Interaction between slug fracturing fluids energized with scCO2 and shale minerals

Accepted Manuscript Interaction between slug fracturing fluids energized with scCO2 and shale minerals Aly A. Hamouda PII:

S1875-5100(17)30196-8

DOI:

10.1016/j.jngse.2017.04.025

Reference:

JNGSE 2164

To appear in:

Journal of Natural Gas Science and Engineering

Received Date: 8 March 2017 Revised Date:

6 April 2017

Accepted Date: 23 April 2017

Please cite this article as: Hamouda, A.A., Interaction between slug fracturing fluids energized with scCO2 and shale minerals, Journal of Natural Gas Science & Engineering (2017), doi: 10.1016/ j.jngse.2017.04.025. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

RI PT

ACCEPTED MANUSCRIPT

Simulated ions and pH

AC C

EP

TE D

M AN U

SC

Of the effluent

Shale core before and after treatment

Increased Pore Size

ACCEPTED MANUSCRIPT Interaction between slug fracturing fluids energized with scCO2 and shale minerals Aly A. Hamouda Department of Petroleum Engineering, University of Stavanger, Ullandhaug, 4036 Stavanger, Norway [email protected]

ABSTRACT

M AN U

SC

RI PT

Large volumes of the fluids (FF) are needed to fracture shale rocks. The driving force for conducting this research was to develop approach to minimize the needed volume of fracturing fluids and the returned FF. Detail energized FF by supercritical CO2(scCO2) interaction with shale rock is lacked in literature. The developed approach of multi-rate injection of scCO2 following a slug of FF was shown to be an efficient fracturing process (no indication of formation damage) with low returned FF, compared to injection of FF alone. Large change in the ion concentration when the approach of slug FF/scCO2 as well as the increased porosity shown by scanning electron microscope (SEM) images indicate enhanced interaction with shale’s minerals. Examined shale cores before and after treatments indicated that fractures occurred between the layers, where weak points exists. SEM images indicated pyrite (FeS2) oxidation and strontium carbonate weathering that released strontium that formed celestite. Gypsum was also observed in the outlet of the core, which may have resulted from dissolved calcium ions with the formed sulfate and / or present in the injection fluid.

Abbreviations FF-Fracturing fluid

EP

scCO2- supercritical CO2

TE D

Keywords: Shale fracturing fluid; Supercritical CO2 (scCO2), Shale minerals; Fracturing approach; Fluid minerals interaction; SEM analysis.

SEM – scanning electron microscope Introduction

AC C

1.

In the past slickwater treatment fluids were often used for fracturing shale reservoirs, however due to environmental concerns and handling the returned volumes. Typical shale gas well injects between 2and 4 million gallons of water into a deep shale reservoir (Scanlon, Reedy, and Nicot 2014). The produced water (flow-back water and shale formation water) is contaminated with the water additives such as hydrochloric acid, gelling agents and others. The produced water has therefore to be treated before disposal via deep re-injection, (Jackson et al. 2013). This has triggered the research interest to minimize the use of water in hydraulic fracturing. This could be possible by supercritical CO2, (Wang, Li, and Shen 2012). Energized fracturing fluids with CO2 does not only reduce the used water and formation damage caused by interaction with some minerals, CO2 enhances oil/gas recovery (Hamouda and Pranoto 2016) . (Shiraki and Dunn 2000), their experiments with sandstone and CO2, revealed slight dissolution of ankerite/dolomite and aluminium silicate, which may help to enhance porosity 1

ACCEPTED MANUSCRIPT

Materials and methods

2.1

Materials

SC

2.

