High-Temperature Core Flood Investigation of Nanocellulose As A Green Additive For Enhanced Oil Recovery

nanomaterials
Article
High-Temperature Core Flood Investigation of
Nanocellulose as a Green Additive for Enhanced
Oil Recovery
Reidun C. Aadland 1, *,† , Trygve D. Jakobsen 2, *,† , Ellinor B. Heggset 3 ,
Haili Long-Sanouiller 1 , Sébastien Simon 2 , Kristofer G. Paso 2 , Kristin Syverud 2,3
and Ole Torsæter 1
1 Department of Geoscience and Petroleum, PoreLab Center of Excellence, Norwegian University of Science
and Technology (NTNU), N-7491 Trondheim, Norway; haili.long-sanouiller@ntnu.no (H.L.-S.);
ole.torsater@ntnu.no (O.T.)
2 Department of Chemical Engineering, Norwegian University of Science and Technology (NTNU),
N-7491 Trondheim, Norway; sebastien.c.simon@ntnu.no (S.S.); kristofer.g.paso@ntnu.no (K.G.P.);
kristin.syverud@rise-pfi.no (K.S.)
3 RISE PFI, N-7491 Trondheim, Norway; ellinor.heggset@rise-pfi.no
* Correspondence: reidun.aadland@ntnu.no (R.C.A.); trygve.d.jakobsen@ntnu.no (T.D.J.);
Tel.: +47-73-59-30-59 (R.C.A.); +47-95-05-04-51 (T.D.J.)
† These authors contributed equally to this work.

Received: 11 March 2019; Accepted: 24 April 2019; Published: 27 April 2019
Abstract: Recent studies have discovered a substantial viscosity increase of aqueous cellulose
nanocrystal (CNC) dispersions upon heat aging at temperatures above 90 ◦ C. This distinct change in
material properties at very low concentrations in water has been proposed as an active mechanism
for enhanced oil recovery (EOR), as highly viscous fluid may improve macroscopic sweep efficiencies
and mitigate viscous fingering. A high-temperature (120 ◦ C) core flood experiment was carried out
with 1 wt. % CNC in low salinity brine on a 60 cm-long sandstone core outcrop initially saturated
with crude oil. A flow rate corresponding to 24 h per pore volume was applied to ensure sufficient
viscosification time within the porous media. The total oil recovery was 62.2%, including 1.2% oil
being produced during CNC flooding. Creation of local log-jams inside the porous media appears
to be the dominant mechanism for additional oil recovery during nano flooding. The permeability
was reduced by 89.5% during the core flood, and a thin layer of nanocellulose film was observed
at the inlet of the core plug. CNC fluid and core flood effluent was analyzed using atomic force
microscopy (AFM), particle size analysis, and shear rheology. The effluent was largely unchanged
after passing through the core over a time period of 24 h. After the core outcrop was rinsed, a micro
computed tomography (micro-CT) was used to examine heterogeneity of the core. The core was
found to be homogeneous.
Keywords: enhanced oil recovery; nanocellulose; petroleum; cellulose nanocrystals; tertiary recovery;
crude oil; nanoparticle; CNC; core flood; high temperature; heat aging; rheology modification
1. Introduction
The average oil recovery factor (RF) on the Norwegian Continental Shelf (NCS) is 47% and on a
global basis the average recovery factor remains less than 40% [1]. In addition, oil production rates at
existing fields are declining and the frequency of new field discoveries has diminished substantially
in the last few decades [2]. Thus, it is essential to target the residual oil remaining in petroleum
reservoirs and achieve higher recovery factors from existing fields. Deployment of new and innovative
Nanomaterials 2019, 9, 665; doi:10.3390/nano9050665 www.mdpi.com/journal/nanomaterials

Nanomaterials 2019, 9, 665 2 of 26
production technologies is necessary to achieve this goal. One of the technologies is called enhanced
oil recovery (EOR) and includes any technique where fluids/material that are not normally present
in a reservoir is injected into it. The three primary categories of EOR include: thermal, chemical
and miscible or solvent injection [3]. Thermal recovery is typically utilized in heavy oil reservoirs.
The method applies heat to reduce oil viscosity, facilitating mobilization of oil towards the production
well. Chemical methods fall into three different groups: polymer flooding, surfactant-polymer (SP)
flooding and alkaline-surfactant-polymer (ASP) flooding. The main idea is to add chemicals to the
water flooding stage to either increase the mobility ratio (polymer), reduce the interfacial tension
(surfactant), or create micro-emulsions (alkali). Miscible or solvent injection involves injection of a
mutually miscible low viscosity solvent to mobilize residual oil trapped in the reservoir rock [4,5].
Chemical flooding with polymer additives is considered the most promising EOR method [6] for
implementation on the Norwegian continental shelf.
The main objective of EOR methods is to increase the overall recovery efficiency (E), which may
be quantified as the product of the macroscopic or volumetric displacement efficiency (EV) and the
microscopic displacement efficiency (ED), as shown in Equation (1) [7]. A schematic illustration of the
recovery process and specific definitions can be found in Figure 1.
E = EV ED (1)
Microscopic displacement efficiency refers to trapping of residual oil at the pore level (Figure 1).
The term ED quantifies how well displacing fluid mobilizes residual oil upon contact. At the pore level,
various factors influence oil displacement, such as interfacial tension, wettability, capillary pressure
and relative permeability. For oil/brine systems, interfacial tension values are typically in the order of
20–30 mN/m [7,8]. However, by adding tailored chemicals to the water phase, it is possible to reduce
the interfacial tension by several orders of magnitude [7]. A lower interfacial tension allows extraction
of more oil, as the oil droplets will become dislodged from the pore throats.
The macroscopic displacement efficiency term may be quantified as the product of the areal (EA)
and vertical (EI) sweep efficiencies, where EA denotes the horizontal area that can be swept and EI
denotes the cross-sectional area that can be invaded (Figure 1). Overall, the EV term describes how well
the injected fluid has traversed the reservoir, i.e., which fraction of the reservoir the displacing fluid
has been able to contact. Several factors affect the macroscopic sweep efficiency, such as mobility ratio,
reservoir heterogeneities and anisotropy, well pattern, and the type of rock matrix imbibed with oil.
The mobility ratio (M) is a measure of how easy one fluid moves through the porous media
relative to another fluid, and is given in Equation (2) [9].
mobility o f displacing f luid krw µo

M= = (2)
mobility o f displaced f luid kro µw
where krw and kro denote the effective permeabilities of the displacing fluid (e.g., water) and displaced
fluid (oil), respectively, and µw and µo denote the corresponding viscosities, respectively. The efficiency
of the volumetric sweep increases as the mobility ratio decreases. Thus, a favorable mobility ratio is
obtained if M < 1. If M > 1 the mobility ratio is considered unfavorable, because the displacing fluid
(water) flows with less viscous resistance than the displaced fluid (oil). This imbalance will result
in the displacement fluid bypassing oil pockets and oil-filled pores because of viscous fingering and
an unstable fluid front (Figure 1). Viscous fingering occurs as a result of the viscosity contrast and
effectively reduces the oil recovery efficiency [7,9].
Figure
Figure 1.
1. Schematic
Schematic illustration
illustration of
of enhanced
enhanced oil
oil recovery
recovery and the overall
and the overall recovery
recovery efficiency.
efficiency.
The displacing
The displacingfluid fluidshould
should ideally
ideally have have
a higha high viscosity
viscosity in to
in order order to be
be able able tothe
to sweep sweep the
reservoir
reservoir more effectively. A favorable mobility ratio can thus be achieved
more effectively. A favorable mobility ratio can thus be achieved by altering the parameters in by altering the parameters
in Equation
Equation (2).(2).
ForFor a givenreservoir
a given reservoirwith
witha aspecified
specifiedoil,oil,the
theparameter
parameterthat thatcancanbebe controlled
controlled is is the
the
viscosity of the injection
viscosity the injection fluid. fluid. However, it is important that the fluid has a sufficiently low initial
sufficiently low initial
viscosity to
viscosity to make it easy to pump into the reservoir. reservoir. The optimal solution is, therefore, to have a low
viscosity during
viscosity during injection,
injection,and andthenthenasas
the fluid
the fluidenters thethe
enters reservoir
reservoirit would increase
it would in viscosity
increase with
in viscosity
time.
with time.
common technique for increasing the displacement fluid viscosity is to add polymers to the
A common
water. Shiran et al. [10] performed
injection water. performed laboratory experiments in which polymer flooding was was
introduced as a tertiary recovery technique (after water flooding). To obtain
introduced as a tertiary recovery technique (after water flooding). To obtain a favorable mobility ratio, a favorable mobility
aratio, a diluted
diluted crude crude
oil withoilawith a viscosity
viscosity of 2.4
of 2.4 cP wascPused.
was Two
used.polymer
Two polymer solutions
solutions with viscosities
with viscosities of 2.6
of 2.6 and 2.2 cP were tested. Three Bentheimer sandstone cores
and 2.2 cP were tested. Three Bentheimer sandstone cores were used with different wettabilities, and were used with different
wettabilities,
all cores had aand all cores had
permeability valuea permeability
of approximately value2ofD.approximately
The quantity of 2 D. The quantity
additional of additional
oil recovered after
oil recovered
water floodingafter water
ranged flooding
from rangedThe
3.2%–12.9%. from low3.2%–12.9%.
percentageThe low was
of 3.2% percentage
achieved ofwhen
3.2% was achieved
the core plug
when
was the core plugwater-wet,
preferentially was preferentially
while thewater-wet, while theof
highest percentage highest
12.9% percentage
was achieved of 12.9%
when was achieved
the core plug
when
was the core plug
preferentially was preferentially
oil-wet. A lower recoveryoil-wet. A lower
efficiency andrecovery
higher efficiency
residual oiland higher residual
is generally observed oil in
is
generallywater-wet
strongly observed porous
in strongly
media.water-wet porous
The reason formedia. The oil
this is that reason
mayfor this isinthat
remain oil may
larger poresremain
where in it
larger pores where it could be snapped off, or a water film may disconnect a
could be snapped off, or a water film may disconnect a continuous mass of oil. In both scenarios the oil continuous mass of oil.
In both scenarios
becomes trapped in thethe
oilpore
becomes trapped
structure in porous
of the the poremedia
structure
[10].of the porous media [10].
Two main polymer types are currently in use commercially; xanthan gum and partially
hydrolyzed polyacrylamide (HPAM) [11]. Both of these serve to reduce the mobility of water by
increasing the viscosity. However, both polymers have substantial performance limitations with
Two main polymer types are currently in use commercially; xanthan gum and partially hydrolyzed
polyacrylamide (HPAM) [11]. Both of these serve to reduce the mobility of water by increasing the
viscosity. However, both polymers have substantial performance limitations with respect to EOR.
Xanthan gum can plug formations, is prone to microbial degradation, and exhibits poor thermal
stability [12]. HPAM, on the other hand, is susceptible to numerous chemical, thermal and mechanical
degradation mechanisms. Thus, the viscosifying capability of HPAM will gradually diminish as it
migrates through porous media [7,12]. As many of the polymers used today are considered toxic and
detrimental to the environment, it is of importance to develop stable and green alternatives to this
EOR technique. This has been a focus area in research over the last decade, where many different types
of green alternatives have been tested [13–22].
Nanocellulose is introduced as another green and environmental friendly alternative to existing
EOR polymers. Nanocellulose is nanoscaled particles composed of the biopolymer cellulose, and is an
abundant and renewable natural resource. Cellulose is the structural component of plant cells [23], and
the two most common cellulosic sources are wood and cotton. According to Klemm et al. [24] cellulose
is almost an inexhaustible source of raw material and thereby perfectly able to supply the increasing
demand for environmentally friendly and biocompatible products. A tree produces 10 g of cellulose
per day and the total production of cellulose all over the world is estimated to be 1.3 × 1010 tons per
year [23]. The majority of forest areas utilized for commercial wood conversion are managed based on
the criterion of sustainable development, meaning that the annual growth of the forests is considerably
greater than what is removed. In addition, since nanocellulose can be produced from unused (today)
remains from paper production, no new consumption of trees may even be necessary. Thus, extensive
use of nanocellulose should not have a large effect on the environment. Another positive attribute
with nanocellulose in terms of offshore EOR applications is that it can be discharged into the ocean.
Nanocellulose would agglomerate in high salinity water and degrade.
Disintegration of wood fibers into cellulose nanofibrils (CNFs) was first described by
Turbak et al. [25] and Herrick et al. [26]. Similar materials called cellulose nanocrystals (CNCs)
were studied as early as 1949 [27] and described as rod-like cellulose micelles by Rånby and Ribi [28],
and the current study focuses on CNC material. CNCs are produced from acid hydrolysis of cellulosic
material, which dissolves amorphous sections, leaving crystalline, rod-like particles intact. Sonication
or similar methods are often required to obtain dispersed particles [29]. The particles are typically
100 to 250 nm in length and 5 to 70 nm wide.
There does not exist much literature on nanocellulose for enhanced oil-recovery applications.
However, preliminary experiments have investigated the potential of nanocelluloses as the EOR agent.
For instance, retention is a commonly encountered problem when developing EOR polymers. Molnes
et al. [19] investigated injectivity of nanocellulose through porous media, and found that the particles
traversed the core. Nevertheless, there were indications that a portion of the CNC particles became
trapped inside the pore matrix [19]. Aadland et al. [20] performed a more extensive retention study on
CNCs in core plugs and sandpacks and observed that salinity had the largest effect on retention. The
study concluded that CNCs dispersed in low-salinity water could have promising potential as EOR
agents [20]. Jakobsen et al. [22] investigated interactions between anionic surfactants and nanocellulose
and found that both surfactants and nanocellulose play a role in lowering interfacial tension between oil
and water, and that the presence of surfactant increased the viscosity of the nanocellulose dispersions
by effectively binding together nanocellulose particles.
Temperature conditions in petroleum reservoirs may be quite high (>100 ◦ C), a potential EOR
agent should, therefore, be thermally stable within the temperature limits of oil reservoir applications.
Heggset et al. [30] performed a temperature stability study where nanocellulose exhibited improved
thermal stability in comparison to guar gum and xanthan gum [30]. Molnes et al. [18] also performed
a heat aging study on CNCs, where shear viscosity was measured after 24, 48 and 168 h of heating.
The study demonstrated that shear viscosity increases drastically after only 24 h of aging at 120 ◦ C.
The results collectively demonstrate that a certain aging period is required under high temperature
conditions to obtain a dramatic increase in shear viscosity for CNCs in low salinity water [18].
In a subsequent study by Molnes et al. [17] the injectivity of CNC dispersions at 60, 90 and
120 ◦ C was investigated using core plugs. They observed no viscosity increase for any of the tests.
A possible explanation for this was that the residence time of the CNCs in the matrix was insufficient
for modification of the CNC structure. For their tested flow rate (4 pore volume (PV)/day), a CNC
particle would take approximately six hours to traverse the core. The heat aging experiment revealed
that the increase in viscosity was not observed before the dispersion had been aged for at least 20 h [18].
They also performed two oil recovery experiments using nanocellulose on Berea sandstone cores.
CNCs in low salinity water was injected as a tertiary recovery technique at 60 ◦ C and 90 ◦ C and the
incremental oil recovery was calculated. It was concluded that CNCs dispersed in low salinity water
might have a certain potential for use in enhanced oil recovery. Additional oil produced attributable to
nanocellulose was 3.4% for the test at 90 ◦ C, while no significant EOR effect was observed for the test
performed at 60 ◦ C [17].
The current investigation is an extension of the heat aging experiments performed by
Molnes et al. [17,18]. The potential of CNCs as an EOR agent was tested by performing a core
flood experiment in low salinity water. The core flood was performed at 120 ◦ C using a long core
(approx. 60 cm) and low flow rate in order for the CNC solution to have a long residence time within
the porous media. The goal was to reproduce the same viscosification effect that was revealed in the
Molnes et al. [18] study, and to determine the amount of incremental oil produced after water flooding.
Permeability measurements were also conducted during the core flood.
2. Materials
2.1. Porous Media

