SPE-200001-MS

Understanding the Mechanisms of Huff-n-Puff, CO2-EOR in Liquid-Rich

Shale Plays: Bakken Case Study

Abdulaziz Ellafi and Hadi Jabbari, University of North Dakota

Copyright 2020, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Canada Unconventional Resources Conference originally scheduled to be held in Calgary, Alberta, Canada, 18 –
19 March 2020. Due to COVID-19 the physical event was postponed until 29 September – 2 October 2020 and was changed to a virtual event. The official proceedings
were published online on 24 September 2020.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
A revolution of unconventional reservoirs is a turning point in the global oil and gas industry since these
resources have massive reserves with large potential in contributing to hydrocarbon production. Previous
EOR laboratory experiments and simulation studies in the literature illustrated promising results in terms of
recovery factor for different EOR applications, such as CO2, surfactant, and natural gas. However, pilot tests
performance reported contrast behavior due to misleading predicting for the EOR physics processes. This
paper presents the experimental work to evaluate the feasibility of CO2-EOR using the huff-n-puff (HNP)
protocol in the Middle Bakken (MB) Formation, the Mountrail County, Williston Basin, ND. We evaluate
the oil recovery from CO2-EOR under several scenarios of operational and well/reservoir conditions. The
parameters considered in the sensitivity study include temperatures, pressure, soak time, and number of
injection cycles to obtain optimum conditions under which the incremental oil recovery from the MB
Formation is increased. The wettability alteration (i.e. contact angle) was also studied using rock-chip
samples before and after the HNP experiment at the Bakken reservoir conditions (present for example
P & T in psi/F). The outcomes indicated on the effect of the reservoir temperature and pressure on the
performance of the CO2, where the recoverable oil increases as the temperature and pressure increase until
reach the optimum. As a previous research outcome, the number of cycling and soaking time are crucial
design parameters for the HNP experiment and on the field as well to let the CO2 time to diffuse into the deep
formation and swell more oil. In addition, the wettability alteration was changed by CO2-EOR as injection
pressures increase and the wetting phase move from the oil-wet toward the water-wet system. As overall
outcomes from this research, the CO2 HNP process has a good potential in the lab, and could be succeeded
economically in field applications that might reduce the need for refracturing stimulation or infill drilling.

Introduction
A revolution of unconventional reservoirs is a turning point in the global oil and gas industry since these
resources have massive reserves with large potential in contributing to hydrocarbon production. In recent
years, the domestic oil production from liquid-rich shale (LRS) reservoirs in North America have shown
2 SPE-200001-MS

immense development, and the production has dramatically increased in the "top producers" American
oil fields: Bakken, Eagle Ford, and Permian Basin. The total U.S oil production from conventional and
unconventional reservoirs 16.5% was produced in 2008 compared to nearly 60% in 2019 (The Energy
Information Administration (EIA), 2019). Currently, these shale plays have the most drilling and completion
activities in the U.S., with the number of wells in each play at over 12,000 producing either oil or gas (EIA,
2020; Drilling Info, 2019; FracFoucs, 2019). EIA outlook data in 2050 shows that the U.S. shale plays’
daily production rate will be extended to 70% of the total U.S. daily oil production. This improvement
in hydrocarbon production is driven by applying modern horizontal drilling and multi-stage hydraulic
fracturing that make it a reality to access low porosity (<10%) and low permeability (<0.1 mD) formations
(Ellafi et al., 2020a; Ba Geri et al., 2019a). In the Bakken Formation, a complex-fracture geometry system
is often generated as a result of significantly distributed in natural fractures. Although breakthroughs in
unconventional technologies have been achieved to create larger stimulated reservoir volume (SRV), the
oil recovery is believed to be less than 8% due to sharply decline in oil production rates after the fractures
depletion with small to no recharge from the ultra-tight matrix blocks (Jin et al., 2017; Sheng, 2015). The
estimated oil reserves in the Bakken Sweet Spots is around 500 billion barrels of oil, and only 30 to 40 billion
barrels of oil can be produced with current technology (Energy & Environmental Research Center (EERC),
2019; Continental Resources, 2018). Therefore, the unrecovered hydrocarbon from tremendous storage is
isolated in tight pores without using unconventional applications, such as improving oil recovery (IOR)
and enhanced oil recovery (EOR) methods. The common questions are asking by people in the industry
and required to answer: what are the methods, how the process works, and can these applications enhance
incremental oil recovery in a commercial way.
The current options that are used and gaining more attention in the industry to revive the performance
in unconventional wells are infill drilling and refracturing treatment applications. The refrac treatment
is performed by injecting fracking fluids, such as high viscosity friction reducers (HVFRs) through the
fractures of the previous job and/or new entry points to create new fracture clusters with smaller fracture
spacing in order to enhance production performance (Ellafi et al., 2020a; Ba Geri et al., 2019b). Based
on the drilling spacing unit (DSU) and imposed government regulations, up to 20 wells can be drilled
and stimulated from a single well pad to produce economically sound and financially profitable oil from
unconventional shale plays (Ahmed and Meehan, 2016). However, these development applications of shale
reservoirs have reached a challenging point, where the operators in North America face problems in terms of
management and environmental issues. For example, the successful restimulation application is required a
careful selection of the design parameters: treatment fluids, proppant type, completion method (diversion or
isolation), design of refrac stages, and the selection of proper well candidate (Ellafi et al., 2020b; Ba Geri et
al., 2020). As a result, the uncertainty is significant and difficult to infer which the most critical factors due to
high cost and unavailable diagnostic tools to improve prediction of post refracture treatment design. On the
other hand, the treatment process is being re-evaluated the long term impacts on environmental perspective,
where the critics claimed that the hydraulic fracturing uses materials would contaminate groundwater
resources and toxic air emissions (Ellafi et al., 2020b; Chen et al., 2014).
There are a number of papers mentioned that refracturing process is crucial methodologies to expand
the oil and gas production from unconventional reservoirs, while others (Jin et al., 2019; Gubian, 2017;
Cipolla and Wallace, 2014) claimed that the increase in the well lateral length, number of fracture stages,
and proppant mass loading will not improve the incremental oil recovery more enough proportionally
to operating cost. Furthermore, seeking a way to keep sustainability and enhance oil recovery in
unconventional reservoirs, such as EOR has become an emerging technology to access more remaining oil
and improve the long-term well productivity from cost-effective operations (Ellafi et al., 2020a; Ellafi and
Jabbari, 2019a; Lashgari et al., 2018; Kurtoglu, 2013; Shoaib and Hoffman, 2009).
According to EIA, increasing in one percent of the oil recovery in LRS reservoirs could lead to increase
the technically recoverable oil by 10 to 25 billion barrels of oil as well as improve the net present value
SPE-200001-MS 3

(NPV) of a field (Jin et al., 2019). Therefore, any effort to improve the recovery factor through an EOR
process, such as gas-injection is worthwhile. The extraction technique would reduce the remaining oil
from the nano-darcy pores media and increase the oil recovery factor up to 10% (Mahzari et al., 2019).
Most recent EOR experimental investigations and numerical simulation studies reported that miscible and
immiscible gas injection processes, such as associated-produced gases (C1, C2, & C3), carbon dioxide (CO2),
and nitrogen (N2) are the most effective EOR agents due to larger injectivity and lower viscosity (Alfarge
et al., 2017; Kurtoglu, 2014). These applications might take tight formations to the level up to 20% of
incremental oil recovery (Wang et al., 2019; Thakur, 2019). In contrast, Bakken Formation makes traditional
EOR processes are extremely challenging to implement due to its lower characterization rock quality. As a
result, normal water flooding or continuous gas injection are not applicable for unconventional reservoirs
due to poor seep efficiency and low injectivity that can take long payback period. The only way can
overcome mitigate the negative effect of fractures in unconventional reservoirs if the huff-n-puff (HNP)
approach is used to inject large volume of gas into high conductivity fracture extend area, then produce oil
through the same well at different time interval (Yu et al., 2014; Gamadi, et al., 2013).
Hawthorne et al. 2013 summarized the gas injection EOR process in unconventional reservoirs in five
conceptual steps as follows: 1) gas flows into and through the fracture networks, 2) injected gas exposed
at the fracture surfaces into rock matrix, 2) the gas penetrates the tight pores carrying some hydrocarbon
inside the rock matrix by pressure, but this step assists to reduce the oil viscosity by swelling mechanism
and extract more oil out of the nano-porous media, 4) by swelling and viscosity reduction mechansims, the
hydrocarbon migrates to micro-porous media (fractures system), and 5) the migrate process caused by gas
injection pressure gradient that becomes smaller, then molecular diffusion mechasim drives oil flow slowly
from nano-porous to the fracture networks. Based on Hawthorne’s explination, unconventional reservoirs
have shown two distinguishale flow regimes: viscous flow in the high permeability fracture network and
diffusion dominated flow in the low permeability rock matrix. Improving incremental oil recovey in tight
formations reiles on enhancing the ability of diffusion mechamesims from the rock matrix with significant
oil recovey to the fracture system (Jin et al. 2017). Understanding the HNP mechanism, which is the focus
of this paper, guide to enhance oil recovery and improve the long-term well productivity. Subsequently, this
process is a profitable operation that might reduce and/or limit the need for refracturing stimulation and/or
infill drilling, especially when the oil prices are going down (Jin et al., 2016; Hawthorne et al., 2013).
However, the pilot tests assessment report in the Bakken Formation showed contrast behavior due
to misleading predicting for the EOR physics processes from lab to field. Today, the industry needs to
know many of the key questions remaining unanswered regarding the optimum HNP operation conditions,
uncertainty factors of unconventional EOR adsorption/diffusion mechanisms in the field scale, solvent-
fluid-rock minerals interaction, nano-confined oil (PVT) effect during depletion and EOR process, damage
because of changing in fluid composition and reservoir conditions (asphaltene precipitation), and the full
economic analysis study compared to current options (i.e. refrac & infill drilling).
This study presents a comprehensive review and experimental investigation to assess the feasibility of
CO2-EOR using the HNP protocol in LRS reservoirs. We first reviewed several publications to gain a deeper
understanding of the CO2-EOR mechanisms and address opportunities and challenges in the Williston
Basin. In the lab, we evaluate the oil recovery from CO2-EOR using the Middle Bakken (MB) formation,
the Mountrail County, Williston Basin, ND core samples under several scenarios of operational and well/
reservoir conditions. The design parameters considered in the sensitivity study include temperatures,
pressure, soak time, number of injection cycles, and depletion pressure to obtain optimum conditions under
which the incremental oil recovery from the MB core samples is increased. Moreover, the wettability
alteration (i.e. contact angle) was also studied using rock-chip samples before and after the HNP experiment
at the Bakken reservoir conditions (present for example P & T in psi/F).
4 SPE-200001-MS

