MEASUREMENT AND INTERPRETATION OF POROSITY AND PORE-SIZE

DISTRIBUTION IN MUDROCKS:
THE HOLE STORY OF SHALES

by
Utpalendu Kuila

 Copyright by Utpalendu Kuila, 2013

All Rights Reserved

A thesis submitted to the Faculty and the Board of Trustees of the Colorado School
of Mines in partial fulllment of the requirements for the degree of Doctor of Philiosophy
(Petroleum Engineering).

Golden, Colorado
Date

Signed:
Utpalendu Kuila

Signed:
Dr. Manika Prasad
Thesis Advisor

Signed:
Dr. Douglas K. McCarty
Thesis Advisor

Golden, Colorado
Date

Signed:
Dr. William Fleckenstein
Professor and Head
Department of Petroleum Engineering

ii

ABSTRACT

The commercial success of unconventional mudrock plays have stimulated interest in understanding their porestructure. Recent development of imaging techniques have provided
valuable insights about the nanometer scale pore structure in the mudrocks, however, quantication of this pore structure still remains a challenge. Both, current industry adopted
GRI technique of measuring total porosity and the pore-size distribution (PSD) measurement using commonly used mercury intrusion porosimetry technique have serious limitations.
The focus of this thesis was to develop a quantitative understanding of the mudrock pore
structure by developing mudrocks specic porosity and PSD measurement techniques and
correlating the pore-structure attributes with composition.
Two measurement techniques were adapted and modied for measuring porosity and PSD
in mudrocks. (i) A total porosity measurement technique in intact mudrock core sample using
an immersion-saturation technique with deionized water is developed. This water immersion
porosimetry (WIP)technique gave reproducible porosity measurements for mudrocks with
an absolute experimental uncertainty of about 0.20.3 porosity unit (p.u.). (ii) Nitrogen gas
adsorption techniques was adapted for pore-size distribution analysis of mudrocks. Crushed
(<40) mesh powder is recommended for such analysis. Inversion of porestructure attributes
of the isotherm data using dierent inversion model and algorithm yields non-unique results.
The inversion method should be kept consistent for comparative purpose. The BJH model
was recommended for mudrock applications.
The pore-structure of two important compositional end members of a mudrock system,
clays and organic matter, were studied using a combination of dierent sample treatment
procedure. The results on pure clay minerals characterized the pore-structure associated
with their multiscale stacked microstructure. The most characteristic feature of a clay PSD
is the presence of incompressible pores with a modal size of 3nm. The results on organic

iii

matter indicate absence of open porosity in thermally immature organic matters. Removal
of organic matter from thermally mature mudrocks indicated presence of open pores <5 nm
diameter. For all natural mudrocks, the clay microstructure is the fundamental control in
nescale pore structure attributes. The PSD of natural mudrocks show the characteristics
3 nm modal peak indicative of the clay pore-structure. Part of the organic matter in the
mudrocks occupies the pore space within the clay microstructure. Thermally immature
organic matter (OM) does not show evidence of open porosity but ne scale pore structure
is seen to develops in OM with thermal maturation. Presence of carbonate cementation can
aect the pore connectivity in the mudrocks both in microscopic and macroscopic scales.
The quantitative measurements and understanding of pore structure in mudrocks have
dierent implications in developing exploration strategies, rock-physics, petrophysics and
ow modeling for mudrock reservoirs. The work performed in this thesis has provided two
pore structure attribute measurement techniques applicable for mudrocks and enhanced our
understanding of mudrock pore-structure.

iv

TABLE OF CONTENTS

ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi
LIST OF SYMBOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
LIST OF ABBREVIATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxv
ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xxvii

DEDICATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxx
CHAPTER 1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1

Challenges in shales/mudrocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2

Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1

Shales / Mudrocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.2.2

Organic Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.2.3

Porosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.2.4

Pore-size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.3

Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.4

Organization of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.5

List of Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

CHAPTER 2 SAMPLE CHARACTERIZATION AND DESCRIPTION . . . . . . . . 12

2.1

Sample Characterization Methods . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.1.1

Quantitative Mineral Phase (QPA) Analysis . . . . . . . . . . . . . . . 13

v

2.2

2.3

2.1.2

Organic Matter Characterization . . . . . . . . . . . . . . . . . . . . . 14

2.1.3

Textural Characterization . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.1.4

Thermogravimetric Analysis . . . . . . . . . . . . . . . . . . . . . . . . 14

Description of Sample Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.2.1

Source Clays

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.2.2

Eastern European Silurian gas shale play . . . . . . . . . . . . . . . . . 16

2.2.3

Haynesville Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.2.4

Niobrara Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.2.5

Assorted set of mudrocks . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

CHAPTER 3 TOTAL POROSITY MEASUREMENT IN MUDROCKS USING

WATER IMMERSION POROSIMETRY (WIP) TECHNIQUE . . . . . 31
3.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3.2

Current Porosity Measurement Techniques for Mudrocks . . . . . . . . . . . . 33

3.3

Total Porosity Measurement by Liquid Saturation and Immersion Techniques . 37

3.4

3.5

3.3.1

Choice of Saturating and Immersing Fluids . . . . . . . . . . . . . . . . 38

3.3.2

Pretreatment Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.4.1

WIP Measurements Calculations . . . . . . . . . . . . . . . . . . . . 41

3.4.2

WIP Measurements Procedure . . . . . . . . . . . . . . . . . . . . . . 43

3.4.3

Reproducibility test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

3.4.4

Other porosity measurements . . . . . . . . . . . . . . . . . . . . . . . 46

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

vi

3.5.1

Reproducibility Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3.5.2

Water Immersion Porosity Measurement Results . . . . . . . . . . . . . 50

3.5.3

Comparison of WIP with GRI measurements . . . . . . . . . . . . . . . 55

3.5.4

Comparison of WIP with MIP measurements

. . . . . . . . . . . . . . 55

3.6

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

3.7

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

CHAPTER 4 N2 GAS ADSORPTION TECHNIQUE FOR MEASURING

PORESTRUCTURE PARAMETERS IN MUDROCKS. . . . . . . . . 67
4.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

