected to reduce the influence of scale effects with XRD, SEM and CT.

Table 2 Reservoir characteristics and mineral composition in different regions.

Before testing, the chemical and physical characteristics of these cores, such as permeability and porosity, were provided, as detailed in Table 3. The selected core type was tight fine sandstone, which is characterized by an extremely tight pore structure, extremely low permeability, and difficulty in saturating with conventional oil. In the NMR experiment, to ensure accurate measurements, the hydrogen signal of any one phase in the oil-water system needed to be removed. Therefore, the oil phase used in the study was fluorinated oil with a viscosity of 1.424 mPa•s, and the water phase was distilled water.

Table 3 Core parameter table.

X-ray diffraction (XRD)

Due to the similarity in particle size and crystal structure, it will affect the accuracy of mineral quantitative analysis. It was known that the main components of tight sandstone are quartz, feldspar, clay minerals and carbonate²². First, the appropriate particle size of the corresponding mineral type was selected, and then the measurement was carried out at different angles for multiple scans to enhance the reliability of the data. This study employed a Bruker AXS D8 Discover X-ray diffractometer and adhered to the industry standard “X-ray Diffraction Analysis Methods for Clay Minerals and Common Non-clay Minerals in Sedimentary Rocks” (SY/T 5163 − 2018) during the experiments. According to Stokes’ settling theorem in fluid mechanics, water suspension separation or centrifugal separation methods were used to extract clay mineral samples with particle sizes less than 10 microns and less than 2 microns, respectively²³. Clay mineral samples with particle sizes less than 10 microns were used to determine the total relative content of clay minerals in the original rock, while clay mineral samples with particle sizes less than 2 microns were used to determine the relative content of various clay minerals in detail²⁴. The “K value method” was applied to obtain the characteristic peak intensity of minerals in unknown samples and calculate the mineral content.

Scanning electron microscopy (SEM)

This study used a Zeiss Evo Ma 15 scanning electron microscope and followed the petroleum and natural gas industry standard “Scanning Electron Microscope Analysis Methods for Rock Samples” (SY/T 5162 − 2014) for the experiments. The samples were milled with an accelerating voltage of 2 kV for 7 h to reduce the damage of pore structure in the sample. Subsequently, the surfaces of the samples were coated with carbon to provide a conductive surface layer. Finally, the samples were observed under a scanning electron microscope.

Micro computed tomography (Micro-CT)

This study used a Zeiss Xradia 620 Versa X-ray microscope with a scanning voltage of 60 kV, a power of 6.5 W, a pixel size of 0.3 μm, and 1014 slices. The samples were made into core columns with a diameter of 3 mm. Slice images were obtained through CT scanning, and post-processing such as noise removal, contrast enhancement, and artifact correction was performed on the images. Avizo software was used for three-dimensional reconstruction to obtain parameters such as pore diameter, porosity, coordination number, and connectivity.

Low-field nuclear magnetic resonance (LF-NMR)

Before analysis, all samples were oven-dried for 72 h at 120 °C and then immersed in distilled water. T2 curves were generated at intervals of 1, 4, 8, 24, 48, 96, and 168 h through repeated calculations. At each designated time, the samples were removed, and T2 spectra were recorded. This procedure was repeated. The NMR experiments were conducted using a Newmai MicroMR12-040 V high-precision desktop NMR unconventional core analyzer, featuring an echo time of 0.1 ms, 12,000 echoes, a repetition sampling interval of 3.5 s, and an accumulation sampling number of 32. In addition, changes in mass were monitored during the testing process.

After the above experiment, all samples were oven-dried for 72 h at 120 °C and then immersed in fluorocarbon oil. Fluorocarbon oil rather than manganese chloride solution or heavy water was used in order to ensure the consistency of the experiment and eliminate the hydrogen signal. It was saturated at a pressure of 20 MPa for 12 h, and then taken out, dried, weighed and tested for saturation²⁵. Whereafter, it was immersed in distilled water. T2 curves were generated at intervals of 1, 4, 8, 24, 48, 96, and 168 h through repeated calculations. At each designated time, the samples were removed, and T2 spectra were recorded. This procedure was repeated.