Bentheimer sandstone was used as the porous media. The core sample had a permeability of
2.0 Darcy and porosity of 21.6%. The pore volume (PV) was 141.8 mL. Length and diameter were
59.2 cm and 3.76 cm, respectively. The absolute permeability was measured using brine.
2.2. Brine
Experiments were performed using low salinity water, consisting of 0.1 wt. % sodium chloride
(NaCl) mixed with de-ionized water (DIW). After mixing, it was filtered using a 0.45 µm Millipore
filter under vacuum. Brine properties are listed in Table 1.
2.3. Oil
During the saturation phase of the core, various types of oil were used and their properties are
listed in Table 1. Isopar L (ExxonMobil Chemical Europe, Mechelen, Belgium) was applied first, then a
mix of 20% isopar L and 80% high-viscous paraffin (HVP) oil (VWR International AS, Oslo, Norway).
At the end, when the desired initial water saturation was reached, the 20/80% isopar L/HVP solution
was replaced with stabilized and degassed crude oil. From the SARA (Saturates, Aromatic, Resin and
Asphaltene) analysis the crude oil consisted of: 66.21 wt. % saturates, 25.78 wt. % aromatics, 7.69 wt. %
resins and 0.32 wt. % asphaltenes. The crude oil was obtained from a field in the Norwegian Sea, and
was filtered twice through a 5 µm Millipore filter under vacuum in order to remove impurities.
2.4. Cellulose Nanocrystals

Nanocellulose was purchased from the University of Maine. The material was manufactured at
the USDA Forest Products Laboratory in Madison, WI, USA (U.S. Dep. of Agriculture). The cellulose
nanocrystals were produced using 64% sulphuric acid to hydrolyze amorphous regions of cellulose
material, resulting in acid resistant crystals [31]. The acid hydrolysis procedure also results in formation
of surface sulphate groups on the CNC crystals. The sulphate charge density is 0.3 mmol/g [30].
The stock-dispersion had a concentration of 12 wt. %.
Table 1. Fluid properties.
Density (g/cm3 ) Viscosity (cP)

Fluid
20 ◦C 60 ◦C 120 ◦C 20 ◦C 60 ◦ C 120 ◦ C
Isopar L 0.76 0.73 — 1.33 0.72 —
20/80% isopar L/HVP oil 0.87 — — 121.50 — —
Crude oil 0.91 0.89 0.85 55.90 12.19 3.32
0.1 wt. % NaCl 1.00 0.98 0.94 0.91 0.47 0.23
1 wt. % CNC in 0.1 wt. % NaCl 1.00 — — 1.40 1.09 —
3. Experimental Methods
3.1. Characterization of Cellulose Nanocrystals (CNC)
3.1.1. High-Temperature Aging of CNC

CNC dispersion was diluted to samples of 0.49, 1 and 2 wt. % containing 0.1 wt. % NaCl, and
40 mL of the sample was transferred to 50 mL Schott bottles with heat-resistant caps. Schott bottles
containing CNC suspension were refrigerated both before and after heat aging. Heat aging was
performed in a heating bath (Julabo Gmbh, Seebach, Germany) containing oil at 120 ◦ C for aging
durations of 2 h to 16.5 weeks. These trials were conducted in an identical manner to a previous study
by Molnes et al. [18], except for the sample amount during aging.
3.1.2. Rheology Measurements and Shear Rate

Rheology measurements were performed using a Physica MCR 301 rotational rheometer from
Anton Paar (Graz, Austria). A 40 mm diameter cone-and-plate geometry was used with roughened
surfaces and a cone angle of 2◦ . The truncation gap size was 0.057 mm. All rheology measurements
were carried out at 20 ◦ C. The rheometric protocol for all samples comprised of the following stages:
(1) a 60 second pre-shear (100 s−1 ) stage, (2) 60 s period of quiescent rest, (3) two consecutive cycles
of increasing and decreasing shear rate from 0.1–1000 s−1 . Each step in the protocol lasted for 600 s,
and the duration of the entire protocol was 42 min. The concentration of samples was not altered
before rheology measurements and each analysis was performed at least twice. For all comparison of
viscosity data, viscosity values were extracted from a shear rate of 1 s−1 .
The shear rate withinthe porous media was calculated using Equation (3) [3]. The calculations
were performed to relate the shear rate inside the core to relevant shear viscosity measurements.
4q
γ= (3)
A 8kϕ
p
In Equation (3), q denotes the volumetric flow rate (m3 /s), A is the cross-sectional area of the
porous media (m2 ), k is the single phase permeability (m2 ) and ϕ is the porosity.
3.1.3. Atomic Force Microscope (AFM) Pictures

Atomic force microscopy (AFM) images were obtained using a Veeco diMultimode V instrument
from Veeco Instruments Inc. (Santa Barbara, CA, USA). ScanAsyst mode was employed and Bruker
AFM probes with silicon tips on nitride levers were used. The scanning probes had a spring constant
of 0.4 N/m. 2 × 2 µm area images were obtained of effluent samples and 5 × 5 µm area images were
obtained of aged CNC samples. Nanoscope Analysis v 1.50 software (Bruker Inc. Santa Barbara, CA,
USA) was used to analyze the images. The CNC and effluent samples were prepared according to
Lahiji et al. [32] Mica discs of 10 mm width were cleaved and a drop of suspension was deposited.
The surface was immediately rinsed with DIW, and compressed N2 gas was used to dry the sample.
Aged CNC samples were diluted with DIW to 0.0125 wt. % prior to deposition onto the mica. Effluent
samples from the core flood were sonicated for 2–4 min and diluted 10 times in DIW.
3.1.4. Interfacial Tension

The interfacial tension (IFT) between crude oil and brine/nanofluid was measured with the pendant
drop method using the drop shape analyzer DSA 100S (Kruss GmbH, 22453 Hamburg, Germany).
The experimental procedure consisted of filling a glass container with the respective fluid (i.e., brine or
nanocellulose) and a J-shaped needle was then immersed in the fluid and an oil droplet was created.
The diameter of the needle was 1.047 mm. The drop shape analyzer DSA 100S uses the software
ADVANCE to measure and analyze the droplet shape. Interfacial tension was measured at five minute
intervals for 12 h, and recorded values were automatically logged in ADVANCE. The experiments
were performed at ambient temperature and pressure conditions. Four measurement replicates were
performed with crude oil and brine, while two replicates were performed with crude oil and nanofluid.
3.2. Core Plug Setup and Cleaning

The core was mounted in a hydrostatic core holder with a Viton® rubber sleeve (Elastomer
Engineering Limited, Lymm, UK). The core was pressure tested at 60 bar for 16 h using nitrogen (N2 )
gas, to ensure that no leakage into the core occurred from outside the sleeve. The nitrogen gas was
subsequently replaced with high-viscous paraffin oil to maintain a constant sleeve pressure during the
experiment. All valves were leak tested prior to initiating the experiment.
The Bentheimer core plug had not been used in previous experiments and was not contaminated
with oil residue. Therefore, a mild cleaning procedure was applied. After the core plug was installed
in the core holder inside the heating cabinet, it was rinsed by injecting methanol overnight. A low flow
rate of 0.01 mL/min was used and the back pressure was maintained at 15 bar. The cleaning procedure
was performed at room temperature.
3.3. Establish Initial Saturation

Saturation of the core was performed using an injection scheme in which the core was mounted in
the core holder on the rig. The cleaned core was first 100% saturated with brine by injecting 0.1 wt. %
NaCl. This brine composition was selected for all experimental stages to avoid concentration gradients
and associated uncertainties in local salinity values. The oil saturation protocol consisted of four stages
and the duration of each stage is listed in Table 2. A synthetic oil (isopar L) was first injected through the
core at a rate of 1 mL/min, resulting in an initial water saturation (Swi ) of 36%. Subsequently, a mixture
of 20% isopar L and 80% high-viscous paraffin oil was injected in two steps. An initial low rate of
2 mL/min provided a Swi value of 22%, and a subsequent high injection rate of 9 mL/min provided
a final Swi value of 15%. The original oil in place (OOIP) was quantified as 120.5 mL. To achieve
an optimal displacement of the fluid front, these first three injection stages were conducted at room
temperature conditions corresponding to a high oil viscosity. In the final step of the oil saturation
procedure, crude oil was injected to replace the 20/80% isopar L/HVP oil. This was done at 60 ◦ C.
The core was then aged for five weeks at 60 ◦ C.
Table 2. The different stages during saturation of the core with oil.
Stage Fluid Temp. Core Flow Rate PV Injected OOIP Swi