RI PT

/permeability (Kaszuba, Janecky, and Snow 2005, Bertier et al. 2006). CO2/water/rock interactions caused by CO2 flooding of reservoirs are complex and not straightforward to simulate numerically. The fluid/rock interactions are highly reservoir specific and must be validated by experiments (Holloway 1997). Zhou et al. (2016) reported significant mineral dissolution with water compared with the case of supercritical CO2/mineral interaction case. They also concluded that the dissolution with the dynamic experiments higher than the static tests. The advantage of using scCO2, it enhances fracturing and fracturing propagation, reduces flow blocking mechanisms, increases the desorption of the adsorbed methane on organic-rich parts of the shale and sequestration (Middleton et al. 2015). The work done here was to develop a method for scCO2 energizing fracture fluid. Friehauf and Sharma (2009)

2.2

Shale fracturing apparatus

M AN U

Shale rocks, fracturing fluid (Acrylate-Acrylamide Copolymer (fracturing fluid for Silurian shale formation).

AC C

EP

TE D

Figure 1 shows a schematic of the testing apparatus of FFs. It consists of core holder, FF piston cylinder and compressed CO2 piston cylinder placed in an oven. It is equipped with pressure gauges sampling to follow the changes in the pH and the ionic composition. The experiments were performed at about 400 bar and ambient temperature. At one stage of this work it was identified by SEM analysis that oxidation process of the mineral took place. It was then decided to remove the air/oxygen from CO2 during the compression stage. The compression stage was, then, divided into two steps. Figure 1b&1c illustrate schematically the two steps. Upstream the compression system was evacuated to reduce the air/oxygen content in the vessel and tubing. The used fracturing fluid in this work was based on AcrylateAcrylamide Copolymer (fracturing fluid for Silurian shale formation). Shale composition is shown in table 1.

2

ACCEPTED MANUSCRIPT Table 1 Shale mineral content Shale W1 (% wt)

Quartz

37.67

Chlorite

8.69

Albite

8.41

Calcite

6.08

Pyrite

5.61

Muscovite

22.56

SC

RI PT

Minerals

M AN U

3. Results

Two approaches, were performed, one by bubbling CO2 (experiment 2) and the other by CO2 flooding to push the FF slug (5ml) into the shale core (experiment4).The consistency of the first approach was improved by monitoring the pH, consistent bubbling rate for the same time.

AC C

EP

TE D

The experimental results in this section is organized to illustrate the monitored pressures during the different approaches and pictures of the cores before and after the FF treatments. Figures from A-D representing 4 experimental approaches. Experiment1, figure A (a, b & c), represents injection of FF alone, experiment 2, figure B (a, b & c) represents FF mixed with CO2 (bubbled), experiment 4, figure C (a, b & c) represents slug injection of FF followed by scCO2 injection and figure D (a, b & c) represents slug injection of FF followed by injecting scCO2 stripped oxygen/air. All the figures include, the monitored pressures during the injection and the sampling points (PV) are indicated, pictures of the cores before and after the injection of FF, respectively for a, b and c. Experiment 4 and 5 were performed by injecting slug of FF (5ml) followed by scCO2 injection. It was interesting to monitor the pressure, developing fractures were indicated by occurrence of sudden pressure drop. This was observed in all the experiments, however, pressure fluctuation was observed more in the first two experiments with FF alone and with FF mixed with CO2. The pictures before and after all the experiments showed clear gaps between some layers, which may suggest that the fractures occurred between the layers where weak points may exist.

3

ACCEPTED MANUSCRIPT

RI PT

a

M AN U

SC

b

AC C

EP

TE D

c

Figure 1: Schematic of the used apparatus: (a) core flooding setup developed to test shale FFs, (b and c) are schematics of the two steps air evacuation processes and compression process, respectively.

Only one sample was obtained as marked in the pressure/PV graph due to insufficient breakthrough of fluid volume in experiments 4 and 5. This approach, then has reduced the returned FF, hence overcome/reduced the handling problem of the returned fluids. The photos after experiments 4 and 5 (slug/scCO2) demonstrated the fluid/rock interaction at the core inlet. The shown small pieces at the face rock were not sure if it happened during retrieving 4

ACCEPTED MANUSCRIPT the core or during the injection. However, there were no indication from the monitored pressure of formation damage in the core. It may be concluded here that the slug/scCO2 approach with multiple injection rates was efficient approach in penetrating the shale layers and reducing the returned fluid. Detail analysis of the retained fluids are addressed at later sections.

