Formation damage due to asphaltene precipitation during CO2 flooding processes with NMR technique

In order to quantitatively evaluate the pore-scale formation damage of tight sandstones caused by asphaltene precipitation during CO2 flooding, the coreflood tests and Nuclear Magnetic Resonance (NMR) relaxometry measurements have been designed and applied. Five CO2 coreflood tests at immiscible, near-miscible and miscible conditions were conducted and the characteristics of the produced oil and gas were analyzed. For each coreflood test, the T2 spectrum of the core sample was measured and compared before and after CO2 flooding to determine the asphaltene precipitation distribution in pores. It is found that, the solubility and extraction effect of the CO2 plays a more dominant role in the CO2-EOR (Enhanced Oil Recovery) process with higher injection pressure. And, more light components are extracted and recovered by the CO2 and more heavy components including asphaltene are left in the core sample. Thus, the severity of formation damage influenced by asphaltene precipitation increases as the injection pressure increases. In comparison to micro and small pores (0.1–10 ms), the asphaltene precipitation has a greater influence on the medium and large pores (10–1000 ms) due to the sufficient interaction between the CO2 and crude oil in the medium and large pores. Furthermore, the asphaltene precipitation not only causes pore clogging, but also induces rock wettability to alter towards oil-wet direction.


Introduction
CO 2 flooding has been proven to be an effective technique to Enhanced Oil Recovery (EOR) through both laboratory experiments and field application for several decades [1][2][3]. The injected CO 2 could interact with crude oil in the reservoirs, leading to significant effects on EOR. The main mechanisms of CO 2 -EOR technique include oilswelling effect, viscosity reduction, light-hydrocarbon extraction and interfacial tension reduction [4]. However, it has been reported that the interaction between CO 2 and crude oil is the determine factor for asphaltene precipitation. For example, in Midale in Canada, no prior asphaltene problem was encountered until CO 2 injection. Asphaltene precipitation also occurred in other CO 2 floods, such as in Little Creek Field, Mississippi and West Texas [5].
Asphaltene precipitation could cause serious damages to formations [5][6][7]. This is because the precipitated asphaltene will deposit on to the reservoir rocks, which may cause reservoir plugging and wettability alteration [8,9]. To investigate asphaltene precipitation during CO 2 flooding processes, extensive coreflood experimental studies have been conducted. Wang et al. [10] found that the degree of permeability reduction is positively correlated with the percentage of asphaltene precipitated through CO 2 core flooding tests. Cao and Gu [11] demonstrated that less amount of asphaltene remained in the cores in immiscible conditions while more asphaltene precipitation is observed under miscible conditions. Wang et al. [12] found that permeability reduction due to asphaltene precipitation mainly occurs in the middle and tail end of the reservoir in the miscible CO 2 Water-Alternating-Gas (WAG) injection after the continuous CO 2 injection through long coreflood experiments. Moreover, the wettability alteration also occurs due to asphaltene precipitation, which has negative influence on the formation. Amroun and Tiab [13] and Escrochi et al. [14] reported that the asphaltene precipitation was the main platform for wettability alteration and the porous media changed towards strongly oil-wet condition. Uetani [15] reported that the productivity dropped immediately and water cut increased from 2-3% to 10-15% in field ''M'', which was caused by the rock wettability altered form water-wet to oil-wet because of asphaltene precipitation.
Although, the asphaltene precipitation during CO 2 flooding could be determined through the measurement of permeability reduction combined with the asphaltene content of the produced oil [16,17]. The pore-scale distribution of asphaltene deposition has been rarely investigated. Srivastava et al. [18] used X-ray CAT (Computer Aided Tomography)-scanning technique to visualize the asphaltene deposition along the length of the core. Song et al. [19] made a microscopic model to observe the distribution of asphaltene deposition in 2-D porous networks. However, these methods can only qualitatively observe the distribution of asphaltene precipitation, and the influence of asphaltene precipitation on the pores cannot be quantitatively evaluated. In order to quantitatively analyze the distribution of asphaltene precipitation in pores, Wang et al. [10] applied Nuclear Magnetic Resonance (NMR) to scan tight artificial cores before and after CO 2 flooding. Besides determining the distribution of asphaltene precipitation, NMR technique is a powerful tool for non-invasively analyzing the wetting state of rock [20,21]. Shikhov et al. [22] studied the wettability change of sandstone cores over aging time with low-field NMR measurements.
In this study, an analysis method reference to Amott method [23] was proposed to quantitatively evaluate the wettability alteration before and after CO 2 flooding combined with determining the distribution of asphaltene precipitation. First, in order to determine the influence of CO 2 -brine-rock interaction on the core samples, two core samples saturated with kerosene which does not contain asphaltene were applied to conduct CO 2 coreflood experiments. Then, five CO 2 coreflood tests were conducted at different injection pressures. The oil recovery factors, the viscosity and the asphaltene contents of the produced oil were measured during these tests. Then, through comparing the difference in the NMR transverse relaxation time (T 2 ) spectrum for the water-saturated cores before and after CO 2 flooding, the distribution of asphaltene precipitation in the pores and throats of core samples was quantitatively evaluated. At the same time, the wettability alteration condition was evaluated by calculating the condition of oil saturation in pores before and after CO 2 flooding. On these bases, this study can help to improve the system of quantitatively evaluating the distribution of the asphaltene precipitation and wettability alteration in pores and throats.