Bakken Petroleum System (BPS), Williston Basin

The Williston Basin is a shared area between the United States and Canada. The basin occupies about
225,000 square miles of the subsurface and covers parts of Eastern Montana, Southern Saskatchewan,
Manitoba, and Western North Dakota. The most productive formation in the Williston Basin is the Bakken
Formation or known as the Bakken Petroleum System (BPS). The initial oil in place is estimated between
100 to 900 billion barrels of oil, and cumulative oil production since 10 years is approximately 1 billion
barrels (bbl) of oil from several reservoirs through over 10,000 wells in Parshall, Sanish, Reunion Bay,
Bailey, Murphy Creek, Antelope, and Elm Coulee Fields (North Dakota Industrial Commission (NDIC),
2019; Jin et al., 2016). This recoverable oil represents approximately 1-2% of the total unconventional
oil reserve with the available technologies today (Sheng, 2015; North Dakota Council, 2012). Fig. 1
demonstrates the Bakken production performance in December 2019 of several counties in North Dakota
State (ND), where our study focuses on Mountrail County as marked by the gold star.

Figure 1—Bakken oil production performance of North Dakota counties in December 2019 (NDIC, 2019).

Bakken Well Performance

Most unconventional wells spend years producing by primary depletion in the transient flow until
hydrofractures around SRV begin to deplete, and the wells switch to boundary-dominated flow (BDF)
(Male, 2019). For instance, Fig. 2 presents the average Bakken wells oil production rate begin with a high
production rate due to high initial reservoir pressure (initial pressure greater than 6,500 psi) and significant
conductivity and connectivity between fractures. Moreover, the rapid decline was observed by around 75%
after two years of begin producing because of large flow resistance at the matrix-fracture interface due to
closure of natural and/or induced fractures. Thus, the increase in net stress leads to a zero-pressure gradient
(formation pressure equal to initial reservoir pressure) of the fluid flow from the rock matrix into the fracture.
SPE-200001-MS 5

Figure 2—Avergare oil production per well in the Bakken formation (EIA, 2019).

Fig. 3 illustrates Male’s research that show water cut maintains constant over entire life of the Bakken
wells at 40%. This means most of the produced water is the formation water that produces as a result of
water level near the reservoir (the overlying Lodgepole Formation), while low percentage of flowback water
caused by treatment fluids. The Permian Basin behaves similar to the Bakken Formation, but with higher
water cut that excess of 70% of total liquid production. On the other hand, the Eagle Ford started with
a higher water cut, which is likely caused by the flowback water at 30%, and then the water depleted to
produce at a constant value of around 10%. The change in water cut over time indicates that the geological
description of the formation plays an important role in the percentage of produced water production. In case
of the Middle Bakken tight formation, the capillary effect is greater than the gravity segregation due to a
stabilized water cut that indicates the pore space contains both free water and oil (Jin et al. 2017).

Figure 3—The change in water cut over time in the Bakken Formation compared to other U.S. shale plays (Male, 2019).

As shown in Fig. 4 gas/oil ratio (GOR) remains constant over entire life of the Bakken wells at average
1,000 Scf/Stb compared to Eagle Ford and Permian Basin, where GOR started at 2,000 SCF/STB and
increased to reach around 4,000 SCF/STB over the first three years of production. GOR behavior in the
Bakken Formation indicates that most of the gas is dissolved into oil and confined in the formation, and the
primary production mechanism is the oil expansion drive with slowly releasing the gas with oil production.
6 SPE-200001-MS

Figure 4—The change in gas/oil ratio over time in the Bakken Formation compared to other U.S. shale plays (Male, 2019).

Tran et al. 2011 published a research study that introduced the types of well production trends in
unconventional wells and the outcomes of the interpretation utilize to calculate original oil in place (OOIP)
and the area of matrix drainage between fractures (Acm). The research is mainly focused on the Bakken
Formation, where 146 wells with twenty years of the production histories in the nine different counties in
North Dakota were analyzed, and the work classified the production performance into three categories using
production decline analysis and semi-analytical of linear dual-porosity Stehfest. Half of the wells are under
Type I, which is defined as the production behavior when the reservoir pressure drops below the bubble
point pressure, and the trend can be recognized on the GOR curve versus time plot. In addition, the type I
is divided into three sub-types production performance as follows: a) production behavior has less support
from the rock matrix, which is shown a fast decline then after a period of time the behavior changes to a
steady and a slow decline in production, b) strong support from the rock matrix, it can be noticed with a
short rapid decline at the beginning of production followed by an almost steady production 6 trend, and
c) the production behavior has not any support, and the oil recovery is only through the fracture network.
The behavior can be recognized as a sharp decline in the oil production trend and continues for the entire
well life, as summarized in Fig. 5.

Figure 5—The production trend for the Bakken wells under Type I (Tran et al. 2011).
SPE-200001-MS 7

Furthermore, Fig. 6 demonstrates the Type II, which is observed for some wells, as a half-slope on the
log-log plot of the oil rate versus time when the reservoir pressure is above the bubble point pressure, and
the produced oil flow is linear from the matrix into the fractures (oil production only from the matrix).
Also, the GOR is almost constant for the entire production life, as illustrated in Fig. 6. Finally, the rest of
the wells are under Type III, which is shown unclear behavior of the production trend due to scattering
data that leads to uncorrected analysis. In 2013, Kumar et al. quantified the production contribution in both
shale layers in the Bakken using reservoir simulation models by considering the effect of adsorption and
diffusion mechanisms to involve the significant mode of fluid storage and recovery mechanism in nanopores
scale. The observation of sensitivity analysis suggested that the Upper and Lower Bakken Shale contribute
from 12% to 52% of the total oil production from the three Bakken Layers. The finding addressed that the
shale layers are in interference with the Middle Bakken Formation and improvements in fracture treatment
application yield to more contribution and support for recovery mechanisms.

Figure 6—The production trend for the Bakken wells under Type II (Tran et al. 2011).

Characteristics of the Middle Bakken Formation

Fig. 7 illustrates the description of the sequence in the Bakken Formation, which is "two, black fissile
shales separated by light grey to grey-brown fine-grained sandstone". The Bakken shale play is subdivided
into three distinct stratigraphic members: Upper Bakken (UB), Middle Bakken (MB), and Lower Bakken
(LB), listed from top to bottom. The MB member acts as a reservoir rock and the target pay-zone for
horizontal wells with a thickness of 85 ft. The formation is characterized by low porosity (<10%) and ultra-
low permeability (<0.01md). The total organic carbon (TOC) content in the middle member is low ranging
from 0.1 to 0.3 wt.%, which is an organic poor compared to upper and lower members, as illustrated in
Fig. 5. Based on these reservoir properties, the MB Layer is classified as an unconventional play (Ellafi et
al., 2019b; Ba Geri et al., 2019c; Assady et al., 2019; Kurtoglu, 2013; Klenner et al., 2014). Furthermore,
geostatistics modeling has indicated that the reservoir properties in the MB Member is highly heterogeneous
throughout the Williston Basin (Jin et al. 2017). Also, the fluid transport is governed through the fracture
networks (natural and artificial fracture) at both the micro and macro-scale (Klenner et al., 2014). History
field data of long-term pressure and rate transient analysis (RTA) of the Bakken wells identify that natural
fractures can play a significant role in unconventional wells performance as well as can have both positive
and negative impacts on the effectiveness of an EOR scheme (Kurtoglu, 2013). In addition, reservoir
8 SPE-200001-MS

properties, especially the wettability of the reservoir is really important and must be investigated, which
has the main role in the relative permeability, capillary pressure, electrical properties, and EOR process
(Zhu et al. 2011).

Figure 7—Schematic of Bakken Petroleum System stratigraphy (Sorensen J., 2016; Klenner, et al., 2014).

To assess and gain a deeper understanding of the CO2-EOR mechanisms in unconventional reservoirs,
rock and oil properties of the Bakken Formation as well as CO2-fluid-rock minerals interaction need
carefully reviewing to address opportunities and challenges in the Williston Basin.

