Probe material choice for nuclear magnetic resonance cryoporometry (NMRC) measurements of the nano-scale pore size distribution of unconventional reservoirs

Abstract

Nuclear magnetic resonance cryoporometry is a newly developed technique that can characterize the pore size distribution of nano-scale porous materials. To date, this technique has scarcely been used for the testing of unconventional oil and gas reservoirs; thus, their micro- and nano-scale pore structures must still be investigated. The selection of the probe material for this technique has a key impact on the quality of the measurement results during the testing of geological samples. In this paper, we present details on the nuclear magnetic resonance cryoporometric procedure. Several types of probe materials were compared during the nuclear testing of standard nano-scale porous materials and unconventional reservoir geological samples from Sichuan Basin, Southwest China. Gas sorption experiments were also carried out on the same samples simultaneously. The K_GT values of the probe materials octamethylcyclotetrasiloxane and calcium chloride hexahydrate were calibrated using standard nano-scale porous materials to reveal respective values of 149.3 Knm and 184 Knm. Water did not successfully wet the pore surfaces of the standard controlled pore glass samples; moreover, water damaged the pore structures of the geological samples, which was confirmed during two freeze-melting tests. The complex phase transition during the melting of cyclohexane introduced a nuclear magnetic resonance signal in addition to that from liquid in the pores, which led to an imprecise characterization of the pore size distribution. Octamethylcyclotetrasiloxane and calcium chloride hexahydrate have been rarely employed as nuclear magnetic resonance cryoporometric probe materials for the testing of an unconventional reservoir. Both of these materials were able to characterize pore sizes up to 1 μm, and they were more applicable than either water or cyclohexane.

Keywords

Nuclear magnetic resonance cryoporometry pore size distribution octamethylcyclotetrasiloxane tight sandstone shale

Introduction

The exploration for and development of unconventional oil and gas resources is topic of considerable interest (Dai et al., 2012; Holditch, 2006). Due to the existence of micro- and nano-scale pore throats, these types of reservoirs exhibit different fluid and seepage mechanics relative to conventional reservoirs (Zou et al., 2012). Micro- and nano-scale pores must be represented in detail during research on unconventional reservoirs. Existing methods used to characterize pore structures can be divided into three types: image analyses, nonintrusive methods and intrusive methods (Maex et al., 2003). Image analyses such as transmission electron microscopy and scanning electron microscopy (SEM) can provide partial micrographs with quantitative information (Desbois et al., 2011). Nonintrusive methods, such as micro-CT (computed tomography) (Bai et al., 2013) and small-angle neutron scattering (Yang et al., 2017), are normally based on radiation theory. Intrusive methods are the most conventional for quantifying the pore size distribution (PSD) and include both gas adsorption (Zhang et al., 2017) and mercury intrusion porosimetry (Lai and Wang, 2015). To address the limitations of each method, several different techniques are usually combined to study geological samples during research efforts (Pan et al., 2017; Schmitt et al., 2015; Zhao et al., 2015). Nuclear magnetic resonance cryoporometry (NMRC), which is based on radiation theory, is a relatively new technique for acquiring PSD measurements that uses intrusive materials, and thus, it is classified as an intrusive method. Compared with the other two methods, gas adsorption can typically characterize only pore sizes between 2 and 300 nm even when using an analytical model with a wider range (Anovitz and Cole, 2015). Moreover, mercury intrusion porosimetry has the advantage of testing the micron-scale PSD; however, this technique may have the potential to modify pore structures with high capillary pressures (especially at the nano-scale), and it tends to probe the pore throat size rather than the pore size itself. Therefore, a method that is otherwise harmless to the pore structure and that constitutes a single overarching test for pore sizes less than 1 μm must be developed. If an appropriate probe material is chosen for NMRC, it will be absorbed into the pores spontaneously, thereby avoiding any damage to the pore structure similar to mercury injection techniques, and thus, NMRC could be complementary to other methods due to its theoretical pore size distribution coverage from 2 nm to 1 μm.