(◦ C) (mL/min) (mL) (%)
1 Isopar L 21 1.3 3.2 91.0 36
2 20/80% isopar L/HVP oil 21 2.1 7.0 110.4 22
3 20/80% isopar L/HVP oil 21 8.5 17.1 120.5 15
4 Crude oil 60 2.0 5.0 120.5 15
3.4. Core
3.4. CoreFlood
FloodExperiment
Experiment
3.4.1. Experimental
3.4.1. Experimental Setup
Setup
Figure22illustrates
Figure illustratesaaschematic
schematicofofthe theexperimental
experimental setup.The
setup. The core
core plug
plug waswas installed
installed vertically
vertically in
in the
the corecore holder.
holder. SleeveSleeve pressure
pressure was at
was kept kept at 75
75 bar bar during
during the primary
the primary experimentexperiment
using highusing high
viscous
viscous oil.
paraffin paraffin oil. Acontaining
A cylinder cylinder containing
high viscoushigh viscous
paraffin paraffin
oil (200 mL) in oilthe(200 mL) and
bottom in the bottom
nitrogen gasand
on
nitrogen gas on the top was placed outside the heating cabinet and was
the top was placed outside the heating cabinet and was used to maintain the system at equilibrium. used to maintain the system
at equilibrium.
The back pressure Thewasbackkeptpressure
at 15.5 bar was keptthe
during at main
15.5 bar during and
experiment the the main experimentwas
temperature andsetthe
to
temperature
◦
120 C. was set to 120 °C.
The injection
The injection fluid
fluid container
container waswas placed
placed on on aa balance
balance (during
(during water
water flooding)
flooding) or or on
on aa magnetic
magnetic
stirrer (during
stirrer (during nanonano flooding).
flooding). The The balance
balance measurements
measurements were were fed
fed toto aa computer
computer for for data
data logging.
logging.
The balance was not able to be used in combination with a stirrer, but
The balance was not able to be used in combination with a stirrer, but the pumping rate was tested the pumping rate was tested
and
and confirmed prior to nano flooding. The sample changer also provided
confirmed prior to nano flooding. The sample changer also provided an opportunity to monitor the an opportunity to monitor
the flow
flow rate rate
on an onhourly
an hourlybasis.basis.
TheThe injection
injection fluidfluid
waswas pumped
pumped intointo the core
the core usingusing a Pharmacia
a Pharmacia P-500P-
500 pump. The differential pressure (dP) across the core was logged, as
pump. The differential pressure (dP) across the core was logged, as well as the inlet pressure and the well as the inlet pressure and
the temperature
temperature insideinside the core
the core holder.
holder. During During saturation
saturation of the
of the core,
core, a cylinder
a cylinder containing
containing crude
crude oil oil
at
at the top and isopar L at the bottom was placed inside the heating cabinet
the top and isopar L at the bottom was placed inside the heating cabinet for the final saturation step for the final saturation
step (replacing
(replacing 20/80% 20/80%
isopar isopar
L/HVP L/HVP oil with
oil with crude crude
oil). oil). Autoclave
Autoclave tubetube fittings
fittings andand valves
valves werewere used
used to
to direct
direct fluid
fluid flow.
flow. AA bypass
bypass line
line waswas installedover
installed overthe thecore
coretotoallow
allowfluid fluiddiversion
diversionduring
during various
various
stages of
stages of the
the experiment.
experiment. For For example,
example, fluidfluid diversion
diversion was was applied
applied to to reduce
reduce the
the dead
dead volume
volume of of the
the
prior fluid during heating and prior to introducing a
prior fluid during heating and prior to introducing a new fluid to the core. new fluid to the core.
Figure 2.
Figure 2. Schematic illustration of the experimental
experimental setup
setup of
of the
the core
core flood.
flood.
3.4.2.
3.4.2. Injection
Injection Scheme
Scheme
After ◦ C, the temperature
Afterestablishing
establishinginitial water
initial saturation
water and aging
saturation the core
and aging forcore
the five weeks at 60
for five weeks at 60 °C, the
was increased to 120 ◦ C. The core flood was initiated after the temperature stabilized at 120 ◦ C. The first
temperature was increased to 120 °C. The core flood was initiated after the temperature stabilized at
stage
120 °C.ofThe
the first
core stage
floodofwasthesecondary
core flood recovery using recovery
was secondary low-salinity water
using (LSW) (0.1
low-salinity wt.(LSW)
water % NaCl).
(0.1
LSW
wt. %was injected
NaCl). LSWatwasa rate of 0.1atmL/min
injected a rate ofuntil no additional
0.1 mL/min oil additional
until no was produced andproduced
oil was the pressure andhad
the
stabilized. Near the end of the water flooding stage, the injection rate was increased
pressure had stabilized. Near the end of the water flooding stage, the injection rate was increased to to 0.5 mL/min for
half a PV to determine
0.5 mL/min for half a PVif any capillary trapped
to determine oil at thetrapped
if any capillary core outlet could
oil at be released
the core prior to
outlet could bethe nano
released
flood.
prior toNanocellulose
the nano flood. fluidNanocellulose
was then injected fluidinwas
several
thenstages as aintertiary
injected severaloilstages
recovery
as aagent. First,
tertiary oil
the core was flooded at the same rate used during water flooding, 0.1 mL/min.
recovery agent. First, the core was flooded at the same rate used during water flooding, 0.1 mL/min. Utilizing a flooding
rate of 0.1a mL/min
Utilizing floodingindicates
rate of 0.1that one PV
mL/min requires
indicates thatapproximately
one PV requires 24approximately
h to complete;2424h htoshould
complete;be
sufficient
24 h should time befor the nanocellulose
sufficient time for thefluid to undergofluid
nanocellulose a substantial
to undergo viscosity increase
a substantial inside the
viscosity core,
increase
inside the core, which hypothetically should give rise to a different flow path than the pure LSW
which hypothetically should give rise to a different flow path than the pure LSW flood. By creating a
different flow path, nanofluid may extract additional oil from the core. However, to provide complete
certainty that the nanofluid was able to properly viscosify inside the core, it was decided to shut-in the
core for 24 h. During the shut-in period, the content of the core was un-extracted oil and nanocellulose
fluid, while the bypass line was flooded with brine to prevent the flow lines from becoming constricted
with viscous nanocellulose fluid. After 24 h, nanofluid injection continued ata low rate of 0.1 mL/min.
The injection continued until no additional oil was produced. Subsequently the rate was increased in
three stages to 0.3, 0.5 and 1.0 mL/min, respectively. It was decided to increase the rate in steps, as there
was uncertainty regarding how shear thinning the nanofluid was. The pressure gauge could only read
values up to five bar before it saturated. Therefore, it was important to build the pressure up gradually
so that it was maintained below this value. During all nano flooding stages the goal was to inject
nanofluid until no additional oil was produced and a stable pressure reading was attained. The final
step of the core flood was a post flush, in which low salinity water was injected at a rate of 0.5 mL/min.
3.5. Permeability
Permeability was measured initially, after each core flood stage, and after rinsing the core. It was
measured by flooding with the respective fluid at four different injection rates until a stable differential
pressure (dP) was obtained, and then it was calculated using the Darcy equation.
3.6. Effluent Characterization

Effluent samples were collected at all times during the different stages of the core flood.
3.6.1. Oil Production

Oil produced during core flooding was measured in one of two ways, either from vials from
the fraction collector or by graded measuring cylinders. The vials in the fraction collector were not
graded; the calibration line was established based on reading the height of volume in centimeters and
measuring the weight for each reading. The volume was plotted against the height for all the reference
samples and a factor was established to correlate fluid height (cm) directly to volume (mL).
For the graded cylinders, it was difficult to estimate the amount of oil when reaching the tail of
the production for each flood. A weight-based method was implemented to obtain a more accurate oil
volume estimation. Toluene was added to the measurement cylinder containing oil and water and
the toluene/oil phase was extracted using a syringe. The mixture was then added to an empty beaker
of known weight, and re-weighed. The beaker was subsequently placed over a heating bath (60 ◦ C)
for some hours and then placed in an oven (80 ◦ C) so that the toluene could evaporate over time.
The beaker was weighed over three days until it stabilized. The amount of crude oil in the beaker was
then scaled using a factor that was determined from a reference set. The reference set (three samples)
consisted of a known amount of oil and toluene added to a beaker and reading the final weight at the
end after evaporation. The difference in weight from start to end was calculated and based on a factor
that was determined on how much oil was lost to evaporation.
3.6.2. Particle Size

Particle size measurements were performed on selected effluent samples. These effluent samples
were collected during the nano flood and the idea was to investigate the size distribution throughout
that stage to see if it could yield additional information concerning the processes occurring inside the
core. Particle sizes were, therefore, measured prior to oil breakthrough, at oil breakthrough, at the end
of first low rate stage, after shut-in, at the end of the second low rate (after shut in), and at two high
rate stages (0.5 mL/min and 1.0 mL/min). The particle size was measured at room temperature using a
Zetasizer Nano ZS instrument (Malvern Instruments Nordic AB, Skallestad, Norway).
Nanomaterials 2019, 9,
Nanomaterials 2019, 9, 665
665 10
10 of
of 26
25
3.7. Characterization of the Core after Core Flood

3.7. Characterization of the Core after Core Flood
3.7.1. Micro
3.7.1. Micro Computed
Computed Tomography
Tomography (Micro-CT)
(Micro-CT)Scan
Scan
A micro
A micro computed
computed tomography
tomography (micro-CT)
(micro-CT) scanscan was
was performed
performed to to determine
determine if if the
the core
core was
was
homogenous or if high permeable flow paths were present in the core.
homogenous or if high permeable flow paths were present in the core. The micro-CT scan was The micro-CT scan was
performedon
performed onaaNikon
NikonXT XTH225
H225scanner
scanner(Nikon
(Nikon Metrology,
Metrology, Tring,
Tring, UK).UK).
TheThe
scanscan
waswas executed
executed afterafter
the
the core flood was completed and the core had been rinsed and dried. The core
core flood was completed and the core had been rinsed and dried. The core was placed in a plastic was placed in a plastic
holder (Figure
holder (Figure A1, A1, Appendix
Appendix A),
A), and
and due
due to
to the
the core
core length
length the
the equipment
equipment couldcould only
only scan
scan half
half of
of
the core at a time. One scan (half the core) consisted of seven sub-scans at 205
the core at a time. One scan (half the core) consisted of seven sub-scans at 205 kV and 110 µA with kV and 110 µA with
exposure time
exposure time 0.14
0.14 s.s. The
The image
image resolution
resolution is is 27.4
27.4 µm/voxel.
µm/voxel. After
After reconstruction,
reconstruction, oneone image
image stack
stack
was generated for each sub-scan. All image stacks were subsequently stitched
was generated for each sub-scan. All image stacks were subsequently stitched together using the together using the
program ImageJ.
program ImageJ.
The image segmentation was
The image segmentation was based
based onon the
the range
range ofof the
the grey
grey value,
value, i.e.,
i.e., black
black represents
represents pores,
pores,
and grey to light grey represents matrix, e.g., quartz. Single threshold method
and grey to light grey represents matrix, e.g., quartz. Single threshold method was applied betweenwas applied between
pores and
pores and matrix
matrix (Figure
(Figure A2,
A2, Appendix
Appendix A). A). The
The threshold
threshold value
value was
was based
based on on visual
visual observation.
observation.
Therefore, if there were any pores smaller than the voxel size they would
Therefore, if there were any pores smaller than the voxel size they would not be captured via not be captured via the
the
micro-CT technique.
micro-CT technique.
3.7.2. Scanning
3.7.2. Scanning Electron
Electron Microscopy
Microscopy (SEM)
(SEM) Imaging
Imaging
Scanning electron
Scanning electron microscope
microscope (SEM)
(SEM) images
images were
were acquired
acquired of
of selected
selected areas
areas of
of the
the core
core to
to
determine how far the nanoparticles were stuck inside the porous media. The core was cut
determine how far the nanoparticles were stuck inside the porous media. The core was cut in three in three
places in
places in total
total where
where each
eachpiece
piecewas
was1.5
1.5cm
cmthick
thick(Figure
(Figure3).
3).
Figure 3. Illustration
Figure 3. IllustrationofofBentheimer
Bentheimercore.
core.The
Thecore
corewas cutcut
was in in
three places
three where
places each
where piece
each waswas
piece 1.5 cm
1.5
thick. These
cm thick. three
These pieces
three were
pieces investigated
were for
investigated fornanoparticles
nanoparticlesusing
usingscanning
scanningelectron
electron microscopy
microscopy
(SEM)
(SEM) imaging.
imaging.
The
The three
three core
core pieces
pieces were
were coated
coated with
with gold
gold and
and investigated
investigated using
using SEM
SEM system
system from
from FEI
FEI (Field
(Field
Electron and Ion) company (model Apreo) in the Norwegian University of Science and
Electron and Ion) company (model Apreo) in the Norwegian University of Science and TechnologyTechnology
(NTNU)
(NTNU) NanoLab.
NanoLab. TheThe same
same SEM
SEM acquisition
acquisition setup
setup was
was applied
applied on
on one
one clean
clean unused
unused Bentheimer
Bentheimer
core.
core. The image of this core sample was used as reference for comparison with the images of
The image of this core sample was used as reference for comparison with the images of the
the cut
cut
core pieces from the flooding experiment.
core pieces from the flooding experiment.
4. Results and Discussion
4. Results and Discussion
4.1. Characterization of CNC
4.1. Characterization of CNC
4.1.1. High-Temperature Aging of CNC
4.1.1 High-Temperature Aging of CNC
After aging at 120 ◦ C, changes in the nature of the CNC suspension were clearly visible, as
After in
displayed aging
Figureat 120 °C, changes
4. While in thesample
the unaged naturewasof the CNC transparent,
slightly suspension were
all agedclearly visible,
samples wereas
displayed in Figure 4. While the unaged sample was slightly transparent, all aged samples
translucent (i.e., not transparent). After 12 h, the suspension had attained a gel form able to sustain were
translucent
its (i.e., not
shape against transparent).
gravity. The CNC After 12 h, the became
suspension suspension had from
altered attained a gel
a pale formcolor
white able at
to 12
sustain
h to
its shape against gravity. The CNC suspension became altered from a pale white
increasingly beige hues after each subsequent day. After 10 days, black and grey tones, indicative color at 12 h to
increasingly beige hues after each subsequent day. After 10 days, black and grey tones,
of charring, became visible, and after 6 weeks, the top half of the sample was completely black, indicative of
charring, became
Undoubtedly, oxygenvisible, and after
gas trapped above6 the
weeks, the topplayed
suspension half of
an the sample
essential rolewas
in thecompletely black,
charring process,
Undoubtedly, oxygen gas trapped above the suspension played
but most likely does not contribute to the observed viscosity increase. an essential role in the charring
process, but most likely does not contribute to the observed viscosity increase.
Nanomaterials 2019,
Nanomaterials 9, 665
2019, 9, 665 11 of
11 of 25
26
Figure 4. Pictures of 2 wt. % cellulose nanocrystals (CNC) in low-salinity water (LSW) after aging at
Figure
120 4.4. Pictures
°C for
Figure Pictures ofperiods
differentof 22wt.
wt. %%ofcellulose
cellulose
time. The nanocrystals
top row shows
nanocrystals (CNC) in low-salinity
samples
(CNC) in low-salinity water
after aging times
water (LSW) after
of 0, after
(LSW) 12, aging
24aging at
and 48
at
120
h. ◦
The°C for
bottom different
row periods
shows of time.
samples The
after 4 top
daysrow
to 6shows
weeks.samples after aging times of 0,
120 C for different periods of time. The top row shows samples after aging times of 0, 12, 24 and 48 h. 12, 24 and 48
h. The
The bottom
bottom rowrow shows
shows samples
samples after
after 4 days
4 days to 6toweeks.
6 weeks.
The color change followed the same color scheme observed by Molnes et al. [18], but over
The
The color
dissimilar time change
color changeThere,
scales. followed
followedthe the the same
samplessamewerecolor scheme
scheme observed
colorcompletely observed
black after by
by Molnes
Molnes
one week, et al.
al. [18],
et while [18], but
but over
6 weeks over
of
dissimilar
dissimilar time
time scales.
scales. There, the samples
samples were
were completely
completely black
black
aging time were required before true black areas were observed. The dissimilar time scales are after
after one
one week,
week, while
while 6
6 weeks
weeks of
of
aging time
aging time
believed were
to be were required
due required before
to differing beforetrue
amounts black
true of areas
black
sample were
areas observed.
in were The
observed.
the bottles, dissimilar
as this The time
dissimilar
is the scales
only known are believed
time difference
scales are
believed
to be due the
between be due toamounts
to differing
samples differing ofamounts
of the current sample studyofthe
in sample
and theinsamples
bottles, the bottles,
as this theas
isfrom thisMolnes
only
the is the only
known et al.known
difference difference
[18]between
study. the
The
between
samples ofthethe samples
current of
studythe current
and the study
samples and
from the
the samples
Molnes
current study used approximately three times as much sample in containers of identical size. from
et al. the
[18] Molnes
study. The et al. [18]
current study.
study The
used
current study
approximately
Therefore, used
afterthree approximately
times
the viscosity as much
increased three
sample andintimes as ofmuch
containers
the rate sample
of identical
diffusion in containers
size.
diminished, Therefore,
a smallerof identical
after size.
the viscosity
volume was
Therefore,and
increased
accessible toafter
thethe
rateviscosity
chemically increased
of diffusion
reactive species. andThis
diminished, the rate of diffusion
a smaller
hypothesis volume was
is based diminished,
accessible
largely on atosmaller
chemically
the color volume was
reactive
gradient
accessible
species. This to chemically
hypothesis
observable across the sample height.isreactive
based species.
largely on This
the hypothesis
color gradient is based
observable largely
across on thethe color
sample gradient
height.
observable
Resultsacross the sample
from Molnes
Molnes et al.height.
et al. [18]
[18] show
showdrastically
drasticallyincreased
increasedviscosity
viscosityofofCNC CNCdispersions
dispersionsafter after2424h
ofof Results
heat aging from
atat120 ◦ C. These
Molnes et al. [18] show
results are drastically increased viscosity of CNC dispersions after 24
h heat aging 120 °C. These results arerepeated
repeatedand and expanded
expanded upon
uponinin the
thecurrent
current study.
study.Figure
Figure 5
h of
shows heat aging
these at 120
viscosity °C. These
results for results
2 wt. %are repeated
samples
5 shows these viscosity results for 2 wt. % samples of CNC, demonstrating of and
CNC, expanded
demonstrating upon in the
that current
the study.
viscosity Figure
increase
5 shows
begun these viscosity
immediately and results
Figure for 2 wt. %
6 shows thatsamples of CNC,
the viscosity demonstrating
reached a maximum thatafter
the viscosity
about 12 increase
h, after
begunthe
which immediately
viscosity was andstable
Figurefor 6 shows
over 16that weeks.the viscosity
The heat reached
aging a
study maximum
was after
subsequently
weeks. The heat aging study was subsequently terminated, about 12 h, after
terminated,
which the
without viscosityany
observing
observing wasviscosity
stable forreduction.
over 16 weeks. Heggset
Heggset Theet heat
et al. aging
al. studythat
[30] found wasduring
subsequently terminated,
high-temperature
without
heat
heat aging
agingobserving
of CNC any andviscosity
CNF a form reduction.
of Heggset et al. [30]
of thermo-oxidative
thermo-oxidative found that
degradation during
called high-temperature
oxidative-reductive
heat aging of CNC
depolymerization, andplace,
takes CNF which a formis of thermo-oxidative
propagated via hydroxyl degradation
radicals. called oxidative-reductive
depolymerization, takes place, which is propagated via hydroxyl radicals.
Figure 5. Shear viscosity curves of 2 wt. % CNC aged at 120 ◦ C up to 12 h. The displayed viscosity
Figure 5. Shear viscosity curves of 2 wt. % CNC aged at 120 °C up to 12 h. The displayed viscosity
values were measured at shear rate 1 s−1 from the first cycle of decreasing shear rate.
Figurewere
values 5. Shear viscosity
measured curves
at shear of 12 swt.
rate % CNC
−1 from agedcycle
the first at 120 °C up to 12shear
of decreasing h. The displayed viscosity
rate.
values were measured at shear rate 1 s−1 from the first cycle of decreasing shear rate.
The persistence
The persistence over
over time
time of
of this
this high
high viscosity
viscosity state
state is
is highly advantageous
advantageous for
for oil
oil reservoir
reservoir
The persistence over time of this high viscosity state is highly
highly advantageous for oil reservoir
applications as
applications as residence
as residence time
residence time in the reservoir may be in the order of many months.
applications time in
in the
the reservoir
reservoir may
may be
be in
in the
the order
order of
of many
many months.
months.
Figure 6.
6. Shear
Shear viscosity
viscosity ofof 22wt.
wt. %
% CNC
CNC aged aged at
at 120
120◦°C instatic
C in
in static containers. The
The displayed
displayed viscosity
viscosity
Figure
Figure 6. Shear viscosity of 2 wt. % CNC aged at 120 °C static containers.
containers. The displayed viscosity
values were
values were measured
were measured at shear rate 1 s−1 from the first cycle of decreasing shear rate.
−1
values measured at at shear
shear rate
rate 11 ss−1 from
fromthe
thefirst
firstcycle
cycleofof decreasing
decreasing shear
shear rate.
rate.
Theshear
The shearviscosity
viscosityvalues
valuesandandshape
shape ofof curves
curves fitfit well
well with
with thethe results
results of Molnes
of Molnes et al.et[18],
al. [18], but
but the
The shear viscosity values and shape of curves fit well with the results of Molnes et al. [18], but
the onset
onset of the
of the viscosity
viscosity increase
increase at static
at static conditions
conditions hashasnotnot previously
previously beenstudied.
been studied.The
Thelong-term
long-term
the onset of the viscosity increase at static conditions has not previously been studied. The long-term
stability of increased
stability increased viscosity
viscosityisisalso
alsoelucidated
elucidatedinin
thethecurrent study.
current study.TheThe
aforementioned
aforementioned study had
study
stability of increased viscosity is also elucidated in the current study. The aforementioned study had
suggested some difference in onset and rate of viscosity increase
had suggested some difference in onset and rate of viscosity increase depending on depending on initial CNC
CNC
suggested some difference in onset and rate of viscosity increase depending on initial CNC
concentrationofof
concentration thethe suspension.
suspension. Due toDue to the importance
the importance of selectingofa selecting a representative
representative CNC concentration CNC
concentration of the suspension. Due to the importance of selecting a representative CNC
concentration in the current study, corresponding to relevant core flooding parameters,
in the current study, corresponding to relevant core flooding parameters, the concentration dependence the
concentration in the current study, corresponding to relevant core flooding parameters, the
concentration
of dependence
the viscosification of the viscosification
phenomenon phenomenon
was also studied. wasare
The results alsoshown
studied. The results
in Figure 7, and areproved
shown
concentration dependence of the viscosification phenomenon was also studied. The results are shown
in Figure
vital 7, and provedthe
for understanding vital for understanding
viscosification the viscosification time scale.
time scale.
in Figure 7, and proved vital for understanding the viscosification time scale.
Figure 7.7. Shear