Experimental section 2.1 Materials
In this study, the Stock Tank Oil (STO) sample was collected from Changqing Oilfield, China. The density and viscosity of the cleaned STO were measured to be 833.4 kg/m 3 and 4.76 mPa s at the atmospheric pressure and 61°C and the MW oil = 229.7 g/mol. The asphaltene content of the cleaned crude oil was measured to be 0.94 wt% (n-pentane insoluble) with the standard ASTM D2007-03 method [24]. The wax content was measured by the Thin-Layer Chromatograph/Flame Ionization Detection (TLC-FID) method and the Saturate, Aromatic, Resin and Asphaltene (SARA) analysis of the crude oil was measured with the standard SATM D4124, as shown in Table 1. The density and viscosity of the kerosene were measured to be 792.0 kg/m 3 and 1.01 mPa s at the atmospheric pressure and 61°C. The Gas Chromatography (GC) compositional analysis of the cleaned crude oil sample and kerosene is given in Table 2.
The reservoir brine sample was collected from the same formation and cleaned. The reservoir brine has the Total Dissolved Solids (TDS) of 30 917.8 mg/L, which was considered to be the water type of calcium chloride. The brine viscosity was measured to be 0.4 mPa s at the atmospheric pressure and 61°C. The purity of the CO 2 used in this study was equal to 99.99% supplied by Beijing Huayuan Gas Chemical Co. Ltd.  Figure 1 shows the schematic diagram of the slim-tube apparatus (CFS-100, Core Lab, Tulsa, OK, USA) for conducting a series of displacement experiments to determine the Minimum Miscibility Pressure (MMP) of the crude oil -CO 2 system in this study. The apparatus was consisted of a displacement pump (260D, ISCO, Lincoln, NE, USA), a stainless steel slim tube packed with silica sands (Shengfa Mining Industry Co. Ltd., China), a back-pressure regulator (Huada, Haian, China) and two pressure transducers to monitor injection pressure and back pressure constantly. A burette was used to collect and measure the produced oil and a gas flow meter to measure the volume of the produced gas.
In this study, the MMP of the CO 2 -crude oil system was determined by six slim-tube tests with different injection pressure at the formation temperature of 61°C. The displacement system was cleaned by petroleum ether and dried by nitrogen several times in preparation of each slim-tube tests. Then, the apparatus was saturated with the crude oil at the reservoir temperature of T res = 61°C with a constant flow rate of 0.2 cm 3 /min and the back pressure should be maintained the desired production pressure in order to prevent the crude oil from degassing. The CO 2 was injected into the slim tube to displace the crude oil with a constant flow rate of 0.1 cm 3 /min at the set injection pressure. The injection and production pressure was continuously monitored and recorded during the entire experiment. The volume of the produced oil and gas was measured at every 0.1 PV of pure CO 2 until 1.2 PV CO 2 was injected. Figure 2 depicts the schematic diagram of the high-pressure coreflood apparatus used for CO 2 coreflood tests. A constant flow pump (260D, ISCO, Lincoln, NE, USA) was applied to displace dead crude oil, brine and CO 2 through the core plug inside a high-pressure stainless steel coreholder (Huada, Haian, China) with the inner diameter and outer diameter of 25 mm and 40 mm. Three high pressure cylinders were applied to store and deliver crude oil, brine and CO 2 , respectively. Another ISCO syringe pump was used to exert the confining pressure which was always kept 2-3 MPa higher than the injection pressure on the core plug. All above mentioned components were placed inside an air bath which was heated by two electronic heat guns. A temperature controller was used to keep the air bath at the reservoir temperature of 61°C. A backpressure regulator (Huada, Haian, China) was used to target the desired production pressure during the coreflood test. A burette was used to collect and measure the produced oil and a gas flow meter to measure the volume of the produced gas.