Inlet pressure vs PV

430

Shale WY-1

IR=0.05 ml/min

Inlet pressure [bar]

420 410

402 barIR=0.05 ml/min IR= 0.1 ml/min sample #2 pH= 8.11

400

380

IR=0.1 ml/min

350 0

10

20

M AN U

370 360

inj. stopped

sample #3

SC

sample #1

390

RI PT

Experiment 1 (a)

30

PV

40

50

sample #4 pH=7,60 60

70

AC C

EP

TE D

Experiment 1 (b&c)

Figure A: Flooded shale core FF alone – (a) monitored experimental pressure, (b & c) visual inspection of the core before and after flooding respectively. Experiment 2a

5

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

Experiment 2 (b & c)

Figure B: Experiment 2, flooded shale core with FF after bubbling CO2 – (a) monitored pressure, (b & c) visual inspection of the core before and after the experiment.

6

ACCEPTED MANUSCRIPT

Shale WY-1 Fracture fluid

Inlet pressure vs time

417,02 bar 412,24 bar

0,1 ml/min

400

IR changed to 376,66 bar 408,76 bar

Fracture fluid

300

RI PT

Inlet pressure [bar]

switched to CO2

IR changed to

500

CO2 injection started

200 100

0.0

100.0

200.0

300.0

400.0

Time [min]

600.0

700.0

M AN U

Experiment 4a

500.0

SC

0

AC C

EP

TE D

Experiment 4 (b & c)

Figure C: Shale flooding, experiment 4 – (a1 and a2) and (b) monitored pressure during the experiment, overview and closely monitored pressure, (c & d) visual inspection before and after the experiment.

7

ACCEPTED MANUSCRIPT Experiment. 5a

inlet vs time

450

IR=0,05 ml/min

400 350

#1 sample

swiched to CO2

300 250 200 150 100 50 0

20.00

40.00

60.00

80.00

100.00

Time [min]

140.00

160.00

180.00

AC C

EP

TE D

M AN U

Experiment. 5b&c

120.00

SC

0.00

RI PT

inlet pressure [bar]

Shale WY-1 C Fracture fluid and CO2

Figure D: Shale flooding, experiment, 5a) monitored pressure during the experiment, (b & c) visual inspection of the core before and after the experiment, respectively 4. Discussion SEM analysis (performed at the Silesian Technical University, Poland) revealed that the compressed CO2 possibly contained oxygen. Pyrite mineral, although in small quantities, but it may have affected the fluid/rock interaction, hence the effluent ion compositions. A 8

ACCEPTED MANUSCRIPT modified process was then, developed to reduce the oxygen content in CO2 during the compression stage (Fig.1b&c).

M AN U

SC

RI PT

Figures 2-5 show the effluent fluid ion compositions for the flooded cores (exp.1) with FF (reference), after bubbling CO2, (exp.2), slug/scCO2 , exp.4 and exp5 (stripped oxygen content). Figure 4 represents the effluent ionic fluid composition (experiment 4), where the FF slug was injected (≈5ml) followed by scCO2 injection between 22-29 PV. Fig.5 the effluent of the ions from experiment 5. Fig.6 is an overview of the percentage change of the different ions between the influent and the effluent for the four approaches. It is interesting to observe that the concentrations of [Na+], [Cl-] and [K+], have the highest effluent concentrations in the cases where the scCO2 was injected (experiment 4) followed by the bubbled CO2 into the FF (experiment 2). In other words, the energized FF with CO2 showed higher interaction with shale minerals. It was also shown that the [SO42-] in the effluent was reduced compared to the influent concentration of FF for experiments 1 and 2, but not in the cases of experiments 4 and 5 (figure 6). scCO2/FF/shale minerals (albite, calcite, quartz, chlorite and muscovite) interactions were simulated using the geochemical simulator, Phreeqc, 3.7.7. The influent ion compositions for FF and FF/CO2 are shown in table.2 Table 2 Ion compositions of FFs (Molal) Ions HCO3-

