Investigation of Properties Alternation during Super-Critical CO₂ Injection in Shale

Abstract

The low recovery of oil from tight liquid-rich formations is still a major challenge for a tight reservoir. Thus, supercritical CO₂ flooding was proposed as an immense potential recovery method for production improvement. While up to date, there have been few studies to account for the formation properties’ variation during the CO₂ Enhanced Oil Recovery (EOR) process, especially investigation at the micro-scale. This work conducted a series of measurements to evaluate the rock mechanical change, mineral alteration and the pore structure properties’ variation through the supercritical CO₂ (Sc-CO₂) injection process. Corresponding to the time variation (0 days, 10 days, 20 days, 30 days and 40 days), the rock mechanical properties were analyzed properly through the nano-indentation test, and the mineralogical alterations were quantified through X-ray diffraction (XRD). In addition, pore structures of the samples were measured through the low-temperature N₂ adsorption tests. The results showed that, after Sc-CO₂ injection, Young’s modulus of the samples decreases. The nitrogen adsorption results demonstrated that, after the CO₂ injection, the mesopore volume of the sample would change as well as the specific Brunauer–Emmett–Teller (BET) surface area which could be aroused from the chemical reactions between the CO₂ and some authigenic minerals. XRD analysis results also indicated that mesopore were altered due to the chemical reaction between the injected Sc-CO₂ and the minerals.

1. Introduction

The boom of Bakken formation in North America is limited by the effective recovery method even with the recent advancement of hydraulic fracturing and horizontal drilling technology utilization [1]. Since then, the CO₂ flooding method has been proposed as the potential way to increase the recovery factor for the unconventional shale reservoir in the Bakken formation, especially combined with the advanced water-alternating-gas (CO₂-WAG) and CO₂ huff and puff technologies [2,3,4,5,6,7]. Meanwhile, the injection of the supercritical CO₂ into the shale and coal bed methane reservoir would assist the emission reduction of greenhouse gases [8,9,10]. However, the Sc-CO₂ injection process may change the mechanical properties such as the Young’s modulus and hardness [11,12,13]. Meanwhile, the physical nature of the rock such as pore structure and pore size will alter with CO₂ adsorption [12], let alone the mineral components varied mainly due to the chemical reaction [14]. Therefore, it is critical to investigate and evaluate the properties’ alteration of the Bakken shale which is being treated with Sc-CO₂ injection.

To date, many researches had focused on the investigation of the effect of Sc-CO₂ to the shale. Raza et al. [14] examined the shale from Gippslan Basin, Australia, and characterized the shale sealing efficiency as the caprock. They found that several chemical reactions between the rock-forming minerals and the brine-Sc-CO₂ system occurred. Busch et al. [15] also investigated the shale from Australia and stated that the geophysical properties, such as the shale porosity, permeability and diffusion nature were significantly affected by the carbonate dissolution behavior by Sc-CO₂ adsorption. Qiao et al. [16] conducted several lab tests to investigate the mechanical properties alteration of a low-clay shale which acquired from Sichuan Basin, China. They illustrated that both the subcritical and supercritical CO₂ soaking processes can decrease the rock strength while enhancing the shale ductility. Lahann et al. [17] investigated the variation in pore structure of the shale treated with Sc-CO₂. They identified that the temperature will have the positive effect on the specific surface area increment of shale under the Sc-CO₂ environment. Xiang et al. [18] studied the structural and chemical properties of the shale by conducting four types of instrumental analysis after the Sc-CO₂ soaking procedure. They believed that the Sc-CO₂ treatment reduced the shale strength and altered the mineral content of the shale. Jiang et al. [19] demonstrated the specific surface area and porosity of the shale samples were changed with the Sc-CO₂ treatment, which relies heavily on the variation of pore structure and permeability caused by the Sc-CO₂ adsorption.