Figure Shear viscosity
viscosity at
at shear rate11ss−1
shear rate −1
of CNC aged at 120 ◦°C,
C, from thethe first
first cycle
cycle of
of decreasing
decreasing
Figure 7. Shear viscosity at shear rate 1 s−1 of CNC aged at 120 °C, from the first cycle of decreasing
shear rate.
shear rate. The
The concentrations
concentrationsof ofCNC
CNCare aregiven
givenininwt.
wt.%.%.
Aging
Agingof 1ofwt. % samples
1 wt. % samples at 6,at126,and 18 h
12 and
shear rate. The concentrations of CNC are given in wt. %. Aging of 1 wt. % samples at 6, 12 and 18 h
was
18 conducted
h was twice.
conducted Error
twice. bars
Error barssignify
signifystandard
standarddeviation
deviationbetween
between shear replicate
shear viscosity replicate
was conducted twice. Error bars signify standard deviation between shear viscosity replicate
measurements.
measurements. The other aging tests were were conducted
conducted only
only once.
once. The 1% lineline follows
follows the
the average
average
measurements. The other aging tests were conducted only once. The 1% line follows the average
between test replicates.
As
Asshown
shownin in
Figure 7, the7,2 wt.
Figure the % 2 CNC
wt. %sample
CNCreached
sample a maximum
reached aviscosity
maximum after viscosity
approximately
after
12 As shown
h of aging time,in Figure 7, the 2 wt. % CNC sample reached a maximum viscosity after
approximately 12 but
h ofthe 0.49 time,
aging wt. %but
sample showed
the 0.49 nosample
wt. % sign of change
showedatno 12 sign
h. Allofthree concentrations
change at 12 h. All
approximately
underwent 12 h of aging
a substantial time, but
viscosity the 0.49
increase wt.
after 24%h,sample
with showed
higher no signasofa change
viscosity functionat of
12higher
h. All
three concentrations underwent a substantial viscosity increase after 24 h, with higher viscosity as a
three concentrations
concentration. underwent
The concentration. a
extreme effect of substantial
heat viscosity
aging effect increase
on theof2 heat
wt. % after
CNCon 24 h, with
samples, higher viscosity
manifested as a
by both
function of higher The extreme aging the 2 wt. % CNC samples,
function
the of higher
immediate concentration. The extreme effect of heatofaging on the increase
2 wt. % CNC orderssamples,
manifested byonset
both of
theviscosity
immediate increase
onsetand
of the magnitude
viscosity increasetheand viscosity
the magnitude (three of
of the viscosity
manifested
magnitude) by both
gave the
cause immediate
for concern onset
that of
the viscosity
nanofluid increase
was and
likely to the
plugmagnitude
the core of the viscosity
completely, with
increase (three orders of magnitude) gave cause for concern that the nanofluid was likely to plug the
increase
the (three orders of magnitude) gave cause forbefore
concern that the nanofluid was likelylateto plug
onsetthe
coreconsequence
completely,ofwith
shutting down the
the consequence experiment
of shutting down results
the could be obtained.
experiment before Theresults could of
be
core completely, with the consequence of shutting down the experiment before results could be
obtained. The late onset of viscosity increase for the 0.49 wt. % CNC samples were also deemed
obtained. The late onset of viscosity increase for the 0.49 wt. % CNC samples were also deemed
problematic, as it is desirable to maintain elevated viscosity through a core section of maximum
problematic, as it is desirable to maintain elevated viscosity through a core section of maximum
viscosity increase for the 0.49 wt. % CNC samples were also deemed problematic, as it is desirable to
maintain elevated
Nanomaterials 2019, 9,viscosity
665 through a core section of maximum length to mobilize the greatest14amount of 25
of trapped oil. A compromise of 1 wt. % was chosen, with the hope of avoiding both the extremes of
as well the
plugging as for theand
core, second
CNChigh rate of through
travelling 0.5 mL/min, the pressure
the core without seemed to be stable. However, the
having aged.
second high rate did not produce any additional oil. Finally, a high injection rate of 1.0 mL/min was
4.1.2. Interfacial
applied Tension until 23.4 PV was reached. The pressure increased slightly during this stage,
and continued
but no extra oil was produced. Overall, 15.5 PV of nanofluid was injected during the nano flood stage,
The interfacial tension between crude oil and brine/nanofluid was measured at ambient pressure
and it resulted in a total oil recovery of 62.2%. This means that the nanofluid contributed to an extra
conditions using the pendant drop method. The experiments were performed at room temperature,
1.2% oil being recovered after LSW flooding.
as the equipment did not have the necessary components to perform measurements at 120 ◦ C. The results
At the end of the water- and nano-flood stage permeability measurements were conducted using
(see Appendix B), demonstrate that the nanofluid exhibited a reduced interfacial tension in comparison
four different flow rates (at 120 ° C). The effective permeability near the end of the water flooding
to stage
low salinity water. At
was calculated to the startmD
be 205.3 of the experiment,
(where µw = 0.233 thecP,IFT for CNC
Table wasthe
1), while 1.4effective
mN/m lower than the
permeability
reference
near thesample
end of theandnano
afterflooding
12 h thestagedifference was 2.3 to
was calculated mN/m for mD.
be 753.9 the two systems.
In order to takeIninto
other words,
account
thethat
interfacial tension reduction is modest. Thus, IFT reduction is, therefore, likely
the CNC had been aged for a certain time inside the core during the experiment, the viscosity of not a dominant
mechanism for incremental
the nanofluid was calculated oil according
production.to the kinetics in Figure 7. Because CNC is a non-Newtonian
fluid, the calculations were based on the viscosity after 10 h, providing a reasonable estimate for the
4.2. Oil-Recovery Experiment
average CNC viscosity inside the core during the experiment (µn = 30 cP, from Figure 7). The effective
permeability
Experimental of oil had been
results fromcalculated during
the core flood arethe oil saturation
presented stage8 of
in Figure theTable
and core and was
3. The 1965.3
core mD
flood was
(where µoat= 120
conducted 3.32 ◦cP, Tablean
C with 1). initial
From these
watervalues, the mobility
saturation (Swi ) ofratio
15%. forThe
water-
coreand
wastheaged
nanofor
flood
fivecould
weeks
be calculated
using crude oil.using Equation
The study was (2), and was 1.49
performed and 0.04,
in three respectively.
flooding stages. Thus,
First,nano flood should
low salinity water give a
(LSW)
favorable mobility ratio, which could indicate that the fluid behaved more
was injected to recover as much oil as possible, then nanocellulose was injected as an enhanced oil like a viscous polymer
front compared
recovery method to to water.
determine if incremental oil could be produced. Finally, a post flush with LSW
The post-flush stage was performed as a final step in which LSW was injected at 0.5 mL/min for
was conducted. The core flood methodology was to continue each stage until no additional oil was
4.8 PV. With this stage, the idea was that the LSW could help break-up nanocellulose agglomerates
produced and the pressure had stabilized. For the water flood 7.4 PV was injected at a low rate of
and push the nanoparticles out from the core. In addition, if the nanocellulose had released some
0.1 mL/min, which resulted in an oil recovery of 60.8%. The rate was then increased to 0.5 mL/min and
additional oil, but was unable to transport it all the way out due to the agglomerates, perhaps LSW
an additional 0.21% oil was produced. This might be capillary trapped oil that was released from the
could help facilitate this oil movement. Injection was continued until the pressure stabilized.
outlet. Figure 8, shows that the pressure stabilized for both water flood stages.
However, no additional oil was produced during this stage.
Figure
Figure 8. 8.
OilOil recoveryinin%%OOIP
recovery OOIPof oflow
low salinity
salinity water
water flood
flood(blue
(blueline),
line),nano
nanoflood
flood(green and
(green orange
and orange
line) and post flush (yellow line) as a function of injected pore volumes (time).
line) and post flush (yellow line) as a function of injected pore volumes (time). The grey lineThe grey line is is
thethe
corresponding
corresponding pressurecurves.
pressure curves.The
Theexperiment
experiment was
was conducted
conducted atat120
120°C.
◦ C.
Table 3. Detailed overview over recovery factor (RF) from each stage and total residual oil saturation
(Sor ).
Rate PV Injected Total RF RF for Each Stage Total Sor