In order to eliminate the influence of environmental temperature, humidity, magnetic field strength and other factors on the results of nuclear magnetic resonance test, and ensure the accuracy and reliability of the test results, the samples under saturated water conditions were calibrated before the test. The calibration of sandstone samples are shown in Fig. 1.

Fig. 1

Calibration of sandstone samples for NMR.

There was a certain deviation between the nuclear magnetie results and the real formation properties in the high temperature and high pressure measurement. Factors had a great influence on the fluidity and mobility of the fluid such as pressure, temperature, and clay content. Relaxation times were corrected²⁶ as follows:

$$\begin{gathered} {T_{1,2}}={\left( {\frac{1}{{{T_{1,2{\text{ref}}}}}} - \frac{1}{{{T_{1,2{\text{bulk}}}}}}} \right)^{ - 1}}\cdot {{\text{e}}^{{k_2}\left( {\frac{1}{{{T_{}}}} - \frac{1}{{{T_{{\text{ref}}}}}}} \right)}} \\ \approx {T_{1,2{\text{ref}}}}\cdot {{\text{e}}^{K\left( {\frac{1}{{{T_{{\text{ref}}}}}} - \frac{1}{{{T_{}}}}} \right)}} \\ \end{gathered}$$

(1)

where T_1,2ref is relaxation time at room temperature in the laboratory, T_ref is room temperature in the laboratory, K is the proportional coefficient, with a value of 0.017 for illite-bearing sandstone.

Nuclear magnetic resonance- cryoporometry (NMR-C)

The experimental instrument used was a Newmai NMRC12-010 V-T NMR low-temperature pore analyzer. The magnet used in the NMR system was a permanent magnet with a main frequency of 11.835 MHz, a probe coil size of 10 mm, and an instrument time of 60 µs. The selection of the basic parameters of the instrument was based on maximizing the signal-to-noise ratio as a priority. The cooling method used liquid circulation cooling, with electronic fluorinated liquid FC110 as the temperature medium and a low-temperature cold bath at -60 °C as the cold source. External circulation was used for temperature control and constant temperature, providing a stable temperature environment for the samples. The temperature control range of the instrument was − 35 °C to 10 °C, with a temperature fluctuation of ± 0.02 °C.

Based on the XRD and SEM results, Y1 and Y2 samples were selected for low-temperature freezing-thawing experiments. The Y1 and Y2 samples were first ground and sieved, and 2–3 g of each were selected. The samples were vacuumed and pressurized at 15 ~ 20 MPa, saturated with water for 12 h, and the surface water was removed using a syringe. The samples were stored for 3 days to ensure uniform distribution of the liquid in the samples. The temperature gradient was set from − 35 °C to 0 °C, with 15–20 temperature points for the experiments, and each temperature point was maintained for at least 15 min. After the temperature program was completed, the pore size distribution at each temperature could be directly observed from the software. The samples after the experiment were dried at 70 °C for 12 h to retain bound water and remove free water. The above experimental measurement process was repeated, and the pore size distribution was recorded.

Result and discussion

Analysis of pore types and mineral composition

The main mineral components of sandstone reservoirs include quartz, feldspar, clay minerals, etc., and their content and proportion affect the mechanical, physical, and oil-bearing characteristics of the reservoir^27,28. Figures 2, 3 and 4 show the mineral composition and pore characteristics of rock samples from different regions obtained through XRD and SEM. The results indicate that Y1 and Y3 samples contain approximately 40% feldspar, which is easily weathered into clay minerals. There are large amounts of chlorite and illite fillings between the grains, with only a small proportion of intergranular and intercrystalline pores, and the overall pore development is poor. The Y4 sample contains 63.2% quartz, and the content of other types of minerals is not high. Since quartz itself is very dense with few pores, the porosity of Y4 is lower compared to other samples. In addition, the Y4 sample has low clay mineral content and “quartz overgrowth” phenomenon, which will lead to low nanopore content, plugging of primary pores, and reduced effective porosity. However, although the Y2 sample also has a quartz content of 53.1%, its calcite content reaches 18.6%, and its dissolution effect will form secondary pore spaces, resulting in increased rock porosity. Moreover, the intergranular spaces are filled with calcite cement, microcrystalline quartz, and other substances, with only a small amount of intergranular pores, intergranular residual pores, and pores within quartz grains. This phenomenon causes the Y2 sample to have a higher porosity, but a strong heterogeneity in pore structure. In summary, feldspar and clay minerals have a strong relationship with nanopore content and core connectivity, while minerals such as heulandite, calcite, and gypsum determine the heterogeneity of the core pore structure.