To the best of our knowledge, research on the corrosion effect in Middle Bakken shale by Sc-CO₂ injection is very scarce, and the pore structure and strength of shale under Sc-CO₂ treatment is not well understood. Therefore, the main objective of this research was to assess mechanical, physical and mineralogical alteration of Middle Bakken shale samples exposed to Sc-CO₂ under various adsorption durations. Instead of the macroscale tests that most researchers applied to analyze the mechanical property changes during the Sc-CO₂ adsorption, in this study, the advanced nano-indentation test was applied to test the Young’s Modulus. The low-temperature N₂ adsorption test was conducted to analysis the pore structure and pore size properties. X-ray diffraction (XRD) was utilized to investigate the mineral content of the samples. This study is significant for the CO₂ storage and migration after being injected into reservoir. Chemical reactions among CO₂, water, and minerals change the pore systems and thus alter rock petrophysical and geomechanical properties, which are not only crucial to the storage capacity of reservoirs but also are vital to the rocks’ ability to trap CO₂ over a long geological time. The study can provide a guideline for storage sites or reservoir selection and can be a reference to predict the evolution of trap capacity. Therefore, necessary measures can be taken to limit or alleviate possible CO₂ seepage.

2. Materials and Methods

2.1. Sample Preparation

The main production unit in Bakken formation was location in the Middle Bakken layer. In this study, two core plugs with different depth (9848 ft and 9850 ft) (as received) were collected from the Middle Bakken formation in order to conduct the primary supercritical CO₂ flooding test. The core samples are from the well located in the Sanish field, North Dakota. Those samples we collected are in the low porosity and permeability range. The porosity values of Sample 1 and Sample 2 are 6.3% and 5.1%, respectively, while the permeability of these two samples was less than 0.01 mD before the test. The schematic diagram of the Sc-CO₂ flooding system is shown in Figure 1. The core plug covered with sleeve was placed into the core holder, then the syringeESCO pump supplied the pressure source to maintain the constant injection pressure (1500 psi) through the core plug from upstream to the downstream. The N₂ was applied to maintain the function of the back-pressure regulator (BPR), which is control the back pressure above capillary pressure (Pc). The operating temperature was set to 40 °C in order to assure that the CO₂ passing through core holder was in the super-critical phase when the operating pressure was maintained at 1500 psi. The different injection durations were set as 0 (Before the CO₂ injection), 10, 20, 30, and 40 days. In each time slot, some small chips were sliced from the Sc-CO₂ treated core plug for the further analysis.

2.2. Properties Analysis Test

Pieces of the core chips were pretreated for the following properties analysis. The nano-indentation test was run to study the rock mechanical properties alteration. In this paper, we applied the TI-700 Ubi1 Nano-mechanical indenter (Hysitron, INC., Minneapolis, MN, USA, 2016) equipped with the fused quartz calibration plaque to conduct the nano-indentation tests. To prepare the sample specimens for the nano-indentation test, the chosen chips were immersed into the resin liquid until the resin get solidified under vacuum conditions. The smoothness of the specimens’ surface is vital for obtaining the accurate results. Since then, the sample surface was rubbed by different grit size sandpaper, then followed with polishing the rough surface by diamond polisher under different grain sizes (6, 3, 1 microns). After being polished, the samples are dried for a few hours. Before the nanoindentation, the air gun was applied to clean the sample surface. Based on our previous experiments, after following these procedures, the surface roughness of the sample surface will be small which can be applied for the nanoindentation. Figure 2b represents the pretreated sample specimens ready for the Nano-indentation test.

The X-ray diffraction test was used to analyze the composition deformation of the Middle Bakken sample. Those series of samples analysis under different injection times were performed by using the Rigaku intelligent X-ray diffraction (XRD) system with the copper X-ray source. The samples were crushed by the McCrone micronising mill which can rapidly reduce samples size by a unique vibratory grinding action. The sample particles within the grinding vessel moves with respect to their neighbors to produce linear contact blows and planar shearing in the vessel. The samples were milled in to the powder-like status to pass through the 80-mesh screen for the future XRD analysis. Figure 2d presents the prepared sample powder after the pretreatments.