NMRC has been used widely in many different fields, such as the testing of bio-medicine polymeric nanoparticles (Gopinathan et al., 2014), ultrafiltration membranes (Jeon et al., 2008) and mesoporous TiO₂ spheres (Ryu et al., 2010). However, NMRC has rarely been used for the testing of unconventional reservoir nano-scale pore sizes, and few studies have used water as the probe material during the testing of porous rocks such as coal and shale (Firouzi et al., 2014; Qian et al., 2016; Webber et al., 2013). NMRC was unable to measure the PSDs of certain coal and shale samples in previous research; this outcome was attributed to the chemical effect of a combined size and pore surface that prevents water from condensing within the pores (Firouzi et al., 2014). Therefore, the selection of an appropriate probe material has great relevance for the accuracy of NMRC tests on geological samples. Distinct probe materials have not only different effects on the wettability of a mineral surface and the NMR signal intensity but also a disparate influence on the crystallization and melting of pore structures. Meanwhile, not all probe materials used during NMRC experiments are able to characterize PSDs up to 1 µm. In this work, several different types of probe materials were used to test the same batch of unconventional reservoir samples with supplementary gas adsorption data for comparison, and the probe materials were compared to select the appropriate substance with which to test unconventional reservoir samples using NMRC.

Theoretical background of NMRC

The Gibbs–Thomson equation, which forms the theoretical basis of NMRC, describes the relationship between the depression of the melting point or changes in the phase transition temperature with the pore sizes of micron- and nano-scale pores. After choosing the appropriate probe material to completely saturate a solid porous material, the pore volume corresponding to different nanometre pore sizes in the PSD can be described by the liquid volume in the pores and can then be detected by the phase transition of the probe material in the porous media. The Gibbs–Thomson equation is shown below (Jackson and McKenna, 1990; Strange, 1994)

Δ T_{m} = T_{m} {-T}_{m} (x) = \frac{K_{GT}}{x} = \frac{4 σ_{s l} T_{m}}{x Δ H_{f} ρ_{s}} \cos φ

(1)

ΔT_m – Melting point depression, K

T_m – Melting point of crystals

T_m(x) – Melting point of pores (with a diameter of x)

K_GT – Gibbs–Thomson constant, Knm

x – Pore diameter, nm

σ_sl – Surface energy at the crystal-liquid interface

ΔH_f – Bulk enthalpy of fusion (per gram of material)

ρ_s – Density of the solid

φ – Contact angle between the liquid and the liquid wall

As shown above, K_GT is an essential calibration constant in NMRC that is determined empirically (Strange et al., 1993). This value can be obtained from samples with known pore dimensions that are filled with absorbate. According to the Gibbs–Thomson equation, the depression of the melting point will be larger (i.e., the melting point will be reduced) with increasingly smaller pore sizes. As the temperature rises, the probe material in the micropores and macropores will gradually melt and the NMR signal intensity will become stronger corresponding to the liquid content. During experiments, the NMR signal intensity I, which corresponds to changes in the temperature T, can be observed (Strange et al., 1993).

V(x) is the pore volume corresponding to x, and the PSD dV/dx is shown as follows

\frac{d V}{d x} = \frac{d V}{d T_{m} (x)} \times \frac{d T_{m} (x)}{d x}

(2)

From the Gibbs–Thomson equation, equation (2) becomes

\frac{d V}{d x} = \frac{K_{GT}}{x^{2}} \times \frac{d V}{d T_{m} (x)}

(3)

The quantity of V(x) replaced by the liquid volume is directly proportional to I, and thus, dV/dx can be obtained from the slope of the original curve of I against T. Subsequently, K_GT is the only constant needed to convert T into x (Strange et al., 1993)

\frac{d V}{d x} \propto \frac{d I}{d x} = \frac{K_{GT}}{x^{2}} \times \frac{d I}{d T_{m} (x)}

(4)

Samples and techniques

Sample preparation

First, 11.5 nm pore diameter silica-alumina samples (Micromeritics Instrument Corp., USA) and controlled pore glasses (CPG; SIGMA corp., USA) of different pore diameters (24 nm, 38 nm and 50 nm) were used as standard porous samples. Deionized and distilled water, cyclohexane (Nanjing Chemical Reagent Co. Ltd, China) and octamethylcyclotetrasiloxane (OMCTS; SIGMA Corp., USA) were used as three different types of probe liquids for testing. The properties of the standard porous samples are listed in Table 1. Calcium chloride hexahydrate (CaCl₂•6H₂O; SIGMA Corp., USA) was also used for the first time as a probe material for the NMRC testing of unconventional reservoir samples.

Table 1.

Properties of the four different standard materials.

