Subscriber access provided by CAL STATE UNIV SAN FRANCISCO
Article
Selection of shale preparation protocol and outgas procedures for applications in low-pressure analysis Randall T. Holmes, Erik C. Rupp, Vikram Vishal, and Jennifer Wilcox Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.7b01297 • Publication Date (Web): 24 Jul 2017 Downloaded from http://pubs.acs.org on July 27, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
Selection of shale preparation protocol and outgas procedures for applications in lowpressure analysis Authors: Randall Holmes1, Erik C. Rupp2, Vikram Vishal2,3, Jennifer Wilcox4* Affiliations: 1Department
of Earth System Science, Stanford University, Stanford, CA 94305 USA
2Department
of Energy Resources Engineering, Stanford University, Stanford, CA 94305 USA
3Department
of Earth Sciences, Indian Institute of Technology Bombay, Mumbai – 400076 India
4Department
of Chemical and Biological Engineering, Colorado School of Mines, Golden, CO 80401
USA Corresponding Author: Jennifer Wilcox, Associate Professor Colorado School of Mines Department of Chemical and Biological Engineering 1613 Illinois St. Golden, CO 80401
Alderson Hall 235 T: (303) 273-3885 E:
[email protected] 1 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Highlights •
Low pressure gas adsorption (LPGA) experiments on shales using CO2, N2 and Ar adsorbates
•
Use of the adsorption branch rather than the desorption branch in isotherms for LPGA
•
Cut-off temperature at 250 °C during outgassing for regeneration of ‘clean’ shale
•
Higher temperatures may lead to structural deformation of kerogen in shale
•
Best characterization achieved when relative pressure (P/Po) is limited to 0.90 and adsorption branch is used
2 ACS Paragon Plus Environment
Page 2 of 38
Page 3 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
Highlight
35
30
25
20
15
10
5
0 0
0.2
0.4
0.6
0.8
1
Relative Pressure (P/P0)
Highlight Caption: Increasing outgassing temperature results in increasing pore capacities as evidenced by adsorption isotherms shifting upward as shown here with a Barnett sample * – Room Temperature; ● – 110 °C; ♦ – 250 °C (with multiple repeats at 250 °C).
3 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1
Abstract
2
The low-pressure gas adsorption (LPGA) method for estimation of pore capacities, pore size
3
distributions, and total surface area using adsorption-desorption isotherms is selected as an
4
effective technique in pore characterization. A recent application of this method is to understand the
5
complex and heterogeneous nature of shales across the globe. The LPGA experiments were
6
conducted on shale samples from Barnett and Eagle Ford formations in the U.S. using CO2 for
7
micropores of 0.3-1.5 nm in diameter, and N2 and Ar as the adsorbates to focus on micropores from
8
1.5-2.0 nm, and the lower range of mesopores above 2.0 nm to 27 nm in diameter. It was
9
hypothesized that a significant error in estimations could occur due to inconsistencies in the shale
10
outgas temperatures. It was observed that lower pore capacities result from lower outgas
11
temperatures, and higher pore capacities result from increasing outgas temperatures. It is
12
hypothesized lower outgas temperatures fail to completely eliminate adsorbed moisture and
13
adsorbed low-molecular weight hydrocarbon species from shale pores, which leaves the pores
14
partially filled, and as such, result in lower values of pore capacity. By increasing the outgassing
15
temperature, the adsorbed species in the pores are completely removed, yielding higher pore
16
capacities. The cut-off temperature of 250 °C during outgassing for regeneration of ‘clean’ shale
17
pores was arrived at by analyzing the LPGA results of samples without any outgassing, and samples
18
outgassed at 60 °C, 110 °C, and 250 °C. The 250 °C maximum outgas temperature is intended to
19
maximize the results of LPGA while minimizing structural changes to shales. Mass stabilization as
20
shown by thermogravimetric analysis and magnetic suspension balance measurements support the 4 ACS Paragon Plus Environment
Page 4 of 38
Page 5 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
1
assertion that the shale is not fundamentally altered by processes such as kerogen cracking until a
2
temperature higher than 250 °C is reached. The kerogen had approximately 3.0% weight loss at 110
3
°C, with an additional 1.3% loss between 110 °C and 250 °C. Likewise, the desorption experiments
4
carried out on clay at 110 °C was approximately 1.3%, with an additional 0.5% loss between 110 °C
5
and 250 °C. Based on the interpretation of pore size distributions using the LPGA method, it was
6
concluded that accurate shale characterization is achieved when the analysis is limited to results
7
from relative pressures (P/Po) less than or equal to 0.90. At higher relative pressures, the sizes of the
8
adsorbate-occupied pores cannot be distinguished.