Fig. 2

Relative content of mineral types in different samples.

Fig. 3

Relative content of clay minerals in different samples.

Fig. 4

Scanning electron microscope images of different samples.

Study of pore size distribution in tight sandstone

As shown in Fig. 5, the pore size distribution curves of all samples exhibit a bimodal distribution. As core permeability increases, the maximum pore diameter also rises, generating the pore-throat distribution curve’s peak to shift to the right while the left peak’s height progressively drops. The pore size distribution patterns measured in this study are consistent with those reported from field measurements by Wang et al. and Yan et al.^29,30. It proves that the results measured in this paper have guiding significance for field scale. In line with the IUPAC pore classification standard, the samples are divided into three categories: micropores (less than 2 nm), mesopores (ranging from 2 to 50 nm), and macropores (greater than 50 nm), as illustrated in Fig. 6³¹. Mesopores account for the vast majority of pores, with a proportion of 71.14 ~ 92.60%. The proportion of macropores in Y1 and Y3 samples is around 10%, less than 1% in Y2, and 28.86% in Y4. This indicates that mesopores only affect the connectivity and effective porosity, while the maximum pore size and its proportion determine the permeability.

Fig. 5

Relaxation curves of different samples.

Fig. 6

Cumulative pore volume percentage curves of different samples.

Currently, it is generally recognized that quartz content is the main factor in controlling the development of macropores. However, there are differing perspectives on the relationship between clay mineral content and pore volume as well as specific surface area^32,33. As shown in Figs. 7 and 8, the nanopore content is not only related to the content of clay minerals but also to the content of easily weathered minerals such as feldspar. The ratio of feldspar to clay mineral content represents the degree of rock weathering. A smaller ratio indicates a higher degree of weathering, with most intercrystalline pores being obstructed by debris or clay minerals. This blockage results in a reduced proportion of pores with diameters less than 50 nm. However, clay minerals are composed of very fine mineral particles and have a flaky structure, which can also form a large number of nanoscale pores. When blocking intercrystalline pores, it does not mean that such pores are completely closed, but are transformed into nanopores of smaller sizes. This relatively slows down the reduction of pore volume while increasing the proportion of smaller nanopores. Tian et al.³⁴ reports micropores are commonly distributed in the clay mineral (especially illite and illite-smectite mixed-layer). It is consistent with the conclusions of this study.

Fig. 7

Correlation between the ratio of feldspar to clay mineral content and the proportion of mesopores.

Fig. 8

Correlation between clay mineral content and the ratio of clay mineral to total feldspar content and the proportion of pores with a diameter < 10 nm.

Figures 5 and 9 demonstrate a significant difference in pore size distribution measured by LF-NMR and NMR-C methods. This discrepancy is attributed to the rock sample size effect, which is influenced by both the changes in pore structure during the crushing process and the fluid-pore connectivity related to the sample size³⁵. Despite the differences in pore size distribution obtained by the two experimental methods, the distribution patterns of different pore size levels remain consistent. Therefore, these two methods can be applied to characterize and cross-validate pore sizes at different scales. In comparison to the LF-NMR results, the NMR-C experimental findings indicate that the four samples contain a substantial number of pores ranging from 1.71 to 10 nm, which constitute 45 to 51% of the total pore volume within the range of 1.71 to 600 nm. Notably, as the pore size decreases, there is a significant increase in pore volume. The pore volumes between 1.71 ~ 3 nm are almost identical for different samples, indicating that the pore distribution below 3 nm is relatively uniform and that the mineral composition has little effect on the pore content below 3 nm. Currently, many scholars are developing “nanometer water“³⁶ to mobilize these pores, which will greatly enhance the connectivity of the cores.