Label	x (nm)	Mesh size	Pore dist.	Pore volume (cc/g)	Surface area (m²/g)
SA115^a	11.5	Particle	±13%	0.99	216
PG240^b	24	120–200	±4.9%	0.96	88.1
PG350^b	38	200–400	±9.0%	1.17	64.7
PG500^b	50	120–200	±2.9%	0.94	51.4

^aProvided by Micromeritics Instrument Corp.; the samples are cylindrical particles.

^bProvided by SIGMA Corp.

Second, real geological samples were used for the experiment. Sichuan Basin has become one of the most important basins for unconventional oil and gas exploration in China. According to the results of PetroChina’s latest oil and gas resources evaluation, the basin contains tight gas of 3.98 × 10¹² m³ and 21.73 × 10¹² m³ shale gas under 4500 m. The tight sandstone samples were acquired from the Upper Triassic Xujiahe Formation in the Guang'an gas field. The formation of this gas field has proven reserves of 1.356 × 10¹¹ m³ (Ma, 2017). The depths of the well core samples, which are labelled GA002 and YING21, are greater than 2 km. The average porosity and permeability of the tight sandstone of the Triassic Xujiahe Formation are 4.2% and 0.35 × 10^–3 µm², respectively (Zou et al., 2012). The Lower Silurian Longmaxi marine shale has been considered one of the most important target plays of shale gas exploration in Sichuan Basin due to its good geological properties and accumulation conditions, with a workable favourable area of 2 × 10⁴ km² and resources of 10 × 10¹² m³ (Ma, 2017). The shale sample YB12 was acquired from an outcrop of the Silurian Longmaxi Formation, Yibin, Sichuan Province (Figure 1).

Figure 1.

Locations of geological samples in the Sichuan Basin, China, and the stratigraphic column of the basin (modified from Li et al., 2018).

The CPG samples were directly and compulsively vacuum saturated for 24 h after weighing and were thereafter stationary under normal pressure for 12 h prior to testing. The geological samples were pulverized down to particle sizes ranging from 32 mesh to 100 mesh. Tight sandstones and shale samples (without oil) were weighed after drying, saturated using different types of probe materials, vacuum saturated for 24 h and then kept stationary under normal pressure for 12 h. As the probe coil diameter we used was 10 mm, samples with volumes from 0.5 ml to 1 ml were recommended to freeze the samples sufficiently.

Experimental apparatus and parameters

A 12-010V experimental apparatus (Niumag Corporation Ltd, China) was used as the NMRC pore analyser. A sketch map of the apparatus is shown in Figure 2. The main frequency of the magnets was 12 MHz, and the main magnetic field intensity was 0.3 T. Dry compressed air was used as the temperature medium for the system; to control the temperature of the dry clean air, a cryogenic liquid tank was used for the cold source and thermal resistance for the heat source. The controlled temperature of the freezing and thawing system ranged from −30°C to 40°C, the temperature control precision was ±0.02°C, and the minimal temperature gradient was 0.1°C.

Figure 2.

Sketch map of the experimental apparatus.

Prior to the test, alcohol liquid (melting point: −114.3°C) was used to test the NMR signal intensity at different temperature points from −20°C to 20°C (which is a frequently used temperature range) to calibrate the influences of temperature changes on the receiver coil in the NMR probe and eliminate the impacts of different temperatures on Zeeman energy level splitting in a magnetic field (see Figure 3).

Figure 3.

Calibration of the receiver coil to eliminate the influence of temperature.

A Carr-Purcell-Meiboom-Gill (CPMG) spin-echo sequence of 90°-1/2σ-180°-σ-180° was used to collect NMR signal intensities at different temperature points. As an example, the optimized CPMG parameters used for water were σ = 0.1 ms, P1 = 3.4 µs, P2 = 7 µs, TW = 1500 ms, NECH = 3000, and NS = 128. The parameters of the other probe materials are listed in Table 2. We then used these CPMG parameters to calibrate the NMR signal intensity with the water quality at the melting point temperature (see Figure 4). The ratio of the signal intensity to the water quality was 234,465.19, and the relationship between the volume and signal intensity can be obtained if the density of the water is known.

Table 2.

NMRC testing parameters for the four probe materials.