4.2

Theory of N2 Gas Adsorption Technique . . . . . . . . . . . . . . . . . . . . . 69

4.3

4.4

4.2.1

Experimental Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . 70

4.2.2

Qualitative Interpretation of N2 Isotherm Prole . . . . . . . . . . . . 72

4.2.3

Specic Surface Area Determination . . . . . . . . . . . . . . . . . . . 75

4.2.4

Total Specic Micropore Volume Determination: tplot Technique . . . 78

4.2.5

Total Specic Pore Volume Determination . . . . . . . . . . . . . . . . 80

4.2.6

PoreSize Distribution Analysis . . . . . . . . . . . . . . . . . . . . . . 80

Considerations on Application of N2 Gas Adsorption Techniques for

Mudrock Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.3.1

Sample Preparation: Intact or Crushed Samples? . . . . . . . . . . . . 87

4.3.2

Choice of Thickness Equation for t-plot and BJH PSD Inversion . . . . 93

4.3.3

PSD Inversion Technique: BJH or DFT? . . . . . . . . . . . . . . . . . 96

4.3.4

Graphical Representation of PSD Data . . . . . . . . . . . . . . . . . . 98

4.3.5

Reproducibility of N2 Gas Adsorption Data . . . . . . . . . . . . . . 100

Comparison of N2 gas adsorption and MIP techniques . . . . . . . . . . . . . 104

vii

4.4.1

Theory of MIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

4.4.2

Phenomenological comparison between MIP and N2 gas adsorption . 108

4.4.3

Methods and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 110

4.4.4

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

4.4.5

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

4.5

Future Recommendations

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

4.6

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

CHAPTER 5 PORESTRUCTURE IN MUDROCK END MEMBER

COMPONENTS: CLAYS AND ORGANIC MATTER . . . . . . . . . 120
5.1

5.2

5.3

Pore Structure in Clay Minerals . . . . . . . . . . . . . . . . . . . . . . . . . 121

5.1.1

PSD Results on Pure Clays and Compacted Pellets . . . . . . . . . . 122

5.1.2

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

Pore Structure in Organic Matter . . . . . . . . . . . . . . . . . . . . . . . . 130

5.2.1

Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

5.2.2

Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

5.2.3

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

5.2.4

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

CHAPTER 6 PORESTRUCTURE IN NATURAL MUDROCKS . . . . . . . . . . 154

6.1

6.2

Case Study I: Assorted set of Mudrocks . . . . . . . . . . . . . . . . . . . . . 155

6.1.1

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

6.1.2

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Case Study II: The Niobrara Formation . . . . . . . . . . . . . . . . . . . . . 162

viii

6.3

6.4

6.2.1

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

6.2.2

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

Case Study III: The Haynesville Formation . . . . . . . . . . . . . . . . . . . 171

6.3.1

Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

6.3.2

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

CHAPTER 7 APPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

7.1

Gas shale Exploration Strategies . . . . . . . . . . . . . . . . . . . . . . . . . 181

7.1.1

Predicting total porosity . . . . . . . . . . . . . . . . . . . . . . . . . 181

7.1.2

Predicting Adsorbed Gas Content . . . . . . . . . . . . . . . . . . . . 183

7.1.3

Completion Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

7.2

Implications in Rock Physics Modeling . . . . . . . . . . . . . . . . . . . . . 188

7.3

Implications in Gas-Flow Modeling . . . . . . . . . . . . . . . . . . . . . . . 190

7.4

Implications in Petrophysical Evaluation: Density of OM . . . . . . . . . . . 195

7.5

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

CHAPTER 8 CONCLUSION AND RECOMMENDATIONS . . . . . . . . . . . . . 201

8.1

8.2

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
8.1.1

Mudrock Pore-Structure Measurement Methodologies . . . . . . . . . 201

8.1.2

Mudrock Pore Structure . . . . . . . . . . . . . . . . . . . . . . . . . 202

8.1.3

Application of the Experimental Results . . . . . . . . . . . . . . . . 204

Recommendations for Future Work . . . . . . . . . . . . . . . . . . . . . . . 204

REFERENCES CITED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

APPENDIX A - SAMPLE CALCULATION WIP . . . . . . . . . . . . . . . . . . . . 223
ix

A.1 Calculation: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

APPENDIX B - MEASURED POROSITY DATASET . . . . . . . . . . . . . . . . . 226
APPENDIX C - STANDARD OPERATING PROCEDURES . . . . . . . . . . . . . 229
C.1 Sample Homogenization and Splitting . . . . . . . . . . . . . . . . . . . . . . 229
C.2 Water Immersion Porosimetry . . . . . . . . . . . . . . . . . . . . . . . . . . 230
C.3 Nitrogen Gas Adsorption Experiments using Micromeritics ASAP 2020 . . 232

LIST OF FIGURES

Figure 2.1

Compositional and textural characterization of the samples from

Silurian gas shale play (SS1) . . . . . . . . . . . . . . . . . . . . . . . . . 18

Figure 2.2

Compositional characterization of the samples from the Haynesville

Formation (SS2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Figure 2.3

Textural variations seen in backscattered electron images (COMPO)of

polished thin section representative samples from the Haynesville
Formation (see text for detailed description) QTZ = Quartz. OM =
Organic Matter. ALB = Albite. . . . . . . . . . . . . . . . . . . . . . . . 20