9
Keywords: Shale, nanoporosity, adsorption, Low Pressure Gas Adsorption, Thermogravimetric
10
analysis, outgassing
11 12
1. Introduction
13
1.1 Background
14
While horizontal drilling and hydraulic fracturing have revolutionized shale gas production1,
15
2,
16
characterization has also become important in efforts to combat climate change, as shales are
17
considered for geological sequestration of anthropogenic CO23-6.
18
implemented to assess the dynamics of fluid flow in shales, gas content, system integrity upon
19
injection and more7-9. An important factor for methane production or CO2 injection in shale is to have
20
reasonable estimate of gas storage capacity, whether that be gas-in-place estimates for extraction, or
21
capacity available for geologic sequestration of CO21,
22
significant momentum in recent years, some fundamental questions on capacity measurements
23
remain.
problems remain in characterizing the fundamental attributes of shale plays. The issue of shale
2, 10-13.
Multiple methods are being
While these studies have gathered
5 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1
Early investigations on shale gas plays have established that the free gas capacity is usually
2
proportionally less in comparison to the adsorbed gas content14, 15. The capacity of gas adsorbed is
3
controlled by several geologically constrained factors such as formation temperature, thermal
4
maturity, constituent grain size, moisture content, proportions of organic content and inorganic
5
matter, rock porosity and pore characteristics, etc.16-19. The standard International Union of Pure and
6
Applied Chemistry (IUPAC) definition of pore domains will be used throughout this study, to include
7
the terms micropores referring to pores less than 2 nm, mesopores from 2 to 50 nm, and the term
8
nanopores and nanoporosity which the IUPAC defines as pores less than 100 nm in diameter20-22.
9
It is well-known that the pores in the organic content of shale are the primary source of gas
10
10,
11
Although the clay component of shale also contains microporosity and mesoporosity, it remains
12
unclear what its contribution is to both gas storage and transport. Diffusive processes dominate gas
13
transport from the nanoscale storage spaces into the permeable pathways of the shale formation
14
networks. With diffusive transport, it is essential to measure pore spaces that are not considered in
15
the multi-phase flow regime of conventional petroleum reservoirs because the nanoscale pore
16
capacities of shale are significant contributors to gas storage.
while natural and hydraulically-induced fractures allow passage of gas through the reservoir23-25.
17 18
1.2 Low-pressure gas adsorption
19
An accepted technique is low-pressure gas adsorption (LPGA), which is now being used to
20
measure and identify the micropores and fine mesopores in shale. This characterization technique
21
has the advantage of being widespread, rapid and easily adapted for a variety of measurements.
22
However, standard protocols must be formally established when using the technique, as varying
23
methods of sample preparation and lack of clearly stated underlying assumptions for models used to
6 ACS Paragon Plus Environment
Page 6 of 38
Page 7 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
1
compute pore capacity estimates have further compounded the confusion caused by various and
2
often conflicting findings in literature dealing with shale pore characterization.
3
Saidian et al. (2014) concluded that subcritical N2 adsorption is a “reliable technique for both
4
clay-rich and organic-rich samples” as the gas phase can access a large range of pore sizes in both
5
the organic and inorganic constituents26-28
6
adsorption to characterize shale, the technique has limitations that are often unacknowledged. The
7
only direct measurement from LPGA is the sorption isotherms, which typically consist of both
8
adsorption and desorption branches. These isotherms show the adsorbate capacity at standard
9
temperature and pressure. Any additional characterization from the experimental data relies on
10
characterization models, which can be used to estimate surface area, pore capacity, and pore size
11
distributions (PSDs). Correct interpretation of the isotherms depends on understanding the
12
assumptions underlying characterization models, and choosing the best approximations for each
13
material and isotherm.
11, 29, 30.
Despite the widespread use of low-pressure
14
Pore size distributions estimated from adsorption data are used to gain information about
15
mesopores and micropores. The standard PSD interpretation for porous materials was proposed by
16
Barrett-Joyner-Halenda (BJH), which is a Kelvin Equation-based calculation of pore capacity that
17
uses the desorption isotherm, often with N2 at 77K, and assumes a cylindrical pore geometry with
18
typical Langmuir assumptions34. The BJH model can be useful for larger mesopores. However, the
19
underlying assumptions of BJH adsorbate layering are not appropriate for micropores and fine
20
mesopores. The simple layering of adsorbates assumed by BJH (and BET) fail to take into account
21
surface chemistry that are essential in micropores and fine mesopores (i.e. from 0.3-10 nm) such as
22
surface-to-adsorbate interactions and other molecular behaviors that change in the fine mesopores
23
and micropores. As such, BJH should not be applied to pores filled at relative pressures below 0.35.