Probe material	Pulse separation time σ (ms)	90° pulse width P1 (μs)	180° pulse width P2 (μs)	Time wait (ms)	Echo number	Cumulative sampling number
Water	0.1	3.4	7	1500	3000	128
Cyclohexane	10.0	3.4	7	2000	2500	64
OMCTS	0.3	3.4	7	2000	2000	128
CaCl₂•6H₂O	0.5	3.4	7	5000	1800	96

Figure 4.

Linear correlation between the water quality and the NMR signal intensity.

The saturated samples were placed in the apparatus with a heating operation model to avoid freezing hysteresis. We initiated the experiment from a minimum temperature of approximately 243 K, which lasted for 1 h to ensure that the liquid in the samples was completely frozen. Then, each temperature step continued for 12 min for thermal balance. As a result, an I-T curve was obtained that represented the liquid signal intensity with the temperature change.

The ASAP2020 HD88-specific surface area and porosity analyser (Micromeritics Instrument Corp., USA) was employed for the gas sorption measurements. This analyser can provide high-quality surface area, porosity and sorption isotherm data for material analyses that satisfy laboratory analytical requirements.

Results and discussion

Standard materials

An appropriate probe material must meet several criteria. First, to more effectively prepare the samples, a probe material should exhibit good wettability in order to spontaneously absorb into the pores of the medium through capillary action, which will avoid damage to the pore structure as a result of the large pressure that presses the liquid into the pores. Second, a probe material with a large K_GT value is preferable. According to the Gibbs–Thomson equation, identically sized pores will cause a larger depression of the melting point (making the instrument easier to identify); alternatively, an equivalent depression of the melting point will be characterized by a higher upper detection limit of the pore size with a greater K_GT value. Third, the NMR intensities of the solid and liquid phases of the probe material should be remarkably distinct, and un-melted solids in the macropores should not disrupt the signal intensity of liquid in the micropores with rising temperatures. In addition, the melting point temperature of the probe material should satisfy the temperature control range of the instrument. Finally, a non-toxic (i.e., harmless) material is preferable since large batches of samples must be tested during research on unconventional reservoirs (Table 3).

Table 3.

Properties of the four different probe materials.

Probe material	Melting point (K)	K_GT (Knm)	Liquid density (g/cm³)	Enthalpy of fusion (kJ/mol)	Molecular size (nm)
Water	273.1	5057.3^a	0.997	6.01	0.4
Cyclohexane	279.5	178^b190.1^c182.5^d	0.779	2.63	0.67
OMCTS	290.5	160^e	0.955	19.7	0.9–1.0
CaCl₂•6H₂O	303.15		1.352	37.24	–

^aWebber et al. (2001).

^bGallegos et al. (1987).

^cDore et al. (2004).

^dRen et al. (2003).

^eVargas-Florencia et al. (2007).

Water has been the most common probe material for NMRC experiments due to its ubiquity, its clear difference in signal intensity between liquid and solid phases and the ease with which it is absorbed into hydrophilic pores. The K_GT value for water was found to be approximately equal to 50 Knm in numerous previous investigations (Mitchell et al., 2008; Petrov and Furó, 2009; Webber, 2010), and a 500 nm upper limit of the pore diameter can be retrieved from back-calculation with a ± 0.1 K temperature control accuracy. In our work, the first tests of the CPG standard materials (i.e., PG240, PG350, PG500) were limited to water, after which the NMR signal only demonstrated the existence of bulk water. The corresponding PSD was unobtainable despite the performance of vacuum saturation for an additional 48 h and centrifuging the sample to ensure saturation. These processes involve potential significant, specific interactions with coated hydrophobic silica surfaces that are different from uncoated CPG materials used in the past. Sample SA115 showed a different wettability to that of CPG on the pore surfaces, and a corresponding peak value appeared between 10 nm and 12 nm after being vacuum saturated for 24 h (see Figure 5).

Figure 5.

PSD of sample SA115 with water as the probe material.