The evaluation of surface area, average pore size, pore volume, and pore size distribution of the analytic samples of powders and other porous solids can be characterized through studying gas adsorption and desorption behavior. Thus, the low-temperature N₂ adsorption tests were conducted and the Quantachrome Autosorb-IQ used for this purpose.

Table 1. Parameters.

Weight of sample taken (g)	0.1
Adsorbate	N2
Bath temperature (K)	77.00
Outgas temperature (°C)	120
Outgas time (h)	10
Relative pressure (P/Po)	0.01–0.99

illustrates the analysis parameters through the low-temperature N₂ adsorption tests. During the test, samples were degassed at a relatively high temperature that can efficiently degas and evaporate the moisture from the samples and will not cause structural change. The evaporates were drawn out by a vacuum pump and condensed in a cold trap tube immersed in liquid nitrogen. A sample can be considered ready for analysis when the sample passes the degas test of no more than 10 microns Hg per minute. Figure 2f shows the sample powder ready for low-temperature N₂ adsorption tests.

3. Results and Discussion

3.1. Mechanical Property Deformation

The mechanical properties alteration during the variable Sc-CO₂ injection times is crucial for the formation of the evaluation process. Since that, we applied a most advanced equipment, the nano-indenter, to measure the Young’s modulus of the samples, which are the main mechanical properties for the rock. Nanoindentation test is considered as a unique technique whereby a sharp indenter tip is used to touch the smooth surface of the sample with a certain mode, then the penetration depth will be recorded along with the applied load [20]. Figure 3a illustrates the typical curve of the indentation test. Generally, the curve consists of three stages, which are the initial loading stage, medium-term holding stage, and following the unloading stage. For the initial loading stage, since the penetration depth get increased with the applied load increases it can be regarded as the association of elastic and plastic deformation; whereas at the unloading stage, it was assumed that the elastic deformation is the only recoverable parameter, which was applied to conduct the following calculation procedure of the rock mechanical properties. Figure 3b identified the standard unloading process of the nano-indentation test. In our test, the peak force that we applied is 6000 uN. The loading time, holding time and unloading time will be 5 s, 5 s, 5 s, respectively.

Figure 4 presents the effect of the Sc-CO₂ injection time on the samples’ Young’s modulus value. It can be observed that the Young’s modulus tends to decrease gradually with the injection time increasing. The original sample from the two different depths, 9848 ft and 9850 ft, had the highest Young’s modulus of 75.91 GPa and 68.37 GPa, respectively. When the samples were flooded in the Sc-CO₂ for 10 days, 20 days, 30 days and 40 days, for the first sample, the Young’s modulus value reduced 2.54% to 73.98 GPa, 10.082% to 68.26 GPa, 14.379% to 65.00 GPa, and 15.446% to 64.03 GPa, respectively. While for the second sample, the Young’s modulus value reduced 2.42% to 68.37 GPa, 10.274% to 61.35 GPa, 13.693% to 59.00 GPa, and 14.429% to 58.50 GPa, respectively. This may be caused by the Sc-CO₂ adsorption behavior. As the samples adsorb more and more Sc-CO₂ through the pore throat, the pore will get swelled, which contributes to the rock strength decrease. With the Sc-CO₂ flooding time increase, the decline in the strength of the two core samples slacks off, which is mainly due to the ultra-low permeability of the Middle Bakken sample. By the curves interpreting, after 30 days of Sc-CO₂ adsorption, both curves were observed to reach a stable status.