Bakken Reservoir Permeability

Assady et al., 2019 conducted an experimental study to characterize the MB core samples using both steady
state and unsteady-state permeability measurements. They explored the range of reservoir permeability of
the core samples from the Mountrail County wells, ND, which are the same wells and depths of our study.
Fig. 8 presents the comparison between the three methods used, and the conclusion showed that the pulse-
decay permeability measurement is in close agreement with the permeability range of the tight formation in
the literature. Kurtoglu, 2013 pointed out that the permeability magnitude in the laboratory is 10-4 to 10-5 mD
while in the field scale measurement is 10-1 to 10-2 mD due to significant contribution of a micro-fracture
network. In the field application, the only way to estimate the reservoir permeability is from a diagnostic
fracture injection test (DFIT), and the research by Melcher et al., 2020 summarized the analysis of thirty-
three DFIT tests and reservoir permeability in the MB Member reported as mean value around 0.022 mD
that represents the upper-bound estimation value.
SPE-200001-MS 9

Figure 8—Comparison between three methods of the Middle Bakken permeability measurements (Assady et al., 2019).

Despite research findings of Assady et al., 2019 observed that the pulse-decay is associated with some
of the errors; its outcomes are still more reliable compared to steady state and oscillating approaches. In
the case of complex pore throat configurations, the steady state and oscillating methods are associated with
high errors since these approaches are used to determine core samples in medium and high permeabilities
ranges using Darcy’s law. On the other hand, the foundation of pulse-decay based on the Darcy-Klinkenberg
Fibonacci calculation model that reduces the uncertainty for permeability measurement for unconventional
reservoirs. The results of the experiment showed a significant effect on the reservoir permeability due to
change in confining pressure and pore pressure. As can be seen in Fig. 8, the permeability in the three
methods is high at low pressure, while the reservoir permeability is gradually decreased as confining
pressure and pore pressure are increased. Therefore, the outcomes using pulse-decay, steady state, and
oscillating methods are decreased by 75%, 68%, and 56%, respectively when the pressure is increased from
1,045 psi to 2,205 psi.
A clear statement was proposed in the work of Assady et al., 2019 that the running of the hysteresis
technique can be modeled the fluid flow during EOR (injection and soak), as well as production depletion in
unconventional reservoirs. The matrix permeability in tight formations is stress-dependent that significantly
impact early and long-term production. Thence, understanding EOR mechanisms in unconventional
reservoirs, such as molecular diffusion rate relies on the permeability hysteresis that highlights the impact of
micro-fractures and nano-pores in tight formations under different conditions of pore pressure and confining
stress.
Teklu et al., 2018 addressed the importance of permeability hysteresis information in reservoir
development strategy by studying of fracture and matrix permeability dependency on stress during loading
cycle (depletion process) and unloading cycle (injection process). They obsereved that the major role of
the stress dependency and hysteresis of permeability is larger in nano-pores than mirco-proes media. The
pore interconnectivity of the effective porosity and permeability of the tight formation depends on the net
effective (σeff) stress that govern the deformation of the rock and can affect material properties (Civan,
2019). This term is defined as the difference between applied confining pressure (external stress, (σ) and
the internal pore pressure, (p), as described in Eq. (1) (Terzaghi, 1943).
(1)
10 SPE-200001-MS

Where α is a porirelastic (Biot’s) coefficient.

For example, Fig. 9 shows during loading cycle that represents production depletion (depressurization
mechanism), the pressure drop in pore pressure is equivelent to increase in applied effective stresses to
the reservoir, which leads to siginificant decrease in the MB core samples permeability up to 37%, and
microcracks could not be open to enhance bulk permeability. On the other hand, during unloading cycle
(injection) when the effective stress is decreased, a siginifcant hysteresis with 24% restored permeability
was observed for the Bakken cores (Teklu et al., 2018). Fig. 10 shows similar trend behavior in the work of
Assady et al., 2019, where the reduction observed is around 31%. It can be concluded that the fluid flow in
unconventional formations depends on reservoir properties and micro-pore structure that influenced through
stress changes. As a result, the reservoir petrophysical properties may change during the production and/or
injection process, and this leads to affect the storage capacity of the formation. Furthermore, unconventional
reservoir properties are not important for only understanding primary depletion process, but also may play
an important role in success in the implementation of CO2-EOR and/or CO2 sequestration applications,
where the stress dependency might cause an altering of permeability magnitude (Wang et al., 2019). Also,
in high reservoir temperature reservoir, such the Bakken reservoir (220 to 240 deg F), the increasing in
temperature enhances the diffusion mechanism, which might lead to increase the effective permeability of
the rock matrix. In fact, the reservoir permeability is much effected by confining pressure than injection
pressure, where minimizing the permeability change due to injection pressure can be done by extending the
injection pressure period until it reaches the optimum time (Wang et al., 2019).

Figure 9—The permeability hysteresis of the Bakken core samples during loading and unloading cycles (Teklu et al., 2018).

Figure 10—Loading/unloading effects on permeability of Bakken core samples (Assady et al., 2019).
SPE-200001-MS 11

Bakken Reservoir Porosity

In unconventional shale reservoirs, the pore structures are a crucial factor that can affect the site for
accumulation and storage capacity of hydrocarbon include chemical properties of the rock. In addition, the
pores provide the flow paths (permeability), which might influence the fluid flow mechanisms during the
production and EOR process. Unlike conventional reservoirs, there are three types of existing porosity:
matrix porosity (inorganic), organic porosity, and fracture porosity with pore sizes ranging from nanoscale
to microscale level. It is necessary to answer the question of the nature of fracture networks in the scale of
macro and micro within the reservoir properties to consider the potential use of CO2 injection in the Bakken.
The high resolution of SEM images (Scanning Electron Microscopy) is the most used method to quantify
porosity in shale reservoirs, where multiscale pore level can provide rock characterization information
such as pore shape, pore size, pore location, and pore connectivity. (Liu et al., 2018; Liu and Ostadhassan,
2017) these authors reported that the matrix and fracture porosity are main dominated hydrocarbon storage
capacity in the MB Formation, while Eagle Ford shale both organic and matrix porosity prevails the major
storage volume. Furthermore, the porosity outcomes for the MB Member is low, most samples that measured
by Liu et al., 2018 reported porosity below 10%. This confirmed by Luo et al., 2019, and Fig. 11 presents
porosity distribution ranges from 2 to 10% over various counties in the Williston Basin, ND. In another
study by Kowalski and Sonnenberg, 2013, the porosity showed varying values by changing the facies,
where they concluded that facies with high calcite cement and illite/smectite (clay) provide low to no
porosity. In contrast, the high porosity presents if the formation contains a high concentration of quartz, as
shown in Fig. 12. The secondary porosity and both induced and natural fractures were observed also using
photomicrographs, (see Fig. 12, Facies E & F). This causes by a large amount of dolomite as a result of
the dissolution process and diagenetic alteration during burial and forming a history of the rock. Overall
reported results from the literature review conclude that the average porosity in the MB Formation is around
6% with significant connected and open pore space, as shown in Fig. 13.

Figure 11—Porosity distribution map in the Middle Bakken, Wiliston Basin (Luo et al., 2019).
12 SPE-200001-MS

Figure 12—Photomicrographs of the Middle Bakken facies, Wiliston Basin (Kowalski and Sonnenberg, 2013).

Figure 13—FIB-SEM analysis of the Middle Bakken lithofacies, Wiliston Basin (Sorensen et al., 2016).

In 2016, Sorensen et al. proposed advanced analytical method to characterize twenty-six samples
from the Bakken Formation using an integral approach, which is a combination of advanced computer
tomography (CT) imaging and scanning electron microscopy (SEM) techniques, including whole-core and
micro x-ray CT imaging, field emission (FE)–SEM, and focused ion beam (FIB)–SEM. The multiscale
workflow was utilized to improve and support the characterization technique of the geomodels analysis for
more representative studies of fluid flow pathways within various lithofacies of the MB Formation. The
observation from this study show evidence that the network of pore structures in the MB Formation could
serve as the main pathway to diffuse CO2 into the rock matrix through the diffusion-dominated flow and then
mobilized more hydrocarbons by viscosity reduction mechanism to keep the viscuss flow in the fractures, as
shown in Fig. 13 significant connected porosity was observed using FIB-SEM analysis. Fig. 14 demonstrates
the laminated zones in the MB Formation, which is geomechanically weaker with higher porosity and
permeability, but also tends to micro-fracturing as a result of fracture treatment job, and characterization
applications approved that these lithofacies are most favorable zones to CO2 EOR more easily than others
(Sorensen et al., 2015).
SPE-200001-MS 13

Figure 14—SEM images and mineral map of the Middle Bakken

lithofacies "laminated zone", Wiliston Basin (Sorensen et al., 2015).

As shown in Fig. 14, might answer the question of the role of fractures of CO2 movement into the rock
matrix. A fracture network in the reservoir can serve as a means of more beneficial effects of CO2 to expose
more oil-saturated rock when the generated fractures are higher density and limited near the wellbore. Then,
the surface area of the formation is a lager contact with CO2, and lower maintained pressure is needed to
optimize more effective miscible CO2. Otherwise, CO2 quickly migrates away from the productive area,
and higher pressure is required to operate CO2-EOR (Sorensen et al., 2015).