Cyclohexane is an organic compound that has been applied in NMRC experiments for many years due to its appropriate melting point (6.5°C) and high K_GT value (greater than 150 Knm) (Dore et al., 2004; Gallegos et al., 1987; Ren et al., 2003). We used a mixed sample containing both PG240 and PG500; after two signal increases were observed, another steep signal rise remained when the temperature was close to the melting point (see Figure 6). Our temperature control instrument was fully functional, which guaranteed that the final signal rise did not originate from the bulk cyclohexane but was, instead, a “plastic crystal” signal. Cyclohexane is disadvantageous because it forms a soft plastic crystal phase and exhibits an extended relaxation time when it approaches the melting point. Cyclohexane has also shown metastable states and complex phase diagrams under confinement (Dore et al., 2004) and anomalous diffusion (Aksnes and Gjerdaker, 1999) conditions. In past experiments, a pulse separation time σ of 10 ms was employed in the CPMG spin-echo sequence to suppress the “plastic crystal” signal (Mitchell et al., 2008). In the present study, an increase in σ partially suppressed but did not eliminate the “plastic crystal” signal from the final steep rise with parameters of σ = 20 ms, P1 = 3.4 µs, P2 = 7 µs, TW = 2000 ms, echo number = 2000, and NS = 32. The echo sequence employed herein (different from the abovementioned) was 90°-1/2σ-180°-σ-180° in low-field NMR. As the experiment time will be extended with an increase in σ, the stability of the gas supply will be influenced; although we did not continue to increase the value of σ, this phenomenon seems feasible.

Figure 6.

I-T curve of cyclohexane as the probe material in the CPGs at σ = 10 ms and 20 ms.

OMCTS is commonly used as a probe material due to its potential to yield PSDs for µm-sized pores and its ability to wet both hydrophilic and hydrophobic substances (Vargas-Florencia et al., 2007). No practical report on the application of OMCTS as a probe material in an NMRC experiment has been published since the 2007 study. During our work, mixed samples of PG240/PG500 and PG240/PG350/PG500 were used for testing. The former showed two steep signal increases, and the bulk OMCTS signal appeared at the appropriate melting point of 17.4°C. The latter sample did not seem to provide equally differentiable signals to those of the former due to the less distinguishable range of pore sizes; however, through a comparison with the PG240/PG500 mixed sample, information regarding the obscured signal curve can still be obtained (Figure 7). We can fit the data based on the reciprocals of the nominal pore sizes (11.5 nm, 24 nm, 38 nm and 50 nm) relative to the depression of the melting point to find the K_GT value. The middle temperature value corresponding to the steep curve of this relationship was treated as the melting point within the pores. An experimental value of K_GT=149.3 Knm was obtained through this linear regression (Figure 8). This value was subsequently used for the testing of the geological samples.

Figure 7.

I-T curve of OMCTS as the probe material for two types of mixed CPG samples.

Figure 8.

Relationship between the nominal inverse pore size and the melting point depression using OMCTS as the probe material for the standard materials.

Most minerals hosted within tight sandstones (e.g., quartz, calcite, mica and clay minerals) are hydrophilic. A probe material with a high K_GT value relative to water and a good wettability on a hydrophilic surface is preferable for these minerals; meanwhile, the probe cannot cause damage to the pore structure during phase changes, as this type of modification must be neglected during our research (see section Unconventional reservoir samples). Inorganic salt hydrate CaCl₂•6H₂O, which has an appropriate melting point (29.9°C), was used for testing in this study. Salt hydrates appear to shrink upon freezing; this feature, which is in contrast to the behaviour of water, will help avoid volume expansion that would otherwise damage the delicate porous structure (Vargas-Florencia et al., 2006). During the experiment, a temperature range of 31°C–32°C was used to melt a CaCl₂•6H₂O crystal into a liquid, which was subsequently used to saturate the samples as excessive heat may cause CaCl₂•6H₂O to forfeit water molecules. We originally considered that the NMR signal might be disrupted by the weak paramagnetism of the calcium ions, but the results showed that a low-field NMR system can overcome this weak influence. A value of K_GT=184 Knm was matched to sample SA115, and Figure 9 shows the I-T curve for sample SA115. Of course, additional standard porous materials must be tested to verify this value.

Figure 9.

I-T curve of CaCl₂•6H₂O as the probe material for sample SA115 at different temperatures.

