As an important kind of unconventional resource, tight oil has been the focus of many researches worldwide^{[1,2,3,4,5,6,7,8,9,10,11,12]}. Generally, those reservoirs with air permeability less than 1.0×10^-3 μm² or formation permeability less than 0.1×10^-3 μm² are considered as tight reservoirs^[11,13]. The diameter of pore throat in tight reservoirs is usually nano-scale, in which oil percolation ability is poor with less phase segregation and is prone to be adsorbed^[14]. Therefore, it is of theoretical and practical significance to investigate tight oil mobility. Currently, the research on the oil mobility in tight reservoirs is at early stage and there are few studies on the pore structure and movable oil quantification, which restricts objective assessment of tight oil and correct understanding of recovery factor.

The study of fluid mobility in tight reservoirs in China is mainly based on high pressure mercury injection analysis^[15] or NMR (nuclear magnetic resonance) centrifugation analysis for oil driving water and water driving oil processes^[16,17], whereas at overseas, based on FESEM (field emission scanning electron microscope) combined with argon icon milling^[18,19], FIB-SEM (focus ion beam scanning electron microscope)^[20] and, Micro-CT or Nano-CT^[21,22], researchers established digital rock models to study pore network and fluid percolation mechanisms in micro space^[23,24].Due to low permeability and small pore throat, conventional water driving oil experiment may be ineffective for nano pores^[25], while digital rock models aren’t surely representative. In this study, taking Gaotaizi and Fuyu tight oil reservoirs in northern Songliao Basin as examples, based on analyses of NMR and high pressure mercury injection, for the first time, supercritical carbon dioxide displacement and extraction is adopted to conduct oil mobility experiments on tight reservoir samples of different lithologies and oil content grades. The oil mobility in nano pores with complex pore throat network and low oil content is quantified and oil saturation distribution and movable oil rate at different tight reservoirs are characterized, based on which the controlling factors of oil mobility in tight reservoirs in northern Songliao Basin are discussed and results are supposed to provide theoretical basis for tight oil exploration and exploitation.

Large-scale lacustrine delta systems are developed in the Upper Cretaceous strata of Songliao Basin deposits, laying the foundation for development of lacustrine tight oil reservoirs^[26]. Studies on sedimentary facies and organic geochemical characteristics show that the Upper Cretaceous Qingshankou and Nenjiang formations have experienced two periods of lake “expansion-shrinkage”, giving rise to multiple source-reservoir-cap assemblages and oil containing layers. The Qingshankou 1^st and 2^nd members deposited during large-scale lacustrine transgression have widespread black shale with high organic content and oil potential; the source rock has high organic matter abundance, moderate maturity and large hydrocarbon generation potential, laying foundation for in-source tight oil (Gaotaizi oil layer) and below-source tight oil (Fuyu oil layer) in Qingshankou Formation. Controlled by source rocks and reservoirs, the Gaotaizi and Fuyu oil layers are mainly distributed in the Qijia-Gulong depression and Sanzhao depression on the plane (Fig. 1).

Tight oil reservoirs in the Fuyu oil layers are in various channel sand bodies deposited in large-scale river-delta system. The sand bodies are small in scale individually, and scattered vertically and discontinuous laterally, but multiple periods of channel sand bodies stack and overlap with each other into vast area of tight reservoirs. Spatially, the tight reservoirs are characterized by sandy layers interbedded with mudstone layers, like a hamburger^[27]. The thickness of oil layers is usually in the range of 4-12 m with the maximum of a single oil layer at 10 m. The reservoir porosity is in the range of 5%-12% with the average at 9.5% and the permeability is (0.03-1.00)×10^-3 μm² with the average at 0.2×10^-3 μm². The burial depth generally ranges 1 700-2 500 m.

Tight oil reservoirs in the Gaotaizi oil layer are mainly distributed in inner delta-front, outer delta-front and shore-shallow lacustrine deposits^[28]. The reservoirs are thin in single layer but wide in lateral distribution like sandwich. The oil layers are 10-30 m thick combined with 3.5 m thick at maximum individually. The reservoirs have a porosity of 4%-12%, 8.3% on average, a permeability of (0.01-0.50)×10^-3 μm², 0.1×10^-3 μm² on average, and a burial depth of 1 900-2 400 m generally.