Fig. 9

Distribution of pore volumes in various samples using NMR-C.

As shown in Fig. 10, according to the analysis of CT three-dimensional reconstruction results, the porosities of Y1, Y2, Y3, and Y4 samples are 1.1128%, 1.1890%, 0.9715%, and 1.9392%, respectively. As illustrated in Fig. 11, the content of micron-scale pores in the core shows a positive correlation with the quartz content. Based on the LF-NMR results, the proportions of effective pores > 0.3 μm in Y1, Y2, Y3, and Y4 samples are 4.83%, 0.91%, 6.22%, and 4.04%, respectively, which can be converted to corresponding effective porosities of 0.47%, 0.095%, 0.54%, and 0.25%. These results indicate that although the micron-scale porosity of the core appears relatively high, most of the pores are isolated and not even connected through nanoscale pores. Due to the high chemical stability of quartz, it will not swell and form cracks when encountering water like clay minerals. Without methods such as like hydraulic fracturing and acidification, it extracting oil and gas from such pores is quite challenging. This difficulty contributes significantly to the high initial oil production followed by a rapid decline in the middle and later stages of tight oil reservoirs after fracturing.

Fig. 10

CT three-dimensional reconstruction of pore distribution.

Fig. 11

Relationship between quartz content and micron-scale pores.

Analysis of the effect of pore structure and displaced phase relative imbibition

To compare the contribution values of pore structure and fluid properties to the displacement efficiency, water/gas imbibition, and water/oil imbibition experiments were carried out, where water/gas imbibition represents the situation of water entering dry pores, and water/oil imbibition represents the process of water displacing oil. Figure 12 presents the distribution of pore size of all samples, as determined by NMR. The peak signal intensity of sample Y2 is significantly lower than other samples, especially when the imbibition time is short (0 ~ 4 h). It still has a large increase in the later stage of imbibition compared with samples Y1, Y3, and Y4. This suggests that sample Y2 has a more concentrated distribution of smaller pores with relatively uniform pore sizes, although its pore structure is the most uneven, with a smaller proportion of large pores. In contrast, samples Y1 and Y3 contain fewer large pores and exhibit a relatively uniform pore distribution. Sample Y4 has a relatively uneven pore distribution, with a significantly higher proportion of large pores compared to samples Y1, Y2, and Y3. This leads to its highest peak signal intensity and strongest imbibition capacity. This phenomenon suggests that the proportion of large pores is the key factor infuencing the imbibition effect.

Fig. 12

Comparative analysis of water/gas imbibition and water/oil imbibition.

As shown in Fig. 12; Table 4, there is almost no significant shift in peak position with the increase of imbibition time. It indicates that mineral dissolution or pore collapse during spontaneous imbibition is relatively weak. This can be attributed to the low hydrodynamic intensity relying solely on capillary forces. The percentage of large pores of sample Y2 increases significantly with time. The pores with relaxation time > 10 ms increase and the pores with relaxation time < 0.05 ms decrease. Because Y2 contains a large amount of calcite, the carbonate dissolves and forms larger pores with the increase of imbibition time. The calcite content of samples Y1 and Y3 is less than 2%. The pores with relaxation time < 0.05 ms shows an increasing trend. And the percentage of large pores is less than 2%. It proves that there is a strong correlation between calcite content and the reduction in micropores as well as the increase in macropores. In addition, carbonate minerals are weak water-wet or oil-wet. Although its dissolution has a good effect on the improvement of pore structure, it will significantly weaken the imbibition.