Figure 2.4

Mineralogical associations in Haynesville samples. (a) Relationship

between calcite and I+S group clay content (wt.%). (b) Relationship
between OM content and I+S group clay content. . . . . . . . . . . . . . 21

Figure 2.5

Compositional characterization of the samples from Niobrara

Formation SS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Figure 2.6

Mineralogical associations in Niobrara samples. (a) Relationship

between calcite and I+S group clay content (wt.%). (b) Relationship
between OM content and I+S group clay content. . . . . . . . . . . . . . 24

Figure 3.1

Flow chart summarizing the analytical procedure steps followed in the

GRI technique. The potential pitfalls associated with each Analytical
step are highlighted in the dark grey boxes. . . . . . . . . . . . . . . . . . 35

Figure 3.2

Whole rock grain density measured by helium pycnometry of the

samples, preheated at 200 C,when kept under a controlled vacuum
chamber (dry) compared to when equilibrated to room conditions. The
dashed line is the tted linear equation for all data points. . . . . . . . . 36

Figure 3.3

TGA curve (black) for a representative mudrock sample from the

Haynesville Formation using a heating protocol as shown by the
temperature curve (grey). Total weight loss from start to 200 C and
weight loss from 110 to 200 C are shown. . . . . . . . . . . . . . . . . . . 42

Figure 3.4

Flow chart summarizing the analytical procedure steps followed in the

WIP technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

xi

Figure 3.5

Measured bulk density, grain density and total porosity values for the
22 test samples from SS1 using WIP for adopted protocol, test 1 and 2.
The errors bars indicate expanded uncertainty of the experiment with a
coverage factor of 2. The results are consistent and repeatable within
the experimental uncertainty. . . . . . . . . . . . . . . . . . . . . . . . . . 48

Figure 3.6

Comparison of measured (a) total porosity (b) grain density and (c)
bulk density on exact same samples between Test 2 and Test 3. . . . . . . 51

Figure 3.7

Compositional controls on total porosity of mudrocks from SS1

(Eastern European Silurian mudrocks). (a) Total porosity as a function
of total clay content (wt.%). The samples are color-coded by their OM
content (wt.%). The dotted line indicates general trend and not a linear
t. (b) Total porosity as function of OM content (wt.%), samples color
coded by their total clay content. (c) Total porosity as a function of
total carbonate content (wt.%) . . . . . . . . . . . . . . . . . . . . . . . . 54

Figure 3.8

Compositional controls on total porosity of Haynesville samples SS2 . . . 56

Figure 3.9

Comparison of (a) measured total porosity and (b) measured grain

density of the samples from SS1 by WIP and GRI techniques . . . . . . . 57

Figure 3.10

Comparison of (a) measured total porosity and (b) measured grain

density of the samples from SS2 by WIP and GRI techniques . . . . . . . 57

Figure 3.11

Comparison of (a) measured total porosity and (b) measured grain

density of the samples from SS1 (circles) and SS2 (triangles) by WIP
and MIP techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

Figure 3.12

Measured WIP data for each of ve randomly selected samples with

propagated expanded uncertainty with coverage factor of 2. The colors
are operator-specic:Op.1 = blue, Op.2 = green, Op.3 = peach. (Dr.
Douglas K. McCarty and Dr. Timothy Fisher, personal communication) . 60

Figure 3.13

Thermogravimetric analysis (TGA) data of isolated kerogen from

Haynesville samples (SS2). The kerogen was equilibrated at 75% RH
before analysis. The sample is heated to 105 C from room temperature
at 5 Cmin1 rate under owing nitrogen gas and then kept for 26
hours under laboratory conditions (50% RH). . . . . . . . . . . . . . . . 61

Figure 3.14

Comparison of RockEvalII S2 pyrogram between a natural mudrock

and its equivalent Dean Stark extracted portion. The Dean Stark
extraction method aects the organic matter and removes the solid
bitumen (low temperature peak/shoulder) and some parts of
pyrolysable kerogen compared to the natural rock. . . . . . . . . . . . . . 64
xii

Figure 3.15

Comparison of RockEval II (a) S2 and (b) S3 parameters measured

before and after preheating at 200 C on thirty samples from Baltic
Basin. (Dr. Arkadiusz Derkowski and Dr. Douglas McCarty, personal
communication) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Figure 4.1

Typical isotherm shape exhibited by (a) purely microporous material

(Type I isotherm prole) (b) non-porous and macroporous material
(Type II isotherm prole) and (c) purely mesoporous materials (Type
IV isotherm prole). Modied from Sing et al. . . . . . . . . . . . . . . . 73

Figure 4.2

The four hysteresis shapes of adsorption isotherm usually found by

subcritical N2 adsorption. Modied from Sing et al. . . . . . . . . . . . . 75

Figure 4.3

N2 adsorption isotherms at 77K of a representative mudrock sample

from Haynesville (Sample SS21). Inset gure is a detailed view
showing the isotherm forced closure in the relative pressure (P/P 0 )
range 0.350.55 due to tensile strength eect. . . . . . . . . . . . . . . . . 76

Figure 4.4

Modied BET technique following Rouquerol et al.applied on isotherm

data from sample SS21 (Figure 4.3). . . . . . . . . . . . . . . . . . . . . 79

Figure 4.5

Schematic example of tplot for two samples with and without

micropores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Figure 4.6

Schematic representation of the relationship between Kelvin radius, rm ,

the core radius rk and the pore radius rp of a cylindrical pore with a
hemispherical meniscus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Figure 4.7

Schematic diagram of constructing a DFT kernel images from][]MM11.