A total of 29 tight reservoir samples from Gaotaizi oil layer and the Fuyu oil layer of 7 wells in northern Songliao Basin were selected (Fig. 1), most of them are fine sandstone, siltstone, and mud-bearing sandstone to argillaceous sandstone. Besides tests of porosity, permeability, high pressure mercury injection and starting pressure, NMR and supercritical carbon dioxide displacement and extraction experiments were also conducted.

Plug samples with the diameter at 25 mm were drilled and then were extracted using Soxhlet extractor to remove residual oil until the fluorescence grade is under 3. The extracted samples were dried to constant weight and saturated with standard saline water under high pressure (20 MPa for at least 24 h)^[29] (the standard saline water was prepared by the formula of NaCL:CaCL₂:MgCL₂·6H₂O with the weight ratio of 7.0:0.6: 0.4). The saturation of 100% was determined by the ratio of saturation capacity versus pore volume.

Water saturated samples were centrifuged for 5h by a high- speed centrifuge to remove water with the centrifugal pressure respectively at 0.690 MPa, 1.725 MPa, 3.105 MPa, 4.485 MPa and 5.175 MPa. The centrifuged samples and corresponding waters were analyzed by Manmr-7 NMR instrument according to petroleum industrial test standard^[30]. Relaxation time T₂ spectra of samples with different water saturation were attained.

The centrifuge experiment shows that when the centrifugal pressure is larger than 4.485 MPa, the T₂ spectrum varies little, implying it is the T₂ spectrum of irreducible water. The irreducible water saturation can be evaluated by cumulating saturation components at different relaxation times. The method can avoid the defect that the irreducible water saturation obtained by T₂ cut-off method cannot reflect the irreducible water distribution in pores^[31]. The T₂ spectrum of movable fluid saturation can be obtained by minus the T₂ spectrum of irreducible water saturated sample from the T₂ spectrum of water saturated sample.

According to the method proposed by Li et al.^[32], that is, evaluating conversion factor C for T₂ relaxation time and pore radius using NMR T₂ spectra combined with high pressure mercury injection data, the factor C was determined between 31.9-71.7 and the radius of pore throat was evaluated by multiplication of the factor C and T₂ relaxation time^[33,34]. Therefore, the relationship between pore throat radius and distributions of movable and irreducible fluids was established.

The experiment equipment is the test device for oil occurrence in unconventional reservoirs (YQMV-12) and the technical specifications are: the flow rate of oil injection and displacement is 0.000 01-25 mL/min and the maximum displacement pressure is 105 MPa, and the accuracy of supercritical displacement and extraction is 1 mg/g. The device can test movable oil and retained oil in tight reservoirs with different lithologies and oil contents.

Sample preparation: plug samples were drilled from tight reservoir cores and then extracted to remove oil and saturated with stand formation saline water.

Method for quantitatively injecting oil to core samples saturated with standard formation saline water. The core samples were encapsulated by lead coat and installed in quantification gripper and the ambient pressure was set at 70 MPa. The modelling temperature was in accord with oil layer temperature and based on reservoir depth and geothermal gradient (4 °C/100 m), the temperature was set in the range of 76-89 °C and time at constant temperature was more than 3 h. The oil was injected at the flow rate of 0.01 mL/min. Based on abnormal pressure gradient (1.7 MPa/100 m) of the Qingshankou Formation ever experienced, the maximum injection pressure was set in the range of 35-42 MPa. The injecting oil was 0.11-0.86 mL depending on the lithology and oil content of the samples. The oil-bearing grades (oil containing, oil immersed, oil patch and oil trace) were determined according to oil content in the reservoirs^[35].

Method of supercritical CO₂ displacing movable oil. The movable oil referred to the oil displaced by supercritical CO₂ from core samples, and represented the recoverable oil using current technology. The movable oil rate is the ratio of movable oil to the total oil (sum of movable oil and unmovable oil). The displacement temperature was set in accordance with oil layer temperature at the range of 76-89 °C. The displacing gas flow rate was set at 2 mL/min. The maximum displacing pressure equals the injecting pressure which ranges 35-42 MPa. The displaced fluid weight was monitored online using a burette combined with an electronic balance with the precision of 0.1mg. The weight of movable oil was determined when the oil displaced stopped increase.