Bakken Rock Mineralogy

The mineral composition of the unconventional reservoir is a key factor for successful IOR/EOR
applications, which is crucial to understand the pore structure and estimate the fluid transport and storage in
the nano and micro scale. Additionally, the rock composition is fundamental that has a main role to assess
the best zone for hydraulic fracturing jobs, thus improving the hydrocarbon oil recovery from tremendous
oi resource, such as Bakken Shale Play (Jarvie et al., 2007). The rock mineralogy directly relates to
geomechanical properties (i.e, Poisson’s ratio, Young’s modulus, ect.), which can reduce or increase the
brittleness of the tight formation and provides a poor or significant fracture network between the wellbore
and micro and nano-pores media. The brittleness index is defined as the ability control to create and maintain
the fracture, and the research findings from previous studies concluded that the MB Formation is more
densely and prone to develop fractures compared to other shale plays due to low clay content (Bhattacharya
and Carr, 2019; Kurtoglu, 2014). For the EOR application, the high brittle formation generates a large
surface area that counts as advantages, and the SRV would be exposed to a large volume of the gas injection
and unlimited solvent oil contact to increase incremental oil recovery.
Li et al., 2019 and Liu et al., 2017 studied the characterization of the Bakken Layers using X-ray
diffraction (XRD) analysis in the laboratory. The research aimed to determine the physical parameters
and deposition environment of the Bakken Shale by experimenting with more than twenty-five samples
from different locations to identify the mineral composition, clay type, and clay abundance of the Bakken
Formation. Their results were gathered and analyzed using the box plot, as shown in Fig. 15, where the
mineralogy analysis in the MB Formation is described as quartz, pyrite, feldspar, clay, dolomite, calcite,
and ankerite with median values, as listed in Table 1. The dominated components in the MB Formation
are quartz, feldspar, clay, and dolomite. The main clay types are: smectite, chlorite, kaolinite, and illite/
mica. All previous studies agreed that the primary clay type is illite in moderate amounts of around 13.1%
and the lithology of the layer is defined as a mix of sandstone and limestone (Kurtoglu, 2014). The tight
14 SPE-200001-MS

oil reservoirs are typically sedimentary rocks, which contain unstable minerals (e.g., calcite, dolomite, K-
feldspar, and albite).

Figure 15—Mineral analysis results of the Middle Bakken Formation samples.

Table 1—Statistical analysis of the Middle Bakken Formation mineral composition.

Quartz Pyrite Feldspar Dolomite Calcite Ankerite

Clay (wt %)
(wt %) (wt %) (wt %) (wt %) (wt %) (wt %)

Minimum 15.24 0.10 5.36 9.36 2.25 0.00 0.00

Q1 17.70 1.39 7.17 14.83 11.35 7.53 0.00
Median 31.14 2.62 10.85 19.19 21.07 9.68 0.00
Q3 38.48 3.57 13.15 25.19 32.92 14.82 0.00
Maximum 42.56 4.98 25.36 62.94 38.57 25.70 0.56
Mean 29.22 2.57 11.37 24.66 20.78 10.93 0.06
Range 27.32 4.88 20.00 53.58 36.32 25.70 0.56
St.D 6.83 1.22 5.00 13.40 9.08 6.43 0.14

Q1: First quartile (25%), Q3: Third quartile (75%), and St. D: Standard deviation.

One of the remaining questions that is required to answer is how CO2 interacts with the rock matrix
mineralogy. Al Ismail and Zobach, 2017 highlighted the point that the contact between CO2 and rock
mineralogy can not be ignored, which is a crucial factor to understand the behavior of permeability
changes, the transport mechanism of CO2 in nanoporous reservoirs, and design the optimum CO2-EOR
implementation pilot test. Beside XRD analysis, the bulk chemistry of the MB Formation is very important
to detect and quantify, where x-ray fluorescence (XRF) was used in the work by Jin et al. 2016, as shown in
Fig. 16. The outcomes clearly show that silicon, calcium, and aluminum are the most chemical elements in
the MB Layer with a significant percentage of calcium as a result of the basin composition of carbonate rock
contains an abundant amount of calcium. The reservoir rock is classified as hydrophilic when the formation
is rich with silicon and aluminum and poor with iron.Then, the wettability of the formation may be induced
by CO2-EOR, which causes to change the petrophysical properties of the whole reservoir (Zhu et al. 2011).
The interaction in the soaking period between CO2 and formation rock can partially dissolve the skeletal
grain minerals and cement of the rock. As a result, the rock mineralogical and pore structural properties
alter and the mineral dissolutions can generate macropores or even microfractures that result in altering the
properties of the rock mineralogy (Li et al. 2020).
SPE-200001-MS 15

Figure 16—X-ray fluorescence (XRF) analysis for the Middle Bakken core samples (Jin et al. 2016).

Heller and Zoback, 2014, conducted experimental work to investigate the relationship between gas
adsorption capacity and TOC and minerals that represents Barnett, Montney, Marcellus, and Eagle Ford
Formations. The results showed that the high TOC value, the maximum absolute adsorption capacities,
where Barnett and Montney have TOC in range of 5.3% and maximum adsorption around 40.8 gmole/g.
Also, the rich shale members contain the kerogen that is the primary organic matter component that has
micro porosity, spanning micrometer to nanometer in scale. On the other hand, the study by Smith et al.
2019 concluded that the MB Formation is considered as lower adsorptive potential with ten times compared
to UB and LB Layers due to its low amount of clay, as shown in Fig. 17.

Figure 17—Carbon dioxide adsorption isotherms on samples from the Middle Bakken Formation (Smith et al. 2019).

In the work by Kurtoglu, 2013, they discussed that the main pathway for solvent injection into a tight
shale play would be through the micro-fracture networks for transporting the solvent into the tight matrix.
Obviously, the abundance and favorable distribution of fracture networks yield better solvent exposure
and control the contact time that the solvent has with the oil and thus higher Bakken wells productivity in
an EOR process. In addition, the Bakken Formation is characterized by the oil-wet system that using the
waterflooding process will not be an effective technique. Therefore, based on the above characteristics of the
MB Formation that approve the formation is under a high-intensity naturally fractured and/or hydraulically
fractured formations, the solvent injection (e.g. CO2) could be an effective EOR technique.
16 SPE-200001-MS

CO2 Huff-n-Puff Protocol in Unconventional Reservoirs

The results from several studies on gas injection EOR, such as carbon dioxide (CO2), nitrogen (N2), rich
natural gases (C1 "70%" and C2+ "30%"), and lean gas (primary methane) proved that the application is a
more feasible EOR technique in shale plays due to their low viscosity and easier transport into matrix that
dissolves in oil and cases oil swelling and reduced viscosity (Alfarge et al., 2017; Sheng, 2015). Furthermore,
most of the recent studies reported that CO2 injection is preferred application due to several reasons (Alfarge
et al., 2018; Zhang et al., 2016; Kurtoglu et al., 2014): 1) CO2 requires lower miscibility pressure to reach
miscibility than other gases, which makes CO2 miscible injection attainable under wide range of reservoir
pressures. 2) The injectivity of CO2 injection is favorable as a result of continuous gas pathways from
fracture and penetrating into the rock matrix to swell oil through the diffusion mechanism. Due to their
ultra-tight permeability in shale plays, the simulation studies and experiment works have proved that the
single-well HNP (cyclic gas injection) process would be an effective EOR approach for LRS reservoirs.
A CO2 flooding process as compared to HNP may take a considerably long time for the injection transient
pressure to propagate towards the producing well. Sheng, 2015 showed that in fractured reservoirs (naturally
and fracked wells), the HNP approach has a better recovery performance compared to the gas flooding
process. They explained that the complex fracture networks around a production well could cause early
breakthrough or viscous fingering which means poor sweep efficiencies or an underperformed flooding
performance. Therefore, the industry has resolved this issue by drilling a single horizontal well that acts
as both production and injection well (Wan and Sheng, 2015). Also, the presence of natural fractures will
benefit the recovery by increasing the contact area of the injected solvent with reservoir rocks. Fig. 18
illustrates the method, which starts with a huff stage by injecting the solvent at high pressure; then the first
or multiple contact miscibility occurs, where the injection fluid mixes with the formation oil, resulting in
a single-phase fluid flow. Next, a soaking period begins when the well is shut-in for a specified period.
In this step, the CO2 diffuses into the rock matrix and interacts with the host rock and fluid. Then, as the
puff step, the well is put on the production for a certain period and under specified depletion pressure so
that the mixture of CO2 and shale oil expand and flow to the surface in one phase. This application has
been successfully implemented for the steam injection process, but it has also been used recently to face
challenges of tight formations during CO2-EOR injection.

Figure 18—Typical gas HNP process for unconventional EOR application (Pankaj, et al., 2018).