24
Application of BJH to pores less than approximately 10 nm for Ar and N2 adsorption measurements
25
can result in underestimating pore size distributions as much as 20-30% 7 ACS Paragon Plus Environment
22, 31, 32.
Detailed
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1
information regarding surface area measurements of microporous materials has been reviewed by
2
Thommes and Cychosz33.
3
The traditional method for determining PSDs in micropores relies on CO2 adsorption
4
isotherms at 273K. Methods proposed by Dubinin and Radakovitch have been used to determine
5
micropore capacity, based on an assumed CO2 liquid density, and have been adapted by Dubinin-
6
Astakov (DA) to create micropore size distributions. While potentially insightful, the method
7
proposed by DA relies on a Gaussian distribution, which may enforce an artificial structure on the
8
microporous region35.
9
Based off of imaging data obtained from the literature, it is known that much of the shale
10
porosity is present in the kerogen, so a model that uses Ar, N2, or CO2 on carbon may be used to
11
determine PSDs in this regime using density functional theory (DFT) methods.33 The actual pore
12
geometries are complex, but based on the geological stresses imposed on shale rocks, assuming
13
basic geometries to approximate the pore structures, such as a cylindrical pore geometry for
14
mesopores, and a slit geometry for micropores, can yield valuable insights for approximating shale
15
pore capacities. The current work explores these models, and the assumptions behind them, in
16
conjunction with experimental data, to help direct future efforts associated with achieving accurate
17
gas storage estimates.
18 19
2. Experimental
20
2.1 Shale Samples
21
Shale samples for this study have been obtained from the Eagle Ford and Barnett shale
22
formations in the southwestern U.S. Composition and depth information, labeled Barnett 1-6, and
8 ACS Paragon Plus Environment
Page 8 of 38
Page 9 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
1
Eagle Ford 1-6 can be found in Table 1. For the purposes of this work, 12 samples were chosen, with
2
a range of clay and total organic carbon (TOC) contents.
3
All of the samples were prepared using a chalcedony mortar-and-pestle to grind smaller
4
shale chips sourced from well-bore cores. The ground shale samples were then sifted through a 200-
5
mesh sieve, ensuring dimensions of 75 µm or less on at least two sides. The sample was then placed
6
in a scintillation vial and shaken vigorously to ensure the particles were thoroughly mixed. Failure to
7
thoroughly mix the samples results in variation in the samples that reduces the likelihood of having
8
reproducible results. Each analysis was performed with an independent sample from each of the
9
ground and mixed powder source shales with a weight of at least 100 mg.
10
2.2 Isolated Kerogen and Illite
11
The isolated kerogen sample was isolated from a Silurian shale by Chevron and provided in a
12
dry powdered form. The illite sample is from the Green River Shale formation obtained from Ward’s
13
Natural Science Establishment, Rochester, NY.
14
2.3 TOC Analysis
15
The TOC of the shale and isolated kerogen samples was measured by removing all inorganic
16
carbon by acid digestion. Approximately 2.00 to 2.30 mg of the Barnett and Eagle Ford powdered
17
shale samples, and approximately 0.65 mg of a powdered kerogen sample from a Silurian shale,
18
were measured and placed into silver receptacles. Inorganic carbonates were removed by
19
repeatedly adding 40 ml HCl (1 N) to the samples until no effervescence was apparent under a
20
microscope. The samples were placed in an oven at 60 °C for approximately four hours in between
21
successive HCl treatments. Once no effervescence was detected, the samples were dried at 40 °C for
22
at least 96 hours. The dried samples were processed on a Carlo Erba NA 1500 Series 2 Elemental
23
Analyzer that combusts the samples at approximately 1020 °C with pure oxygen.
9 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
1
Page 10 of 38
2.4 X-ray Diffraction
2
The Barnett shale, Eagle Ford shale, and Green River Shale illite compositions were
3
determined by collecting X-ray diffractograms on a Rigaku model CM2029 powder x-ray
4
diffractometer. Diffractograms were collected over the 2q range 5° to 70° using a Cu Ka x-ray source.