Method of supercritical CO₂ extracting retained oil. The retained oil referred to the oil extracted by supercritical CO₂ from core samples after displaced by supercritical CO₂, representing unrecoverable oil with current technology. The core samples after displacing movable oil were crushed and extracted by supercritical CO₂. The extracting kettle was set at temperature of 50 °C and pressure of 20 MPa, and separation kettle was set at 40 °C and 10 MPa. The extracted fluid weight was monitored online using a burette combined with an electronic balance with the precision of 0.1mg. The weight of retained oil was determined when the extracted oil stopped increase.

Tight reservoir samples saturated with standard saline water were centrifuged to remove water under different centrifugal conditions and then were analyzed by NMR to determine fluid saturation distribution (Fig. 2, Table 1). The irreducible water saturation spectrum at different pore throat radii were integrated to get the unmovable fluid saturation (irreducible water saturation), and the difference between the total saturation (100%) and the unmovable fluid saturation is the movable fluid saturation. Experimental data statistics show that under the experimental conditions of 76-89 °C and 35-42 MPa, fine sandstone or siltstone reservoirs with better physical properties have a porosity of 8.5%-12.4%, permeability of (0.12- 0.46)×10^-3 μm², an average pore throat radius of 81-268 nm, radius of pores occupied by movable fluid of 20-11 851 nm with the peak at 163-637 nm, movable fluid saturation of 41%-71%, radius of pores occupied by irreducible fluid of 2-542 nm with the peak at 39-68 nm and an irreducible fluid saturation of 29%-59%. In contrast, mud-bearing and argillaceous siltstone reservoirs with moderate physical properties have a porosity of 6.1%-8.4%, permeability of (0.03-0.09)× 10^-3 μm², an average pore throat radius of 54-56 nm, a radius of pores occupied by movable fluid of 15-2 031 nm with the peak at 96-189 nm, movable fluid saturation of 16%-23%, radius of pores occupied by irreducible fluid of 2-501 nm with the peak at 33-40 nm, and an irreducible fluid saturation of 77%-84%. Argillaceous siltstone reservoirs with poor physical properties have a porosity of 4.8%-5.2%, permeability of 0.02×10^-3 μm², average pore throat radius of 10-35 nm, radius of pores occupied by movable fluid of 15-254 nm with the peak at 28-59 nm, movable fluid saturation of 6%-9%, radius of pores occupied by irreducible fluid of 1-201 nm with the peak at 11-34 nm, and an irreducible fluid saturation of 91%-94%.

It is noted that in tight reservoirs: (1) Pores occupied by movable fluid have a wide range of radius from micrometer to nanometer, indicating that movable fluid can be accumulated in pores of various scales; (2) the pores occupied by irreducible fluid are all nano-scale with a maximum radius of less than 550 nm and the peak of less than 70nm, indicating that in all types of tight reservoirs, nano pores have adsorption and retaining effect on fluid^[11,14].

Oil was injected into tight sandstone samples saturated with standard saline water quantitatively to get experimental samples of different lithologies, oil content grades and oil saturations (the ratio of injected oil volume to effective pore volume). The movable oil weight, retained oil weight and movable oil rate were obtained by supercritical CO₂ displacement and extraction. The results (Table 2) show that, the oil containing, oil immersed and oil trace siltstone reservoir samples with better physical properties have a porosity of 8.9%-12.2%, permeability of (0.12-0.96)×10^-3 μm², an average pore throat radius of 100-204 nm, oil saturation of 9.61%-70.27%, a movable oil content of 1.40-13.27 mg/g, retained oil content of 2.00-9.49 mg/g, and movable oil rate of 41.64%-58.48%. In contrast, the oil containing, oil immersed and oil trace argillaceous siltstone and siltstone reservoirs with moderate physical properties have a porosity of 5.2%-8.0%, permeability of (0.03-0.08)×10^-3 μm², an average pore throat radius of 41-137 nm, oil saturation of 9.80%-52.59%, a movable oil content of 0.02-4.50 mg/g, retained oil content of 2.22-9.15 mg/g and movable oil rate of 0.62%-32.95%. The oil patch/trace calcium-bearing and argillaceous siltstone reservoirs with poor physical properties have a porosity of 3.7%-3.9%, permeability of (0.01-0.02)×10^-3 μm², an average pore throat radius of 17-18 nm, oil saturation of 19.96%-30.59%, a movable oil content of 0, retained oil content of 3.21-5.31 mg/g, and movable oil rate of 0.