An example model of Ar on Argon on Carbon at 87.3 K is chosen for
illustration. (a) DFT approach used to model the realistic equilibrium
density distribution for the conned uid, Ar on carbon surface at
P/P 0 =0.5 and T = 87.3K; (b) The model isotherm (quantity adsorbed
versus pressure) generated for a slit pore of 4 nm; (c) Kernel q(p, H)
or collection of model isotherms with dierent pore sizes (H). This
kernel q(p, H) is used to deconvolve the experimental isotherm to
obtain distribution function of pore size H. . . . . . . . . . . . . . . . . . 86

Figure 4.8

Experimental results of N2 gas adsorption runs on chips of a mudrock

sample from Eastern European Silurian gas shale play. Note the raw
isotherm data with at adsorption and desorption branches indicating
signicant problem of N2 to diuse into the pore structure. . . . . . . . . 89

xiii

Figure 4.9

Incremental volume percent particle size distribution of asreceived

SWy2 and hand-ground mudrock < 424 m powders measured by laser
prole size analysis in suspension in isopropyl alcohol (black curve with
lled circles). The cumulative external specic surface area (blue line)
of the powder is calculated assuming spherical grains. Note the
reported particle size distribution is that of a crushed powder and not
the actual grain size distribution of the sample. . . . . . . . . . . . . . . . 92

Figure 4.10 t-plot transform plots of sample SS21(solid plus symbols) and SS25
(open circles) using dierent thickness equations. The best-t straight
line shows for the extrapolation of the linear region on the y-axis, which
gives the specic micropore volume. . . . . . . . . . . . . . . . . . . . . . 94
Figure 4.11

PSD obtained from inversion of N2 gas adsorption data for two samples
SS21 and SS25 using dierent thickness equation in the BJH method.
The pore-size distribution data shows the similar peaks but the
absolute volume diers depending upon the thickness equation used. . . . 95

Figure 4.12

Comparison of PSD results obtained by using dierent DFT kernels

with N2 as adsorptive for inversion of the adsorption data for sample
SS2-1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

Figure 4.13 dV /dD representation of PSD for two samples Berea sandstone (blue)
obtained by MIP and mudrock SS22 obtained by N2 gas adsorption
(red) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Figure 4.14 dV /d log D representation of PSD for two samples Berea sandstone
(blue) obtained by MIP and mudrock SS22 obtained by N2 gas
adsorption (red) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Figure 4.15

Comparison of PSD for dierent samples measured at two dierent

labs, Golden CO and Houston TX ) . . . . . . . . . . . . . . . . . . . . 104

Figure 4.16

Raw experimental data obtained by MIP in a representative mudrock

showing show strong hysteresis pattern between the intrusion and
extrusion curves. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

Figure 4.17

: Deviation in the inverted pore size distribution as a function of the

mercury contact angle . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Figure 4.18

Comparison of total specic pore volume measured by N2 gas

adsorption and MIP techniques on (a) the Haynesville samples (SS2),
and (b) on pure SWy-2 clays and other mudrocks. The solid line is the
line of equal values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

xiv

Figure 4.19

Comparison of PSD of as-received SWy-2 clays obtained by inverting

the dierence branches of (a) N2 isotherm data and (b) MIP data.
Note the yaxis is broken for MIP data plot. . . . . . . . . . . . . . . . 113

Figure 4.20

Comparison of PSD of as-received SWy-2 clays and the compacted

pellets (a) obtained from N2 desorption and MIP intrustion data and
(b) obtained from N2 adsorption and MIP extrustion data. Note the
change in scales of the pore-size axis in (a) and (b) . . . . . . . . . . . 113

Figure 4.21

Comparison of PSD obtained from N2 adsorption (gray, N2 desorption

(blue) and MIP intrusion (red) branch from four mudrock formations.

114

Figure 4.22

Comparison of PSD from Clear Creek Sandstone measured using MIP

and optical microscope stereological image analysis techniques
from][]dullien74. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

Figure 5.1

N2 gas adsorption isotherms of as-received powders of Wyoming

Montmorillonite SWy-2 and Kaolinite KGa-1b. The isotherm shapes
indicate the kaolinite powders are mostly macroporous, with negligible
or non-existent micropores and mesopores. Montmorillonite powders
shows presence of signicant volumes micropore, mesopores and
macropores. (See text for detailed description). . . . . . . . . . . . . . 124

Figure 5.2

PSD of as-received source clay powders: Montmorillonite (SWy2)

and Kaolinite (KGa1b) obtained from inverting N2 gas adsorption
isotherm. SWy2 have bimodal distribution with a major peak between
6080 nm with a minor peak around 3 nm. KGa1b shows unimodal
pore distribution with peak around 100 m. . . . . . . . . . . . . . . . . 124

Figure 5.3

(a) Specic surface area obtained by modied BET technique on N2

adsorption isotherm; (b) Total pore-volume, obtained from N2 gas
adsorption (black) and MIP (blue) techniques, on pure SWy-2 samples,
as a function of with increasing axial load. eSSA does not show a
denitive trend with compaction. The total pore volume decreases with
compaction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

Figure 5.4

(a) PSD of as-received SWy-2, compacted clay pellets Mont4K,

Mont8K and Mont10K obtained from N2 gas adsorption technique,
(b) PSD of as-received SWy-2, compacted clay pellets Mont4K,
Mont8K and Mont10K obtained from MIP technique. . . . . . . . . . 127

xv

Figure 5.5

Comparison of the PSD obtained from N2 gas adsorption and MIP for
pure SWy-2 clays and compacted pellets. The three modes observed in
these distribution can be correlated with multi-scale porosity associated
with clay structures especially montmorillonite. Mode A corresponds to
intra-tactoid pores, Mode B corresponds to intertactoid or
intra-aggregate mesopores and Mode C corresponds to macropores
between aggregates. (See text for detailed description). . . . . . . . . . 129

Figure 5.6

Bulk XRD pattern of (a)Concentrated OM obtained from commercial

kerogen isolation (b) Isolated OM from sample SS22 using method of
Ibrahimov & Bissada. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