0.00

SO42Mg2+

EP

Ca2+ Na+

0.0578

0.0001

0.0001

0.0615 0.0002 0.0002 0.0178

0.0152

0.0009

0.0008

AC C

K+

FF mixed with CO2 (bubbled) 0.00

0.0058

0.0052

TE D

Cl

-

Pure FF (Molal)

9

ACCEPTED MANUSCRIPT

Exp 1 FF 0

0.0052

0.06

0.0001 0.0604 0.0152 0.0008

0.0324

0.05

Molal

0.0578 0.0001

0.0317

0.0422

0.04

0.0322 0.0004 0.0300 0.03 0.0234 0.0164 0.02 0.0008 0.0007 0.0007 0.01

0.0457 0.0398

0.0375

0.0031 0.0296

0.0022 0.0016 0.0010 0.0005 0.0008 0.0033 0.0013

0.00 HCO3-

Cl-

SO42-

Mg2+

Ca2+

Na+

Ions Sample 2

Sample 3

Sample 4

K+

FF

SC

Sample 1

0.0014 0.0011 0.0014 0.0012

RI PT

0.07

0.16

Exp 2 Mixed with CO2 (bubbled) 0.0058

0.0000

0.14

0.1270

0.12

0.0615

0.0002

0.0002

0.1456 0.0178

0.0009

0.098 0.0880 0.083 0.031 0.0022 0.0744 0.0696 0.002 0.08 0.0267 0.0301 0.0032 0.06 0.0016 0.0216 0.0461 0.0022 0.0294 0.004 0.0137 0.04 0.013 0.0017 0.0015 0.002 0.0025 0.0107 0.02 0.0064 0.0035 0.00 HCO3ClSO42Mg2+ Ca2+ Na+ K+ 0.10

TE D

Molal

M AN U

Figure 2: Effluent ion concentration during flooding with FF, experiment 1. Red values represent the influent concentrations.

EP

Ions

Sample 1

Sample 2

Sample 3

Sample 4

FF_CO2 mix

AC C

Figure 3: Effluent ion concentration during flooding with FF after being mixed with CO2, experiment 2. Red values represent the influent concentrations.

10

ACCEPTED MANUSCRIPT

Exp 4 FF Injected CO2 0.90 0.7991 0.80 0.70

0.7265 0

0.0052

0.0578

0.0001

0.0152 0.0008

0.0001

0.50

RI PT

Molal

0.60

0.40 0.30 0.20 0.10

0.0834

0.0007

0.0622

0.0401

0.0088

HCO3-

Cl-

SO42-

Mg2+

Ions FF

Ca2+

Na+

K+

M AN U

Exp.4

SC

0.00

Figure 4: Effluent ion concentration during flooding with FF followed by CO2 injection, experiment 4. Red values represent the influent concentrations.

Exp.5 _stripped O2 experiment 5

Pure FF2

TE D

0.09 0.08 0.06 0.05

[VALUE]

0.04 0.03 0.02 0.01

[VALUE] 0.003

AC C

0

0.001

0.056

EP

Molal

0.07

0.077

HCO3-

Cl-

0.002

[VALUE]

[VALUE] [VALUE] 0.018

0.004

0.002

0.003

SO42-

Mg2+

Ca2+

Na+

K+

Ions

Figure 5: Effluent ion concentration during flooding with FF followed by CO2 injection, experiment 5. Red values represent the influent concentrations.