The experimental results show that the movable fluid saturation of tight reservoir is closely related to the porosity, permeability, and average pore throat radius of the tight reservoir (Fig. 3). From Fig. 3a-3c, the porosity, permeability, and average pore throat radius when the movable oil rate is over 0 are the corresponding lower limit of them. From Fig. 3 we can see that under the experimental conditions of 76-89 °C and 35-42 MPa the lower limit of porosity, permeability and average pore throat radius at which the movable oil can be produced are 4.4%, 0.015×10^-3 μm², and 21nm respectively. Additionally, the experiments also show: (1) Movable oil rate and oil saturation have no apparent correlation, indicating that oil content is not the key factor controlling oil mobility. For example, siltstone reservoirs in Gaotaizi oil layer of Well Jin 341 (sample No. 15) and Jin 28 (sample No. 16) with better physical properties and larger average pore throat radius, have an oil saturation of less than 25%, but movable oil rate of more than 49%. However, calcium-bearing siltstone from the Fuyu oil layers in Well Gu 616 (sample No. 28) with poorer physical properties and smaller average pore throat radius has an oil saturation of greater than 30% but movable oil rate of 0, indicating that oil mobility in tight oil reservoirs is mainly closely related to pore structure but not affected by oil content strongly. (2) Movable oil rate is positively correlated with and generally smaller than movable fluid saturation. For example, sample No. 28, 29 from Well Gu 616 have a porosity of less than 6%, permeability of less than 0.03×10^-3 μm², and movable oil rate of less than 1%, but the same type of reservoir has movable fluid saturation close to 10% (Table 1). This is because the viscosity of experimental oil (13.9 mPa•s) is higher than that of formation water, and water in tight reservoirs dominated by nano pores and pore throats has higher mobility than oil, indicating that oil viscosity or oil gas ratio have great impact on recovery factor.

NMR and CO₂ displacement and extraction experiment show fluid mobility in tight reservoirs is closely related to reservoir physical properties and pore structure. The relationships between movable oil rate, movable oil saturation, starting pressure, reservoir porosity, permeability and average pore throat radius (Fig. 3), all show a pattern of “three sections”, based on which tight reservoirs are divided to 3 types. Type I tight reservoirs are dominated by fine sandstone and siltstone, and have a movable fluid saturation of more than 40%, movable oil rate of more than 30%, and starting pressure gradient of 0.3-0.6 MPa/m, which vary little with the increase of porosity, permeability and average pore throat radius. The corresponding porosity, permeability and average pore throat radius are 8-12%, (0.1-1.0)×10^-3 μm² and 100-300 nm respectively. Type II tight reservoirs are mostly siltstone, mud- bearing and argillaceous siltstone, and have a movable fluid saturation of 10%-40%, movable oil rate of 5-30%, and starting pressure gradient of 0.6-1.0 MPa/m; with the increase of porosity, permeability, average pore throat radius, the fluid saturation and movable oil rate increase rapidly, whereas the starting pressure gradient decreases significantly. The corresponding reservoir porosity, permeability and average pore throat radius are 5%-8%, (0.03-0.1)×10^-3 μm², and 50-10 nm. Type III tight reservoirs are dominated by argillaceous siltstone and calcium-bearing siltstone and have a movable fluid saturation of generally less than 10%, movable oil rate of less than 5%, and starting pressure gradient of greater than 1.0 MPa/m, which vary little with the increase of porosity, permeability and average pore throat radius. The corresponding reservoir porosity, permeability and average pore throat radius are less than 5%, less than 0.03×10^-3 μm², and less than 50 nm respectively.