Figure 5.7

Comparison of RockEval II S2 pyrograms of natural and bleached

sample of (a) SS2A, (b) SS2B. The bleaching removes majority of
OM from the samples but some high molecular number bituminous
organic matter remains in the bleached samples. . . . . . . . . . . . . . 138

Figure 5.8

Bulk XRD patterns of untreated natural (blue) and bleached (red)

sample from (top) SS115 and (bottom) SS2C. Note the lack of pyrite
peaks in the bleached samples. Q = Quartz. Cal = Calcite Py =
Pyrite. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

Figure 5.9

N2 adsorption isotherm (at LN2 temperature) for the immature OM

samples. (a) Utah Gilsonite, (b) Gulf of Mexico (GOM) Bitumen, (c)
Mahogany Fm. shale natural and Mahogany shale cooked at 350 C, (d)
Green River (GR) isolated OM. . . . . . . . . . . . . . . . . . . . . . . 141

Figure 5.10

PSD of the immature OM samples. . . . . . . . . . . . . . . . . . . . . 142

Figure 5.11

(Left)N2 gas adsorption isotherm (at LN2 temperature) for the

concentrated and isolated OM. (Right) Zoomed image showing detailed
isotherms for concentrated and isolated OM from Green River (GR),
SS11 and SS21. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

Figure 5.12

(a) PSD of the concentrated OM from SS115 and SS2A. PSD from
isolated Green River OM is shown for comparison. (b) PSD of the
isolated OM from SS2B. . . . . . . . . . . . . . . . . . . . . . . . . . . 143

Figure 5.13

Comparison of PSD between the natural sample and its bleached

equivalent for (a) sample SS115 (b) sample SS2A (c) sample SS2B
(d) sample SS2C. PSD from the OM separates from these samples
(blue) are shown for comparison. . . . . . . . . . . . . . . . . . . . . . . 146

xvi

Figure 5.14

FESEM image of ion milled sample from Eastern European Silurian

Gas Shale showing microstructural feature where OM hosting a sponge
like pore system lls the intraaggregate clay porosity. Image
Courtesy: Dr. Kitty Milliken, Bureau of Economic Geology, Jackson
School of Geosciences, University of Texas at Austin, Austin, Texas . . 149

Figure 6.1

N2 adsorption-desorption isotherms obtained for the assorted set of

mudrocks samples. Based on the isotherm shapes, the nature of the
pore-structure for the dierent samples is inferred (See text for detailed
description). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

Figure 6.2

PSD of the natural mudrock samples obtained from N2 gas adsorption

technique. Note the yaxis scale for the North Sea Shale is dierent
from other plots. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

Figure 6.3

PSD of the natural mudrocks up to 10 nm pore diameter, colorcoded

by clay content with hotter colors indicating higher clay content. . . . . 159

Figure 6.4

(a) SEM backscatter image of the Cox Argillite . The rock shows clay
dominated continuous microstructure with distributed angular detrital
grains, such as dolomite and quartz. (b) SEM backscatter thin section
image from the Middle Bakken sample . The microstructure in this
sample is controlled by extensive calcite cementation destroying
porosity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

Figure 6.5

BJH cumulative pore volumes between 2 to 5 nm diameter of the

natural mudrock samples except for the North Sea Shale (lled circles)
are plotted as a function of their I+S clay content. A linear trend with
zero intercept (equation y = bx) is tted to the data points. Note that
the result from the pure montmorillonite (open triangles) is not used in
the regression and is shown only for reference. . . . . . . . . . . . . . . 161

Figure 6.6

Typical N2 gas adsorption isotherm prole for the Niobrara Formation

(sample Niobrara Chalk B 3048 ft. sample SS39 Table 2.3). Inset
gure is a detailed view showing the isotherm forced closure in the
relative pressure (P/P 0 ) range 0.35-0.55 due to tensile strength
eect. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

xvii

Figure 6.7

(a) Detailed view of N2 adsorption isotherm for a typical Niobrara

sample with emphasis on the forced closure behavior of the desorption
branch. The yaxis is the quantity adsorbed in cm3 /g STP (QA). The
TSA parameter is dened to quantify the extent of the forced closure,
as the volume desorbed per unit pressure between the P/P 0 of
0.410.49. (b) The TSA parameter increases with increasing I+S clay
contents of the sample. This indicates samples with higher clay content
have a signicantly higher amount of ne mesopores <4 nm. . . . . . . 163

Figure 6.8

eSSA of the Niobrara samples as a function of their illite+smectite

group clay content. Each color indicates separate lithostratigraphic
units. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

Figure 6.9

PSD of all the Niobrara samples; color coded based on their clay
content where red indicates highest clay content. . . . . . . . . . . . . . 166

Figure 6.10

PSD of the Niobrara samples up to 10 nm poresize with (a) low OM

(<1 wt.% OM) content, and (b) high OM (>1 wt.% OM) content.
Hotter colors indicate higher clay contents. . . . . . . . . . . . . . . . . 166

Figure 6.11

Comparison between eSSA and I+S clay content for the Niobrara
samples based on groups with low OM and high clay content, high
organic and high clay content, low organic and low clay content, and
high organic and high clay content . . . . . . . . . . . . . . . . . . . . 168

Figure 6.12

Relationship between eSSA and OM content for the high OM content

(>1 wt.%) samples. The labels indicate the I+S clay content (wt.%). . 168

Figure 6.13

Relationship of small mesopore distribution (1.7-10 nm) for the

Niobrara samples based on groups with (a) low OM and high clay
content, (b) high OM and high clay content, (c) low OM and low clay
content, and (d) high OM and high clay content. . . . . . . . . . . . . 170