Because the minerals with higher feldspar and clay exhibit abundant dissolution pores and well-developed pore-throat connectivity such as samples Y1 and Y3, they eventually show great imbibition ability among the cores representing the four geological characteristics. The sample Y4 exhibits the highest imbibition efficiency, despite having the lowest feldspar and clay mineral content. It is attributed to its high quartz content of 63.2%, resulting in a large pore volume proportion of approximately 20%. It leads to the lowest seepage resistance. However, it also causes the lowest effective porosity.

Comparing the water/gas imbibition and water/oil imbibition schemes, it can be found that the peak between 0.01 and 1ms is generally significantly reduced and shifted to the right; the proportion of pores around 10ms is reduced, and the peak is shifted to the left. This is because the binding effect of nanopores < 2ms on oil is significantly increased, the increase of capillary force is less than the increase of viscous force, while in the large pores around 10ms, the capillary action is reduced, and the decrease of capillary force is less than the decrease of viscous force. This indicates that as the viscosity of crude oil increases, the pore radius in the main imbibition region also tends to become larger and more concentrated. By equating the derivatives of capillary force (2) and viscous resistance (3), Eq. (4) is derived, representing the median pore radius in the main imbibition region. Furthermore, evidence suggests that a continuous reduction in interfacial tension results in a gradual decline in the mobilization of nanoscale pores.

$${P_c}=\frac{{2\gamma \cos \theta }}{R}$$

(2)

$${P_\nu }=\frac{{8{\mu _o}\left( {L - l} \right)}}{{{R^2}}}\frac{{{\text{d}}l}}{{{\text{d}}t}}+\frac{{8{\mu _w}l}}{{{R^2}}}\frac{{{\text{d}}l}}{{{\text{d}}t}}$$

(3)

$$R=\frac{{8\left[ {{\mu _o}\left( {L - l} \right)+{\mu _w}l} \right]}}{{\gamma \cos \theta }}\frac{{{\text{d}}l}}{{{\text{d}}t}}$$

(4)

Where, P_c is the capillary force; P_v is the viscous resistance.

The maximum peak reduction of Y1, Y2, Y3 and Y4 are 24.60%, 38.19%, 11.78% and 29.34%, respectively, and the mesopore pore volume reduction are 11.16%, 9.21%, 9.77% and 17.48%, respectively, indicating that the worse the rock pore structure, the greater the influence of fluid properties on imbibition efficiency. Although sample Y4 exhibits the highest imbibition efficiency, its decline is second only to Y2. This is because, despite having good connectivity, the pore size distribution is uneven. As a result, connected flow channels form in certain pores during the imbibition process, causing the capillary forces at the inlet and outlet of these channels to reach equilibrium. This leads to the formation of dead oil zones, which impede the effective displacement of fluids. In addition, during water/gas imbibition, the proportion of large pores steadily increases as imbibition time progresses. In contrast, during water/oil imbibition, the distribution of large pores follows a more irregular pattern. The greater the viscosity of the displaced phase, the greater the seepage resistance and the fewer the pores affected.

Table 4 Different types of pore volume proportion table.

The method to evaluate connectivity is based on the spontaneous imbibition curve obtained from NMR data²⁷, as shown in Fig. 13. The tangent slopes corresponding to the curves at 4 h and 96 h are selected respectively, and the time corresponding to the intersection point of the two tangent lines is used as the demarcation point between the early and late stages of spontaneous imbibition, and the tangent slope is used instead of the spontaneous imbibition velocity to compare the imbibition differences of different cores. For the same sample, the imbibition rate of the water/gas imbibition scheme is much faster than that of the water/oil imbibition scheme within 0 ~ 1 h and 8 ~ 48 h, and the average imbibition rate of the water/gas and water/oil imbibition schemes basically remains the same or slightly decreases within 1–8 h and 48–168 h. This phenomenon is because the gas viscosity is low, the seepage resistance is small, and the surface tension between water and gas is smaller, and the capillary force is greater. It can be considered that the water/gas imbibition process mainly reflects the impact of complex pore structure on the behavior of imbibition, that is, the residual oil that cannot be displaced only by changing the interaction force between oil and water or lowering the crude oil’s viscosity. In addition, the area between the blue line and the red line reflects the residual oil caused by factors such as the higher seepage resistance of the oil phase and the binding effect of nanopores on the oil phase. These phenomena show that throughout the tight oil reserve development phase, even if efficient oil displacement strategies such as CO₂ flooding or the addition of imbibition agents are adopted, it is still difficult to achieve high recovery due to the existence of these complex pore structures.