Unconventional reservoir samples

In the final PSD results, we omitted the outlying final point due to its large span and adopted only data for the pore sizes below 1 µm for comparison. A density functional theory model was used in the gas sorption experiment to enable a broader, more effective characterization. The results from the testing of the tight sandstones are shown in Figure 10(a) to (d). Larger pore volumes were obtained from the NMRC method corresponding to different pore sizes than from the gas sorption experiment. This difference was most likely the result of pore shrinking/collapsing induced by sample drying during the N₂ sorption experiment (Zhao et al., 2017). The peaks of the three different probe materials were all located in the range from 100 nm to 200 nm. Differences were detected among the results from using the different probe materials; for instance, the pore volume obtained using water as the probe material was greater than that retrieved using either OMCTS or CaCl₂•6H₂O. The pore volumes always presented an increasing trend with increasing pore sizes using water as the probe material; meanwhile, complete and distinct peak values in the PSDs were observed upon using OMCTS and CaCl₂•6H₂O. The OMCTS and N₂ sorption experiments exhibited many more peaks, and the results from these two tests correlate well with regard to the trend of the curves. Compared with the relatively monotonically increasing trend with water, the PSDs using OMCTS and CaCl₂•6H₂O are more reasonable and closer to geological reality.

Figure 10.

The PSDs and cumulative pore volumes of the different NMRC probe materials used in the characterization of the unconventional reservoir samples.

The tests of the shale samples revealed different results from those of the tight sandstones. The PSD characterizations using OMCTS and CaCl₂•6H₂O were relatively consistent between the two lithologies; however, water did not provide the same results for the tight sandstones, although the total pore volume was still greater than those obtained from using OMCTS and CaCl₂•6H₂O (Figure 10(f)). The pore volume, which corresponds to different pore sizes, was not always as high as those using the other two types of probe materials in the tight sandstones. Water revealed a greater pore volume for pore sizes below approximately 10 nm (Figure 10(e)), while its curve was between those of OMCTS and CaCl₂•6H₂O for pore sizes between 7 nm and 20 nm, and its ability to characterize the pore volume decreased for pore sizes larger than 20 nm. Compared with the tight sandstones, the shale samples contained more organic matter and clay minerals. A large proportion of the nano- and micro-scale pores were hosted in these spaces that water molecules could not easily enter due to the hydrophobic nature and weak wettability of organic kerogen surfaces. This condition led to a smaller pore volume for pore sizes larger than 10 nm. In contrast, when water encountered smectite and illite-smectite mixed layers, the volume of the clay minerals expanded, which led to a decrease in the pore size, but the pore size increased when the water froze within the pores. The dual function of the interaction of water with minerals in the shales resulted in smaller pore volumes relative to the other two probe materials for a corresponding pore size. This effect is different from the testing results for tight sandstones.

Water exhibited larger pore volumes during the testing of the unconventional reservoir samples. This result does not mean, however, that water is more effective than the other types of probe materials; rather, we attribute this difference to the reconstruction of the pore structure upon the freezing of water. Based on the first test, we conducted a second experiment on the geological samples with water as the probe material and contrasted the results of the first test with two freeze-melting tests. Figure 11(a) to (d) shows that the water volumes in the pores became larger with increases in the pore size during the second test. For the tight sandstones, this occurred for pore sizes 10 nm and larger (see Figure 11(a) and (b)), while this reconstruction within the shale samples was not obvious until the pore sizes exceeded 100 nm (see Figure 11(c) and (d)). This finding also demonstrates that only minor quantities of water molecules were able to enter the pores with sizes from 10 nm to 100 nm in addition to smaller pore volumes than the characterizations of sandstone using the other types of probe materials in Figure 10(e). The effects of water freezing on the pore size reconstruction during the second test was more distinct for pore sizes that are larger than 100 nm both within tight sandstones and shale.

Figure 11.

Comparison of the first test with the two freeze-melt tests in the unconventional reservoir samples with water as the NMRC probe material.

Comparison between standard materials and the geological samples

Both standard materials and unconventional reservoir samples were investigated using the NMRC method in our research. The properties of the pore structure in the standard materials are approximately the same, i.e., the same material, pore size, pore shape and surface texture (Figure 12(a)). Thus, the standard material with a single pore size has a uniform K_GT value when using the same probe liquid. However, in geological samples, the pores are present in different minerals and among the mineral particles (Figure 12(b)), causing them to exhibit different interfacial shapes and variable interfacial energy. Additionally, the crystalline solid thermodynamic property of a solid-liquid system may change among different geological samples, as can the K_GT value.

Figure 12.

SEM images of standard materials and geological samples. (a) PG 350 with the same composition and similar pore structure; (b) tight sandstone sample with different mineral compositions and pore structure; the boundary separates pore 1 from pore 2 with different pore structures on particles and clay mineral.