Figure 6.14

Variation in eSSA of the Haynesville Formation samples with (a) I+S

group clay content, (b) total carbonate mineral content, and (c) OM
content. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

Figure 6.15

Variation observed in PSD in the ne mesopore range < 10 nm for the

Haynesville Formation samples. Some samples show (a) very prominent
3 nm modal peak characteristics of I+S microstructure, while (b) high
OM content samples show a very uniform distribution of pore volumes.
(b) BJH cumulative pore volumes between 2 to 5 nm diameter of the
Haynesville samples are plotted as function of their I+S clay content. . 174

xviii

Figure 6.16

(a) Comparison of 25 nm pore volume trend with I+S clay content in

the Haynesville samples (lled circles) with the trend observed in Case
Study I, Figure 6.5. The numbers indicate the OM content (wt.%) in
the samples. The results from the pure montmorillonite (open
triangles) are shown for reference. (b) BJH cumulative pore volumes
between 2 to 5 nm diameter of the Haynesville samples are plotted as
function of their OM content. . . . . . . . . . . . . . . . . . . . . . . . 177

Figure 6.17

Eect of cementation on PSD of the Haynesville samples.

Figure 6.18

Combined thermogravimetric and mass spectrometry (TGA-MS)

analysis of a highly cemented Haynesville sample. The MS data shows
evolved CH4 (mass 16) signal at high temperatures beyond the
decompostion temperature of OM and calcite. . . . . . . . . . . . . . . 179

Figure 7.1

Compositional correlation between total porosity measured by WIP

technique for mudrocks from (a), Eastern European gas shale formation
(SS1), (d), the Haynesville Formation (SS2), and the Niobrara
Formation (SS3). The dashed line are not best t linear trendlines, but
are shown to illstrate general correlation. . . . . . . . . . . . . . . . . . 184

Figure 7.2

Comparison of supercrtical methane adsorption at 25 C between the

natural and OM removed sample for (a) SS115 and (b)SS2C. . . . . 187

Figure 7.3

(a) FESEM image of ion milled sample from Eastern European Silurian
Gas Shale showing microstructural organo-clay packages (Image taken
by Dr. Kitty Milliken). (b) Eect of OM and its thermal maturity on
the ne mesopore structure in mudrocks. . . . . . . . . . . . . . . . . . 191

Figure 7.4

Flow constants as a function of pore-radius at dierent pressures . . . . 193

Figure 7.5

The modeled contribution of Knudsen diusive ow to the total ow as

a function of measured total porosity in the Haynesville samples. . . . 195

Figure 7.6

Comparison of modeled liquid permeability from Poiseuilles equation

with reported measured permeability data on mudrocks in (a) linear
scale, (b) logarithmic scale. The solid line indicate the tted linear
trend. Note the linear trend show a curve in log-log plots. The Mancos
B data (red) is not included in the t. . . . . . . . . . . . . . . . . . . . 196

Figure 7.7

Whole rock grain density sensitivity analysis due to the variation in

assumed OM density variation. . . . . . . . . . . . . . . . . . . . . . . . 197

xix

. . . . . . . 178

Figure 7.8

Histogram of calculated OM densities, using Equation 7.7, for (a) the

Eastern European gas shale samples SS1, (b) the Haynesville
Formation, and (c) the Niobrara Formation. The red bar in the
histogram on OM density from the Niobrara are treated as outliers. . . 200

xx

LIST OF TABLES

Table 2.1

Mineral and OM content (wt.%), CEC and RockEval II parameters of

samples from Eastern European Silurian gas-shale play (SS1) . . . . . . 26

Table 2.2

Mineral and OM content (wt.%), CEC and RockEvalII parameters of

samples from Haynesville gas-shale play (SS2). . . . . . . . . . . . . . . . 27

Table 2.3

Mineral and OM content (wt.%), CEC and RockEvalII parameters of the

Niobrara (SS3) samples. The concentration values reported are wt.%
normalized to 100% obtained after combining the mineral phases wt.%
(from QXRD) and OM wt.% (from LECO). Carbon wt.% in OM is
assumed to be 83% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Table 2.4

Major mineral and organic matter content of the studied samples from
the assorted set of mudrocks. . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Table 3.1

Penetration coecient of commonly used saturating uid for

liquid-saturation and immersion techniques, assuming the contact angle
to be zero. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Table 3.2

Test Conditions for the reproducibility test.

Table 3.3

Results of the repeatability test of WIP technique. The details of the test
conditions are provided in Table 3.2. . . . . . . . . . . . . . . . . . . . . . 49

Table 3.4

Results of repeated-measure ANOVA test comparing porosity values for

the same samples under dierent test condition (AP: Adopted protocol,
Test 1 and Test 2, Table 3.2). . . . . . . . . . . . . . . . . . . . . . . . . . 52

Table 3.5

Results of the paired-samples t-test conducted to compare the bulk

density (BD), grain density (GD) and porosity (WIP) values of the
samples under test condition Test 2 and Test 3 . . . . . . . . . . . . . . . . 53

Table 4.1

eSSAN 2 for sample SS21 and SS25 with dierent equilibration

interval. C = BET constant; Vm = monolayer capacity . . . . . . . . . . . 91

Table 4.2

t-plot inversion results of sample SS21 and SS25 using dierent

thickness equations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Table 4.3

Results of reproducibility test for samples ve samples between two

laboratories located in Golden, CO and Houston, TX . . . . . . . . . . . 103
xxi

. . . . . . . . . . . . . . . . . 46

Table 5.1

Modied BET eSSA,open specic surface area obtained from t-plot and
micropore specic surface area of the samples based on N2 adsorption
isotherms. The uncertainty estimate is only calculated from the mist of
the regression t to the actual data point. Other experimental
uncertainties are not evaluated. . . . . . . . . . . . . . . . . . . . . . . . 125

