Chip Flood ( vs ) Core Flood — Assessment of Flowback and Oil Productivity in Oil-Wet Hydraulic Fractured Rocks

New developments in lab technologies help us to explore problems that were less understood in the past due to the limitations and technological constraints. One such problem of assessing the formation damage created by the invasion of fracture fluid into the matrix at lab scale is the visualization of fluid saturation distributions inside the matrix. According to the current understanding, the high capillarity contrast between the fracture and the matrix creates a non-uniform saturation distribution of invaded fluid phase during flowback, with the saturations mostly concentrated at the fracture face. With the advent of microfluidics, their application has become more feasible to visually analyze the effectiveness of surfactants to mitigate the invasion-created formation damage and understand the impact of depth of invasion on the characteristics of flowback and oil productivity. Through our previous work, we have successfully demonstrated the capability of this new visualization tool in studying the factors of the presence of surfactant in the fracture fluid and its depth of invasion, to understand the flowback efficiencies and later oil productivities in oil-wet fractured formations. Since the substrate for flooding was a proxy model of an actual rock, the chip flooding results need to be validated with conventional core flooding experiments. In contemporary times, when the new advancements in technology are driving the research progress in all industries, it is mandatory to take a well informed decision by imposing a comparative check on the results with accessible conventional means, wherever possible. The success of validation of chip flooding approach with the core flooding approach in this work instates a strong belief over the application of microfluidics to pursue more research in related fields of oil recovery.