11

ACCEPTED MANUSCRIPT +ve

- ve

HCO3-

Cl-

SO42-

Mg2+

Ca2+

Ions

Exp.1

Exp.2

Exp. 4

RI PT

100 90 80 70 60 50 40 30 20 10 0 -10 -20 -30 -40 -50 -60 -70

Na+

K+

SC

% Difference

Overview of the fluid/shale interaction

Exp.5

M AN U

Figure 6: Percentage change in influent ion concentrations. Experiments 4 and 5 have the same approach except for experiment 5 where O2 was stripped/reduced. 4.1 Simulation – ion tracking and possible reactions

The input minerals to the simulator were, quartz (SiO2), muscovite (KAl2(AlSi3O10)(OH)2 ), albite (NaAlSi3O8 ), calcite (CaCO3) and chlorite

TE D

(Mg5Al2Si2O10(OH)8 ), representing the major minerals of the used shale. The simulated and average ion concentrations were compared at simulated injection of 1PV. The simulation was done based on experiment 4 (slug FF and CO2 injection), due to its shown potential.

AC C

EP

It was difficult to compare the simulated and experimental ion concentration based on pore volume profile. This may be due to the FF and scCO2 path, which was through shale layers rather than only cutting across the shale layers from the inlet to the outlet. In addition the small fluid volume breakthrough was not enough to carry out many analysis. However, the simulation gave better understanding of the possible ion exchange (interaction) and may qualitatively suggest the order of the mineral dissolution as the fluid was injected. It was therefore decided to use simulated 1PV injection for the comparison between experimental and simulation data. In the simulation results showed that at one PV peak concentration of the ions appeared almost around that point. Experimental and simulated [Na+], [Ca2+] and [K+] were compared. The experimental sodium ion concentration was shown to be far much higher than the influent [Na+] (figure 5) and the simulation data. This may indicate dissolution of mineral-bearing sodium, such as albite, during flooding with FF/scCO2. Albite dissolution may be presented by the following equations, which as one possible source for [Na+]. CO2 + H2O = HCO3-+ H+ (1)

12

ACCEPTED MANUSCRIPT NaAlSi3O8 (albite) + CO2 +5.5 H2O = HCO3- +H+ + Na+ +2 H4SiO4 + 0.5 Al2Si2O5(OH)4 (kaolinite) (2) Figure 7 shows that all the simulated ion concentrations are higher than the experimental ones except for [Ca2+] and [SO42-].

RI PT

Ca

0.12

0.10

K

Mg Exp.4,(0.064)

0.06

Na

SO4 Exp.5 (0.004)

Mg SO4

0.00 0.00

Ca Exp.5(0.02)

M AN U

Ca Exp.4, (0.041)

0.04

0.02

SC

SO4 Exp.4,(0,084) Na Exp.5(0.08)

0.08

1.00 Ca-CH_3

2.00 Mg_CH_3

Ca exp.4

K Exp.5

Na Exp.5

Mg Exp.5

3.00

PV

4.00 K_CH_3

K Exp.4(0.0093) Mg Exp.5(0.005) Mg Exp5 (0.00350 K.Exp. 5(0.002)

5.00 SO4_CH_3

Mg Exp.4

K Exp.4

Ca Exp.5

SO4 Exp.5

TE D

[Ca2+], [Mg2+], [Na+], [SO42-] and [K+] _Molal

0.14

6.00

7.00 Na_CH_3

SO4 Exp.4

Figure 7: Comparison of experimental (dots) and simulated ion concentrations (line) for, [Na+], [Ca2+], [Mg2+], [SO42-], [K+] and [SO42-]. CH_3 denote the simulation data profiles.

AC C

EP

[K+] in both experiments 4, 5 and the simulation show concentrations of about, 0.009, 0.002 and 0.024 molal, respectively. The simulation showed higher [K+] (about 10 times) than the obtained by the experiments. The discrepancy between the simulation and the experiments may be due to the heterogeneity of mineral distribution in the core, while in the simulation it was assumed homogenous distribution. Calcium ions showed highest concentration from the simulation, it was the first to reach peak concentration followed by sodium and magnesium. This may indicate the order of the mineral dissolution. In the case of [K+], the simulated concentration profile was different from the other ions. A smooth concentration decline profile, compared to other ions, which reached maxima before they decline. The following equation represents the possible mica dissolution as a potassium bearing mineral. KAl3Si3O10(OH)2 (mica) +H2CO3 +3/2 H2O = 3/2 Al2Si2O5(OH)4 (kaolinite) +K + + HCO3(3) 13