An appropriate K_GT value should be selected for application in geological samples based on the value calibrated by the standard samples. According to the principle of the NMRC method, the choice of K_GT value will influence the pore size distribution spectrum and different K_GT values will cause a lateral shift of the PSD curve. Sample YING21 was used as an illustration; the increase in K_GT causes the corresponding pore size to increase (Figure 13) and vice versa.

Figure 13.

The influence of different K_GT values of water on the PSD spectrum of sample YING21.

Conclusions

The PSDs for standard porous materials and unconventional reservoir geological samples have been obtained using the NMRC method with different types of probe liquids. We do not recommend water as a probe material for the testing of unconventional reservoir samples, as the testing of standard CPG materials and the reconstruction of the pore structures were limited by the wettability of water and the freezing of water. Water caused the pore volume to expand to varying extents for different pore sizes within the tight sandstones and exhibited a dual effect on the reconstruction of the shale nanopores. In addition, appropriate experimental parameters must be applied to eliminate the signal influences of complex phases during cyclohexane melting, which results in a longer testing time. We also note that cyclohexane is volatile during sample saturation.

Due to its larger K_GT value, we recommend OMCTS as a probe material to measure pore sizes up to 1 µm in NMRC for the testing of unconventional reservoir samples. OMCTS is able to wet both hydrophilic and hydrophobic surfaces (Vargas-Florencia et al., 2007) and has a smaller impact on the pore structure, enabling more realistic testing results. CaCl₂•6H₂O is also recommended for use in testing, although inorganic salt hydrates will be unstable and forfeit water molecules during melting and recrystallization processes.

Footnotes

Acknowledgements

The authors sincerely thank Zhenmeng Sun of the School of Earth Sciences and Engineering, Nanjing University, for the help with the gas sorption test. We also appreciate the anonymous reviewers for the constructive suggestions, which considerably improved the paper.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was financially supported by National Science and Technology Major Project (grant no. 2016ZX05002006-005).

References

Aksnes DW and Gjerdaker L (1999) NMR line width, relaxation and diffusion studies of cyclohexane confined in porous silica. Journal of Molecular Structure 475: 27–34.

Anovitz

Cole

(2015) Characterization and analysis of porosity and pore structures. Reviews in Mineralogy and Geochemistry 80(1): 61–164.

Bai

Zhu

et al . (2013) Multi-scale method of nano(micro)-CT study on microscopic pore structure of tight sandstone of Yanchang Formation, Ordos Basin. Petroleum Exploration and Development 40(3): 354–358.

Dai

(2012) Tight gas in China and its significance in exploration and exploitation. Petroleum Exploration and Development 39(3): 277–284.

Desbois

Urai

Kukla

et al . (2011) High-resolution 3D fabric and porosity model in a tight gas sandstone reservoir: A new approach to investigate microstructures from mm- to nm-scale combining argon beam cross-sectioning and SEM imaging. Journal of Petroleum Science and Engineering 78(2): 243–257.

Dore

Webber

Strange

et al . (2004) Phase transformations for cyclohexane in mesoporous silicas. Physica A: Statistical Mechanics and Its Applications 333: 10–16.

Firouzi

Rupp

Liu

et al . (2014) Molecular simulation and experimental characterization of the nanoporous structures of coal and gas shale. International Journal of Coal Geology 121: 123–128.

Gallegos

Munn

Smith

et al . (1987) A NMR technique of pore structure: Application to materials with well-defined pore structure. Journal of Colloid Interface Science 119: 127–140.

Gopinathan

Yang

Lowe

et al . (2014) NMR cryoporometry characterisation studies of the relation between drug release profile and pore structural evolution of polymeric nanoparticles. International Journal of Pharmaceutics 469(1): 146–158.

10.

Holditch

(2006) Tight gas sands. Journal of Petroleum Technology 53: 86–93.

11.

Jackson

McKenna

(1990) The melting behavior of organic materials confined in porous solids. The Journal of Chemical Physics 93(12): 9002–9011.

12.

Jeon

J-D

Kim

Kwak

S-Y

(2008) 1H nuclear magnetic resonance (NMR) cryoporometry as a tool to determine the pore size distribution of ultrafiltration membranes. Journal of Membrane Science 309(1–2): 233–238.

13.