Coreflood tests
The core samples used in this experiment are tight cores collected from Changqing Oil Field, China. It is noted that the core samples with nearly the same gas permeability pore size distribution are selected, which are from the same formation. The properties of cores are listed in Table 3.
The general procedure for the CO 2 coreflood tests is briefly described as follows: 1. Prior to each test, the core plugs were thoroughly cleaned by using a DeanÀStark extractor (SXT-02, Shanghai Ping Xuan Scientific Instrument Co., Ltd., China) for 20-30 days. After the core plugs were cleaned and dried at 100°C. The gas permeability and porosity were measured with nitrogen (High-Pressure Gas Permeameter/Porosimeter, Temco, Tulsa, OK, USA). 2. The core plug was placed in the high-pressure coreholder and vacuumed for 24 hours. Then the formation brine was injected at the flow rate of 0.2 cm 3 /min to saturate the core plug. Then, the NMR apparatus was used to measure T 2 transverse relaxation time of the core sample under initial water-saturated condition. 3. The core was displaced with the MnCl 2 solution (15 000 mg/L) of 5 PV. And then the saturated core was scanned again by NMR apparatus to make sure the hydrogen signal of the brine eliminated. 4. After that, 3.0 PV of the crude oil was pumped through the core plugs at a constant rate of 0.1 cm 3 /min until no water was produced to achieve the connate water saturation (S wc ) the initial oil saturation (S oi ) at the reservoir temperature of 61°C. The physical properties of core plugs were listed in Table 3. The T 2 spectrum was measured again after the core had been saturated with crude oil. 5. In each test, 2.0 PV CO 2 was pumped into the coreholder to displace the crude oil at the desired injection pressure and reservoir temperature of 61°C. The injection and production pressure was continuously monitored and recorded during the entire test. The cumulative produced oil volume was recorded by a video camera and the cumulative volume of the produced gas was measured and recorded by using the gas flow meter. The produced oil and gas were collected during each coreflood test and the components of the produced oil and gas were analyzed by GC technique. 6. After the CO 2 coreflood test, the core samples, were cleaned by a Soxhlet Extractor with the solvent of petroleum ether which cannot dissolve asphaltene [25] and dried for 12 h at 100°C. The gas permeability core samples were measured by permeameter with nitrogen. 7. The cleaned core sample was conducted the same treatments from step 1 to step 4.

NMR tests
NMR refers to the response of atomic nuclei to magnetic fields. The NMR apparatus (Mini-MR, Niumag, Suzhou, China) used in this study detects the transverse relaxation motion of 1 H of fluids in the pores, which produces a relatively strong signal compared to other elements in earth formations [26]. The magnetic intensity, gradient value control precision and frequency range of the NMR apparatus are 0.5 T, 0.025 T/m, 0.01 MHz and 1-30 MHz, respectively. As for the NMR transverse relaxation time of fluid in the pore is given as [27,28]: T 2S : the surface relaxation time (ms); T 2D : the relaxation time as induce by diffusion in magnetic gradients (ms); T 2B : the bulk relaxation time of the pore-filling fluid (ms).
Because T 2B is much larger than T 2 for fluid in porous media, T 2B is usually neglected. T 2D is reasonably neglected, when the magnetic field is thought to be uniform with a quite small field gradient and echo time is small enough. Furthermore, T 2S is associated with specific surface of a pore. Then, 1 q 2 : the surface relaxation rate (um/ms); S : the interstitial surface area (um 2 ); V : the pore volume (um 3 ).
S/V can be written as a function of the dimensionless shape factor of a pore, F S , and pore radius, r (um), as follows, Combining equations (2) and (3), then, C is considered to be a constant for equations (5) and (6) so the T 2 response is proportional to the pore radius. In our work, 0.1-1 ms of T 2 is defined as micro pores, 1-10 ms defined as small pores, 10-100 ms defined as medium pores and 100-1000 ms as large pores.
3 Results and discussion