ACCEPTED MANUSCRIPT Concentrations of [Ca2+] and [HCO3-] in the effluent increased by more than 90 per cent and about 100 per cent, respectively. There are numerous sources of HCO3- as presented by the above equations, which may explain the large relative increase of the concentrations in the effluent. The following is the overall calcite dissolution equation, which, contributes to both [Ca2+] and [HCO3-]: +

H2O

+CO2

=

HCO3-

2

+

Ca2+

RI PT

CaCO3 (4)

The above reaction is driven to the right by an increase of CO2 partial pressure and removal of HCO3- by the fluid.

SC

It is interesting to observe that calcium ions were higher in experiment 4 than in experiment 5, with an average of about 81%, the highest difference was in the case of SO42- ions, about 95%. This will be explained in the next section, by the pH and SEM images.

AC C

EP

TE D

M AN U

Figure 8 shows the simulated pH profile resulted from FF/scCO2/mineral interaction. The simulated pH may be divided into three regions. The first region, a rapid decline (-9 pH/PV) of the pH from about seven to 6.1 (PV from 0.038 to 0.138) followed by a transition with a slower decline rate (-1.8 pH/PV) to 5.4 over a PV of 0.4, and finally a pH declined at 0.58pH/PV, as the pH was reaching 3.7. It took over about 2.9 PV before it almost attained a steady state of about 3.7. The rapid rate of pH decline in the simulation indicates the low buffer capacity by the simulation compared to that obtained by the experiments. The second and third (less declined rate) region of the pH profile may indicate steady alkalinity/acidity interactions before reaching a near equilibrium state in the third region. As showed earlier, figure 7, early breakthrough of Ca ions, indicating calcite dissolution, then as the fluid pH decreases, during the flooding with scCO2 more minerals were dissolved. There are perhaps two main possible explanations to the change of the pH profile and the difference form the measured pH. The first is that, in a complex mineral blend, the affinity of certain minerals to interact with FF/CO2 influences other mineral interactions and thereby affects the dissolution rates of other minerals as indicated by the ion concentrations peaks and the pH profile produced by the simulation (figure 7). A second possibility is related to shale rock morphology. As demonstrated earlier, the fracturing fluids flow between layers hence different interaction results between the present minerals in each layer with FF/CO2. The monitored pressures for experiments 1 and 2 (figures A (exp.1a) and B (exp.2a) clearly shows more pressure fluctuations than that for experiments 4 and 5, where CO2 was injected. The increased of the pore size as shown in SEM images (figure 10) as the fluid was injected may reduce pressure fluctuations. The third possible explanation is combination of the first two possibilities. In other words, the abundance of minerals having different affinity to the FF/CO2, and the heterogeneity from one layer to the other in the same core, would influence the interactions, hence the resulted pH obtained by the experiments and the simulation. This again explain the different effluence between the ion concentrations obtained by experiments and the simulation i.e. the simulation reflects ideal behaviour, hence could be used as a reference.

14

ACCEPTED MANUSCRIPT

pH Sim

pH exp.4

pH exp.5

7.5 Exp 4 (7.3)

7.0 6.5

Exp.5 (6.5)

5.5 5.0 4.5 4.0 3.5 2

PV

4

6

8

SC

0

RI PT

pH

6.0

Figure 8: pH profile during fracturing experiments (4&5).