Stage
(mL/min) (%) (%)
Water flood (low rate) 0.1 7.4 60.80 60.80 0.33
Water flood (high rate) 0.5 0.5 61.00 0.21 0.33
Nano flood (low rate) 0.1 7.5 61.93 0.92 0.32
Nano flood (low rate, after shut-in) 0.1 2.0 62.07 0.14 0.32
Nano flood (high rate, pump fail) 0.3 0.4 62.07 0.00 0.32
Nano flood (low rate, after pump fail) 0.1 4.0 62.16 0.09 0.32
Nano flood (high rate 1) 0.3 0.5 62.19 0.02 0.32
Post flush 0.5 4.8 62.19 0.00 0.32
The nano flood scheme consisted of several stages. CNCs were first injected at a similar low
rate as LSW flooding, until no additional oil was produced. From this stage, it was observed that
the pressure was considerably higher than for water flooding. By comparing the pressure at the
end of water flooding to the end of nano flooding, the pressure was 65.1 times higher for a nano
flood. Furthermore, the pressure also kept increasing and did not stabilize. However, the last oil was
produced around 14.4 PV and nanofluid injection continued until 15.4 PV was reached. Although
the differential pressure continued to increase, the increased pressure differential did not give rise to
additional oil production. Merely 0.92% incremental oil was recovered during this stage. To ensure
that the nanocellulose had sufficient time to viscosify inside the core, the core was shut in for 24 h (this
event is denoted by the red dotted line in Figure 8). After 24 h, CNC was injected through the core at
a low rate of 0.1 mL/min, and an incremental oil recovery of 0.14% was achieved. During this stage
two PVs were injected, even though oil was produced only during the first 0.21 PVs. The pressure
was spiky during this stage, and this might be attributable to log-jams building up and breaking free.
Log-jam formations occurs when particles accumulate at pore throats, causing pore constrictions and
increased differential pressure. Log-jams are a form for mechanical entrapment of particles [20]. At
approximately 16.4 PV, the pressure dropped somewhat, which might be caused by a main flow path
reopening after being constricted by nanocellulose agglomerates. The pressure never stabilized, but no
additional oil was produced so it was decided to proceed with the next flooding stage.
High-rate nano flooding was performed in three stages in order to prevent that the pressure gauge
was capped. In the first attempt for the high rate (0.3 mL/min) the pump malfunctioned at 17.7 PV, and
the core was shut in while repairing the pump. The pump was quickly repaired, but it was decided
to inject additional PV using the low rate (0.1 mL/min) to ensure proper functioning of the pump.
The low rate was continued until 21.7 PV was reached and during this stage 0.09% incremental oil
was produced. The pressure kept increasing and never stabilized as was previously observed for the
other nano flood stages, but with no large spikes of pressure as was seen before. High-rate injection
of 0.3 mL/min was then re-initiated and 0.02% incremental oil was produced. At this rate, as well as
for the second high rate of 0.5 mL/min, the pressure seemed to be stable. However, the second high
rate did not produce any additional oil. Finally, a high injection rate of 1.0 mL/min was applied and
continued until 23.4 PV was reached. The pressure increased slightly during this stage, but no extra oil
was produced. Overall, 15.5 PV of nanofluid was injected during the nano flood stage, and it resulted
in a total oil recovery of 62.2%. This means that the nanofluid contributed to an extra 1.2% oil being
recovered after LSW flooding.
At the end of the water- and nano-flood stage permeability measurements were conducted using
four different flow rates (at 120 ◦ C). The effective permeability near the end of the water flooding
stage was calculated to be 205.3 mD (where µw = 0.233 cP, Table 1), while the effective permeability
near the end of the nano flooding stage was calculated to be 753.9 mD. In order to take into account
that the CNC had been aged for a certain time inside the core during the experiment, the viscosity of
the nanofluid was calculated according to the kinetics in Figure 7. Because CNC is a non-Newtonian
fluid, the calculations were based on the viscosity after 10 h, providing a reasonable estimate for the
average CNC viscosity inside the core during the experiment (µn = 30 cP, from Figure 7). The effective
permeability of oil had been calculated during the oil saturation stage of the core and was 1965.3 mD
(where µo = 3.32 cP, Table 1). From these values, the mobility ratio for water- and the nano flood could
be calculated using Equation (2), and was 1.49 and 0.04, respectively. Thus, nano flood should give a
favorable mobility ratio, which could indicate that the fluid behaved more like a viscous polymer front
compared to water.
The post-flush stage was performed as a final step in which LSW was injected at 0.5 mL/min for
4.8 PV. With this stage, the idea was that the LSW could help break-up nanocellulose agglomerates
and push the nanoparticles out from the core. In addition, if the nanocellulose had released some
additional oil, but was unable to transport it all the way out due to the agglomerates, perhaps LSW
could help facilitate this oil movement. Injection was continued until the pressure stabilized. However,
no additional oil was produced during this stage.
A recovery factor of 61% during water flooding is a high recovery factor. Thus, it is challenging
to try to achieve a higher recovery factor after this stage. Generally, polymers can provide excellent
incremental oil recovery factors only when the recovery factor of the initial water flooding stage is poor
(0–40%) [33,34]. In the current core flood, water flooding appears to have provided excellent sweep
efficiency of the core plug. Thus, the remaining oil proves difficult to extract using a high-viscosity
aqueous front (i.e., nanocellulose fluid).
4.3. Permeability
At 100% water saturation, the core permeability was measured to be 1980 mD. After the
nanofluid core flood experiment, the core was rinsed with methanol and toluene in alternate sequences.
The permeability was then measured using methanol and was calculated to be 192.4 mD. The substantial
permeability reduction provides strong evidence of nanoparticle retention inside the core. Reverse
flooding was therefore implemented to extract large nanoparticle agglomerates that are unable to
transverse the core. Such large agglomerates would be immobilized near the core inlet, and reverse
flooding was implemented to remove the nanoparticle agglomerates from the core inlet and recover
the initial core permeability. After reverse flooding was conducted, the permeability was tested in the
reverse direction using methanol yielding a permeability value of 207 mD. This was a small increase in
permeability of 14 mD, but still represents a substantial deterioration from the initial core permeability.
Thus, the porous media suffers a significant permeability impairment during nano flooding.
The core was dismantled after the flood, and the inlet side and outlet side were further investigated.
By visually observing the inlet side one could see a thin white nanocellulose layer, functioning as a
filter cake (Figure A4, Appendix C). From the outlet side it was not possible to observe any visible
nanocellulose layer.
4.4. Characterization of the Core after the Core Flood

The micro-CT scan shows that the core plug was quite homogenous (see Appendix C, Section C.2).
SEM Imaging
In Figure 9, there are four SEM images with resolution 5.4 nm/pixel. Images of a, b, c in Figure 9
correspond to the pieces of 1, 2, and 3 of core cuttings (Figure 3). Image d is from clean core without
nanoparticles. The SEM images focus on the pore space in the core pieces. The white arrows in image
a, b, and c point out some of the particles which cannot be seen in image d. Because the nanoparticles
can be seen in both piece 1 (close to inlet) and piece 3 (close to outlet), it indicated that the nanoparticles
traversed the entire core, and were immobilized in all core regions, even though a filter cake was
formed at the inlet side (Appendix C, Figure A4). From the images it was observed that there were
more nanoparticles at the inlet side than the outlet side. However, the amount cannot be determined
quantitatively using this method, as this is a qualitative approach. The observation of nanoparticles
determined quantitatively using this method, as this is a qualitative approach. The observation of
present within the core, even after rinsing with toluene and methanol, supports the hypothesis of
nanoparticles present within the core, even after rinsing with toluene and methanol, supports the
the effluent concentration being significantly lower than the injection fluid concentration. However,
hypothesis of the effluent concentration being significantly lower than the injection fluid
selective filtration of CNC particles based on particle size is more difficult to ascertain based on the
concentration. However, selective filtration of CNC particles based on particle size is more difficult
SEM images. In comparison to the CNCs observed in AFM images in Figure A7 (see Appendix D),
to ascertain based on the SEM images. In comparison to the CNCs observed in AFM images in Figure
the nanoparticles in the SEM images seemed to be of roughly the same length range. However, the
D2 (see Appendix D), the nanoparticles in the SEM images seemed to be of roughly the same length
characteristic elongated shape of CNC particles was not seen.
range. However, the characteristic elongated shape of CNC particles was not seen.
Figure 9. Image
Figure 9. Image a,
a, b, and c:
b, and c: SEM
SEM image
image of piece 1,
of piece 1, 2,
2, and
and 33 with
with resolution
resolution 5.4
5.4 nm.
nm. White
White arrows
arrows
pointed out
pointed out some
some of
of the
the nanoparticles
nanoparticles in
in pore
pore space.
space. Image
Image d:
d: SEM
SEM image
image of
of clean
clean reference
reference core
core with
with
resolution 5.4 nm/pixel. There is no similar particles in pore space showing in the
resolution 5.4 nm/pixel. There is no similar particles in pore space showing in the image.image.
Effluent Characterization
4.5. Effluent Characterization
4.5.1. Particle
4.5.1. Particle Size
Size
CNC particles
CNC are very
particles are very anisotropic
anisotropic as as seen
seen inin AFM
AFM images
images (Figure
(Figure A7,D2, Appendix
Appendix D). D). Dynamic
Dynamic
light scattering (DLS) allows determination of an average diameter of a
light scattering (DLS) allows determination of an average diameter of a sphere with the sphere with the same diffusion
same
coefficient as the CNC particles.
diffusion coefficient as the CNC particles.
Based on
Based onthe
theparticle
particle size measurements,
size measurements, it was established
it was that the
established thatinjection fluid had
the injection an average
fluid had an
particle diameter of 139 ± 2.0 nm (Figure A6, Appendix D). Throughout
average particle diameter of 139 ± 2.0 nm (Figure D1, Appendix D). Throughout the nano the nano flood effluent
flood
samples samples
effluent were collected and the particle
were collected and thesize was size
particle measured on selected
was measured on samples
selected to see how
samples to itsee
changed
how it
during the
changed flood.theThe
during firstThe
flood. particle size measurement
first particle size measurementwas taken at 8.6at
was taken PV,
8.6which was was
PV, which before any
before
oil had been produced during the nano flood stage. The effluent particle size
any oil had been produced during the nano flood stage. The effluent particle size was slightly larger was slightly larger
(225 ±
(225 4.4 nm)
± 4.4 nm) than
than what
what was was measured
measured for for the
the unaged
unaged injection
injection fluid. The apparent
fluid. The apparent disparity
disparity in in
particle size indicates that a change in particle characteristics occurs in the porous
particle size indicates that a change in particle characteristics occurs in the porous media, because ofmedia, because of
increased temperature,
increased particle agglomeration,
temperature, particle agglomeration, and/or
and/or size-selective
size-selective filtration
filtration processes.
processes. At At 9.5
9.5 PV,
PV,
oil was
oil was produced
produced fromfrom thethe core
core and
and the
the particle
particle size
size reached
reached its
its maximum
maximum value value during
during thethe nano
nano
flood, 262 ±
flood, 262 ± 5.5 nm. In other words, it seemed like the particles were slightly increasing in size until
5.5 nm. In other words, it seemed like the particles were slightly increasing in size until
oil was
oil was being
being produced.
produced. At At 14.6
14.6 PV,
PV, the
the production
production of of oil
oil had
had stopped,
stopped, andand the
the particle
particle size
size was
was
reduced to 178 ± 2 nm, but it was still higher than the initial size. After the core had been shut in for
24 h, the nanocellulose fluid inside the core should have increased in viscosity. From the particle size
reduced to 178 ± 2 nm, but it was still higher than the initial size. After the core had been shut in for
24 h, the nanocellulose fluid inside the core should have increased in viscosity. From the particle size
measurement afterward, an increase in particle size was observed compared to the value before shut
in. However, the particle size after shut-in was not as high as it was during the oil breakthrough. One
measurement was also performed towards the end of the second low rate after shut in (20.6 PV), and
the size was approximately the same as immediately after shut in, 180 ± 13 nm. For the two high rates
at the end of the nano flood, the size seemed to have stabilized around 167 nm, these were also the
smallest sizes that were observed during the nano flood.
The particle size observations collectively suggest that the particle agglomeration occurs in the
initial stages of the nano flood, as evidenced by the increased size for the two first measurements.
Initially, all the flow paths that the water has flown through should in principle also be open for the
nanofluid. This reasoning suggests that larger agglomerates of particles have the ability to traverse
the core, as the biggest pore throats should be open. Over time, as additional particles traverse the
core, they could start to build up at the pore throat, creating a log-jam. A log-jam means that two or
more particles with sizes slightly smaller than a pore throat arrive at the pore throat together, blocking
the path, thus making it hard for large particle agglomerates to traverse the core. Over time, upon
continuous nanofluid injection, the particles will continue to block off pores and pore throats in the
core. It is believed that this occurs inside this core, and this is further supported by the pressure data
in Figure 8. During the different stages of the nano flood the pressure kept increasing, which is an
indication that either oil is being mobilized inside the core, and/or pores are being blocked off. After
the big oil production in the beginning of the nano flood, there appear to be a filtering of large particles
inside the core, as differential pressure continues to increase and particles exiting the core are smaller
in size. During the final stage of the nano flood, the particle size seemed to be relatively constant and
the pressure was also more stable. These observations confirm that the porous media was filtering the
large particles, but not blocking off more pore throats, and there were some flow paths open where the
small sized particles could traverse the core.
Formation of log-jams in high permeable channels may divert the fluid into un-swept low
permeable channels, however a premise for fluid diversion is that the subsequent fluid contains
particles that are small enough to be able to pass through the narrower pore throats. This could be
a mechanism of producing more oil, as the fluid would be introduced to new flow paths inside the
core. However, if the particles are large (which they seem to be in this case), the fluid will not have
the possibility to sweep the narrower flow paths. Therefore, the idea behind the post flush after the
nano flood was that it would potentially sweep a new area of the core. Nevertheless, it seemed like the
nanocellulose fluid had blocked off most of the core at that point, because there was no extra oil being
produced during that stage and the core had a permanently impaired permeability.
4.5.2. Rheology Measurements