Generally speaking, understanding of the key parameters controlling the CO2 HNP process requires
testing the recovery under several scenarios with different well, reservoir (rock, fluid, rock-fluid, etc.), and
operational parameters. The oil viscosity reduction ratio with CO2 injection and oil swelling factors as well
SPE-200001-MS 17

as recovery at different conditions (e.g. p, T, soak period, # of injection cycles, etc.) are summarized below
in this work. The HNP process was firstly evaluated by Wan et al. 2013, where they used numerical reservoir
simulation to quantify the oil recovery from this EOR process in shale oil plays. They concluded that fracture
networks are crucial for improving oil recovery in ultra-tight shale formations. Furthermore, Hawthorne et
al. 2013 conducted experiments using core samples size 10 mm diameter in 40 mm long, to determine the
effects of CO2 exposure time on recovering hydrocarbons from the UB, LB, and MB formation at conditions
of 110°C and 5,000 psi. Their results indicated that CO2 injection is a promising method for enhancing oil
recovery from both source and reservoir rocks of the Bakken if the operations meet the two conditions of
long exposure time and wide contact area.
Moreover, Gamadi et al. 2013 performed an experimental study of the HNP process using N2 as the
injection fluid on core plugs of the Barnett, Marcos, and Eagle Ford shales. They studied the impact of
operating pressure, shut-in time, and a number of cycles on the N2 injection performance. Their results
indicated that the peak recovery factor can be reached if the injection pressure is near miscibility. The work
of Hawthorne et al. in 2013, 2014 and 2017 gave the observation about the effect of minimum miscibility
pressure (MMP) on the incremental oil recovery, where they approved in fractured tight reservoirs with
light hydrocarbons like Bakken, the recovery process does not rely on the flushing mechanism but strongly
controlled by solubility/diffusion mechanisms in a soaking period. The reason behind that the injection
solvent would favor lighter hydrocarbon components due to its higher solubility, then the diffusion rate is
improved. In contrast, Tovar et al. 2018 discovered that in organic-rich shale, increasing pressure beyond
the MMP leads to increase oil recovery factors. This major change in operation philosophy compared to
inorganic formations due to differences in mechanisms that taking place during CO2 injection. Moreover,
the number of cycles is also critical design parameter, and the observed outcomes from previous works
concluded the majority of oil production is in the first and second cycles, and then the recovery stabilization
has reached. In addition, the depletion condition appeared to be a crucial factor as well to design the HNP
operation in order to enhance oil recovey application.
In 2014, Gamadi et al. repeated the previous experiments using CO2 injection and the same core samples
and operation conditions. The results supported the hypothesis that higher pressures than MMP may not be
an effective strategy to increase the oil recovery from tigh reservoirs. Tovar et al., 2014 discussed the results
of two experiments using the HNP CO2 method on Barnett core samples. They modified a Hassler core
holder to simulate CO2 HNP process by surrounding the core samples by glass beads (to emulate fractures)
and plugging both ends by two Berea sandstone to allow the high-pressure CO2 always in contact with the
matrix (cores) and prevent the glass beads to escape the chamber. Therefore, the high permeability media
surrounding the core samples was saturated with the solvent at constant pressure (1,600 and 3,000 psi)
and temperature (150 °F) during the experiment. They reported that the high permeability media (the glass
beads) provided a high surface area to perform CO2 injection, while it is not possible when CO2 is injected
directly into the core sample. Such a new design would resolve the problem of low to zero injection into a
tight core from a shale play. The significant improvement in the incremental oil recovery was observed, and
the estimated recovery was between 10 to 55% of the pore volume of the core samples. The methodology
used in the experiment (x-ray computed tomography, [CT]) indicated that the increase in oil volumes was
driven by diffusion and reduction in capillary forces.
Lately, in 2017, Jin et al. conducted an experiment on CO2 exposure for some Bakken core samples
with small dimensions (1.1 cm diameter and 4 cm in length) at the pressure and temperature of 5,000
psi and 230 °F, respectively. They aimed at a better understanding of the microstructure and diffusion-
dominated flow in ultra-low permeability formations during a CO2-EOR process. The results concluded
that CO2 is able to extract more oil recovery as high as 68% during 24 hrs exposure time from the UB
and LB formation. Both layers content generally high content of total organic content (TOC) in a range of
10-15 wt%, and small pore-throat size in a range of 3-7 nm, known as mesopores with important number
18 SPE-200001-MS

of micropores. These factors can impact the oil recovery due to their roles in residual oil trapping. The
shale formations with a high amount of TOC contain kerogen, which is organic matter, and its surface
is oil-wet with complex pore structure and confining oil inside. As a result, CO2 could not diffuse and
displace hydrocarbon molecules easily due to large capillary pressure. The presence of organic matter in
unconventional reservoirs has significant impact on the EOR mechanisms and reservoir depletion behavior.
On the other hand, the core samples from MB and TF formations have larger pore sizes (>50 nm), known
as macropores and low TOC level, which assists in a more favorable flow for both CO2 and hydrocarbon
molecules. The ultimate oil recovery reported is 99% of the total pore volume during 24 hrs of CO2 exposure
under Bakken conditions. In the work by Tovar et al., 2018, CO2 injection was studied to investigate the
operation philosophy and understand the recovery mechanisms. They used a similar Hassler core holder to
simulate the behavior of fractures around the core (i.e. rock matrix). Their work provides some experimental
observations on the effects of pressure, soak time, rock transport, and oil and injection gas compositions
on recovery mechanisms.
CO2-EOR technologies have the potential to add millions of barrels of incremental production to the
Williston Basin oil recovery. In general, CO2 injection is a fast method with promising potential that might
succeed economically compared to refrac and infill drilling applications. CO2 operations include capture,
compression, and transportation has started to consider as economic application as a result of the extension
and expansion of Federal 45Q tax credits, which provides $50/ton for CO2 stored for 12 year period in saline
aquifers, while $35/ton for CO2 captured in depleted formations during CO2-EOR. To provide economy of
scale and, potentially, additional subsidy for saline aquifer injection through CO2 sales, CO2-EOR likely
needs to be part of the system (Holubnyak et al., 2019). However, there is a wide range of uncertainties
associated with operating such a process in shale plays, which need to be addressed prior to meaningful
pilot tests in the field.

Methodology Details
Experimental Setup
Fig. 19 shows the schematic of the experimental setup used in this work to simulate the HNP process.
The setup consists of high-pressure vessels containing up to three rock samples together in each run. As
illustrated in the figure, the core and chip samples were numbered and loaded as a stack. The space between
the vessel and core samples represents the fractures with high permeability to allow CO2 to contact with the
whole core samples. The stainless-steel vessel is designed for high pressure and high temperature conditions
to mimic the operational parameters in the Bakken formation. The vessel is connected to a CO2 accumulator
and both are placed in an oven with a maximum operating temperature of 300 °C. The gas is supplied from
a commercial CO2 cylinder with the purity of 99.99% under a maximum pressure of 900 psi. If pressures
higher than 900 psi is desired, a syringe pump can be used to increase the pressure depending upon the
operational conditions of the experiment. The syringe pump works by injecting distilled water to fill up the
accumulator and push the piston up in order to increase the CO2 pressure in the system. All components of
the system are connected to a data acquisition system to monitor and control the temperature and pressure
by using two transducers in the vessel and CO2 accumulator. This system is a modified design of the original
coreflood setup in order to simulate the HNP process in a tight shale play (see Fig. 19).
SPE-200001-MS 19

Figure 19—The experimental setup used for testing a typical huff-n-puff process on Bakken core samples.

The materials used in the experiments are as follows:

1. Dead oil Bakken fluid was used to fully saturate the rock and chip samples, where the properties of
the fluid are listed in Table 2.

Table 2—Bakken crude oil properties.

Oil Type Density, gm/cc Viscosity, cp

Dead oil 0.86 2.46

2. Six core and chip samples from the MB formation from two different wells and depths were studied to
evaluate their recoveries and to assess the proceeding wettability alteration by measuring the contact
angle before and after each HNP run. Table 3 shows the well number, formation depth, dimensions,
surface areas, and bulk volumes of the samples, respectively.

Table 3—Core sample information, dimensions, surface areas, and bulk volumes.

Formation
Sample, # Well, # L, in D, in SF-A, in2 BV,CC
Depth, ft
1 24779 10,242.6-10,248.4 1 1 1.5708 12.8704
2 24779 10,242.6-10,248.4 1.5 1 2.3562 19.3056
3 24779 10,242.6-10,248.4 2 1 3.1416 25.7408
4 25688 10,645.5-10,680.0 1 1 1.4945 12.7648
5 25688 10,645.5-10,680.0 1.5 1 2.4752 19.4167
6 25688 10,645.5-10,680.0 2 1 3.2468 25.8436
C 25688/24779 10,242.6-10,680.0 Chip size = 0.394 in × 0.394 in

3. A precise scale was utilized to weigh the core samples before saturation, after saturation, and at each
CO2-EOR experiment run in order to determine the incremental oil recovery.
20 SPE-200001-MS

4. A wettability tester was used to determine the contact angle of the chip samples in order to study the
wettability alteration from CO2-EOR. In this research, the contact angle was measured to assess the
wettability in different scenarios in order to identify the conditions under which wettability would act
in favor of a CO2 HNP process.

Experimental Protocol
Core Saturation Process. This study was started with cleaning all core samples using the Dean Stark
extraction, where the mixture of toluene and methanol was used as cleaner solvent at low temperatures to
remove the fluid and salt contents by vaporizing the solvent mixture. The cores are placed in such a fluid
mixture for almost a week until the color of the mixture shows no more change. Next, the cores are placed
in the oven at 70 °F overnight to dry and are weighed by using a high accuracy point scale. Then, the same
apparatus, as shown in Fig. 20 is utilized to saturate the core samples with the Bakken dead oil.

Figure 20—Saturated the MB core samples.