Introduction
Hydraulic fracturing in horizontal wells improves the drainage area of production by creating high conductive passageway for hydrocarbon to flow from far-off regions into the wellbore.Almost 51% of US daily production is accounted to hydraulic fractured wells [1] and these production trends, in general, are predominantly dominated by fracture flow with the late-time production being supported by the matrix surrounding the fractures [2].Due to the high injection pressures of operation, the leak-off of fracture fluids into the matrix can easily cause the constriction to the production of hydrocarbons.Moreover, these fluids could even lead to the swelling of clay particles found in the formation causing more formation damage if the flowback process is inefficient [3].
Many methods of mitigation of the leak-off or invasion-induced formation damage have been proposed by researchers.High production rates reduce the capillary entrapment of invaded fluids due to capillary desaturation and hence improve the flowback volumes.But due to the operational constraints of the surface and downhole equipment, very high production rates cannot be achieved in the field.Shut-in operation has also been explored as a possible formation damage mitigation strategy.In water-wet rocks, it is observed that the water block formed by the invaded fracture fluid predominantly exists at the matrix-fracture interface due to the capillary discontinuity between the high-permeable fracture and low-permeable matrix space surrounding it [4] [5].Bertoncello et al. (2014) [6], Liang et al. (2015) [7], Odumabo et al. (2014) [8] and Yan et al. (2015) [9] have shown that with sufficient duration of shut-in period the water-block would imbibe away from the fracture face due to capillary redistribution resulting in improvement of relative permeability to hydrocarbon.But very long shut-in times could affect the operator adversely as the expenditures increase due to the high day-rate services offered by the service companies.Moreover, spontaneous capillary redistribution phenomenon is prominently observed in water-wet rocks and not in oil-wet rocks [10], thus raising concern over the practicality of shut-in strategies.In the case of gas reservoirs, evaporation of water block due to gas expansion has also been considered as another method of the removal of formation damage caused due to invasion [4] [5], but the time-scale of expansion-driven evaporation is significantly longer than that of displacement-driven flowback mechanism.The application of surfactants to reduce the capillary entrapment of invaded fluid has been a plausible solution suggested by Liang et al. (2016) [11] and Kim et al. (2016) [12], who demonstrated surfactant's effectiveness by conducting laboratory experiments.
Due to the dominance of high capillary pressures in fractured tight formations, the macro-scale flow dynamics are also influenced by the capillary pressure gradients acting across the fluids distributed within the formation.Longoria et al.
(2015) [13] have shown that such a factor of capillary gradient across the invaded and uninvaded regions of low permeability matrix impacts on the reduction of formation damage as well as the flowback volumes.Similarly, the exis-Open Journal of Yangtze Gas and Oil tence of high contrast in capillarity between the fracture and the matrix, aids the displacement of fracture and in-situ fluids during the production phase.Since most of the published research is concentrated on water-wet formations, the current work is catered to oil-wet fractured formations only.As discussed before, the practical strategy of mitigating the formation damage by improving the efficiency of flowback in an oil-wet formation is by using either a water-based or surfactant-based fracture fluid, rather than using shut-in procedures.Since the capillary gradient across the matrix and fracture is affected by the average invaded fluid saturation in the matrix, the invaded volume or the depth of invasion is a very crucial factor of investigation that has a direct impact on the flowback characteristics and the oil productivity from the fractured formation.
During recent times, new state-of-the-art microfluidics based tools are being used in the petroleum industry to understand the oil recovery mechanisms by visualizing the macro-scale and pore-scale fluid flow behavior in transparent microchips.Chip-scale displacement patterns, sweep efficiencies of secondary oil recovery methods using water-based fluids [14]- [19] and tertiary oil recovery methods using surfactant and polymer based fluids [20]- [29] are studied with ease using these new microfluidic based tools.He et al. (2017) [22] have incorporated this approach for the first time to understand the relation of flowback and invasion volumes but performed experiments with a constant rate of production, rather than using a more practical and industry-wide practice of constant pressure drop during production.Moreover, they have not considered the varying depths of invasion as another prominent factor that has a big role to play on both the lab-scale and field scale.In our previous work, Tangirala and Sheng (2018) [30] have presented microfluidic experimental results to show the impact of fracture fluid invasion on oil flow rates and flowback as a function of depth of invasion and the usage of surfactant, when invaded into an oil-wet porous space located adjacent to a fracture.It was observed that as the depth of invasion increases, the flowback efficiency decreases, irrespective of the presence of surfactant in the fracture fluid.Another important implication of that study is that for shallow invasions, the fracture fluid in the presence of surfactant does not mitigate the invasion-created formation damage better than that of the fluid without the surfactant.The applicability of these findings in the industry depends on the validity of the results in a more proven lab-scale procedure, i.e. using an actual core block and conducting the equivalent procedures in the domain of core flooding.
In the current work, we designed an experimental core flooding process with operating procedures similar to that used in Tangirala & Sheng (2018) [30] chip flooding method.The results are presented adjacently for both the methods to validate the microfluidic results with the core flooding results.The success of this validation methodology is an indication of the rising prominence of chip-flooding procedures which could give rise to a new technique of preliminary assessment in labs without having to work with expensive cores drilled from the subsurface.Open Journal of Yangtze Gas and Oil

Experimental Materials
Six Berea sandstone core samples named A to F are used for core flooding whose dimensions and pore characteristics are mentioned in Table 1.The porosity of these samples is obtained by the weight measurement method of determination of pore volume.Permeability is also obtained by the steady-state method of flowing liquid through a 100% liquid-saturated core sample.The detailed procedures of obtaining these measurements are highlighted in the succeeding section.The core is covered with teflon tape on the curved portion to form a perfect non-invasive seal to the rubber sleeve in which it is placed.The rubber sleeve is placed inside the hassler core holder with the core tightly enclosed using the metal caps on both sides of the core holder.Besides these metal caps acting as uniform fluid distributors across the cross-sectional area of the core, they even act as a proxy for the high conductive fracture space adjoining the relatively low conductive matrix region of the porous media.Similarly, the microfluidic chip used for chip flooding is a 45 mm × 15 mm borosilicate glass chip, with a uniform channel porous network etched over it having a footprint of 20 mm × 10 mm.The chip is manufactured by Micronit Microtechnologies B.V., Netherlands.At the entry to the uniform pore network, a single wide pore channel distributes the injected fluids evenly across the whole cross sectional area of the pore network.This portion of the chip represents the fracture region where the fluid is collected first before entering into the matrix represented by the porous media channels of the chip.
The contrast in the physical dimensions of the core and the microchip is depicted in Figure 1 as well as Table 1 below.
Besides the core holder, the other equipment used for the core flooding process is as per the requirements such as accumulators, back pressure regulator, quizix pumps and syringe pump.But for the microfluidic experiment, the chip is  purchased from Fluigent Inc.The flow rates in the flow lines are measured using the Flow Rate Platform (FRP) also acquired from Fluigent Inc., having a mea-