Shear viscosity at low shear rate was investigated for core flood effluent from different stages of
the core flood. The results are displayed in Figure 10. Effluent samples were acquired from the same
flooding stages as those displayed in Figure A6 (Appendix D), but slightly different times. As each
hour of flooding only produced 6 mL of effluent, only 2–3 measurement replicates could be measured
for each sample. The two high-rate sample data points are missing error bars because one test replicate
was discarded from each of them due to unreasonably high viscosity values, perhaps because of
particulate matter in the sample.
Figure 10. Shear viscosity at shear rate 1 s−1 of injection fluid and core flood effluent. The dark bars are
from increasing shear rate curves and light bars are from decreasing shear rate curves.
Figure 10. Shear viscosity at shear rate 1 s−1 of injection fluid and core flood effluent. The dark bars
are
Thefrom increasingofshear
magnitude rate curves
viscosity and light
differences bars are from
observed heredecreasing
were quite shear ratewhich
small, curves.stands in stark
contrast to the evolution of viscosity observed for 2 wt. % CNC during static aging. All the effluent
The magnitude
viscosity values wereofinviscosity
the samedifferences
order of magnitudeobservedashere the were quite
injection small,
fluid. which deviations
Standard stands in stark were
contrast to the evolution of viscosity observed for 2 wt. %
substantial in relation to the difference between samples. The only effluent sample to show CNC during static aging. All theaeffluent
distinct
viscosity
deviationvalues was the were in the extracted
sample same order of magnitude
after the injection as the
of 8.5injection
PV. Thisfluid. Standard
sample was deviations
extracted in werethe
substantial in relation to the difference between samples. The only
beginning of the nano flood, and before additional oil production was observed and had slightly lower effluent sample to show a distinct
deviation
viscosity thanwas both the sample
injectionextracted
fluid andafter otherthe injection
effluent of 8.5 From
samples. PV. This
Figuresample was extracted
A6 (Appendix in the
D), a higher
beginning
particle size of can
the nano flood, and
be recorded, before additional
so aggregation, oil production
i.e., reduction was observed
of number of particles,and might
had slightly
be the
lower viscosity than both injection fluid and other effluent samples.
explanation. Dilution from the preexisting water in the core after water flood might also have resulted From Figure D1 (Appendix D),
ainhigher
lower concentration, causing viscosity to be lower. As explained in Section 4.3, a large reductionbe
particle size can be recorded, so aggregation, i.e. reduction of number of particles, might of
the explanation. Dilution from the preexisting water in the core
permeability was observed after the core flooding was finished, and retention of particles inside the after water flood might also have
resulted
pores was inassumed
lower concentration,
to be the cause. causing
Retention viscosity to be lower.
of particles inside the As explained
core wouldinlead section 4.3, a large
to a reduction of
reduction
particle concentration in the effluent, and could partly explain the low viscosity values. Theofamount
of permeability was observed after the core flooding was finished, and retention particles of
inside the pores was assumed to be the cause. Retention of particles
nanocellulose in the core and the concentration of particles in the effluent is unknown, so a quantitative inside the core would lead to a
reduction of particle concentration in the effluent, and could partly
conclusion cannot be drawn on the effect of particle retention on viscosity. The similarity in viscosity explain the low viscosity values.
The amount
between of nanocellulose
injection fluid and effluentin the core
samples andwas the concentration
unexpected, given of particles in the effluent
the evolution displayed is unknown,
in Figure 7,
so a quantitative conclusion cannot be drawn on the effect of
which shows an onset of viscosity increase between 6 and 12 h aging time at 120 C, with an eventual particle retention ◦ on viscosity. The
similarity in viscosity between
viscosity increase of approximately a factor of 100.injection fluid and effluent samples was unexpected, given the
evolution
Fromdisplayed
these results,in Figure 7, which
it seems that shows
the CNC an onset
had not of viscosity
been aged increase
withinbetween
the core6 in andany 12 h aging
manner
time at 120 °C, with an eventual viscosity increase of approximately
similar to the static samples aged in Schott-bottles. A plausible account for the discrepancy is that a factor of 100.
From
the real these results,
residence it seems
time within thethat the CNC
reservoir is nothad not been
known. If theaged within
entire porethe core in
volume of any manner
the core had
similar to the static samples aged in Schott-bottles. A plausible
been accessible to fluid flow, the residence time should be 24 h; however, much of the pore volumeaccount for the discrepancy is that the
real residence time within the reservoir is not known. If the entire
is most likely inaccessible, or at least not preferred flow paths. After the experiment was finished pore volume of the core had been
accessible
the residual to fluid flow, the residence
oil saturation was 32%, time which should
may be be 24 h; however,
a good much
estimation ofofinaccessible
the pore volume is most
pore volume,
likely inaccessible,
corresponding to aorresidence
at least not time preferred flow paths.16
of approximately After the experiment
h. Figure was finished
7 shows, however, thatthetheresidual
1 wt. %
oil saturation was 32%, which may be a good estimation of inaccessible
CNC viscosity should have increased by more than one order of magnitude after 12 h and two orders pore volume, corresponding
to
of amagnitude
residence after time 18of approximately
h. Had the CNC16nanofluid h. Figurereacted
7 shows, however,
similarly thathigh-temperature
to the the 1 wt. % CNCconditions viscosity
should have increased by more than one order of magnitude after
as the statically aged samples, higher viscosity of the effluent should have been observed, provided 12 h and two orders of magnitude
after
that the18 concentration
h. Had the CNC nanofluid
of effluent was not reacted similarly
significantly to thedue
reduced high-temperature conditions
to retention of particles as core.
in the the
statically aged samples,
An equally plausible higher viscositywhich
explanation, of thecould
effluent should relevant,
be equally have been is observed, provided that
that high-temperature the
aging
concentration of effluent was not significantly reduced due
of CNC flowing through a porous system, may elapse in a completely different manner than agingto retention of particles in the core.
An equally
performed plausible
in static explanation, which could be equally relevant, is that high-temperature
bulk conditions.
agingRegardless
of CNC flowing of thethrough
true causalitya porous system,
of the may elapse
viscosity in a completely
discrepancy, the lack of different
viscosity manner
increase thanof
aging performed in static bulk conditions.
CNC during core flooding is important, and can be evaluated in light of Equation (2) (mobility ratio).
Regardless of the true causality of the viscosity discrepancy, the lack of viscosity increase of CNC
during core flooding is important, and can be evaluated in light of Equation (2) (mobility ratio). The
The viscosity ratio between displacing fluid and oil has been calculated and is displayed in Table 4.
This can be used as a simplified alternative to mobility ratio, which requires additional data.
The shear rate inside the porous media was calculated for the various injection rates used during
the nano flood, see results in Table 5.
Table 4. Viscosity ratio between displacing fluid and crude oil (µw /µo ) at 20 ◦ C. Aged and unaged
CNC was of concentration 1 wt. %. All values are from the first cycle of decreasing shear rate.
Flow Rate Shear Rate Unaged CNC Aged CNC Effluent

(mL/min) (s−1 ) 0h 18 h 24 h Lowest visc. Highest visc.
0.1 3.1 0.04 6.24 5.07 0.02 0.04
1 32 0.03 0.96 0.98 0.03 0.06
Table 5. Calculated shear rate in core at different flow rates.
Flow Rate Shear Rate in Core

(mL/min) (s−1 )
0.1 3.2
0.3 9.7
0.5 16.1
1.0 32.2
The calculated porous media shear rates were then compared against the shear viscosity
measurements in order to predict what viscosity to expect inside the core. All the shear viscosity
values displayed in the current study are obtained at shear rate 1 s−1 , and viscosity values at this value
are quite similar to (but slightly lower than) the calculated shear rate at the flow rate chosen for core
flooding, which was 3.2 s−1 .
Viscosity of injection fluid should be larger than viscosity of crude oil in order for the additive to
be truly effective, i.e., the viscosity ratio should be larger than one. 1 wt. % CNC initially has viscosity
not much higher than pure water, but during high-temperature aging the viscosity ratio increased by
two orders of magnitude. The effluent did not exhibit higher viscosity as a result of the 24 h spent at
120 ◦ C, and the viscosity ratio was not any higher than that of the injection fluid, and well below unity.
4.5.3. Atomic Force Microscopy (AFM)