The samples were loaded in the high-pressure vessel and the accumulator filled with certain amount of
the Bakken oil. Then, the oil was injected from the accumulator to the high-pressure vessel using a syringe
pump. In case of unconventional rocks, the saturation pressure is adjusted between 1,500 to 2,000 psi, just
not too high to avoid breaking the core samples. This process took almost five days to allow the oil to
penetrate deep into the rock samples. Finally, the samples are saturated, as shown in Fig. 20, then weighed
and the pore volume can be estimated from Eq. (2). The results of this step are shown in Table 4 where
the cores, taken from different depths, present different pore volumes due to the high heterogeneity in the
MB formation.

Table 4—The saturation process results.

Sample, # Wt. D, gm Wt. Sat, gm PV, CC

1 37.2840 37.5995 0.3668

2 49.2208 49.7635 0.6309
3 62.3863 63.8770 1.7330
4 36.911 37.1112 0.2328
5 50.1234 50.7384 0.7151
6 62.9674 64.4776 1.7560
SPE-200001-MS 21

(2)

Where:
PV: Pore volume, cc
Wtsat: Weight of the core sample when saturated, gm
Wtdry Weight of the core sample when dry, gm
ρoil Bakken oil density, gm/cc
Experiment Procedure of Huff-n-Puff CO2-EOR. The huff-n-puff experiments were conducted on the
six MB core samples and rock chips with different characteristics. As shown in the schematic of the
experimental setup (Fig. 19), the whole surface area of the cores is exposed to the injected CO2. Before the
experiment begins, the original wettability of the cores (chips) were measured through the contact angle.
Next, the core samples were loaded in the high-pressure vessel. Then, a four-step procedure was followed:
a) inject CO2 injection at a certain rate and temperature until it reaches the desirable pressure, b) close the
system and let it soak with CO2 for a specified soak period so that the CO2 can diffuse into the cores, dissolve
in oil, and finally reduce the oil viscosity, c) depressurize the system gradually to simulate depletion process
to assist the CO2 to swell the oil out of rock matrix, d)open the vessel and measure the weight of each core
sample and determine the oil recovery using Eq. (3).

(3)

Where:
RF at eachcycle i: Recovery factor, fraction.
Wti: Weight of the core sample after each CO2 injection cycle, gm.
In addition, the factor of shut-in time was investigated, where the samples were soaked for three, four,
ten, twelve, twenty-four hours, and two days. Furthermore, seven injection cycle were applied to study its
effect on the recovery factor. Moreover, five steps were used to show the impact of depletion pressure on
the incremental oil, where the core sample was depleted during one hour from 3,500 psi to 1,800 psi.
Contact Angle Measurements. Different methods have been introduced the wettability measurements,
where the approaches are classified as quantitative and qualitative methods. The most typical quantitative
method is used to measure the wettability of a rock-aqueous phase fluid system is the contact angle as present
in the Young equation Eq. (4) or using the force acting on the balance, as written in Eq. (5). The contact
angle is the angle formed between the liquid and solid interfaces. In this study, the wettability alteration
from CO2-EOR was investigated through the measurement of contact angle for each rock chip of MB wells.
By using, the equipment in our laboratory can handle the chip size with the dimensions, as listed in Table
3. First, the chip sample of the MB formation is placed in a cell before injecting CO2. Next, distilled water
(1 µL) drop is injected into the cell on the top of the measured surface followed by oil in the air at room
conditions. A high-resolution camera is used to record the oil drop evolving on the rock surface. After
that, Image analysis is performed on the drop formed using the provided software. Then, the contact angle
measured between the edge of the oil drop and the rock surface. Fig. 21 a & b presents large contact angle
values that indicate less water wet occur on the rock surface when the angle measurement is higher than 90
degrees. On the other hand, lower contact angle values when the contact angle results below 90 degrees and
show a more water-wet surface. The main factors that can affect the wettability of the reservoir formations
are the complexity of the rock, reservoir temperature, reservoir pressure, gas properties, liquid properties,
and the rock surface properties, including rock mineralogy (Craig, 1971). This experiment was repeated
several times at the Bakken temperature and under different CO2 operation pressures.
22 SPE-200001-MS

Figure 21—Contact angle measurements (a) and wettability conditions for different rock samples (b).

(4)

(5)

Where:
γwg: The interfacial between the aqueous phase and the gas phase, N/m.
γsg: The interfacial between the surface and the gas phase, N/m.
γsw: The interfacial between the surface and the water phase, N/m.
θ: The contact angle, degree.
F: The measured vertical force, N.
Fb: The buoyancy force, N.
P: The wetted length, m.
γL: The surface tension of the test liquid, N/m.

Sensitivty Runs to Resuce Uncertainty

Effect of Temperature on Recovery
Sensitivity runs were performed to study the impact of temperature on oil recovery from a HNP CO2
injection under a constant injection pressure of 875 psi (i.e. gas phase). As depicted in Fig. 22, three
temperatures were considered: 70°F (lab temperature), 120°F (mid-range), and 220°F (average temperature
in the Middle Bakken).
SPE-200001-MS 23

Figure 22—The effect of temperature on oil recovery factor at 875 Psi; duration of CO2 huff-n-puff.

The results indicate that there is a relatively high increase in oil recovery (3.56%) from 70°F to 120°F
while it reaches a plateau at high temperatures, i.e. from 120°F to 220 °F. The reason is that the injected
CO2 will not be miscible at those lower temperatures while it becomes first-contact miscible at the injected
pressure and high temperatures. The improvement in oil recovery was observed from low to medium
temperatures till the recovery factor reaches the plateau and the stable conditions beyond which we may
not observe substantial increase in recovery factor.
Fig. 23 presents the effect of temperature under a higher injection pressure of 1,600 psi, but it is still
below the miscibility pressure. It shows a similar trend as observed in Fig. 22. The incremental oil recovery
is increased by 4.46 when temperature increases from 70°F to 120°F. The incremental oil recovery increase
is again higher from low to medium temperatures compared to medium to high temperatures. Furthermore,
24 SPE-200001-MS

Figure 23—The effect of temperature on oil recovery factor at 1,600 Psi; duration of CO2 huff-n-puff.

Fig. 24 presents similar results for the cases under 3,500 psi injection pressure (above MMP). This graph
demonstrates that the impact of temperature at higher pressures is more substantial since the system reaches
the miscibility under medium to high temperatures. The higher improvement in the oil recoveries in this case
is mainly due to the miscibility of CO2 which diffuses better into the rock sample and lowers the viscosity
and swells the oil

Figure 24—The effect of temperature on oil recovery factor at 3,500 Psi; duration of CO2 huff-n-puff.

Effect of Injection Pressure on Recovery

The core samples from two Bakken wells were acquired and used in this study to evaluate the effect of CO2
injection pressure (phase behavior) on the incremental oil recovery in one HNP cycle and at the Bakken
Formation temperature (~220 °F). Fig. 25 and Fig. 26 show that the injection pressure has a substantial
SPE-200001-MS 25

impact on the incremental recovery, especially, when it is elevated to MMP and beyond. Obviously, at
those high pressures CO2 is miscible and will diffuse more easily into the rock samples that reduces the
viscosity and causes oil swelling. In the cases with near miscibility injection pressure, the experiments
show promising improvements in oil recovery compared to that under 714 psi injection pressure. The
ultimate recovery from both samples under similar operating conditions show almost equal results being
50% recovery factor after three cycles.

Figure 25—The effect of injection pressure on recovery at 220 °F for sample #1.

Figure 26—The effect of injection pressure on recovery at 220 °F for sample # 2.

Effect of Number of Huff-n-Puff Cycles

In this study, the recovery factor was investigated from one cycle and seven cycles of huff-n-puff. This
helps us to better understand the effect of the number of cycles on both recoverable oil and to identify
optimal number of cycles. Samples were used in the huff-n-puff experiments at the Bakken Formation.
26 SPE-200001-MS

temperature (220 °F) and under the pressure 2,000 psi (near miscibility). The results from both samples
show improvements in recovery factor, but the number of cycles needs to be optimized (see Fig. 27).

Figure 27—effect of number of huff-n-puff cycles on oil recovery at 220 °F and 2,000 psi.

Effect of Soaking Time on Recovery

The effect of soaking time was studied for two samples with different bulk volumes (i.e. contact area) and
pore volumes. The oil recovery increases as high as 6% only from soaking sample 24 hrs longer. A higher
soaking time means more time for the injected CO2 to diffuse into matrix and interact more with the oil and
thus higher recovery factor (see Fig. 28 and Fig. 29). However, this parameter needs to be optimized since
there is a tradeoff between higher soaking time and lower rate of recovery which impacts the economy of
the EOR project.

Figure 28—The effect of soaking time on CO2 huff-n-puff recovery at 220 °F and 3500 psi.
SPE-200001-MS 27

Figure 29—The MB core samples after CO2 injection.

Effect of Wettability on the Recovery from CO2 Huff-n-Puff

Generally speaking, the wettability of the Bakken Formation tends to be more oil-to intermediate-wet. The
wettability of some Bakken rock-chip samples were measured using a contact-angle measuring instrument
that also confirmed this observation. The wettability was measured before and after rock exposure to CO2
from which the impact of CO2 on the state of wettability was examined. When the contact angle of an oil
droplet is lower than 90°, the rock is considered to be intermediate-to oil-wet (Fig. 30). The contact angle
of the sample was measured before applying CO2 (i.e. original wettability) and turned out to be ~99° which
indicates more oil wet.