O2

+ 4Fe2+ + 4H+

TE D

M AN U

The mineral affinity for the interaction with fluids is extremely complex, where the interaction with one could influence the others as the fluid composition changes. In general, the produced H4SiO4, K+ and OH- increase local alkalinity. This may explain the higher pH and K+ in experiment 4 compared to experiment 5. Figure 11a shows SEM images (performed at Silisian Technical University). The figures show celestite (SrSO4) needles, which could be formed due to oxidation of FeS2 and the release of strontium from the carbonate (SrCO3) forming celestite. Celestite disappeared from SEM image Fig.11b when CO2 lean O2 was used. The dissolution of calcite (increase of pore size at the inlet), figure 10 may explain the difference in the pH between experiment 4 and 5. The following equations illustrate the oxidation of FeS2 and its influence on pH, (Bohn, McNeal, and O'connor 1985) =

(5)

αFeOOH (s) geothite +2 H+

(6)

EP

Fe2+ + 0.25 O2 + 1.5 H2O =

4Fe3+ +2H2O

AC C

Fe2+ was first liberated from pyrite (FeS2), then was oxidized to Fe 3+ and hydrolysed according to (Eq.6) causing production of protons, then the process continued, i.e. selfsustaining reaction. This signify that the presence of oxygen may lead to formation damage. In the first glance, it seems that contradict the experimentally obtained pH, however, low pH from the oxidation or due to the CO2 partial pressure, will dissolve calcite, hence increase the pH. That is perhaps explain the reason for higher pH in experiment 4 than that for experiment 5. Figure 9 gives an overall ions’ concentration profiles relative to pH profile. The source for Na+ is most likely to be from the dissolution of albite. The simulated saturation index was taken at the peak concentration, which is approximately at pH 5 and at pH 4 for the declined concentration profile. Saturation index (SI) for albite at pH5 and 4.0 were about -1.1 and 6.14. Similar analyses were done for Mg2+ and Ca2+, where, chlorite and calcite were considered to be the primary sources of Mg2+ and Ca2+, respectively. SIs for chlorite were ≈ +0.51 and ≈ –4.1, respectively at pH5 and pH4. SIs for calcite were -0.12 and -2.3, 15

ACCEPTED MANUSCRIPT respectively at pH5 and pH4. This may explain the reduced slope of mineral concentration at pH <5 to lower. Chlorite dissolution increases below pH5, as it can be seen in figure 9, magnesium concentration peaked at lower pH compared to the other ions. 1.2E-1

7.2

Ca Mg K Na pH

RI PT 6

6E-2

5.4

4E-2

0 0

1E+0

M AN U

2E-2

2E+0

3E+0

4E+0

5E+0

pH

8E-2

6.6

SC

[Ca2+], [Mg2+], [Na+] and [K+]

1E-1

4.8

4.2

3.6 6E+0

PV

Figure 9: Simulated ions’ concentrations; [Na+], [Ca2+], [Mg2+], [K+] and the pH profile.

EP

TE D

Increased pore size

Figure 10: Inlet core flooded, experiment 4

AC C

(SEM analysis was performed at Silisian Technical University, Poland).

16

RI PT

ACCEPTED MANUSCRIPT

Celestite needles

Barit

5. Summary and Conclusions

M AN U

SC

Figure 11: a&b show SEM images after the experiment 4 and 5 (outlet side of the cores, respectively: a) Barite and celestite minerals precipitated due to FeS2 oxidation and combined with Ba and Sr from carbonates weathering, b) Gypsum was not observed in the inlet side due to transport of Ca ions. (SEM analysis was performed at Silisian Technical University, Poland).

TE D

Fracturing approach of slug and multi-rate injection of scCO2 was shown to be promising. The work has also indicated that removal of air from the compressed CO2 may reduce the risk of formation damage. However, it is dependent on the type and amount of minerals that are susceptible to oxidation, such as pyrite.