AFM images were obtained of the unaged CNC injection fluid and of the effluent acquired from
five different stages of the nano flood phase. The results are displayed in Figure A7, Appendix D.
The images before-, at peak- and after oil production appear similar to each other, which is expected as
they are subject to the same flow rate and residence time. By comparing the image with the longest
residence time (after shut-in period) against the image with much lower residence time (high rate,
1.0 mL/min) there are no distinct differences between them. Thus, there does not seem to be a clear
relationship between residence time of CNC in the core at high temperature and aggregation states.
However, the shut-in period seem to be a marker where the effluent changes, as the images after shut
in are much more similar to each other compared to the three samples obtained before shut in.
5. Conclusions
A high-temperature core flood experiment has been carried out in order to determine the potential
of cellulose nanocrystals (CNC) as a green EOR agent for tertiary recovery. The experiment was
conducted on a 60 cm Bentheimer core plug at 120 ◦ C, and data regarding pressure and produced
volumes were obtained.
Prior to the core flood, a high-temperature aging study was performed on CNC. The study showed
that the viscosity increased substantially after 24 h of heating at 120 ◦ C, with higher viscosities obtained
as a function of higher particle concentration. Based on this study it was decided to use 1 wt. % CNC
in the core flood, and the flow rate was chosen to 0.1 mL/min corresponding to 24 h flooding time for
1 PV.
An initial low salinity water flood provided an oil recovery of 61%, which is considered a high
recovery factor. For the nano flood, 1.2% incremental oil was recovered. During the different stages of
the nano flood the differential pressure kept increasing and it was also significantly higher than during
the water flood. The increasing pressure could be an indication of log-jam formation inside the core.
Log-jamming may provide a mechanism for increased oil recovery and it is likely the cause of the oil
production that was observed during the nano flood. Particle size measurements were also performed
on the effluent nanofluid at certain times during the flood, and the results obtained correspond with
the log-jamming hypothesis. In total, 15.5 PV of nanofluid were injected, which ultimately resulted in
a large portion of the pore volume being constricted by agglomerated nanoparticles. This constriction
was also evident from the permeability measurement after the flood, in which the permeability had
decreased by 89.5% from the initial value. A thin nanocellulose filter cake was also observed at the
inlet side of the core plug, which contributes to the decreased permeability. In addition, SEM images
show immobilized nanoparticles throughout the core plug, with a larger portion of particles being
retained at the inlet side of the core in comparison to the outlet side.
An effective EOR method should yield a minimum of 5% OOIP extra in recovery in order to be a
viable candidate for an industrial operations. This means that the oil recovery in this experiment should
ideally have been increased from 61 to 66%. However, since the recovery factor for the water flood is
so high, it is challenging to extract the remaining oil. From the IFT measurements, a lower IFT value
was observed for nanofluid in comparison to brine. Nevertheless, the IFT reduction was sufficiently
small so that it would most likely not have a big effect on the overall oil recovery. For a recovery factor
as high as 61%, the IFT would need to be reduced to an ultralow value (using surfactants) to be able to
achieve a higher oil recovery after the water flood. In other words, in this experiment the water flood
was highly successful in sweeping the core for oil. From micro-CT imaging of the core plug, it was
concluded that the core was quite homogenous. It is believed that the effect of the nano flood would
have been substantially improved in a heterogenic sample (layered).
The CNC nanofluid does not seem to have increased appreciably in viscosity during flooding,
as expected based on heat aging trials of CNC under static conditions. The actual nanofluid residence
time within the reservoir is unknown, but even if it is half of the maximum residence time, the increase
in viscosity observed under static conditions should be sufficient to provoke an increase in incremental
oil recovery. It is likely that the mechanism of heat aging of CNC is very different in static conditions
compared to porous flow conditions.
The above conclusions are based on one single core flood, as it was a complex experiment and
therefore not possible to do several replicates. The water flood stage contributed to a high oil recovery,
which in turn resulted in a very low recovery with nanofluid. Based on this, it is difficult to conclude
whether CNC may represent a good EOR candidate, as more core flood experiments should have been
conducted. Future experiments should, therefore, try to implement the method earlier during water
flooding, such that the recovery factor is not too high for that stage, or alternatively use a heterogeneous
sample. The findings from this study can, therefore, be looked upon as preliminary results and based
on the observations obtained it is believed that CNC has the potential as an EOR additive if tested
under slightly different conditions.
Author Contributions: Conceptualization, R.C.A., T.D.J., E.B.H., K.S. and O.T.; Data curation, formal analysis
and methodology, R.C.A., T.D.J. and H.L-S.; Funding acquisition and project administration, E.B.H., K.S. and O.T.;
Investigation, visualization and writing original draft, R.C.A. and T.D.J.; Resources, E.B.H., S.S., K.G.P., K.S. and
O.T.; Supervision, E.B.H., S.S., K.G.P., K.S. and O.T.; Validation, R.C.A., T.D.J., E.B.H., H.L.-S., S.S., K.G.P., K.S. and
O.T.; Writing — review and editing, R.C.A, T.D.J., E.B.H., H.L.-S., S.S., K.G.P., K.S. and O.T.
Funding: This research was funded by Research Council of Norway through grant 244615/E30 in the Petromaks2.
Program and through the Centres of Excellence funding scheme, project number 262644.
Nanomaterials 2019,
Nanomaterials 2019, 9,
9, 665
665 21 of
21 of 25
25
Acknowledgments: The authors would like to thank the Research Council of Norway for their financial support
through
Excellence thefunding
GreenEOR project
scheme, (grant
project 244615/E30)
number in the
262644. The
The Petromaks2
authors program,
would also toand
like to thankthrough theJohnsen
Per Olav
Olav Centresand
of
Excellence
Excellence funding scheme,
funding scheme, project
project number
number 262644.
262644. The authors
authors would
would also
also like
like to thank
thank Per
Per Johnsen and
Olav Johnsen and
Birgitte H.
Birgitte H. McDonagh
McDonagh for for acquiring
acquiring the
the AFM
AFM images.
images. A A big
A big thank you
big thank
thank you to
you to Martin
Martin Raphaug
Raphaug and
Raphaug and Torleif
Torleif Holt
Holt at
at
SINTEF Petroleum
SINTEF Petroleum for
for guidance
guidance and
and help
help during
during the
the core
core flood
flood experiment.
experiment. Thanks
Thanks
Thanks toto Ole
to Ole Tore Buset
Ole Tore Buset from
from the
the
department
department of
department of physics
of physics at
physics NTNU for
at NTNU for obtaining
obtaining the
the micro-CT
micro-CT images
images atat their
their X-ray
X-ray laboratory.
laboratory. Lastly,
Lastly, the
laboratory. Lastly, the authors
the authors
authors
would
wouldlikeliketoto
tothank Amin
thank AminHossein
HosseinZavieh at NTNU
Zavieh Nanolab/NorFab
at NTNU
NTNU for acquiring
Nanolab/NorFab the SEMthe
for acquiring
acquiring images.
SEM The Research
images. The
would like thank Amin Hossein Zavieh at Nanolab/NorFab for the SEM images. The
Council
ResearchofCouncil
Norway
Council ofis acknowledged
Norway for the support
is acknowledged
acknowledged the to
for the the Norwegian
support to the Micro- and
the Norwegian
Norwegian Nano-Fabrication
Micro- Facility,
and Nano-Fabrication
Nano-Fabrication
Research of Norway
NorFab, project number 245963/F50. is for support to Micro- and
Facility, NorFab,
Facility, NorFab, project
project number
number 245963/F50.
245963/F50.
Conflicts of Interest: The authors declare no conflict of interest.
Conflicts of
Conflicts of Interest:
Interest: The
The authors
authors declare
declare no
no conflict
conflict of
of interest.
interest.
Appendix A
Appendix A
Appendix A
Figure A1.
Figure A1. Inside
Inside view
view ofof micro
micro computed
computed tomography
tomography (micro-CT)
(micro-CT) scanner.
scanner. On
On the left
On the left is
is the
the x-ray
x-ray
source. The sample is sitting on the rotating stage. After one sub scan, the sample will be moved
source. The sample is sitting on the rotating stage. After one sub scan, the sample will be moved one one
step up
step up for
for another
another sub
sub scan.
scan.
Figure A2.
Figure A2. On
On the
the left
left is
is one
one micro-CT
micro-CT slice.
slice. The
The black area area around
around the
the image
image is is air.
air. The
The grey
grey value
value
Figure A2. On the left is one micro-CT slice. The black
black area around the image is air. The grey value
represents matrix
represents matrix
matrixin in the
inthe sample.
thesample.
sample.TheThe black
Theblack area
blackarea in
areainin the slice represents pore space. The top image in
represents thethe slice
slice represents
represents pore
pore space.
space. TheThe
toptop image
image in
in the
the middle
the middle is one zoom in area indicated with a black square in micro-CT slice, to give better
middle is oneiszoom
one in zoom in area indicated
area indicated with
with a black a black
square square in
in micro-CT micro-CT
slice, slice, visualization
to give better to give better of
visualization of
visualization of sample
sample components.
components. TheThe histogram
histogram of of this
this micro-CT
micro-CT slice
slice is
is in
in the
the middle.
middle. TheThe red
red
sample components. The histogram of this micro-CT slice is in the middle. The red rectangle indicates
rectangle indicates
rectangle indicates the grey grey value
value range
range ofof pore
pore space.
space. The bottom
bottom image in in the
the middle
middle is is the
the
the grey value rangethe of pore space. The bottom image in theThe middle is theimage
segmented image when the
segmented
segmentedvalueimage
image when
when thethe threshold value applied. The complete segmented result of this micro-CT
threshold applied. Thethreshold
completevalue applied.
segmented The of
result complete segmented
this micro-CT slice result
is givenof on
thisthe
micro-CT
right.
slice is
slice is given
given onon the
the right.
right.
Appendix B
Appendix B
Appendix B
Results of Interfacial Tension Measurements
B.1. Results
B.1. Results ofof Interfacial
Interfacial Tension
Tension Measurements
Measurements
Interfacial tension between crude oil and brine (0.1 wt. % NaCl) was measured as a reference case,
Interfacial tension
Interfacial tension between
between crude
crude oil
oil and
and brine
brine (0.1
(0.1 wt.
wt. % NaCl) was was measured
measured as as aa reference
reference
so that the IFT to nanofluid and oil could be compared against%it.NaCl)
Four replicate measurements were
case, so that the IFT to nanofluid and oil could be compared against it. Four replicate measurements
case, so that the IFT to nanofluid and oil could be compared against it. Four replicate measurements
were performed
were performed for for the
the reference
reference sample
sample and
and the
the average
average IFT
IFT of
of those
those isis plotted
plotted in
in Figure
Figure B1B1 (grey
(grey
line). The average droplet size was 53 ± 2.3 µL and the average temperature was 23 ±
line). The average droplet size was 53 ± 2.3 µL and the average temperature was 23 ± 1 °C. From the1 °C. From the
performed for the reference sample and the average IFT of those is plotted in Figure A3 (grey22line).
Nanomaterials
Nanomaterials2019,
2019,9,9,665
665 22of
of2525
The average droplet size was 53 ± 2.3 µL and the average temperature was 23 ± 1 ◦ C. From the graph,
it is observed
graph,
graph, itit is that IFT that
is observed
observed did
thatnot
IFTstabilize
IFT did
did notnotover the duration
stabilize
stabilize over theofduration
over the the experiment
duration ofof the (12 h). These(12
the experiment
experiment experiments
(12 h).
h). These
These
will, therefore,
experiments
experimentswill, show the
will,therefore, general
therefore,show trend
showthe for
thegeneral IFT;
generaltrendthe starting
trendforforIFT; point,
IFT;the how
thestarting steep
startingpoint, the
point,how decrease
howsteep
steepthe was and
thedecrease
decreasethe
IFT
was value
wasandandthe obtained
theIFTIFTvalue after 12
valueobtainedh. For
obtainedafter the
after12 reference
12h.h.For
Forthesample, the
thereference IFT
referencesample, started
sample,the at
theIFT 22.4 ± 0.3
IFTstarted
startedatmN/m
at22.4 and
22.4±±0.3 after
0.3mN/m
mN/m 12
hand
andit was
after19.2
after 12 hh±itit0.3
12 wasmN/m.
was 19.2
19.2 ±± 0.3
0.3 mN/m.
mN/m.
For
For
For thethe nanofluid
the nanofluid
nanofluid and and crude
and crude oil,
crude oil, two
oil, two replicates
two replicates were
replicates were run,
were run, and
and the
run, and the average
the averageis
average isshown
is shownin
shown inFigure
in FigureA3
Figure B1
B1
◦ C.
(blue
(blue line).
(blue line).
line). The The average
The average droplet
average droplet size
droplet size was
size was
was 47 47 ±±
47 1.7 µL
± 1.7
1.7 µL and
µL and the
and the average
the average temperature
average temperature
temperature was was
was 26 26 ±±
26 0.5 °C.
± 0.5
0.5 °C.
The
The same
The same
same trendtrend
trend was was observed
was observed
observed in in this
in this case,
this case, with
with aaa decrease
case, with decrease
decrease in in IFT
in IFT over
IFT over time
over time and
time and the
and the value
the value
value did did not
did not
not
stabilize
stabilize over
stabilize overover 12 12
12 h. h. However,
h. However,
However, for for this
for this experiment
this experiment
experiment IFT IFT started
IFT started
started at at 21.0
at 21.0 ±
21.0 ±± 0.70.7 mN/m,
0.7 mN/m, which
mN/m, which
which is is a lower
is aa lower
lower
value
value than
value than
than whatwhat
what was was observed
was observed for
observed for the
for the reference
referencesample.
the reference sample. After
sample. After
After 1212 h,
12 h, the
h, the IFT
the IFT was
IFT was 16.9 ±
16.9
was 16.9 1.2
±± 1.2 mN/m.
1.2 mN/m.
mN/m.
Figure A3. Interfacial tension over time. Grey line is the average interfacial tension (IFT) of brine and
Figure
FigureB1.
B1.Interfacial
Interfacialtension
tensionover
overtime.
time.Grey
Greyline
lineisisthe
theaverage
averageinterfacial
interfacialtension
tension(IFT)
(IFT)of
ofbrine
brineand
and
crude oil (n = 4), while blue line is the average IFT for nanofluid and crude oil (n = 2). The experiments
crude
crudeoil
oil(n
(n==4),
4),while
whileblue
blueline
lineisisthe
theaverage
averageIFT
IFTfor
fornanofluid
nanofluidand
andcrude
crudeoil
oil(n
(n==2).
2).The
Theexperiments
experiments
lasted for 12 h and were performed at room temperature.
lasted
lastedfor
for12
12hhand
andwere
wereperformed
performedat atroom
roomtemperature.
temperature.
Appendix C
Appendix
Appendix C
C
Figure
Figure A4.
Figure C1. Left side:
C1. Left side: Inlet
Inlet with
with thin
thin nanocellulose
nanocellulose
nanocellulose filter
filter cake. Right side:
cake. Right side: Outlet
Outlet with
with no
no visible
visible
visible
nanocellulose
nanocellulose
nanocelluloselayer
layeron
onit.
it.
Results
C.2. of Micro-CT Scan
C.2. Results
Results of of Micro-CT
Micro-CT Scan Scan
The total
The total porosity ofof the core
core calculated based
based on micro-CT
micro-CT image is is 18.63% (Figure
(Figure A5), which
which
The total porosity
porosity of the the core calculated
calculated based on on micro-CT image
image is 18.63%
18.63% (Figure C2),
C2), which
is
is 3% less than measured value in the lab. Due to the limit of micro-CT technique, any component
is 3%
3% less
less than
than measured
measured value value in
in the
the lab.
lab. Due
Due to
to the
the limit
limit of
of micro-CT
micro-CT technique,
technique, any
any component
component
smaller
smaller than the resolution (27.4 µm/voxel) cannot be visualized. Therefore, the porosity via micro-CT
smaller than
than the
the resolution
resolution (27.4
(27.4 µm/voxel)
µm/voxel) cannot
cannot bebe visualized.
visualized. Therefore,
Therefore, thethe porosity
porosity via
via micro-
micro-
CT
CT can
can only
only be
be used
used asas aa reference
reference inin this
this work.
work. However,
However, thethe distribution
distribution ofof porosity
porosityalong
along the
the core
core
indicates that the sample is quite homogeneous
indicates that the sample is quite homogeneous (Figure C2). (Figure C2).
can only be2019,