Classification standard of tight reservoirs was established based on comprehensive study of reservoir lithology, movable fluid saturation, movable oil rate, starting pressure gradient, pore structure and physical properties (Table 3). Based on current exploration practice, for type I tight reservoirs with small starting pressure gradient, high moveable fluid saturation, high moveable oil rate, and better reservoir physical properties, commercial productivity can be obtained by fracturing in vertical wells. As to type II tight reservoirs with great variation of starting pressure gradient, moveable fluid saturation and moveable oil rate, and poor physical properties, commercial productivity can be obtained by massive volumetric fracturing in horizontal wells. For type III tight reservoirs with high starting pressure, low movable fluid saturation and moveable oil rate, and poor physical properties, carbon dioxide or nitrogen mass volumetric fracturing is to be tested to get commercial productivity.

Sandstone reservoirs in Songliao Basin are in the middle diagenetic stage A at the burial depth of less than 2 000 m, and in the middle diagenetic stage B at the burial depth of greater than 2 000 m^[36]. As the depth increases, the tight reservoir enhances in diagenesis, reduces in storage space, porosity and permeability, and the fluid mobility in the tight reservoir shows staged variations (Fig. 4).

Reservoirs of fine sandstone, siltstone and mud-bearing siltstone at the burial depth of less than 2 000 m are mostly type I, only a few reservoirs of calcium-bearing fine sandstone, and argillaceous sandstone are of type II and type III, implying that the fluid mobility in them is good generally. The movable fluid saturation tests (Table 1) show that, reservoirs of fine sandstone and siltstone less than 2 000 m deep (samples No. 1, 2, 3) are of type I reservoir, which have a fluid saturation of greater than 40%. The sample No. 10 is an exception with a movable fluid saturation of less than 10, representing type Ⅲ. The movable oil rate tests (Table 2) show that the movable oil rates of siltstone reservoirs less than 2 000 m deep have a movable oil rate of greater than 30% generally (sample No.13,14, 17, 18, 19, 21) and are mainly type I. Some samples (sample No. 22, 24, 25, 27, 28, 29) are type II or Ⅲ, with movable fluid saturation of less than 30% due to mud or calcium content. In summary, of tight reservoirs at diagenetic stage A, type I reservoirs with high fluid mobility take the majority. For tight oil exploration and development, reservoirs with shallower depth and lower diagenesis have an apparent edge.

The reservoirs more than 2 000 m deep have much poorer physical properties and much smaller pore throat radius, due to advanced diagenesis, and they are mainly of type I or II. But different lithologies of them show different characteristics of fluid mobility. The movable fluid saturation tests (Table 1) show that reservoir samples of fine sandstone and siltstone more than 2 000 m deep (sample No. 3, 4, 6) have movable fluid saturations of greater than 40% and are of type I. Reservoir samples of argillaceous siltstone (sample No. 8, 9, 11, 12) have a movable fluid saturation of less than 20% generally and are type II or III. The movable oil rate tests (Table 2) show that siltstone reservoirs more than 2 000 m deep (sample No. 16, 20) have a movable oil rate of more than 40% and are type I, whereas the mud-bearing siltstone samples (sample No. 23 and 26) have a movable oil rate of less than 30% and are type II or type III. In summary, for tight reservoirs at middle diagenetic stage B, lithology has a significant impact on fluid mobility, so it is necessary to study sedimentary characteristics to find out “sweet spots” of type I tight reservoirs of fine sandstone and siltstone and sort out exploration targets.

The tight sand reservoirs of Gaotaizi and Fuyu oil layers in northern Songliao Basin are mainly in inner delta-front, outer delta-front and shore-shallow lake. Taking the Gaotaizi oil layer as an example, in different sedimentary facies, the tight reservoirs have different lithology and vertical assemblages, leading to different fluid mobility.

Tight reservoirs in inner delta-front are dominated by siltstone and argillaceous siltstone with some fine sandstone. Siltstone and argillaceous siltstone account for 41.0% and 16.2% of formation thickness respectively and are 1.52 m and 1.16 m thick in single layer at most. Fine sandstone only take up 5.3% of the formation thickness and is 1.44m thick in single layer at most. The movable fluid saturation tests (Table 1) show that reservoirs of siltstone and fine sandstone in inner delta-front are of type I reservoir (sample 1-6) with a movable fluid saturation of greater than 40%, whereas those of argillaceous or mud-bearing siltstone are of type II or type III (sample No. 7-10) with a movable fluid saturation of 8%-23%. The movable oil rate tests (Table 2) show that reservoirs of siltstone and fine sandstone in outer delta-front (sample No. 13-20) are of type I with a movable oil rate of greater than 40%. Only sample No. 23 and 24 (mud-bearing siltstone) are exception, which are type II with a movable oil rate of less than 30%. In summary, reservoirs of siltstone and fine sandstone in outer delta-front have better fluid mobility and thus are important targets for tight oil exploration and development.