Table 5.2

Total pore volume of the pure clay samples measured by N2 gas

adsorption and MIP. Micropore volume obtained by using tplot
technique on N2 isotherm data. . . . . . . . . . . . . . . . . . . . . . . . 125

Table 5.3

Mineral and OM content (vol.%) and RockEval II parameters of the

natural samples used in OM pore-structure characterization study. . . . . 134

Table 5.4

RockEval II data of the Haynesville Formation samples before and after

removal of OM by bleaching. . . . . . . . . . . . . . . . . . . . . . . . . . 138

Table 5.5

Pore structure parameters obtained by gas adsorption technique on

samples used for OM pore structure characterization experiments. . . . . 144

Table 5.6

Relative proportion of pores with < 5 nm pore size in OM compared to

clays, obtained from comparison of tplot micropore volume and
BJHHJ ne mesopore (25 nm)volume between natural samples and
bleached samples. The bleached sample with only have clay pores while
the natural samples will have both the clay pore and OM pores. . . . . . 150

Table 6.1

Pore-structure parameters for the assorted set of natural shale samples,

obtained using N2 gas adsorption techniques. eSSA is obtained from
modied BET technique. Open SSA and micropore volume are
obtained using tplot inversion. The uncertainty estimate is only
calculated from the mist of the regression t to the actual data point.
Other experimental uncertainties are not evaluated. . . . . . . . . . . . . 157

Table 6.2

eSSA of the Niobrara samples reported as per recommendations of I.S.O

9277:2010(E). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

Table 6.3

Porestructure parameters of the Haynesville Formation mudrock

samples obtained from N2 gas adsorption technique. . . . . . . . . . . . . 175

Table 7.1

BestROCK derived grain densities of indiviual mineral for each sample

set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

Table A.1

WIP measurement data for sample SS11

Table B.1

Measured bulk density, grain density and porosity of all the samples by
dierent techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
xxii

. . . . . . . . . . . . . . . . . 223

LIST OF SYMBOLS

B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bulk Density
G . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Grain Density
H2 O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Water Density
air

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Air Density

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Porosity
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Standard Deviation
P0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Saturation Pressure
P/P0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relative Pressure
Vm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monolayer Capacity
V

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quantity of Adsorbed Gas

am . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular crosssectional area

L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Avogadro constant
C

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BET constant

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liquid Surface Tension

VL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liquid Molar Volume
R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Universal Gas constant
T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Temperature
rm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kelvins radius
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contact Angle
Kn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xxiii

Knudsen Number

 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mean free path of the gas

Rh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Characteristic pore diameter
KK

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flow constant for Knudsen ow

KP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flow constant for Poiseulle ow

r . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pore radius
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tortuousity
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Viscosity
M . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular weight

xxiv

LIST OF ABBREVIATIONS

API R.P.40 . . . . . . . . . . . American Petroleum Institute Recommended Practices 40

ASS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Assorted Set of Sample
BET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Brunauer, Emmett, and Teller
BJH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Barett, Joyner, and Halenda
DFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Density Functional Theory
CEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cation Exchange Capacity
eSSA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . equivalent Specic Surface Area
FESEM . . . . . . . . . . . . . . . . . . . . Field-Emission Scanning Electron Microscope
GRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gas Research Institute
HJ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Harkins-Jura
I+S . . . . . . . . . . . . . . . . . . . . . . . . . . Illite+smectite group of clay minerals
ICP-MS . . . . . . . . . . . . . . . . . . . Inductively coupled plasma-mass spectrometry
IUPAC . . . . . . . . . . . . . . . . . International Union of Pure and Applied Chemistry
LN2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liquid Nitrogen
MIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mercury Intrusion Porosimetry
OM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organic Matter
p.u. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . porosity unit
PSD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Pore--Size Distribution

QXRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quantitative X-ray diraction

SS1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample Set 1
xxv

SS2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample Set 2

SS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample Set 3
TGA-MS . . . . . . . . . . . . . . . . Thermogravimetric analysis and mass spectrometry
TOC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Total Organic Carbon
TSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tensile Strength Eect
Unconventional Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UCR
Water Immersion Porosimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . WIP

xxvi

ACKNOWLEDGMENTS

The last ve and half years has being quite a journey for me and I owe my sincere
gratitude to many people who have helped me grow both academically and personally.
I owe a lot of gratitude to my advisor, Prof. Manika Prasad. I am grateful to her for the
independence she gave in my research and allowing me to nd my own comfortable niche. I
have truly cherished countless discussions with her: technical, professional, socioeconomical
and personal. She has always been supportive to me during my times of distress and failures.
Thank you, Manika, for believing in me!
Thanks to my coadvisor, Dr. Douglas McCarty, for introducing me to the fascinating
world of clay mineralogy during my internships with Chevron. He always motivated me to
see the application side of the science, to think about how it can solve real-life problems and
also to present in a way that will be accessible to the greater audience.
I would also like to thank Prof. Mike Batzle, for been such an inspiration throughout my
stay in CSM. His philosophy about experimental science always motivated me and adore him
for his sense of humor. I would like to thank my other committee members Prof. Hossein
Kazemi, Prof. Ramona Graves, Prof. Mike Mooney and Dr. Je Kenvin for their signicant
contribution to my education in Mines. Advice from Prof. Kazemi helped me understand
the signicance of ow modeling in my project. I extend my deep appreciation to Dr. Je
Kenvin for introduction to the theories of gas adsorption.
A very special thanks goes to Dr. Arkadiusz Derkowski for the long but immensely
helpful discussion sessions we had during my internship of Chevron on various topics of clay
and shale characterization. I learned a lot while working with you. I would also like to thank
Dr. Timothy Fischer and Dr. Luca Duranti for all the discussion we had. I would also like to
acknowledge Prince Eziburo, Billy Renfro, Ferrell Garner, Bill Lawrence and Steve Yurchick
for all their help in during my lab studies. Also I thank Sheven Poole, Liwei Ou and Lemuel