Experimental identification of CO 2 /brine/rock interactions with NMR technique
In the CO 2 flooding process, the solid precipitation was partly generated due to CO 2 /bine/rock reactions. The solid precipitation and clay particles would migrate in the pore and possibly cause a blockage in the pore throat [29]. So that, the CO 2 /bine/rock reactions would induced permeability reduction of the cores [30]. Nevertheless the studies on the permeability reduction due to CO 2 /bine/rock reactions are basically about aquifers. It is necessary to investigate the influence of the CO 2 /bine/rock reactions on permeability reduction of oil reservoir formation before the experimental study on the impact of asphaltene precipitation on permeability. Therefore, instead the crude oil, the kerosene without asphaltene was used to conduct the coreflood experiments first. As shown in Figure 3, the T 2 spectrum for the initial water-saturated cores 1-1 and 1-2, and the T 2 spectrum for the water-saturated cores after CO 2 flooding, were measured. It can been seen from Figures 3a and 3b that the T 2 spectrum for the water-saturated core after CO 2 flooding did not deviate from the T 2 spectrum for the initial watersaturated cores. Thus, in the case of experimental error, the distribution of pores was considered unchanged. The  k: gas porosity of the core plugs; u: absolutely gas permeability of the core plugs; S oi : initial oil saturation; S wc : initial connate water saturation. experimental result is different from similar experimental results in other literature that the CO 2 /bine/rock reactions occur and change the pore structure in the cores during CO 2 coreflood tests [31,32]. That is because that the reaction time of these coreflood tests is too short in comparison with experiments in other literature. Therefore, the CO 2 /bine/rock reactions are considered to have no influence on the pore distribution of the sandstone cores saturated with oil during the CO 2 flooding process.

CO 2 -oil MMP
In this study, the slim-tube tests at six different injection pressures under a constant reservoir temperature of 61°C were conducted to determine the MMP of the crude oil sample. The measured Oil Recovery Factors (ORF) versus Pore Volume (PV) of injected were illustrated in Figure 4. As expected, the ORF increased with the injection of CO 2 at each injection pressure, and the growth rate of the ORFs decreased rapidly after 0.6 PV of injected CO 2 because of the CO 2 breakthrough. There was no more oil obtained in each test at 1.  Figure 5 to determine the MMP. Figure 5 shows that the first three points and the last three points are linear respectively. The intersection point of two fitting curves is regarded as the MMP of the CO 2 -crude oil system measured by slim-tube test, which is 17.02 MPa.   under the reservoir temperature of 61°C. As expected, the ORF increased with the injection of the CO 2 until no more oil was produced at 2.0 PV of injected CO 2 . And, the ORFs increased faster before the injected CO 2 of 0.30 PV, 0.55 PV, 0.85 PV, 1.00 PV and 1.10 PV corresponding to the injection pressure of P inj = 9.1 MPa, 13.5 MPa, 16.2 MPa, 19.5 MPa and 22.1 MPa, respectively. More specifically, the ORF of lower injection pressure was greater than that of higher injection pressure in the initial period of process. Because a less portion of injected CO 2 was dissolved into the light crude oil at a lower injection pressure due to the lower solubility and a larger portion of the injected CO 2 played a major role in displacement at the same injection rate. With the growth of the injection pressure, the ultimate ORF at the terminal 2.0 PV increased due to the stronger interaction ability between the CO 2 and the crude oil. Figure 7 shows the oil recovery, asphaltene content and viscosity of the produced oil for five coreflood tests at different injection pressures. The oil recovery factor increased significantly with the increasing injection pressure until reaching the MMP = 17.02 MPa. This is because the viscosity of the crude oil and the InterFacial Tension (IFT) between the crude oil and CO 2 decreased at higher injection pressure [11]. In addition, it can been seen from Figure 7 that the viscosity of the produced oil decreased from 10.18 mPa s to 3.72 mPa s when the injection pressure increased from 9.1 MPa to 22.1 MPa. And, the asphaltene content of the produced oil decreased from 0.78 wt% to 0.58 wt% as the injection pressure increasing from 9.1 MPa to 16.2 MPa, while the asphaltent content barely changed when the injection was higher than the MMP (Fig. 7). The aphaltene content of the original oil is 0.94 wt%, which is always higher than that of the produced oil. This result means there is asphaltene precipitation left in the core during CO 2 coreflood process. The higher is the asphaltene content of the produced oil, the more is asphaltene left in the core sample.