5
The JADE diffraction software package was used to analyze the data (Materials Data, I
6
2002). Distinct diffraction peaks were used to match given mineral phases with peak identifications
7
from the National Institute of Standards and Technology (NIST) database. The peaks were analyzed
8
using the JADE least-squares diffractogram peak fitting and “EZ Computation”. The compositions are
9
found in Table 1. The XRD peaks for illite marked < 5% were calculated as 0% illite by the JADE
10
software. It cannot be concluded with these heterogeneous materials that there is no illite, but rather
11
the illite signature is below the detection limit of approximately 5%.
12
2.5 Thermogravimetric Analysis
13
Thermogravimetric analysis was performed using a Netzch Instruments STA 449 F3 Jupiter
14
Simultaneous Thermal Analyzer (TG-DSC/DTA). Samples were prepared as described in Section 2.1,
15
with a sample size ranging from 3.6 mg for the Silurian kerogen to 15-40 mg for the shale samples.
16
Instrument resolution is 1 μg. Each sample was run under a N2 atmosphere, with a temperature
17
increase ramp of 5 °C/min from 20 °C to 400 °C, followed by a temperature decrease ramp of 5
18
°C/min to approximately 100 °C. Each run included a blank run on the sample crucible, followed by
19
the sample being run in the same crucible.
20
2.6 Magnetic-Suspension Balance Analysis
21
Weight loss over time for the Silurian kerogen and illite clay was measured on a Rubotherm
22
magnetic-suspension balance. Samples were prepared as described in Section 2.1, with a sample size
23
of 21.6 mg for the Silurian kerogen, and 1.35 g for the illite clay sample. Instrument resolution is 1
24
μg. Each sample was run under continuous vacuum, with a final pressure of 0.06 bar. The 10 ACS Paragon Plus Environment
Page 11 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
1
temperature increase was ramped at a maximum of 5 °C/min from approximately 20 °C to 250 °C.
2
The weight was measured continuously every two minutes for approximately four days.
3
2.7 Low-Pressure Gas Adsorption Analysis
4
The LPGA analysis was performed on all samples using a commercially available
5
Quantachrome Autosorb iQ2. Any adsorption analysis requires two steps, outgassing and isotherm
6
analysis, both of which can be performed using the Autosorb equipment. All samples were outgassed
7
for final analysis at the maximum outgassing temperature of 250 °C. Outgassing on the Autosorb is
8
under vacuum of as low as 10-3 torr (0.133 pa). The 250 °C outgassing protocol was programmed to
9
ramp at 2 °C min-1 from ambient temperature to 110 °C and was held at 110 °C for 3 hours, followed
10
by a ramp of 5 °C min-1 to 250 °C and was held for 6 hours. Further discussion of the effects of outgas
11
temperature is in Section 3.1. Two samples were selected to be outgassed at progressively higher
12
temperatures to document the increase in pore capacities resulting from higher outgas
13
temperatures. Three different outgas temperatures for the Barnett 1 and Barnett 6 samples were
14
considered in this investigation: 60 °C, 110 °C and 250 °C. The 60 °C outgassing protocol was
15
programmed to ramp at 2 °C min-1 from ambient room temperature to 60 °C and was held at 60 °C
16
for at least 9 hours. The 110 °C outgassing protocol was programmed to ramp at 2 °C min-1 from
17
ambient room temperature to 110 °C and was held at 110 °C for 3 hours. The final outgassing at 250
18
°C was ramped from ambient temperature at 5 °C min-1 and held for 6 hours so as to be comparable
19
to the other 250 °C outgassed samples.
20
Low-pressure adsorption analysis was performed with 3 different adsorbate gases: CO2 at
21
273 K, N2 at 77 K and Ar at 87 K. The N2 and Ar were both research grade 5.0 gases and CO2 was
22
research-grade 4.8 gas, obtained from Praxair. The sample temperature was maintained using liquid
23
N2 (for 77 K), liquid Ar (for 87 K) and a Julabo circulating cooler with a 50/50 mixture of ethylene
24
glycol and water (for 273 K). Adsorption and desorption isotherms were obtained when using N2
11 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 38
1
and Ar as the adsorbate, while only adsorption isotherms were obtained when CO2 was used as the
2
adsorbate.
3
2.8 Model Selection
4
The N2 and Ar isotherm measurements were used to calculate pore capacities and pores size
5
distributions with Quantachrome’s quenched solid density functional theory (QSDFT) models,
6
assuming carbon as the adsorbent and cylindrical pores.