Furthermore, studies have shown that higher permeability and porosity do not necessarily lead to higher recovery, as this correlation is only significant during the imbibition early stage. In the early phase, greater permeability results in a faster spontaneous imbibition rate, and an earlier completion time. However, as the viscosity of the displaced phase increases, this difference diminishes, the effect of permeability on imbibition decreases, and the impact of pore structure becomes more prominent.

Fig. 13

Connectivity of different samples.

Quantitative study of nanopore water film

The evaluation in the preceding section indicates that tight oil reservoirs contain a significant abundance of nanopores and exhibit complex relationships in their pore-throat configurations, resulting in extremely limited seepage areas. Porous media, especially cores with nanoscale pores, will limit liquids to these tiny spaces. This is because clay minerals exhibit negative electricity, and water molecules have strong polar characteristics, forming a thickness between a few water molecules (water molecule diameter is about 0.4 nm) to several hundred water molecules³⁷ The T2 spectrum at different temperatures is used to characterize the proportion of bound water and free water. However, the transformation radius is done by assuming S/V = a/r. In this way, the pore surface area remains the same and the pore fluid volume decreases after the free water vaporizing. It converts too small pore radius. The Gibbs-Thomson equation, based on the NMR-C method, is closely linked to the capillary effect. It describes how interfacial curvature, resulting from interfacial tension, influences the free energy of a material. This relationship is directly associated with factors such as pore diameter, pore shape, and wettability. Combined with the fact that water is the substance in direct contact with the pore walls, we can consider that the presence or absence of free water does not affect the measured pore size. In addition, this method is based on the phenomenon that “as the temperature rises, the liquid in the pores will gradually melt from smaller pores to larger pores”. Through this process, the pore volume corresponding to different diameters can be calculated. Therefore, we can use this method to quantify the content of free water and bound water in nanopores. The temperature thresholds for the removal of different types of water are generally identified as 75 °C ± 5 °C, for free water, 140 °C ± 5 °C for weakly bound water, and 210 °C ± 5 °C for strongly bound water. Nevertheless, experimental findings indicate that there is not a significant difference in evaporation temperatures between weakly bound water and free water³⁸. Therefore, in this paper, by means of long-term low-temperature evaporation, it is ensured that all strongly bound water is retained under the premise of removing free water.

$$\Delta {T_m}\left( x \right)={T_{mB}} - {T_m}\left( x \right)={T_{mB}}\frac{{4{\sigma _{sl}}}}{{{H_f}{\rho _s}x}}$$

(4)

Where H_f is the melting entropy, ρ_s is the density of the crystal state, T_mB is the bulk melting temperature, and σ_sl is the solid-liquid interfacial energy.

Fig. 14

Pore volume distribution before and after low temperature evaporation of Y1 sample.

Fig. 15

Pore volume distribution before and after low temperature evaporation of Y2 sample.

As shown in Figs. 14 and 15, in the pore diameter range of 3.33 ~ 100 nm, the free water content is relatively small. Large amounts of free water are present in pores smaller than 3.33 nm. This phenomenon is mainly because the capillary force in the pores of a few nanometers is greater. According to the Kelvin equation, the evaporation and condensation process of the liquid is affected by the surface tension and the liquid radius. Therefore, it leads to the close evaporation temperature of bound water in pores of a few nanometers and free water in pores of tens of nanometers. It can be observed that the relative content of free water in pores above 3.33 ~ 5 nm is significantly reduced, so this method is suitable for quantitative evaluation of free water and bound water in different pore sizes above 3.33 nm.

$$\ln \left( {\frac{p}{{{p_0}}}} \right)=\frac{{2\gamma {V_m}}}{{rRT}}$$

(5)

Where, p and p₀ indicate the actual and saturated vapor pressures, respectively; ϒ denotes the surface tension; V_m is the liquid’s molar volume; R represents the universal gas constant; r is the curved liquid’s radius; T is the temperature.