Experimental Setup & Procedure
Both of the core flooding and the chip flooding experiments are conducted at equivalent operating conditions, so that the results obtained from them are comparable.These operating conditions and the detailed procedures of operation are highlighted below.

Core Flooding Experimental Procedure
The schematic for the core flood experimental set up is shown in Figure 2. Open Journal of Yangtze Gas and Oil The methodology of conducting the experiment is according to the following sequence: • Initial saturation: The dry core is weighed, W d and vacuum saturated with soltrol oil (density, ρ o ) for 12 hrs in a pressurized core vessel.The weight of the oil saturated core is measured, W o .The porosity is hence obtained according to the Equation ( 1).The results are tabulated in Table 1 with the obtained porosities lying in the range of 17% ~ 19%.
• Wettability assessment: The wettability of the core is assessed by using the sessile drop method with Krűss Drop shape analyzer (DSA-25).The oil saturated core is placed in a transparent glass flask filled with oil and the wettability of the injecting fluid (either water or surfactant) is assessed by placing a droplet size portion of the fluid on the top flat surface of the core.Thereby the contact angle is obtained with oil as surrounding medium.At least 30 such instances are measured at various locations of the core to account for the heterogeneity in texture and the sample statistics thus obtained，which are enlisted in Table 2. Based on the results from the wettability analysis, since the mean contact angles are above 120˚ with a reasonable standard deviation, it can be safely inferred that the wettability of the rocks for both the water and the surfactant fluids is oil-wet.
• Permeability assessment: The teflon wrapped core is placed inside the core holder along with a rubber sleeve.Confining pressure of 1500 psi is applied by pumping water from quizix pump-1.A back pressure of 900 psi is exerted on the core using a syringe pump so as to maintain a constant pressure drop of around 75 psi by pumping soltrol oil from another quizix pump-2.The stabilized steady-state flow rate of oil is measured by collecting downstream fluid from a measuring cylinder.The effective permeability of the core is hence calculated using the Darcy's equation and the obtained results are tabulated in Table 1, with the range of permeability being 50 ~ 65 mD.
• Invasion: The forced invasion of fluid, either water or surfactant, is carried from the side B of the core holder at a constant pressure drop of 50 psi, pro-Open Journal of Yangtze Gas and Oil ducing rates of around 0.1 ~ 1.5 cm 3 /min.Since the invasion is moderately unstable displacement (viscosity ratio = 0.42), the rates are maintained, which are low to prevent any viscous fingering causing early-time breakthrough.The invading volume is varied for different cores so as to observe a trend across these varying conditions.At the end of the period of respective volume of invasion, the core is removed from the core holder and weighed again for its fluid invaded weight, W inv .The invasion efficiency, inv% is hence obtained which is given by Equation ( 2), where V p indicates the calculated pore volume.
• Flowback: Subsequent to the invasion process, the core is placed again inside of the core holder and the oil is pumped at a constant pressure drop of 300 psi from side A of the core holder for about 10 pore volumes.The invaded fluid is ejected out as flowback and amounts pertaining to residual saturation are attained approximately after 10 PV, as observed from the measured stable oil flow rates.The pressure drop acting across the cores dictate these later oil production rates and the pressure is decided so as to obtain a comparable capillary number with that of the microfluidic experiment.At the end of the flowback phase, the core is once again removed from the core holder to measure the final weight of the core, W flb .
Both the residual saturation of invaded fluid, S w2 and the flowback efficiency, flb% are calculated according to the Equations ( 3) and ( 4).