7
adsorption branch to avoid metastability issues (such as cavitation) present in the desorption
8
branch at a partial pressure of approximately 0.5, as evidenced by the characteristic desorption
9
curve drop as shown in Figure 3 for the Barnett samples, and Figure 5 for the Eagle Ford samples.
10
The use of the adsorption curves to interpret these samples has been corroborated and supported
11
by Dr. Matthias Thommes of Quantachrome (M. Thommes, personal communication, April 17, 2015).
12
The CO2 adsorption isotherms were used with Quantachrome’s non-local density functional theory
13
(NLDFT), which assume slit pore geometry on a carbon adsorbent.
14
3. Results and Discussion
15
3.1 Samples
All PSDs were calculated from the
16
Samples were ground to a powder due to previous experimental experience that attempted
17
to use shale core chips. The experimental attempts with the chips were abandoned because results
18
indicated large chips of shale did not provide consistent and repeatable isotherms. Additionally,
19
analyzing chip samples gives estimates only of interconnected pores and permeable pathways,
20
which leaves a large portion of the unconnected pore capacity unmeasured. It is important to note
21
that it is unclear how available these originally inaccessible pores will become upon stimulation or
22
fracturing, but they are of key significance as technologies improve for gas extraction and geologic
23
CO2 sequestration. Understanding the pore capacities and PSDs are important for understanding
24
potential capacities for CO2 sequestration in depleted gas wells, which will have the required 12 ACS Paragon Plus Environment
Page 13 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
1
fractures for injection purposes. However, the fractures themselves will not be responsible for
2
significant quantities of CO2 storage in the long term, but rather the pores in the geologic formation.
3
While the amount of extractable gas based on fracture networks is useful for estimates of
4
recoverable gas, it does not give estimates of total gas in place because it fails to account for the pore
5
capacities that exist, but are not connected. The inability to account for unconnected pore capacity is
6
unsuitable for shale gas reservoirs as it is standard practice to induce connectivity with techniques
7
such as hydraulic fracturing. Thus, the gas-in-place estimates for shale gas reservoirs should be
8
based on both disconnected and connected pore networks. Once a gas-in-place estimate is
9
determined, the portion of the gas that is extractable would decrease to the degree to which the
10
pores can be accessed, which would be dependent on the fracturing techniques and methods
11
employed by the reservoir engineers. As diffusive processes dominate the transport of gas out of the
12
storage space into the permeable pathways of the formation networks, it is essential to measure
13
pore spaces that are not considered in conventional petroleum reservoirs, and to include them in
14
gas-in-place estimates36. As such, these pore capacities in this case are considered a part of the total
15
porosity, which requires fully removing water and other adsorbed low-molecular-weight
16
hydrocarbon species that are considered irreducible saturations in conventional petroleum
17
reservoirs.
18
3.2 Outgas Parameters
19
The impact of the outgas procedure on low-pressure adsorption results must be considered
20
before analysis and interpretation of the adsorption results. This is true for all materials, but
21
especially in the case of shale, where there may still be CO2, H2O, and other distilled hydrocarbons
22
adsorbed. Since the interpretation techniques rely on a “clean” surface, meaning free of non-
23
structural, adsorbed species, and assume physisorption of the adsorbate, any adsorbed species that
24
remain on the surface may prevent further gas adsorption, thereby decreasing the total observed
13 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 38
1
pore capacity. This leads to inconsistency within samples and an inability to compare results across
2
multiple samples, since the initial state of the sample surface should be considered as an unknown.
3
Figure 1 shows the effect the outgas temperature (following methods described in Section 2.4) has
4
on the isotherms in the Barnett 6 sample and on the Eagle Ford 4 sample.
5
There is a clear increase in total adsorbed gas at STP for both the Barnett and Eagle Ford
6
samples as shown in Figure 1. For the Barnett 6 sample, when the sample is not outgassed
7
(maintained at ambient room temperature prior to analysis), and Eagle Ford 4 sample at 60 °C, there
8
is little porosity observed at relative pressures (P/P0) below 0.5. When the maximum outgas
9
temperature is raised to 110 °C for the Barnett 6 sample, a temperature which is expected to desorb
10
water from the pore system, there is a slight increase observed in porosity at relative pressures
11
(P/P0) below 0.25, and approximately a doubling of gas adsorbed from relative pressures (P/P0)
12
0.25 to 0.75 as compared to when the sample was not outgassed. When the maximum outgas
13
temperature is raised to 250 °C, for both samples, where most adsorbed species are expected to be
14
desorbed, there is an increase in total gas adsorbed not only (P/P0) below 0.25, but yet again an
15
approximate doubling of total gas adsorbed from relative pressures (P/P0) 0.25 to 0.75 compared to
16
when the sample was outgassed at 110 °C. This general increase in total gas adsorbed with
17
increasing outgassing temperature is represented in the isotherm plot as a shift of the entire
18
isotherm upward along the y-axis. This shift upward is interpreted as a result of desorbing gases,
19
rather than other structural changes in the shale itself.37 This is an indication that outgassing at 60
20
and 110 °C are incomplete, and unable to fully remove adsorbed species in the smallest mesopore
21
region, even if species in the larger mesopores were effectively desorbed.