Figure 30—Wettability conditions of different rock samples (Teklu et al., 2015).

The contact angles of the sample measured before and after applying CO2 huff-n-puff to study the impact
of CO2 on the state of wettability. It is clear from Fig. 31 that the contact angle decreases (i.e. becoming more
intermediate- to water-wet) as the CO2 injection pressure increases from 700 psi to 3500 psi (at reservoir
temperature) as a result of CO2 interaction with the host rock. In fact, the recovery increases as CO2 diffuses
into the matrix, alters the wettability in favor of oil flow (i.e. less oil wet), and reduces oil density and
viscosity from CO2 miscibility and oil swelling.
28 SPE-200001-MS

Figure 31—Contact angle measurements for the Middle Bakken Fm. at different conditions before and after applying CO2-EOR.

Conclusion
This work studies the potential of CO2 huff-n-puff in the Bakken Formation, Williston Basin and examines
the factors affecting the recovery performance. The production from the huff-n-puff experiments we
conducted on Bakken core samples resulted up to 50% ultimate recovery factor under different reservoir
conditions. The following results can be drawn from this work:
▪ The outcomes indicated on the effect of the reservoir temperature on the performance of the CO2,
where the recoverable oil increases as the temperature increase until reach the optimum depends on
the injection pressure phase. On the other hand, high injection pressure yielded higher amount of
produced oil.
▪ The surface contact area is a crucial factor for the diffusion mechanism, which plays a main important
role to contact the CO2 flow to whole core samples, then the oil production increases as the contact
surface area increases.
▪ As a previous research outcome, the number of cycling and soaking time are crucial design parameters
for the huff-n-puff experiment and on the field as well to let the CO2 time to diffuse into the deep
formation and swell more oil.
▪ The wettability condition behavior in the Middle Bakken Formation is generally found as the oil wet
to intermediate wetting phase, where the contact angle obtained bigger than 90 deg.
▪ The wettability alteration was changed by CO2-EOR as injection pressures increase and the wetting
phase move from the oil wet toward the water wet system.
▪ As overall outcomes from this research, the CO2 huff-n-puff process has a good potential in the Lab
and could be succeeded economically in field applications.

Acknowledgement
The authors would like to acknowledge the North Dakota Industrial Commission (NDIC), Petroleum
Research Fund for their financial support of this work, through the contract NDIC G-045-89. The financial
support of the North Dakota Industrial Commission (NDIC) is highly appreciated. Also, the authors would
like to thank the American Chemical Society Petroleum Research Fund (ACS-PRF) for their financial
support: through the contract of PRF No. 58746 for the grant: A New Approach to Quantifying Adsorption/
SPE-200001-MS 29

Diffusion Characteristics of Shale Formations. We also acknowledge the North Dakota Geological Survey,
Wilson M. Laird Core Library for providing us with core samples, and special thanks to Oasis Petroleum
Co. for providing the Bakken oil sample.

References
Ahmed, U, and Meehan, N., 2016. "Unconventional Oil and Gas Resources: Exploitation and Development", Taylor
Francis Group, Baker Hughes 2016, https://www.crcpress.com/Unconventional-Oil-and-Gas-ResourcesExploitation-
and-Development/Ahmed-Meehan/p/book/9781498759403.
Al Ismail, Maytham, and Mark Zoback. "CO 2-Based Technologies in Unconventional Resources: Impact of Rock
Mineralogy on Adsorption." SPE Kingdom of Saudi Arabia Annual Technical Symposium and Exhibition. Society
of Petroleum Engineers, 2017.
Alfarge, D., M. Alsaba, M. Wei, and B. Bai. 2018. Miscible Gases Based EOR in Unconventional Liquids Rich Reservoirs:
What We Can Learn. In Proceedings of the SPE Improved Oil Recovery Conference, Tulsa, Oklahoma, USA, 14-18
April 2018.
Alfarge, D., M. Wei, and B. Bai. 2017. IOR methods in unconventional reservoirs of North America: Comprehensive
review. In Proceedings of the SPE Western Regional Meeting, Bakersfield, California, 23-27 April 2017.
Assady, A., Jabbari, H., Ellafi, A. M., & Goudarzi, B. (2019, August 28). On the Characterization of Bakken
Formation: Oscillating-Pulse, Pulse-Decay Permeability Measurement & Geomeachanics. American Rock Mechanics
Association.
Ba Geri, M., Ellafi, A., Flori, R., Belhaij, A., & Alkamil, E. H. K. (2019c, November 11). New Opportunities and
Challenges to Discover and Develop Unconventional Plays in the Middle East and North Africa: Critical Review.
Society of Petroleum Engineers. doi: 10.2118/197271-MS
Ba Geri, M., Flori, R., Ellafi, A., Noles, J., Essman, J., Kim, S., & Alkamil, E. H. K. (2019b, November 11). Correlated
Friction Reduction and Viscoelastic Characterization of Utilizing the Permian Produced Water with HVFRs during
Hydraulic Fracturing. Society of Petroleum Engineers. doi: 10.2118/197748-MS
Ba Geri, Noles, J., Kim, S., & Ellafi, A. (2020). New Developed Mathematical Model for Predicting Viscosity Profile
and Proppant Transport Utilizing HVFRs Dosage with Produced Water. Society of Petroleum Engineers. doi:
10.2118/201433-MS
Bhattacharya, Shuvajit, and Timothy R. Carr. "Integrated data-driven 3D shale lithofacies modeling of the Bakken
Formation in the Williston basin, North Dakota, United States." Journal of Petroleum Science and Engineering 177
(2019): 1072-1086.
Chen, C., Balhoff, M.T., and Mohanty, K.K. 2014. Effect of Reservoir Heterogeneity on Primary Recovery and CO2
Huff ‘n’ Puff Recovery in Shale-Oil Reservoirs. SPE Res Eval & Eng14(03): 404-413. SPE-164553-PA. http://
dx.doi.org/10.2118/164553-PA
Cipolla, Craig and Wallace, Jon. 2014. Stimulated reservoir volume: a misapplied concept? Proc., SPE Hydraulic
Fracturing Technology Conference.
Civan, F. (2019, October 1). Effective Correlation of Stress and Thermal Effects on Porosity and Permeability of Naturally
Fractured Formations by a Modified Power Law. Society of Petroleum Engineers. doi: 10.2118/198893-PA
Continental Resources. 2018, https://clr.com/
DrillingInfo website. 2019, from https://info.drillinginfo.com/
EIA. 2019. Annual Energy Outlook 2019 with projections to 2050, U.S. Department of Energy, https://www.eia.gov/
outlooks/aeo (January, 2019).
Ellafi, A., & Jabbari, H. (2019a, August 28). Coupling Geomechanics with Diffusion/Adsorption Mechanisms to Enhance
Bakken CO2-EOR Modeling. American Rock Mechanics Association.
Ellafi, A., Jabbari, H., Ba Geri, M., & Alkamil, E. (2019b, November 11). Can HVFRs Increase the Oil Recovery in
Hydraulic Fractures Applications? Society of Petroleum Engineers. doi: 10.2118/197744-MS
Ellafi, A., Jabbari, H., Tomomewo, O., Mann, M., and Ba Geri, M. (2020b) ‘Future of Hydraulic Fracturing Application in
Terms of Water Management and Environmental Issues: A Critical Review ', Society of Petroleum Engineers (SPE),
(SPE-199993-MS)
Ellafi, A., Jabbari, H., Wan, X., Rasouli, V., Ba Geri, M., & Al-Bazzaz, W. (2020a). How Does HVFRs in High TDS
Environment Enhance Reservoir Stimulation Volume? International Petroleum Technology Conference (IPTC) 2020
IPTC-20138
Energy & Environmental Research Center (EERC). 2019, https://undeerc.org/
FracFocus website. 2019, from http://fracfocus.org/.
Gamadi TD, Sheng JJ, Soliman MY, Menouar H, Watson MC, Emadibaladehi H. An experimental study of cyclic CO2
injection to improve shale oil recovery. In: SPE IOR Symposium; 2014
30 SPE-200001-MS