Chip Flooding Experimental Procedure
The schematic for the chip flood experimental set up is shown in Figure 3.The image thus attained is processed further to obtain the saturation of the invaded phase, S w1 .
• Flowback: Soltrol oil is reinjected into the chip, but from the opposite direction indicated by the green dashed arrows in Figure 3 at the same pressure of 80 mbar.After approximately 10 PV (pore volumes) of injection, the invaded phase saturation attains residual level, upon visual inspection, and immediately the final stabilized oil rate, Q o is measured using the FRP.The later oil production rates indicate the relative amount of reduction of formation damage caused due to the invasion phenomenon.The justification for applying such low injection pressures could be obtained after evaluating the capillary numbers from these production rates that lie in the order of 10 −5 .
These capillary numbers are typical for the formations where capillary forces begin to show their dominance [23].The valves at the both ends of the chip are closed once again and the image of the whole chip is acquired to calculate the final saturation of the invaded fluid, S w2 .Based on the measured saturations, the invasion efficiency and the flowback efficiencies can be directly calculated from Equations ( 5) and ( 6).

( ) w1
Invasion efficiency • Image processing: The stitched image of the entire chip is cropped at the boundary of the porous media and is converted to a binary image format using Image J software.The saturations of the fluids inside the chip are thus computed by analyzing the histogram of the black and the white colored pixels of the binary image.

Results and Discussion
The experimental results for both the chip-flooding and core-flooding methods are tabulated in Table 3 and Table 4 respectively.The trends obtained from both of the methods are evaluated simultaneously to qualitatively validate the observations arising from the similar experimental procedures, as mentioned in the previous section.With reference to the previous results obtained for chip-flooding from Tangirala & Sheng (2018) [30], the addition of surfactant to the fracture fluid shows a contrasting behavior to the fluid without the surfactant, at varying depths of invasion.These factors are hence dealt individually in the following section to better understand the effects of their variabilities.

Effect of Depth of Invasion in the Absence of Surfactant
The processed microchip images, depicted in Figure 5, after both the invasion and the flowback phases with water as the invading fluid, can be visually analyzed for saturation information i.e. S w1 and S w2 .But for core flooding experiments, such information is derived from the weight measurements due to the lack of availability of CT scanner with better resolving capabilities.As discussed in the previous section, the invasion and flowback efficiency for chip flooding experiment are derived from Equations ( 2) and ( 3), whereas for the core flooding experiment, they are derived from Equations ( 5) and ( 6).   the mentioned parameters beforehand.Even though the invasion efficiency is a volumetric quantity, it can be fairly translated to the depth of invasion due to the low injection pressure application in both the methods preventing viscous fingering.Such an approximation could even be supported by the uniform sweeping pattern of invasion across the cross sectional area of the porous network of the chip visualized in Figure 5.
From Figure 6, it can be seen that both the processes of chip flooding and core flooding exhibit similar direct proportionality of relation of S w2 with invasion efficiency.It can be attributed to the increase in the capillary trapping of the invading fluid within the pores as a larger amount of fracture fluid is injected.
Such increase in the residual saturations of the fracture fluid results in a reduction in the relative permeability of oil which is reflected as formation damage and hence the later oil production rates could also be severely affected.This trend is clearly depicted in Figure 7 for both the processes of chip flooding and core flooding where a decreasing trend for Q o is observed as invasion efficiency is increased.Henceforth, these results could be interpreted to state that as the depth of invasion of the fracture fluid into the oil-wet matrix increases, the tendency for the reduction of formation damage due to flowback decreases.Open Journal of Yangtze Gas and Oil To better understand the characteristic of the flowback process, the invasion efficiency is cross plotted against flowback efficiency, as shown in Figure 8.