For tight reservoirs in outer delta-front, siltstone and argillaceous siltstone still take the majority, but account for a lower proportion of formation thickness of 27.1% and 10.8%, and are 0.76 m and 0.42 m thick in single layer at most, respectively. The movable fluid tests (Table 1) show that the argillaceous sandstone (sample 12) in outer delta-front is of type III reservoir with a movable fluid saturation of less than 10%. The movable oil rate tests (Table 2) show that the movable oil rate of reservoirs (sample No. 21, 22, 24, 25, 27, 28 and 29) of siltstone and fine sandstone in outer delta-front range 0-32.95% and are mostly less than 30%, indicating that reservoirs in outer delta-front are mainly type II or type III. As there are some type I reservoirs in this facies, sedimentary facies need to be investigated to sort out “sweet spots” .

Of tight reservoirs in shore-shallow lacustrine facies, argillaceous siltstone takes dominance, accounting for 22.5% of the formation thickness with the maximum thickness of single layer of 3.85 m. Mud-bearing siltstone accounts for 7.2% of the formation thickness with the maximum thickness of single layer of 0.89 m, indicating limited accumulation capacity. There are a few mud-bearing fine sandstone and carbonate layers, which account for 7.2% and 6.9% of the formation thickness and are 0.89 m and 0.31 m per single layer at maximum respectively. These layers can enhance accumulation space and formation brittleness of reservoirs. The movable fluid saturation tests (Table 1) show that the argillaceous siltstone reservoirs (sample No. 11) in shore-shallow lacustrine facies have a movable fluid saturation of less than 10% and are type III ones. It is necessary to reinforce research on exploration and exploitation techniques in order to realize a breakthrough in productivity in this field.

Pores occupied by movable fluid in tight reservoirs cover a wide range of radius and movable fluid can exist both in micro and nano-pores, indicating that pores of multiple scales can be accumulation space for movable fluid. At the temperature of 76-89 °C and pressure of 35-42 MPa like the Qingshankou Formation in northern Songliao Basin, the lower limit of porosity, permeability and average pore throat radius at which oil in tight reservoirs can move are 4.4%, 0.015× 10^-3 μm², and 21 nm respectively. In contrast, pores occupied by unmovable fluid in tight reservoirs varies little, and are less than 550 nm at most and peak at less than 70 nm, indicating that nano-pores in various tight reservoirs have the same effect of adsorption or retention on fluids.

For tight reservoirs, oil saturation and oil mobility have no close correlation, rather physical properties and pore structure are main controlling factors of movable oil rate. The movable fluid saturation is generally greater than the movable oil rate, and the reason is that the formation water used in the centrifugal experiments is more mobile than the oil used in the displacement experiments, indicating that the oil viscosity or gas oil ratio has a great impact on oil recovery.

Tight reservoirs can be divided to three types. Type I reservoirs refer to those with a movable fluid saturation of greater than 40%, movable oil rate of greater than 30%, and starting pressure gradient of less than 0.3-0.6 MPa/m. Type II reservoirs refer to those with a movable fluid saturation of 10%-40%, movable oil rate of 5%-30% and starting pressure gradient of 0.6-1.0 MPa. Type III reservoirs refer to those with a movable fluid saturation of less than 10%, movable oil rate of less than 5% and starting pressure gradient of greater than 1.0 MPa/m.

The fluid mobility in tight reservoirs is affected by diagenesis and sedimentation. Reservoirs less than 2 000 m deep are dominated by type I reservoirs, whereas reservoirs more than 2 000 m deep include type I and type II. Fine sandstone and siltstone in inner delta-front are mainly type I reservoir, whereas with increase of the content of argillaceous siltstone, reservoirs in outer delta-front and shore-shallow lacustrine facies are dominated by type II or type III, while type I reservoirs are fewer. Apparently, tight reservoirs less than 2 000 m deep and those more than 2 000 m deep in inner delta-front are favorable tight oil exploration and development targets.