Nanomaterials
Nanomaterials used
2019, as665
9, 9,
665 a
reference in this work. However, the distribution of porosity along the core
2323
of of
2525
indicates that the sample is quite homogeneous (Figure A5).
Figure A5. The plot above is the porosity profile through the whole core plug from inlet to outlet.
Figure
Figure C2.
C2. The
The plot
plot above
above is is the
the porosity
porosity profile
profile through
through the
the whole
whole core
core plug
plug from
from inlet
inlet toto outlet.
outlet. InIn
In the middle is an image of the 2D (XZ) profile along the 3D micro-CT image. At the bottom is the
themiddle
the middleis isananimage
imageofofthe
the2D2D(XZ)
(XZ)profile
profilealong
alongthe
the3D3Dmicro-CT
micro-CTimage.
image.AtAtthe
thebottom
bottomis isthe
the
corresponding segmented image of the profile image in the middle.
corresponding segmented image of the profile image in
corresponding segmented image of the profile image in the middle. the middle.
Appendix D
Appendix
Appendix DD
Figure
Figure D1.
D1. Particle
Particle size
size atat different
different stages
stages during
during the
the nano
nano flood.
flood. The
The red
red dotted
dotted line
line illustrates
illustrates thethe
Figure A6. Particle size at different stages during the nano flood. The red dotted line illustrates the
time
time when
when thethe core
core waswas shut
shut in in for
for 2424
h. h.
TheThe green
green and
and orange
orange color
color corresponds
corresponds to to the
the colored
colored lines
lines
time when the core was shut in for 24 h. The green and orange color corresponds to the colored lines in
inin the nano floodinin Figure 9.
thethe nano
nano flood
flood Figure
in Figure 9. 9.
D2.
Results
D2. Resultsofof
of AFM
Results AFM
Images
AFM Images
Images
AFM
AFM images
images were
were obtained
obtained ofof the
the unaged
unaged CNCCNC injection
injection fluid
fluid and and ofof the
the effluent
effluent acquired
acquired from
from
several
several stages
stages
stages ofof ofthe
the thenano
nano nanoflood
flood floodphase.
phase. phase.
The Theare
The
results results
results aredisplayed
are
displayed indisplayed
Figure A7. ininTheFigure
Figure D2.
particle D2. Theparticle
The particle
concentration
concentration
concentration
difference difference
seen difference
between seenthe
(a)seen
and between
between (a)(a)and
other images and
is the
duetheother
toother images
images
different is isdue
methods due todifferent
differentmethods
oftopreparation methods
for ofof
imaging.
preparation
preparation
For for
for
images (e) imaging.
imaging.
and Forimages
For
(f) it proved images(e)to
difficult (e)andand(f)
find (f)it itproved
sample proved difficult
difficult
sections totofind
with individual findsample
sample
crystals sections
sections with
with
and crystal
individual
individual crystals
crystals andand crystal
crystal aggregates
aggregates such
such asas seen
seen inin images
images (b),
(b), (c)(c) and
and (d).
(d). It It
is is unclear
unclear whether
whether
this
this phenomenon
phenomenon corresponds
corresponds totoanan actual
actual tendency
tendency towards
towards aggregation.
aggregation.
aggregates such as seen in images (b), (c) and (d). It is unclear whether this phenomenon corresponds
to an actual2019,
Nanomaterials tendency
9, 665 towards aggregation. 24 of 25
Figure
Figure A7. Atomic force
D2. Atomic force microscope
microscope (AFM)
(AFM) images
images of injection fluid
of injection fluid (a) and diluted
(a) and diluted core
core effluent
effluent from
from
the nano flood stage (b–f).
the nano flood stage (b–f).
Images (b), (c) and (d) appear similar to each other, which is expected as they are subject to the
Images (b), (c) and (d) appear similar to each other, which is expected as they are subject to the
same flow rate and residence time, the only difference between the samples being to which degree the
same flow rate and residence time, the only difference between the samples being to which degree
particles may have been in contact with crude oil. From these images there does not seem to be a clear
the particles may have been in contact with crude oil. From these images there does not seem to be a
relationship between residence time of CNC in the core at high temperature and aggregation states.
clear relationship between residence time of CNC in the core at high temperature and aggregation
Particles in image (e) have much longer residence time than those in (b), (c) and (d), while (f) would
states. Particles in image (e) have much longer residence time than those in (b), (c) and (d), while (f)
have had much lower residence time. Despite this, (e) and (f) are much more similar to each other than
would have had much lower residence time. Despite this, (e) and (f) are much more similar to each
to the other three samples.
other than to the other three samples.
References
References
1. Resource Report 2017: Even More to Gain; Norwegian Petroleum Directorate, 2017; Available online: www.
1. Resource Report 2017: Even More to Gain; Norwegian Petroleum Directorate, 2017. Avaliable online:
npd.no (accessed on 14 July 2017).
www.npd.no (accessed on 14 July 2017).
2. Sun, X.; Zhang, Y.; Chen, G.; Gai, Z. Application of Nanoparticles in Enhanced Oil Recovery: A Critical
2. Sun, X.; Zhang, Y.; Chen, G.; Gai, Z. Application of Nanoparticles in Enhanced Oil Recovery: A Critical
Review of Recent Progress. Energies 2017, 10, 345. [CrossRef]
Review of Recent Progress. Energies 2017, 10, 345.
3. Lake, L.W. Enhanced Oil Recovery; Prentice-Hall, Inc.: Upper Saddle River, NJ, USA, 1989.
3. Lake, L.W. Enhanced Oil Recovery; Prentice-Hall, Inc.: Upper Saddle River, NJ, USA, 1989.
4. Alvarado, V.; Manrique, E. Chapter 2 – Enhanced Oil Recovery Concepts. In Enhanced Oil Recovery: Field
4. Alvarado, V.; Manrique, E. Chapter 2 – Enhanced Oil Recovery Concepts. In Enhanced Oil Recovery: Field
Planning and Development Strategies; Gulf Professional Publishing: Boston, MA, USA, 2010; pp. 7–16.
Planning and Development Strategies; Gulf Professional Publishing: Boston, MA, USA, 2010; pp. 7–16.
5. Sheng, J. Modern chemical enhanced oil recovery: theory and practice; Gulf Professional Publishing: Boston, MA,
5. Sheng, J. Modern chemical enhanced oil recovery: theory and practice; Gulf Professional Publishing:
USA, 2010.
Boston, MA, USA, 2010.
6. Kamal, M.S.; Sultan, A.S.; Al–Mubaiyedh, U.A.; Hussein, I.A. Review on Polymer Flooding: Rheology,
6. Kamal, M.S.; Sultan, A.S.; Al–Mubaiyedh, U.A.; Hussein, I.A. Review on Polymer Flooding: Rheology,
Adsorption, Stability, and Field Applications of Various Polymer Systems. Polym. Rev. 2015, 55, 491–530.
Adsorption, Stability, and Field Applications of Various Polymer Systems. Polym. Rev. 2015, 55, 491–530,
[CrossRef]
doi:10.1080/15583724.2014.982821.
7. Terry, R.E. Enhanced oil recovery. In Encyclopedia of Physical Science and Technology; Academic Press: Tarzana,
7. Terry, R.E. Enhanced oil recovery. In Encyclopedia of Physical Science and Technology; Academic Press:
CA, USA, 2001; Volume 18, pp. 503–518.
Tarzana, CA, USA, 2001; Volume 18, pp. 503–518.
8. Nenningsland, A.L.; Simon, S.; Sjöblom, J. Surface properties of basic components extracted from petroleum
8. Nenningsland, A.L.; Simon, S.; Sjöblom, J. Surface properties of basic components extracted from
crude oil. Energy Fuels 2010, 24, 6501–6505. [CrossRef]
petroleum crude oil. Energy Fuels 2010, 24, 6501–6505.
9. Lyons, W.C.; Plisga, G.J. Standard Handbook of Petroleum and Natural Gas Engineering, 3rd ed.; Gulf Professional
9. Lyons, W.C.; Plisga, G.J. Standard Handbook of Petroleum and Natural Gas Engineering, 3rd ed.; Gulf
Publishing: Waltham, MA, USA, 2011.
Professional Publishing: Waltham, MA, USA, 2011.
10. Shiran, B.S.; Skauge, A. Wettability and Oil Recovery by Polymer and Polymer Particles. In Proceedings of
the SPE Asia Pacific Enhanced Oil Recovery Conference, Kuala Lumpur, Malaysia, 11 August 2015; p. 12.
11. Chang, H.L. Polymer Flooding Technology Yesterday, Today, and Tomorrow. J. Petroleum Technol. 1978,
30, 1113–1128.
12. Raney, K.H.; Ayirala, S.; Chin, R.W.; Verbeek, P. Surface and Subsurface Requirements for Successful
Implementation of Offshore Chemical Enhanced Oil Recovery. SPE Prod. Operat. 2012, 27, 294–305.
10. Shiran, B.S.; Skauge, A. Wettability and Oil Recovery by Polymer and Polymer Particles. In Proceedings of
the SPE Asia Pacific Enhanced Oil Recovery Conference, Kuala Lumpur, Malaysia, 11 August 2015; p. 12.
11. Chang, H.L. Polymer Flooding Technology Yesterday, Today, and Tomorrow. J. Petroleum Technol. 1978, 30,
1113–1128. [CrossRef]
12. Raney, K.H.; Ayirala, S.; Chin, R.W.; Verbeek, P. Surface and Subsurface Requirements for Successful
Implementation of Offshore Chemical Enhanced Oil Recovery. SPE Prod. Operat. 2012, 27, 294–305.
[CrossRef]
13. Jain, T.; Sharma, A. New Frontiers In EOR Methodologies By Application Of Enzymes. In Proceedings of the
SPE EOR Conference at Oil and Gas West Asia, Muscat, Oman, 16–18 April 2012.
14. Mukherjee, S.; Garrido, G.I.; Prasad, D.; Behr, A.; Reimann, S.; Ernst, B. Injectivity, Propagation and Retention
of Biopolymer Schizophyllan in Porous Media. In Proceedings of the SPE EOR Conference at Oil and Gas
West Asia, Muscat, Oman, 26–28 March 2018.
15. Loukopoulos-Kousis, V.; Chrysikopoulos, C.V. Use of GreenZyme®for remediation of porous media polluted
with jet fuel JP–5. Environ. Technol. 2018, 1–10. [CrossRef]
16. Wei, B.; Li, Q.; Jin, F.; Li, H.; Wang, C. The Potential of a Novel Nanofluid in Enhancing Oil Recovery. Energy
Fuels 2016, 30, 2882–2891. [CrossRef]
17. Molnes, S.N.; Mamonov, A.; Paso, K.G.; Strand, S.; Syverud, K. Investigation of a new application for cellulose
nanocrystals: a study of the enhanced oil recovery potential by use of a green additive. Cellulose 2018, 25,
2289–2301. [CrossRef]
18. Molnes, S.N.; Paso, K.G.; Strand, S.; Syverud, K.J.C. The effects of pH, time and temperature on the stability
and viscosity of cellulose nanocrystal (CNC) dispersions: implications for use in enhanced oil recovery.
Cellulose 2017, 24, 4479–4491. [CrossRef]
19. Molnes, S.N.; Torrijos, I.P.; Strand, S.; Paso, K.G.; Syverud, K. Sandstone injectivity and salt stability of
cellulose nanocrystals (CNC) dispersions—Premises for use of CNC in enhanced oil recovery. Ind. Crops
Prod. 2016, 93, 152–160. [CrossRef]
20. Aadland, R.; Dziuba, C.; Heggset, E.; Syverud, K.; Torsæter, O.; Holt, T.; Gates, I.; Bryant, S.J.N. Identification
of Nanocellulose Retention Characteristics in Porous Media. Nanomaterials 2018, 8, 547. [CrossRef]
21. Sveistrup, M.; van Mastrigt, F.; Norrman, J.; Picchioni, F.; Paso, K. Viability of Biopolymers for Enhanced Oil
Recovery. Journal of Dispersion Science and Technology 2015, 1160–1169. [CrossRef]
22. Jakobsen, T.D.; Simon, S.C.; Heggset, E.B.; Syverud, K.; Paso, K.G. Interactions between Surfactants and
Cellulose Nanofibrils for Enhanced Oil Recovery. Ind. Eng. Chem. Res. 2018. [CrossRef]
23. Dufresne, A. Nanocellulose: From Nature to High Performance Tailored Materials; Walter de Gruyter: Berlin,
Germany, 2012.
24. Klemm, D.; Heublein, B.; Fink, H.P.; Bohn, A. Cellulose: Fascinating biopolymer and sustainable raw material.
Angew. Chem. Int. Ed. 2005, 44, 3358–3393. [CrossRef] [PubMed]
25. Turbak, A.F.; Snyder, F.W.; Sandberg, K.R. Microfibrillated cellulose, a new cellulose product: Properties,
uses and commercial potential. J. Appl. Polym. Sci. 1983, 37, 815–827.
26. Herrick, F.W.; Casebier, R.L.; Hamilton, J.K.; Sandberg, K.R. Microfibrillated cellulose: morphology and
accessibility. J. Appl. Polym. Sci. Appl. Polym.Symp. 1983, 37, 797–813.
27. Rånby, B.G. Aqueous Colloidal Solutions of Cellulose Micelles. Acta Chem. Scand. 1949, 649–650.
28. Rånby, B.G.; Ribi, E. Über den Feinbau der Zellulose. Experientia 1950, 6, 12–14. [CrossRef]
29. Klemm, D.; Kramer, F.; Moritz, S.; Lindström, T.; Ankerfors, M.; Gray, D.; Dorris, A. Nanocelluloses: A New
Family of Nature–Based Materials. Angew. Chem. Int. Ed. 2011, 50, 5438–5466. [CrossRef]
30. Heggset, E.B.; Chinga–Carrasco, G.; Syverud, K. Temperature stability of nanocellulose dispersions. Carbohydr.
Polym. 2017, 157, 114–121. [CrossRef]
31. Dong, X.M.; Revol, J.-F.; Gray, D.G. Effect of microcrystallite preparation conditions on the formation of
colloid crystals of cellulose. Cellulose 1998, 5, 19–32. [CrossRef]
32. Lahiji, R.R.; Xu, X.; Reifenberger, R.; Raman, A.; Rudie, A.; Moon, R.J. Atomic Force Microscopy
Characterization of Cellulose Nanocrystals. Langmuir 2010, 26, 4480–4488. [CrossRef]
33. Abidin, A.; Puspasari, T.; Nugroho, W.J.P.C. Polymers for enhanced oil recovery technology. Proc. Chem.
2012, 4, 11–16. [CrossRef]
34. Seright, R. Potential for polymer flooding reservoirs with viscous oils. Soc. Petroleum Eng. 2010, 13, 730–740.
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).

High-Temperature Core Flood Investigation of Nanocellulose As A Green Additive For Enhanced Oil Recovery

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

High-Temperature Core Flood Investigation of Nanocellulose As A Green Additive For Enhanced Oil Recovery

Uploaded by

Copyright:

Available Formats

nanomaterials

Nanomaterials 2019, 9, 665; doi:10.3390/nano9050665 www.mdpi.com/journal/nanomaterials

mobility o f displacing f luid krw µo

2.1. Porous Media

2.4. Cellulose Nanocrystals

Table 1. Fluid properties.

Density (g/cm3 ) Viscosity (cP)

3.1. Characterization of Cellulose Nanocrystals (CNC)

3.1.1. High-Temperature Aging of CNC

3.1.2. Rheology Measurements and Shear Rate

3.1.3. Atomic Force Microscope (AFM) Pictures

3.1.4. Interfacial Tension

3.2. Core Plug Setup and Cleaning

3.3. Establish Initial Saturation

Stage Fluid Temp. Core Flow Rate PV Injected OOIP Swi

3.6. Effluent Characterization

3.6.1. Oil Production

3.6.2. Particle Size

3.7. Characterization of the Core after Core Flood

Figure 7.7. Shear

Rate PV Injected Total RF RF for Each Stage Total Sor

4.4. Characterization of the Core after the Core Flood

4.5.2. Rheology Measurements

Flow Rate Shear Rate Unaged CNC Aged CNC Effluent

Table 5. Calculated shear rate in core at different flow rates.

Flow Rate Shear Rate in Core

4.5.3. Atomic Force Microscopy (AFM)

can only be2019,

You might also like