Gamadi TD, Sheng JJ, Soliman MY. An experimental study of cyclic gas injection to improve shale oil recovery. In: SPE
ATCE; 2013.
Geri, M. B., Ellafi, A., Ofori, B., Flori, R., & Sherif, H. (2019a, July 31). Successful Implementation of High Viscosity
Friction Reducers from Laboratory to Field Scale: Middle Bakken Case Study. Unconventional Resources Technology
Conference. doi: 10.15530/urtec-2019-447
Gubian, Emilie. Changes in shale well design: Reaching the limits? IHS Markit, https://ihsmarkit.com/research-analysis/
changes-in-shale-well-design-reaching-the-limits.html.
Hawthorne, S. B., C.D Gorecki, J.A. Sorensen, E.N. Steadman, J.A Harju, and S. Melzer. 2013. Hydrocarbon mobilization
mechanisms from Upper, Middle, and Lower Bakken reservoir rocks exposed to CO2. In Proceedings of the SPE
Unconventional Resources Conference, Calgary, Alberta, Canada, 5–7 November 2013.
Hawthorne, S. B., Miller, D. J., Grabanski, C. B., Sorensen, J. A., Pekot, L. J., Kurz, B. A., Gorecki, C. D., Steadman,
E. N., Harju, J. A., Melzer, S. (2017, February). Measured crude oil MMPs with pure and mixed CO2, methane, and
ethane, and their relevance to enhanced oil recovery from middle Bakken and Bakken shales. In SPE Unconventional
Resources Conference. Society of Petroleum Engineers.
Heller, Robert, and Mark Zoback. "Adsorption of methane and carbon dioxide on gas shale and pure mineral samples."
Journal of unconventional oil and gas resources 8 (2014): 14-24.
Holubnyak, Y., Dubois, M., Hollenbach, J., & Hasiuk, F. (2019, September 23). Challenges and Opportunities for
Commercial-Scale Carbon Capture and Storage in Kansas. Society of Petroleum Engineers. doi: 10.2118/196186-MS
Jarvie, D. M., Hill, R. J., Ruble, T. E., and Pollastro, R. M., 2007, "Unconventional shale-gas systems: The Mississippian
Barnett shale of north-central Texas as one model for thermogenic shale-gas assessment," AAPG Bulletin, v. 91, no.
4, pp. 475–499.
Jin, L., Sorensen, J. A, Hawthorne, S. B., Smith, S. A, Bosshart, N. W., Burton-Kelly, M. E., … Harju, J. A (2016,
February 24). Improving Oil Transportability Using CO2 in the Bakken System - A Laboratory Investigation. Society
of Petroleum Engineers. doi :10.2118/178948-MS
Jin, L., Sorensen, J. A., Hawthorne, S. B., Smith, S. A., Pekot, L. J., Bosshart, N. W., … Harju, J. A. (2017, August 1).
Improving Oil Recovery by Use of Carbon Dioxide in the Bakken Unconventional System: A Laboratory Investigation.
Society of Petroleum Engineers. doi: 10.2118/178948-PA
Jin, X. (Jacob), Pavia, M., Samuel, M., Shah, S., Zhang, R., & Thompson, J. (2019, July 31). Field Pilots of Unconventional
Shale EOR in the Permian Basin. Unconventional Resources Technology Conference. doi: 10.15530/urtec-2019-506
Klenner, R. C. L., Braunberger, J. R., Sorensen, J. A, Eylands, K. E., Azenkeng, A, & Smith, S. A (2014, August 25).
A Formation Evaluation of the Middle Bakken Member Using a Multimineral Petrophysical Analysis Approach.
Unconventional Resources Technology Conference. doi: 10.15530/URTEC-2014-1922735
Kumar, Amit, Ron Glen Dusterhoft, and Shameem Siddiqui. "Completion and production strategies for liquids-rich
wells in ultra-low-permeability reservoirs." SPE Annual Technical Conference and Exhibition. Society of Petroleum
Engineers, 2013.
Kurtoglu B. Integrated reservoir characterization and modeling in support of enhanced oil recovery for Bakken [Doctoral
Dissertation]. Colorado School of Mines; 2013.
Kurtoglu, B., H. Kazemi, R. Rosen, W. Mickelson, and T. Kosanke. 2014. A Rock and Fluid Study of Middle Bakken
Formation: Key to Enhanced Oil Recovery. In Proceedings of the SPE/CSUR Unconventional Resources Conference,
Calgary, Alberta, Canada, 30 September – 2 October 2014
Lashgari, H.R., A. Suna, T. Zhanga, G.A. Pope, and L.W. Lake. 2018. Evaluation of carbon dioxide storage and miscible
gas EOR in shale oil reservoirs. Journal of Fuel.
Li, C., Kong, L., Ostadhassan, M., & Gentzis, T. (2019). Nanoscale Pore Structure Characterization of Tight Oil Formation:
A Case Study of the Bakken Formation. Energy & Fuels, 33(7), 6008–6019.
Li, Chunxiao, et al. "Multi-scale evaluation of mechanical properties of the Bakken shale." Journal of materials science
54.3 (2019): 2133-2151.
Liu, K.; Ostadhassan, M.; Zhou, J.; Gentzis, T.; Rezaee, R. Nanoscale Pore Structure Characterization of the Bakken Shale
in the USA. Fuel 2017, 209, 567–578.
Luo, G., Tian, Y., Bychina, M., & Ehlig-Economides, C. (2019, August 1). Production-Strategy Insights Using Machine
Learning: Application for Bakken Shale. Society of Petroleum Engineers. doi: 10.2118/195681-PA
Mahzari, P., Oelkers, E., Mitchell, T., & Jones, A. (2019, June 3). An Improved Understanding About CO2 EOR and CO2
Storage in Liquid-Rich Shale Reservoirs. Society of Petroleum Engineers. doi: 10.2118/195532-MS
Male, F. (2019): "Using a segregated flow model to forecast production of oil, gas, and water in shale oil plays," Journal
of Petroleum Science and Engineering, 180, 48–61.
Melcher, H., Mayerhofer, M., Agarwal, K., Lolon, E., Oduba, O., Murphy, J., … Weijers, L. (2020, January 28). Shale Frac
Designs Move to Just-Good-Enough Proppant Economics. Society of Petroleum Engineers. doi: 10.2118/199751-MS
North Dakota Council. 2012. https://www.ndoil.org/
SPE-200001-MS 31

Oil and Gas Division-North Dakota Industrial Commission (NDIC). 2019. https://www.dmr.nd.gov/oilgas/
Pankaj, P., Mukisa, H., Solovyeva, I., and Xue, H. 2018. Enhanced Oil Recovery in Eagle Ford: Opportunities Using Huffn-
Puff Technique in Unconventional Reservoirs. Presented at SPE Liquids-Rich Basins Conference - North America,
Sheng, J. J. Enhanced oil recovery in shale reservoirs by gas injection. J. Nat. Gas Sci. Eng. 2015, 22, 252–259.
Shoaib, S., and B.T. Hoffman. 2009. CO2 flooding the Elm Coulee Field. In Proceedings of the Rocky Mountain Petroleum
Technology Conference, Denver, Colorado, 14-16 April 2009.
Sorensen, J. A., Braunberger, J. R., Liu, G., Smith, S. A., Hawthorne, S. A., Steadman, E. N., & Harju, J. A. (2015, July 20).
Characterization and Evaluation of the Bakken Petroleum System for CO2 Enhanced Oil Recovery. Unconventional
Resources Technology Conference. doi: 10.15530/URTEC-2015-2169871
Sorensen, J., B. Kurz, S. Hawthorne, L. Jin, S. Smith, and A. Azenkeng. 2016. Laboratory characterization and modeling
to examine CO2 storage and enhanced oil recovery in an unconventional tight oil formation. Elsevier Ltd, 5460-5478.
Teklu, T. W., Alameri, W., Kazemi, H., & Graves, R. M. (2015, July 20). Contact Angle Measurements on
Conventional and Unconventional Reservoir Cores. Unconventional Resources Technology Conference. doi:
10.15530/URTEC-2015-2153996
Teklu, T. W., Li, X., Zhou, Z., & Abass, H. (2018, June 1). Experimental Investigation on Permeability and Porosity
Hysteresis of Tight Formations. Society of Petroleum Engineers. doi: 10.2118/180226-PA
Terzaghi, K. 1943. Theoretical Soil Mechanics. New York: Wiley.
Thakur, Ganesh. "Enhanced Recovery Technologies for Unconventional Oil Reservoirs." Journal of Petroleum
Technology 71.09 (2019): 66-69.
Tovar FD, Eide O, Graue A, Schechter DS. Experimental investigation of enhanced recovery in unconventional liquid
reservoirs using CO2: a look ahead to the future of unconventional EOR. In: SPE Unconventional Resources
Conference; 2014.
Tovar, F. D., Barrufet, M. A., & Schechter, D. S. (2018, April 14). Gas Injection for EOR in Organic Rich Shale. Part I:
Operational Philosophy. Society of Petroleum Engineers. doi: 10.2118/190323-MS
Tran, Tan, Pahala Dominicus Sinurat, and Bob A. Wattenbarger. "Production characteristics of the Bakken shale oil." SPE
Annual Technical Conference and Exhibition. Society of Petroleum Engineers, 2011.
Wan, T., and Sheng, J. 2015. Compositional Modelling of the Diffusion Effect on EOR Process in Fractured
Shale-Oil Reservoirs by Gasflooding. J Can Pet Technol 54(02): 107-115. SPE-20214-1891403-PA. http://
dx.doi.org/10.2118/2014-1891403-PA
Wan, T., Sheng, J. J., & Soliman, M. Y. (2013, August 12). Evaluate EOR Potential in Fractured Shale Oil Reservoirs by
Cyclic Gas Injection. Unconventional Resources Technology Conference. doi: 10.1190/urtec2013-187
Wang, Leizheng, and Wei Yu. "Mechanistic simulation study of gas Puff and Huff process for Bakken tight oil fractured
reservoir." Fuel 239 (2019): 1179-1193.
Yu, W., Lashgari, H., & Sepehrnoori, K. (2014, April 17). Simulation Study of CO2 Huff-n-Puff Process in Bakken Tight
Oil Reservoirs. Society of Petroleum Engineers. doi: 10.2118/169575-MS
Zhang, K., Kong, B., Zhan, J. et al. 2016. Effects of Nanoscale Pore Confinement on CO2 Immiscible and Miscible
Processes. Presented at the SPE Low Perm Symposium, Denver, Colorado, USA. 10.2118/180256-MS.