Selection of Shale Preparation Protocol and Outgas Procedures for

Jul 24, 2017 - Selection of Shale Preparation Protocol and Outgas Procedures for Applications in Low-Pressure Analysis. Randall Holmes† , Erik C. Ru...
2 downloads 3 Views 1MB Size
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: wilcox@mines.edu

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