Lai

Wang

(2015) Fractal analysis of tight gas sandstones using high-pressure mercury intrusion techniques. Journal of Natural Gas Science and Engineering 24: 185–196.

14.

Jiang

Wang

et al . (2018) Porosity-preserving mechanisms of marine shale in Lower Cambrian of Sichuan Basin, South China. Journal of Natural Gas Science and Engineering 55: 191–205.

15.

(2017) A golden era for natural gas development in the Sichuan Basin. Natural Gas Industry B 4(3): 163–173.

16.

Maex

Baklanov

Shamiryan

et al . (2003) Low dielectric constant materials for microelectronics. Journal of Applied Physics 93(11): 8793–8841.

17.

Mitchell

Webber

Strange

(2008) Nuclear magnetic resonance cryoporometry. Physics Reports 461(1): 1–36.

18.

Pan

Peng

Wan

et al . (2017) Pore structure characteristics of coal-bearing organic shale in Yuzhou coalfield, China using low pressure N₂ adsorption and FESEM methods. Journal of Petroleum Science and Engineering 153: 234–243.

19.

Petrov

Furó

(2009) NMR cryoporometry: Principles, applications and potential. Prog Nucl Magn Reson Spectrosc 54(2): 97–122.

20.

Qian

Yanhui

Shaoqian

et al . (2016) Nuclear magnetic resonance cryoporometry as a tool to measure pore size distribution of shale rock. Chinese Science Bulletin 61(21): 2387–2394.

21.

Ren

Stapf

Kühn

et al . (2003) Molecular mobility in fixed-bed reactors investigated by multiscale NMR techniques. Magnetic Resonance Imaging 21(3–4): 261–268.

22.

Ryu

S-Y

Kim

Jeon

J-D

et al . (2010) Pore size distribution analysis of mesoporous TiO2 spheres by 1H nuclear magnetic resonance (NMR) cryoporometry. The Journal of Physical Chemistry C 114(41): 17440–17445.

23.

Schmitt

Fernandes

Wolf

et al . (2015) Characterization of Brazilian tight gas sandstones relating permeability and Angstrom-to micron-scale pore structures. Journal of Natural Gas Science and Engineering 27: 785–807.

24.

Strange

(1994) Cryoporometry: A new NMR method for characterising porous media. Nondestructive Testing and Evaluation 11(5): 261–271.

25.

Strange

Rahman

Smith

(1993) Characterization of porous solids by NMR. Physical Review Letters 71(21): 3589–3591.

26.

Vargas-Florencia

Petrov

Furo

(2006) Inorganic salt hydrates as cryoporometric probe materials to obtain pore size distribution. The Journal of Physical Chemistry B 110(9): 3867–3870.

27.

Vargas-Florencia

Petrov

Furo

(2007) NMR cryoporometry with octamethylcyclotetrasiloxane as a probe liquid. Accessing large pores. J Colloid Interface Sci 305(2): 280–285.

28.

Webber

JBW

(2010) Studies of nano-structured liquids in confined geometries and at surfaces. Progress in Nuclear Magnetic Resonance Spectroscopy 56(1): 78–93.

29.

Webber

JBW

Corbett

Semple

et al . (2013) An NMR study of porous rock and biochar containing organic material. Microporous and Mesoporous Materials 178: 94–98.

30.

Webber

JBW

Strange

Dore

(2001) An evaluation of NMR cryoporometry, density measurement and neutron scattering methods of pore characterisation. Magnetic Resonance Imaging 19: 395–399.

31.

Yang

et al . (2017) Applying SANS technique to characterize nano-scale pore structure of Longmaxi shale, Sichuan Basin (China). Fuel 197: 91–99.

32.

Zhang

Xiong

et al . (2017) DFT modeling of CO2 and Ar low-pressure adsorption for accurate nanopore structure characterization in organic-rich shales. Fuel 204: 1–11.

33.

Zhao

Ning

Wang

et al . (2015) Petrophysical characterization of tight oil reservoirs using pressure-controlled porosimetry combined with rate-controlled porosimetry. Fuel 154: 233–242.

34.

Zhao

Sun

Liu

et al . (2017) Pore structure characterization of coal by NMR cryoporometry. Fuel 190: 359–369.

35.

Zou

Yang

Tao

et al . (2012) Nano-hydrocarbon and the accumulation in coexisting source and reservoir. Petroleum Exploration and Development 39(1): 15–32.