Effect of Depth of Invasion in the Presence of Surfactant
The processed microchip images after both the invasion and the flowback phases with surfactant as the invading fluid can be visualized in Figure 9.
A non-ionic surfactant is added to the fracture fluid, due to its non-aggressive nature in changing the wettability of the chip and the core, as evidenced from Table 2. Most of the commercially available non-ionic surfactants reduce the

Effect of Surfactant Addition to Fracture Fluid at Fixed Invasion Depth
The previous analyses highlight the influence of the invasion efficiencies or invasion depths of the fracture fluid, considering the absence and the presence of the surfactant independently.They establish the overall behavior of the parameters of later oil production rates, residual saturation of invaded fluid and the flowback efficiency for an oil-wetting surface.The capillary gradient across the matrix and the fracture is considered accountable for the decreasing flowback efficiency trend with respect to the invasion depth.Since the strength of the capillary gradient depends on the capillary pressure in the matrix, the presence of the surfactant could depreciate such an effect where it is prominent.Therefore, the performance of the two types of fracture fluids i.e. water and the surfactant are compared together at fixed invasion efficiencies to prove their contrasting beha-Open Journal of Yangtze Gas and Oil vior.Such an analysis is once again studied for both the core flooding as well as chip flooding experiments and it is based on the similarity of results obtained in preceding sections.It is expected that both these methods yield similar qualitative results.
For the ease of comparative analysis, new parameters of X, Y and Z are defined which give the difference in the values of S w2 , flowback% and Q o , respectively, between the water and surfactant, normalized by their respective maximum value of differences.These parameters of X, Y, and Z are defined by Equations ( 7)-( 9), which are obtained at fixed invasion efficiencies with the aid of linearly regressed models of S w2 , Q o and flowback% parameters obtained from The values of X, Y, and Z for the selected fixed invasion efficiencies are shown in Figure 13  With respect to both the methods of experimentation, it can be observed that there is a certain critical amount of invasion below which the performance of surfactant fluid in reducing the formation damage is lower than that of the fluid without the surfactant.For chip flooding experiment, the critical invasion efficiency lies between 20% and 30% and for the core flooding experiment, it lies between 30% and 40%.Even though the determination of factors that influence this critical point of invasion is not the current scope of study, qualitatively both the methods of experimentation evidently show the existence of such a critical point.The influence of the previously discussed matrix-fracture capillary gra-Open Journal of Yangtze Gas and Oil dient could be seen to be prominent for the invasions below this critical point.
Hence, at shallow invasions, as the capillary gradient is reduced by the presence of the surfactant in the fracture fluid, the quantities of X, Y and Z representing the relative superiority in the performance of the surfactant are negative.But at deep invasion, as the strength of the capillary gradient across the matrix and the fracture is lesser pronounced, the X, Y and Z are all observed to be positive, indicating that the surfactant is more beneficial to be included in the fracture fluid than be excluded.It could be attributed to the better mobilization of residual saturations by capillary desaturation effect of the surfactant in the fracture fluid, as it increases the capillary number by approximately one order of magnitude (Table 3 and Table 4) and hence reduces the residual saturation of invaded fluid trapped in the pore space better than the fluid without the surfactants.
The resemblance of the trend in results for both the chip flooding and core flooding experiments indicates that customized microfluidic experiments could in fact be used as a precursor of core flooding experiments to understand the qualitative behavior of interaction of fracture fluids with both the in-situ fluids as well as the complex porous network of subsurface in terms of sweep efficiency, oil recovery, flowback potential etc.The recent advances in fabrication of chips [29] also expose a whole new domain to improve the scope of research so as to understand the flow dynamics in unconventional formations such as shales.