Some researchers obtained the relationship between the Young’s modulus (E) and the adsorption time for the Sc-CO₂ injection to the sample with variable lithological character [16,23]. Based on the previous research, we fitted the data for the two samples with a Boltzmann sigmoid relationship between the Young’s modulus (E in GPa) and the Sc-CO₂ injection time (t in days) by the curving fitting tool of OriginPro:

$$E=65.0457+11.1906/(1+e^{\frac{t-16.0421}{4.5314}}) (\mathrm{Sample} \mathrm{Depth}: 9848 \mathrm{ft})$$

(1)

$$E=57.9459+10.8331/(1+e^{\frac{t-17.7597}{5.2977}}) (\mathrm{Sample} \mathrm{Depth}: 9850 \mathrm{ft})$$

(2)

The R-squares are 0.9934 for the sample at depth of 9848 ft and 0.9968 for the sample at depth of 9850 ft, which shows the models and the experimental data are fitting very well. In general, a universal relationship between the Young’s modulus and Sc-CO₂ injection time is organized as the following equation:

$$E=b+(a-b)/(1+e^{\frac{t-c}{dt}})$$

(3)

where, E is the Young’s modulus, t is the Sc-CO₂ injection time. a, b and c are the constant which can be derived from curve fitting.

3.2. Pore Structure Alteration

The low-temperature N₂ adsorption isotherms of the samples with varied Sc-CO₂ injection time (0 day, 10 days, 30 days) are shown in Figure 5. According to the isotherm classification from the International Union of Pure and Applied Chemistry (IUPAC) [24,25], all the curves obtained are following the Type III Adsorption Isotherm curve. At the beginning under the extremely low P/Po applied, the pores of the both samples present micro-pore filling behavior and micro pore volume is related to the adsorption amount. Then, with the relative pressure(P/Po) building up, the isotherm adsorption curve illustrates that the multi-molecular layer adsorption occurred. At higher P/Po, the fluid is at the liquid phase and thus the capillary condensation exists. Comparing the isotherm curves under different Sc-CO₂ injection times for the two samples separately, we proposed several interpretations of the effect on the pore structure. Firstly, with the Sc-CO₂ flooding time increasing, the adsorption amount trend to increase for the first depth sample while the reverse tendency occurred for the sample at the depth of 9850 ft. Also, the similar isotherm curve for all time slots means that the pore shape did not vary significantly during the CO₂ injection, and the slit pore and cylindrical pore dominate the main pore shapes. While the hysteresis loop which illustrated the meso-pore existence, was divided into two scenarios dramatically. For the samples from the depth at 9848 ft (Figure 5a), the width of the hysteresis loop gets wider with the Sc-CO₂ injection time extension, which means that after the Sc-CO₂ treatment, the presentation of the meso-pore becomes abundant. Meanwhile, the wideness tendency represents that the other pore type shapes intent to the silt pore conversion. For the samples from the depth at 9850 ft (Figure 5b), the hysteresis loop starts to narrow with the Sc-CO₂ treatment time increasing, which reservedly indicated that the percentage of the meso-pore tends to reduce. It is also indicated that the pore shape tends to be converted to the plate core from the other types of the pore shape. The plate-shape pores in the Middle Bakken and the slit-shape pores in the Upper and Lower Bakken is advantageous for the flow of the hydrocarbon due to their excellent openness. Both the plate pore and slit pore are advantageous for the hydrocarbon flow in the porous media due to their excellent openness.

The Brunauer–Emmett–Teller (BET) method was applied to characterize the surface area change during the varied Sc-CO₂ treatment period. The BET equation is:

$$\frac{1}{W\left(\left(\frac{P_o}{P}\right)-1\right)}=\frac{1}{W_{m}C}+\frac{C-1}{W_{m}C}(\frac{P}{P_o})$$

(4)

In order to acquire the surface area, two main parameters were required. The slope, which is,

$$s=\frac{C-1}{W_{m}C}$$

(5)

and the intercept, which is,

$$i=\frac{1}{W_{m}C}$$

(6)

Combining the Equations (5) and (6), we have

$$W_m=\frac{1}{s+i}$$

(7)

Then the BET surface area can be calculated as:

$$S_t=\frac{W_m N A_{cs}}{M}$$

(8)

where:

• N: Avogadro’s number (6.0221415 × 1023 molecules/mol);
• Wm: weight of adsorbate constituting a monolayer of surface coverage;
• Acs: molecular cross section of the adsorbate molecule, (16.2 Å2 for N₂ at 77 K);
• M: molar mass (molecular weight) of the adsorbate N₂ gas (28.0123).