22
The Barnett 6 sample has a high compositional fraction of both clay and TOC, so the effect of
23
the maximum outgas temperature may be outsized compared to samples with lower clay and TOC.
24
Figure 1 shows a similar trend for the Eagle Ford 4 sample outgassed at 60 °C and 250 °C. It is clear
25
that the presence of higher amounts of TOC affects the increase in the smallest mesopores observed, 14 ACS Paragon Plus Environment
Page 15 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
1
but both samples show an increase that can be attributed to native gas species desorbed from the
2
smallest mesopores when brought to 250 °C.
3
There is a concern when working with natural samples that the high outgas temperatures
4
may be fundamentally altering the shale and previous literature has typically taken the stance that
5
the maximum outgas temperature for shale should be around 100 °C, in order to ensure the removal
6
of adsorbed water. The liquid to vapor transition for water at 60 °C is at approximately 20 kPa (150
7
torr). Thus under a 0.01 torr vacuum typical of the outgas settings for the vacuum on the Autosorb
8
iQ2, water and heavier species will be desorbed from the shale. However, much of the scientific
9
value from low-pressure adsorption relies on the assumption of a clean surface. Any desorbed
10
species that remain after outgassing will alter the final interpretation. As such, desorption of
11
adsorbed species such as water, methane, etc., is required for LPGA, leading to the need for the
12
highest possible temperature that minimizes any fundamental alterations to the structure of the
13
shale, such as loss of kerogen mass from “cracking” that results in evolved hydrocarbons. To
14
contribute to verification that the 250 °C outgassing temperatures result in minimal structural
15
changes, we have tested each sample using TGA, as described in Section 2.3 with results shown in
16
Figure 2.
17
The key aspect of focus in the TGA results shown in Figure 2 is the initial decrease in weight
18
starting around 60 °C and ending around 110 °C. This weight loss occurs in both the isolated
19
kerogen sample and the illite sample, both of which can be attributed to the desorption of water.
20
This supports outgassing at temperatures above 110 °C. In the kerogen sample, there is then a
21
relatively stable portion between 110 °C and 200 °C, with a slight decrease from 200 °C to 250 °C.
22
This stable region is supported by the literature, which indicates that evolution of hydrocarbons
23
from kerogen does not begin with significance until temperatures greater than 250 °C. The purpose
24
of outgassing at the highest possible temperature is to maximize the LPGA results by gaining access
25
to the cleanest pores possible. Using 250 °C as an approximation for the maximum outgassing 15 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 16 of 38
1
temperature is recommended across all shale samples, as it is should lie in the vicinity of the “pre-
2
generation trough” minimum located between the S1 and S2 pyrolysis curves for most samples18.
3
The second important aspect of these results is the continual decrease in weight of the illite
4
and shale samples as the temperature of the sample increases above 110 °C. This steady decrease in
5
weight, while minor (the maximum total decrease in weight for any of the shale samples between
6
110 °C and 250 °C is 0.39%, equivalent to 6 μg) is important, and the results of this decrease can be
7
seen as the increase in capacity in the smallest mesopores as shown in Figure 4 and Figure 5. While
8
Figure 2 indicates a continuous decrease in weight with increasing temperature, it should not be
9
confused with a continuous weight loss with increased time during outgassing because the actual
10
outgassing maintains the sample at 250 °C for 6 hours, whereas the TGA is continuously ramping the
11
temperature up with time, resulting in additional weight loss when the temperature exceeds 250 °C.
12
The Rubotherm magnetic-suspension balance results illustrate the stability of the mass of the
13
kerogen and clay components of shale over several days at 250 °C. Disregarding the first 12 hours
14
while the sample is brought up to temperature allowing the balance suspension arm to mechanically
15
stabilize, the weight of both the kerogen and illite remain constant for the next 3.5 consecutive days.