AC C

EP

1. Injection stability and consistency were shown to be achieved by a multi-rate injection scheme, where injection was adjusted in accordance to the monitored pressure to avoid fluid bypass of the core or forced damage the rock. 2. Energized FF with CO2, has shown to be an efficient in reducing the used and returned FF. It was no indication of formation damage. 3. Presence of air during compression of CO2 could enhance the chance for formation damage. In this work celestite and gypsum which was not detected in the original before the treatment. In addition gypsum was not observed at the inlet, as primary origin, so may have then been precipitated at the outlet as the dissolved calcium ions reacted with the formed sulfate as a result of oxidation process of pyrite and/or the present sulfate ions in the FF. 4. Large change of ion concentration between the influent and the effluent was observed in experiment 4 (slug FF/scCO2 injection), compared to the other experiments. This may indicate larger interaction areas between scCO2 and the minerals compared to that with only FF. 5. The pH profile from the simulation showed three distinct stages reflecting various interactions that may have taken place. It was attempted to rank the order of the mineral dissolutions based on the injected PV of the respective concentration peaks as a function of pH. 17

ACCEPTED MANUSCRIPT Acknowledgements

RI PT

The author acknowledge the support of ENFLUID project and from Dr. K. Labus for his help in carrying SEM analysis at Silisian Technical University, Poland as well as SEM image analysis. I acknowledge the laboratory assistant by Krzystof Nowicki in executing the setup design and carrying the experimental flooding. Thanks to Inger Johanne for ordering needed parts for the setup.

References

SC

Bertier, P., R. Swennen, B. Laenen et al. 2006. Experimental identification of CO2–water–rock interactions caused by sequestration of CO2 in Westphalian and Buntsandstein sandstones of the Campine Basin (NE-Belgium) (in Journal of Geochemical Exploration 89 (1–3): 10-14. //www.sciencedirect.com/science/article/pii/S0375674206000446.

M AN U

Bohn, HL, BL McNeal, GA O'connor. 1985. Soil chemistry. ed (in John Willey.

Friehauf, Kyle E, Mukul M Sharma. 2009. A New Compositional Model for Hydraulic Fracturing With Energized Fluids (in SPE Production & Operations 24 (04): 562-572. Hamouda, Aly Anis, Arif Pranoto. 2016. Synergy Between Low Salinity Water Flooding and CO2 for EOR in Chalk Reservoirs. Proc.

TE D

Holloway, Sam. 1997. An overview of the underground disposal of carbon dioxide (in Energy Conversion and Management 38: S193-S198.

EP

Jackson, RE, AW Gorody, B Mayer et al. 2013. Groundwater protection and unconventional gas extraction: the critical need for field-based hydrogeological research (in Groundwater 51 (4): 488510.

AC C

Kaszuba, John P, David R Janecky, Marjorie G Snow. 2005. Experimental evaluation of mixed fluid reactions between supercritical carbon dioxide and NaCl brine: Relevance to the integrity of a geologic carbon repository (in Chemical Geology 217 (3): 277-293. Middleton, Richard S, J William Carey, Robert P Currier et al. 2015. Shale gas and non-aqueous fracturing fluids: opportunities and challenges for supercritical CO 2 (in Applied Energy 147: 500-509. Scanlon, Bridget R, Robert C Reedy, J-P Nicot. 2014. Comparison of water use for hydraulic fracturing for unconventional oil and gas versus conventional oil (in Environmental science & technology 48 (20): 12386-12393. Shiraki, Ryoji, Thomas L Dunn. 2000. Experimental study on water–rock interactions during CO 2 flooding in the Tensleep Formation, Wyoming, USA (in Applied Geochemistry 15 (3): 265-279. Wang, Haizhu, Gensheng Li, Zhonghou Shen. 2012. A feasibility analysis on shale gas exploitation with supercritical carbon dioxide (in Energy Sources, Part A: Recovery, Utilization, and Environmental Effects 34 (15): 1426-1435.

18

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Zhou, Cheng, Alvin I Remoroza, Kalpit Shah et al. 2016. Experimental study of static and dynamic interactions between supercritical CO 2/water and Australian granites (in Geothermics 64: 246-261.

19

ACCEPTED MANUSCRIPT Highlights Injection of fracturing fluid slug Followed by supercritical CO2 Fluid/rock interaction

AC C

EP

TE D

M AN U

SC

RI PT

• • •