Conclusions
The microfluidic results published in the previous work [30], indicate that irrespective of the presence of moderate IFT-reducing surfactant in the fracture fluid, an increment in the depth of invasion into the matrix has yielded a decrement in the later oil production rates and flowback efficiencies.Thus, the reduction of invasion-created formation damage during flowback is more difficult when the invasion is deeper.Moreover, the microfluidic results also indicate that moderate IFT-reducing surfactants are beneficial to reduce the formation damage only when the invasion is deep; and their formation damage is even higher than water when the invasion is shallow.These results were only obtained previously for the case of a chip which acted as a proxy for the physical subsurface rock.
The present study thus validates these microfluidic results by following similar procedures of invasion and flowback under equivalent operating conditions in a core flooding set-up, with the chip being replaced by an actual core plug.The results obtained for the core flooding experiments qualitatively resemble the results obtained for the chip flood experiments, both in terms of the effect invasion depth as well as the effect of a moderate IFT-reducing surfactant in the fracture fluid.
Hence, it could be safely said that the assessment of flowback and oil productivities for an oil-wet hydraulic fractured formation could be performed through chip flooding before core flooding and the results could be relied upon with equal confidence.
Figure 1.(a) Microfluidic chip (b) Core sample, represented along with their typical physical dimensions.
surable range of 0 ~ 7 μl/min.Before allowing the fluids to pass through the chip, they are filtered using a 2 μm in-line PEEK filters provided by IDEX corporation.The oil used in both the experiments is soltrol-130 supplied by Chevron Phillips Chemical Company having a dynamic viscosity of 2.37 cp and a specific gravity of 0.74.Deionized water (DI water) is used for conducting immiscible displacements of oil, where the IFT between soltrol oil and DI water, measured using a Du-Nuoy Ring tensiometer is 32.92 ± 0.27 mN/m.A 0.05% non-ionic surfactant supplied by ChemEOR Inc., is used for miscible displacement of soltrol oil in the chip.The observed interfacial tension (IFT) for the surfactant is 1.64 ± 0.15 mN/m, measured by the M6500 Spinning drop tensiometer purchased from Grace Instrument Company.A Fluorescein dye that is soluble only in the aqueous phase is used to distinguish between the multiple colorless fluids when they are viewed under the microscope in the microfluidic experiment.On the contrary, the saturation computation is only weight based in the core flooding experiment and hence there is no requirement of a dye to differentiate fluids for the case of core flooding experiment.

Figure 2 .
Figure 2. Schematic of core flooding set-up.

Figure 3 .
Figure 3.A schematic of chip flooding set-up.

Figure 4 .
Figure 4.A representative instance of (a) Contact angle of water (left) and surfactant (right) against the oil media and the chip surface (b) Contact angle between the water droplet (left) and surfactant droplet (right) against the core surface.

Figure 5 .
Figure 5. Water as invaded fluid (represented by black color) for an oil-wet chip, showcasing the stitched microchip images after invasion (left column) and after flowback (right column) for different invasion efficiencies.

Figure 6 .
Figure 6.Relationship between S w2 and invasion efficiency for (a) Chip flood (b) Core flood, in the absence of surfactant.

Figure 7 .
Figure 7. Relationship between final stabilised oil rate after flowback and invasion efficiency for (a) Chip flood (b) Core flood, in the absence of surfactant.

Figure 8 .
Figure 8. Relationship between flowback efficiency and invasion efficiency for (a) Chip flood (b) Core flood, in the absence of surfactant.

Figure 9 .Figure 10 .
Figure 9. Surfactant as invading fluid (represented by black color) for an oil-wet chip, showcasing the stitched microchip images after invasion (left column) and after flowback (right column) for different invasion efficiencies.

Figure 11 .
Figure 11.Relation between final stabilised oil rate and invasion efficiency for (a) Chip flood and (b) Core flood, in the presence of surfactant.

Figure 12 .
Figure 12.Relation between flowback efficiency and invasion efficiency for (a) Chip flood and (b) Core flood, in the presence of surfactant.
(a) and Figure 13(b) for chip flooding and core flooding experiments, respectively.

Figure 13 .
Figure 13.Relative performance of fracture fluid with surfactant against the one without surfactant for both (a) Chip flood and (b) Core flood experiments.

Table 1 .
Physical characteristics of core and chip.

Table 2 .
Contact angle measures for wettability assessment.

Table 3 .
Summary of microfluidic experimental results.

Table 4 .
Summary of core flooding experimental results.