16
The kerogen weight loss in the first 12 hours is responsible for a 5.1% decrease from the initial
17
weight. An additional 0.3% of the kerogen’s initial weight is lost in the subsequent 3.5 days. The illite
18
weight loss in the first 12 hours is responsible for a 1.1% decrease from the initial weight. An
19
additional 0.08% of the illite weight is lost over the subsequent 3.5 days (See Supporting
20
Information). The stabilization of the mass for several days indicates that the length of outgassing
21
time is secondary to (if not inconsequential) the outgassing temperature. Aspects of clay
22
dehydration and dehydroxilation can be found in the literature.38 39 40 39 41
23 24
3.3 Selection of Adsorption Branch
16 ACS Paragon Plus Environment
Page 17 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
1
As discussed in Section 1, there are a number of options for determining the pore size
2
distribution. Based on the limits of the traditional methods (BJH, DR/DA), DFT-based methods have
3
been chosen for the PSD interpretation. However, before the PSD distribution can be considered,
4
both adsorption and desorption isotherms should be evaluated before selecting either branch.
5
Figures 14 and 15 show the adsorption isotherms, using N2 as the adsorbate for the Barnett and
6
Eagle Ford shale samples, respectively. Both sets of adsorption isotherms are indicative of materials
7
with complex micro- and mesoporous pore networks, leading into larger macropores. This is
8
observable in the steady increase in the adsorbed capacity over the entire range of P/P0 and the
9
sharp increase in adsorbed capacity around P/P0 = 0.9. The sharp increase is representative of bulk
10
pore filling, where the adsorbate is forming a bulk phase in the macropores, but not necessarily
11
filling them. Without a plateau in the isotherm as it approaches P/P0 = 1.0, it is not possible to
12
determine the size of the pores being filled. As such, pore characterization of shales using low-
13
pressure adsorption must be limited to a region lower than P/P0 = 0.9.
14
3.4 Interpretation of PSDs
15
Using CO2 as the adsorbate, two samples, Barnett 1 and Barnett 6, were analyzed for changes
16
in micropore (the Quantachrome CO2 model is limited to pores 0.3-1.5 nm for CO2) capacities and
17
PSDs resulting from changes in outgassing temperature. The samples were outgassed at 60 °C and
18
250 °C, but neither sample showed significant changes in the micropores for pore capacities. Barnett
19
1 had CO2 pore capacity of 0.002 cc/g at 60 °C, and a pore capacity of 0.001 cc/g with the 250 °C
20
outgassing. Barnett 6 had CO2 pore capacity of 0.005 at 60 °C, and a pore capacity of 0.005 with the
21
250 °C outgassing. The samples showed only slight variations in the PSDs, as shown in Figure 18 and
22
Figure 19. As such, the remaining samples were only analyzed at the 250 °C outgassing temperature
23
for the cumulative pore capacities reported in Table 2. Additional plots of the cumulative pore
24
capacities were created by splicing CO2 and N2 results together. The spliced CO2 and N2 PSDs and
25
cumulative pore capacities of all 12 samples are shown in the SI. The pore size distributions for N2 17 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 18 of 38
1
are characteristic of those reported in the literature30, and are shown in Figure 4 for the Barnett
2
samples and Figure 6 for the Eagle Ford samples. The cumulative pore volumes increase with
3
increasing TOC and clay content, which is consistent with the literature.
4
4. Conclusions
5
Different methods to estimate the pore size distributions and total pore capacities in shales
6
yield different estimates and make a reasonably accurate estimation difficult. The low-pressure gas
7
adsorption method is a viable option to characterize shale due to its adaptability, widespread utility
8
and coverage over a large range of pore sizes relevant to gas shales. This study reviews the different
9
methodologies and identifies potential deficiencies in the gas-in-place estimates even in the best-fit
10
technique of LPGA. Experimental investigation was conducted on samples of different depths from
11
Eagle Ford and Barnett formations. The TOC content of the Eagle Ford shales varied from 2.4% to
12
5.3% while that of the Barnett shales ranged from 1.1% to 6.3%.