The effect of asphaltene precipitation on permeability
In this study, the percentage of permeability reduction was obtained by comparing the gas permeability of the core before and after CO 2 flooding, as in the following equation: Pr : the permeability reduction percentage of the core sample, %; K gb : the gas permeability of the core sample before CO 2 flooding, mD; K ga : the gas permeability of the core sample after CO 2 flooding, mD. Figure 8 plots the percentage of permeability reduction of the core samples and asphaltene content of the produced oil at different injection pressures. The asphaltene left in the pores could be inferred from the content of the produced oil. If the asphaltene content of the produced oil was higher, it indicated that there was less asphaltene precipitation in the pores of core sample. It can be seen from Figure 8 that the asphaltene content of the produced oil decreased with the increased injection pressure. But, when the injection pressure approached the MMP, the asphaltene content was almost unchanged.
Similarly, the percentage of permeability reduction significantly increased from 2.4% to 7.41% as the injection pressure increased in the immiscible stage from 9.1 MPa to 16.2 MPa. When the injection pressure reached the MMP, the percentage of permeability reduction still increased with the injection pressure increasing, but changed slowly compared with that in the immiscible stage.  The results were consisted with the asphaltene content of the produced oil at different injection pressures. As a result, more asphaltene would precipitate and block the pores with increasing pressure until the injection pressure reached the MMP.

The effect of asphaltene precipitation on pore structure
As mentioned in the coreflood experimental procedure, the T 2 spectrum for the initial water-saturated cores and the T 2 spectrum for the water-saturated cores after CO 2 flooding were measured and compared. Figure 9 illustrates the T 2 spectrum distributions for cores 2-1, 2-3 and 2-5. The NMR spectrum of the core samples was typical bimodal distribution as shown in Figures 3 and 9. It could be seen from Figure 9 that the T 2 spectrum measured for the water-saturated core after CO 2 flooding moved a slightly lower position compared to that measured for the initial watersaturated core. Because the petroleum ether was used to clean the cores after CO 2 flooding and asphaltene cannot dissolve in the petroleum ether [12]. The reduced amplitude of the T 2 spectrum indicated the pores were clogged due to the asphaltene precipitation and deposition, which could not be saturated with water. The initial water saturated in the pores and the water saturated in the pores after CO 2 flooding is defined as S wb and S wa respectively, the severity of formation damage due to asphaltene precipitation could be calculated as follows: D w : the severity of formation damage due to asphaltene precipitation; S wb : the summation of the water saturated in the pores of the core before CO 2 flooding; S wa : the summation of the water saturated in the pores of the core after CO 2 flooding.
The severity of formation damage due to asphaltene precipitation of the cores 2-1, 2-3 and 2-5 was 3.67%, 8.20% and 13.75%, respectively. And the permeability reduction corresponding to the three cores was 2.4%, 7.41%, 8.32%. The severity of formation damage due to asphaltene precipitation increased as the injection pressure  increased. Moreover, the pore distribution influenced by asphaltene precipitation expanded in the miscible stage ( Fig. 9). Because CO 2 could expand sweep area and enter smaller pores to interact with the crude oil at higher injection pressure.
The amplitude variation of the micro (0.1-1 ms) and small pores (1-10 ms) was smaller than the amplitude variation of the medium (10-100 ms) and large pores (100-1000 ms). That indicated that the asphaltene precipitation had a greater influence on the medium and large pores. The interaction between the CO 2 and crude oil in the medium and large pores was sufficient, so that the asphaltene precipitation in the medium and large pores was more serious than micro and small pores (Fig. 9). When the asphaltene precipitated in the tight sandstone reservoirs, the larger particles could block up the pore throat directly [33], while the smaller ones could cause an obstacle or blockage in the pore or pore throat [8]. On the other hand, although part of the micro and small cores were blocked by precipitated asphaltene particles, the medium and large pore radius decreased due to asphaltene precipitation. The medium and large pores transformed into the micro and small pores. Thus, the amplitude of micro and small cores changed little. Table 4 shows the T 2 spectrum distribution which corresponds to pore distribution for the three core samples. The proportion of medium and large pores of the cores after CO 2 flooding decreased while the proportion of micro and small pores increased, compared to the initial water-saturated cores. The pore distribution of tight cores after CO 2 flooding overall changed to the direction of pore radius reduction after CO 2 flooding. Moreover, the proportion changed more greatly at higher injection pressure.