Figure 6 clearly demonstrates the parameters acquired for the surface area calculation. It is obvious that the slope gets increase with the Sc-CO₂ treatment time increasing. Furthermore, the surface of the sample after the CO₂ treatment tends to decrease with the flooding time increasing.

Table 2. Surface area vs. various Sc-CO₂ injection time.

	Sample ID
	9848 ft		9850 ft
Time, Day	Surface Area, m2/g	Deduction Per., %	Surface Area, m2/g	Deduction Per., %
0	5.595	-	6.85	-
10	5.238	6.38%	4.998	27.04%
30	4.055	27.52%	3.865	43.58%

lists the detailed surface area alteration corresponding to the varied Sc-CO₂ injection time. For the test sample from the depth of 9848 ft, the surface area was 5.595 m²/g initially, while decreases to 5.238 m²/g for the 10 days’ treatment, then to 4.055 m²/g as the Sc-CO₂ flooding time up to 30 days. For the test sample from the depth of 9850 ft, the surface area value deducted 27.04% to 4.998 m²/g for 10 days’ treatment, and 43.58% deduction to 3.865 m²/g with the 30 days’ Sc-CO₂ injection. The main reason for the surface area deduction gradually with the Sc-CO₂ flooding time increment may due to the corrosive property of dissolved CO₂. The acidizing behavior of some minerals will connect some micro pore to the meso-pore type.

The pore size distribution (PSD) was analyzed for the samples with different Sc-CO₂ flooding times. In this paper, based on density functional theory (DFT), the adsorption branch of the isotherm curve was chosen for the PSD analysis. Figure 7 represent the PSD vibration corresponding to the Sc-CO₂ flooding time variation. It was observed that the PSD curves of all samples present the multimodal characteristic. Several observations were obtained through the comparison of the PSD curve via time duration.

The Sc-CO₂ injection mainly affect the meso-pore volume rather than the micro and macro pore. The PSD vibration mostly located at the meso-pore zone, where the pore diameter is from 10 nm to 50 nm roughly.

The Sc-CO₂ flooding time variation has a significant effect on the meso-pore size distribution. Figure 7a illustrates that with the Sc-CO₂ injection time increases, the percentage of the meso-pore accrescent, and the porosity was increased from 6.3% to 8.5% roughly. While for the samples from the depth of 9850 ft (Figure 7b), the increment of the injection time demonstrated a negative effect on the meso-pore size distribution, and the porosity dropped slightly from 5.1% to 4%.

In order to prove that the pore structures of the shale samples changed after the injection. Some SEM images were taken before and after the injection. Figure 8 shows some images. We can clear see some pores appeared after the CO₂ injection (Figure 8b,d) compared with the images before the injection (Figure 8a,b)

3.3. Mineralogical Alterations

The whole-rock XRD analysis was applied to detect the composition of the initial and Sc-CO₂ treated samples individually. We assume that the samples hold the similar mineral components since the chips were sliced from the same core plug and then powdered.

Table 3. Mineral content vs. the varied Sc-CO₂ flooding time.

Sample ID	Treatment Time (day)	Mineral Content (%)
		Quartz	Analbite	Calcite	Dolomite	Muscovite	Clay
9850	0	29.21	40.4	24.2	4.9	-	1.29
	10	30.93	39.7	24	4.41	-	0.96
	20	32.11	39.5	23.8	3.96	-	0.63
	30	33.13	39.4	23.7	3.5	-	0.27
9848	0	47.85	13.6	6.5	21	6.5	4.55
	10	49.96	12.4	6.1	20.3	7.2	4.04
	20	52.03	11.3	5.6	19.6	7.7	3.77
	30	54.51	9.9	5.1	18.8	8.3	3.39

presents the constituent alteration with the varied Sc-CO₂ exposure duration. From the list, the major mineral components are quartz, with around 50%. Dolomite and K-Feldspar also account for a large proportion for the samples from the depth at 9848 ft, weighing 21% and 13.6 respectively. While for the samples from the depth at 9850 ft, the following minerals, analbite and calcite, weigh heavily on the mineral constituent. Minor amounts of other clays were observed in the samples. Comparing two different depth samples, we note that the sample at the depth of 9848 ft has the extra muscovite component with a proportion of 6.5% initially.