13
An important aspect of this study was to define an appropriate outgas procedure to
14
maximize the adsorbed gas estimates. It was found that while no outgassing yielded low pore
15
capacity estimates, raising the outgas temperatures to 250°C led to maximum desorption of
16
adsorbed species as indicated by maximum uptake of gas during the sorption phase. High
17
temperatures therefore complete the process of outgassing, giving better gas-in-place estimates
18
when the adsorbed gas can gain access to micropores and mesopores in shales. The commonly used
19
60°C and 110°C for outgassing are not recommended as the release of water and adsorbed distilled
20
hydrocarbons in the smallest nanopores is not complete. Based upon TGA results, a temperature of
21
250°C is recommended as a standard procedure for outgassing as raising the temperatures further
22
may alter the fundamental characteristics or clay and kerogen in shales. Repeatable isotherms for
23
samples outgassed at 250°C further support this claim. While the 250 °C maximum outgas
24
temperature is intended to maximize the results of LPGA while minimizing structural changes to
18 ACS Paragon Plus Environment
Page 19 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
1
shales, further spectroscopic techniques should be employed to verify any structural shifts that may
2
occur even without mass loss, as there is the potential for other kerogen deformation or
3
transformation to occur at the 250 °C maximum not accounted for even with a stable mass. As such,
4
250 °C maximum is intended as general guideline for LPGA.
5
Pore size distribution obtained from the LPGA analysis showed the characteristic drop in the
6
desorption isotherm around a relative pressure of 0.4-0.5. This drop may lead to misinterpretation
7
of the PSDs, and as such, the adsorption branch is recommended for use with shales. When the
8
isotherms end in an upward “spike” approaching relative pressure of 0.995, it is suggested that a
9
maximum pore diameter of 15nm should be taken when using this method to obtain consistent
10
results using both N2 and Ar. Samples tested on both Ar and N2 were in agreement, and either could
11
be used for LPGA on shales, although Ar is preferred as recommended by IUPAC, although N2 may
12
continue to predominate as Ar can be cost prohibitive and harder to obtain.
13 14
Acknowledgements: We thank Dr. Adam Douglas Jew for help and training on the X-ray diffraction.
15
We thank Dr. Doug McCarty and Chevron, as well as Prof. Mark Zoback and ConocoPhillips, for the
16
sample contributions. Lastly, an extended thank you to Dr. Matthias Thommes and Dr. Katie Cychosz
17
at Quantachrome Instruments for the extra time and effort spent lending their incredible scientific
18
mentorship.
19 20 21 22 23 19 ACS Paragon Plus Environment
Energy & Fuels
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
List of Tables: Table 1: XRD compositional data. Table 2: Pore capacities. List of Figures: Figure 1: Barnett 6 (left) adsorption branch isotherms as a function of maximum outgas temperature * – Room Temperature; ● – 110 °C; ♦ – 250 °C (with multiple repeats at 250 °C). Eagle Ford 4 (right) Adsorption branch isotherms as a function of maximum outgas temperature ● – 60 °C; ♦ – 250 °C.
Figure 2: Thermogravimetric analysis results for Barnett01 (magenta 5-point star), Barnett04 (black asterisk), Barnett06 (red left arrow), Barnett06_repeat (green left arrow), EagleFord03 (blue 6-point star), EagleFord05 (magenta square), illite (black right arrow), isolated kerogen (black diamond), temperature profile (black circles).
Figure 3. Adsorption -•- and desorption -∘- isotherms for Barnett samples (1-6) measured with N2. Notice drops in pore capacity at relative pressures around 0.5 in the desorption branches.
Figure 4: Cumulative pore capacity and pore size distribution for Barnett 1-6 samples ● – Cumulative Pore Capacity ○ – Change in capacity as function of pore diameter.
Figure 5. Adsorption -•- and desorption -∘- isotherms for Eagle Ford samples (1-6) on N2. Notice drops in pore capacity at relative pressures around 0.5 in the desorption branches.
20 ACS Paragon Plus Environment
Page 20 of 38
Page 21 of 38
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Energy & Fuels
Figure 6: Cumulative pore capacity and pore size distribution for Eagle Ford 1-6 samples ● – Cumulative Pore Capacity left side y axis ○ – dV(d) right side y axis.
Figure 7. The Barnett 1 (left) and Barnett 6 (right)_samples outgassed at 60 °C and 250 °C. The PSDs change slightly with only a slight change in cumulative pore capacity in the micropores.
Table 1: XRD compositional data and TOC.
Name Barnett01 Barnett02 Barnett03 Barnett04 Barnett05 Barnett06 EagleFord01 EagleFord02 EagleFord03 EagleFord04 EagleFord05 EagleFord06
Depth (m) 2616 2609 2615 2613 2625 2632 3864 3863 3893 3887 3915 3890
Quartz Calcite Pyrite Illite 34.4 64.5 0.0