The effect of asphaltene precipitation on wettability
In consistent with the variation of T 2 spectrum in Figure 9 due to the asphaltene precipitation, the T 2 spectrum measured for oil saturated core after CO 2 flooding also deviated slightly lower from that for initial oil-saturated core, as shown in Figure 10. The amount of saturated oil decreased more in medium and large pores (10-1000 ms) than in micro and small pores (0.1-10 ms), as well. However, compared with the difference between the initial watersaturated cores and the water-saturated cores after CO 2 flooding (Fig. 9), the difference between the initial oil saturated in pores and the oil saturated in the pores after CO 2 flooding is smaller (Fig. 10). S ob and S oa respectively stands for the initial oil saturated in the pores and the oil saturated in the pores after CO 2 flooding. The relative variation of the T 2 spectrum in Figure 10 due to asphaltene precipitation could be calculated, as follows: D o : the relative variation of the T 2 spectrum due to asphaltene precipitation; S ob : the summation of the oil saturated in the pores of the core before CO 2 flooding; S oa : the summation of the oil saturated in the pores of the core after CO 2 flooding; I WA : the wettability alteration index of the core before and after CO 2 flooding.
The pores occupied by asphaltene deposition are constant, so that the D o should have been same as the D w , theoretically. However, the D o is less than the D w , which means that some pores could be saturated more oil than water after the CO 2 flooding compared with that before CO 2 flooding, relatively. The wettability alteration was assumed to occur due to asphaltene precipitation. The wettability of the rock after CO 2 flooding altered to the oil-wet direction. The index of the wettability alteration can be represented by I WA . When the I WA is zero, there is no wettability alteration. When the I WA is higher than zero, the wettability changes to oil-wet direction and the larger I WA means stronger oil-wet alteration. The I WA of the core samples after CO 2 flooding at different pressures are presented in Table 5. Therefore, the reduction of water permeability was caused by pore clogging and wettability alteration. If the CO 2 -EOR technique is applied, it is necessary to injection chemical inhibitor into the reservoir to reduce the risk of asphaltene precipitation.

Conclusion
In this paper, five CO 2 coreflood tests were conducted at immiscible, near-miscible and miscible conditions. For each test, the ORF, the viscosity and the asphaltene contents of the produced oil were analyzed. Then, the distribution of asphaltene precipitation in the pores and wettability alteration was quantitatively evaluated. It is found that, the extraction effect of the CO 2 played a more dominant role in the CO 2 -EOR process with higher injection pressure. So that more light components are extracted and recovered by the CO 2 and more heavy components including asphaltene were left in the core at higher injection pressure. And, the asphaltene precipitated in the core had little increase in the CO 2 miscible flooding stage.
The severity of formation damage influenced by asphaltene precipitation increased with the increasing of injection pressure. And, the asphaltene precipitation had a greater influence on the medium and large pores due to the sufficient interaction between the CO 2 and crude oil. Furthermore, the asphaltene precipitation not only caused pore clogging, but also induced rock wettability alteration towards oil-wet direction. If the CO 2 -EOR technique is applied, it is necessary to injection chemical inhibitor into the reservoir to reduce the risk of asphaltene precipitation.  Fig. 10. Comparison of T 2 spectrum for the initial oil-saturated cores and for the oil-saturated cores after CO 2 flooding and cleaning.