Table 5 Fluid volume of different pore sizes.

As shown in Table 5, the proportion of free water in Y1 is 56.13%, and the proportion of free water in Y2 is 17.28%. It is found that the free water content in Y2 is 38.85% less than that in Y1. This phenomenon is attributed to the fact that the proportion of illite-smectite mixed layers and illite in Y2 exceeds that in Y1 by 29%. Since montmorillonite has high expandability and large specific surface area, and illite consists of a layered structure of alternately stacked silicon-oxygen tetrahedra and aluminum-oxygen octahedra, and contains a large number of hydrophilic functional groups, making them have a better ability to adsorb water molecules. In contrast, the interlayer bonding of chlorite and kaolinite is tighter, which limits their ability to adsorb water molecules. This difference shows that although clay minerals contain abundant nanopores, the type of clay minerals determines the seepage ability of pores. If the pores are composed of montmorillonite and illite, and the original fluid is oil phase, it is easier to be displaced. On the contrary, if the original fluid is water phase, it will reduce the connectivity of the rock.

Effect of surfactant and low salinity water flooding on imbibition

In order to study the effect of surfactant on imbibition with different mineral contents, water and surfactants were selected as the displacement phases to compare their imbibition efficiency. The experimental results are shown in Fig. 16, surfactant flooding can significantly enhance the imbibition recovery. The sample with the highest enhanced oil recovery is Y2, with an improvement of 23.09%. Because it contains a large amount of calcite and clay minerals, the wettability of the rock is oil wet. The sample Y4 contains a large amount of quartz, resulting in water-wet. After excessively reducing the interfacial tension, the capillary force is significantly reduced. It leads that its enhanced oil recovery is much lower than Y2.

In order to study the effect of salinity on imbibition with different mineral contents, the salinity of the displacement phase was selected to be 0 and 62,750 mg / L ( CaCl₂ water type ) to compare their imbibition efficiency. The experimental results are shown in Fig. 17, the increase of salinity will lead to the decrease of imbibition recovery and imbibition rate. The samples Y1 and Y3 shows higher enhanced oil recovery using high-salinity water. The sample Y4 is the lowest. Because clay minerals have negative charge and high specific surface area, they can adsorb salt through ion exchange process. It results in water retention, which subsequently inhibits imbibition capacity. The dissolution of carbonate and sulfate minerals reacts with sulfate and chloride ions in water to form insoluble precipitates. It reduces pore connectivity and imbibition capacity. Quartz has a weak adsorption of salt and has little effect on imbibition.

Fig. 16

Comparison of water flooding and surfactant flooding in different cores.

Fig. 17

Comparison of low salinity water flooding and stratum water flooding in different cores.

Conclusion

In this work, the relationship between mineral composition and pore types was analyzed through XRD and SEM. The pore size distribution of tight sandstones was characterized by NMR-C, LF-NMR, and micro-CT. Using NMR-C and low-temperature evaporation methods, the proportions of free water and bound water in pores with diameters ranging from 3.33 to 600 nm were quantitatively characterized. The reason of limited imbibition distance in tight oil reservoir was clarified. The detailed conclusions can be drawn as follows:

(1)Small and mesopore concentrations are crucial to evaluating the connectivity quality. In contrast, the maximum pore size level and its proportion are the main determinants impacting permeability, according to NMR-C-LF-NMR-micro-CT investigations. The percentage of large pores and the viscosity of the displaced phase have the greatest impact on imbibition efficiency, followed by pore distribution homogeneity and connectivity.

(2) Cores with a high concentration of quartz have more micron-scale pores, whereas those with a high concentration of feldspar and clay minerals have more nanoscale pores.

(3) While clay minerals enhance the connectivity and capillary force of cores, due to the existence of bound water, they also reduce the pore size and connectivity during imbibition, and even the free water content in pores can be less than 20%.