As shown in Figure 9, the mineral compositions tend to change due to the varied Sc-CO₂ adsorption duration treatment. the concentrations of some certain type of minerals, such as K-Feldspar, calcite, dolomite and clay, start to decrease with the Sc-CO₂ flooding time increasing. While the quartz content gets increased with the duration extension, mainly due to the weigh portion of other minerals decrease and some amount of transformation from other components. Another characteristic phenomenon is that the percentage of muscovite get increase dramatically with the Sc-CO₂ flooding time increasing. It was believed that the generation of muscovite occurred during the solid-liquid contact process. The mineral components alteration was supported by the chemical reaction between the contents and the carbonic acid. When the Sc-CO₂ is dissolved into the water in the pore, the weaken carbonic acid will generate and then start to react with some minerals, then alternate the weight proportion in the sample. A series of chemical reaction between the minerals and carbonic acid were initiated and mainly included as follows:

CO_{2 (aq)} + H₂O ↔ H₂CO₃ ↔ H⁺ + HCO₃⁻

(9)

2NaAlSi₃O₈ + H₂CO₃ → Na₂CO₃ + 2HAlSi₃O₈

(10)

CaCO₃ + H₂CO₃ ↔ Ca²⁺ + 2HCO₃⁻

(11)

Ca_xMg_1−xCO₃ + H₂CO₃ ↔ xCa²⁺ + (1 − x)Mg²⁺ + 2HCO₃⁻

(12)

3K[AlSi₃O₈]_(s) + 2H₂CO₃ + 12H₂O ↔ K⁺ + 2HCO₃⁻ + 6H₄[SiO₄] + KAl₂[OH]₂/AlSi₃O₁₀]_(s)

(13)

4. Conclusions

In this paper, the mechanical, pore structure and mineralogical alternation that occur in the Middle Bakken sample injected with Sc-CO₂ were investigated and analyzed. A series of tests including the nano-indentation test, low-temperature N₂ adsorption test, and whole-rock XRD analysis were conducted to evaluate the properties change of the samples with varied Sc-CO₂ injection duration. Some conclusions were drawn as follows:

The nano-indentation test illustrated that the Young’s modulus of the samples tends to decrease gradually with the injection time incrementally which agrees with the results by Qiao et al. [16]. and Xiang et al. [18]. The Sc-CO₂ adsorption behavior was noted as the main reason for the pore swelling, which contributes to the rock strength decrease. Also, a universal relation between the Young’s modulus and the Sc-CO₂ injection time was proposed.

Low-temperature N₂ adsorption results showed that Sc-CO₂ treatment affects the meso-pore volume efficiently rather than the micro and macro pore, and the PSD results also give evidence of the dominating meso-pore alteration. The surface area of the samples with the varied Sc-CO₂ treatment was observed as tending to deduct with the Sc-CO₂ treatment time increasing.

The whole-rock XRD analysis revealed that the mineral components of the sample except quartz and muscovite were decreasing corresponding to the Sc-CO₂ flooding duration increasing. The chemical reaction between the minerals and carbonic acid explained the mineralogical alteration of the samples which has also been proved by Raza et al. [14].

This fundamental research investigated the effect of the Sc-CO₂ injection process to the shale formation from the aspect of geomechanical variation, geophysical alteration and geochemical vibration, and can provide a guide for CO₂ injection to tight shale formation and the CO₂ storage process.