xxvii

Godinez for their help in collecting some of the data used in this thesis. Special thanks to
Mark Talarico, for all his help with the gas adsorption instrument. I thank Terri Synder,
Patti Hassen and Denise Winn-Bower for who visibly and invisibly made my stay smoother
by helping in terms of administrative hassles. I also thank Al Sami for his help with my
measurements and laboratory work.
I would like to thank the members of OCLASSH Consortium, Fluids and DHI Consortium
and the Norwegian Research Council for their support during my entire Ph.D. stay. Im also
indebted to ConocoPhillips and CSM GSA for their nancial support in form of scholarship
in these ve years.
My education in Earth Sciences started in Presidency College, Kolkata, India and my
professors in my alma mater, especially Dr. Ananda Kumar Chakrabarty, Dr. Prabir Dasgupta, Dr. Aniskumar Ray, Dr. Pradyut Bandopadhyay, Dr. Arunava Basu and Dr. Joydip
Mukhopadhyay, laid the foundation for what knowledge I have acquired today. Thank you
Sir-s. A very special thanks to my professor Dr. Arunava Basu, who encouraged me to
pursue a doctorate degree.
I am grateful to all my colleagues in Center of Rock Abuse and OCLASSH group for
sharing time during my stay at CSM: Agni, Ritu, Arpita, Ravi, Jesse, Alejandra, Fernando,
Kene, Sunny, Patricia I, Patricia II, Mohammed, Farnoush, Weiping, Milad, Saul, Saeed,
Lemuel, Liwei, Stuti, Wendy and Chelsea. I especially thank Agni and Ritu for lots of things,
other than their guidance ease out the rigors of Ph.D. days. They have been my closest
support system at all times and making these ve and half years so memorable. I enjoyed
the memorable friendship of my fellow Indian students: Kumar, Sanjeev, Paritosh and may
more. They have selessly supported me when I needed them the most. A very special
thanks to Sagnik Dasgupta, for all the entertaining and thought provocating discussions we
had and his help during my stay in Houston.
Any amount of acknowledgement would be insucient for my family: my mother late
Smt. Supriti Kuila, my father Shri Bimalendu Kuila, my sister Smt. Saswati Ghosh and

xxviii

my brother-in-law Shri S.P. Ghosh. Without their inspiration and support nothing would
have been possible. They have always stood by me and dealt with all of my absence from
many family occasions with a smile. Finally, and above all, I am indebted to my best friend,
Nabamita, who has been lending her ear, over the last twelve years.

xxix

Dedicated to my parents late Supriti Kuila and Bimalendu Kuila and my sister Saswati
Ghosh.
All I have and will accomplish are only possible due to their love and sacrices.

I also dedicate this thesis to my best friend, Nabamita, who has always helped me in
believing that I could do it.

xxx

CHAPTER 1
INTRODUCTION

Thirty spokes unite in one nave;

the utility of the cart depends on the hollow centre in which the axle turns.
Clay is moulded into a vessel;
the utility of the vessel depends on its hollow interior.
Doors and windows are cut out in order to make a house;
the utility of the house depends on the empty spaces.
Thus, while the existence of things may be good,
it is the non-existent in them which makes them serviceable.
- Lao Tzu, Tao Te Ching: Chapter 11
Translated by Lionel Giles: The Sayings of LaoTzu [1905]

A pore is the void space embedded within the solid matrix of a porous media. The
presence of majority of this pore structure within naturally occurring porous materials cannot
be perceived by naked eye (or apparently non-existent) but is manifested in the behavior of
the porous materials to external stimuli (Dullien, 1991). Almost every mechanism occurring
in the porous media is directly or indirectly controlled by its pore structure. This nonexistent part of the rock must be understood rst in order to understand the behavior of
the porous media. However, for scientic and engineering purposes, mere perception or
qualitative understanding is not enough, it needs to be quantied. Accurate quantitative
understanding of pore structure is the basic requirement for all eld of scientic studies
dealing with porous media.
For the upstream hydrocarbon industry, quantication of the pore structure has both scientic and economic implications. The bottom line of the upstream hydrocarbon industry is
to produce hydrocarbon from the natural resources. The hydrocarbons are stored within the

pores and ows out of the rocks (reservoirs) through the connected porestructure. Therefore, the pore structure attribute are the fundamental reservoir property that needs to be
quantied. Proper estimation of total pore volume, or porosity, is essential for quantication the total hydrocarbon (oil or gas or both) in place and has both legal and economic
implications.
Pore structure in regular reservoir rocks, such as sandstones and carbonates, are well
enough understood. shales/mudrocks are the most dominant sedimentary rock in any petroliferous basin. In spite of their ubiquitous presence, characterization of shales/mudrocks,
in general including pore structure characterization, have been largely overlooked by the
upstream hydrocarbon industry in the last century as there is little economic incentive to
study the non-pay zone lithologies. This is also true for the academic community, as it
was pointed out by Schieber & Zimmerle (1998) who reported the dominance of published
literature on sandstones and carbonates over shale in the 80s and 90s. Most of the existing understanding of shaly systems are based on studies conducted by civil engineers to
understand soil behavior and ODP studies on recent sediments, however, the transfer of this
knowledge to the hydrocarbon exploration community has been limited. Apart from the lack
of economic incentives, the lack of proper tools to characterize the ne scale pore structure
of shales/mudrocks also attributed to this negligence. Experimental studies in shales/mudrocks were dicult because of their negrained texture, presence of extremely small pore
size with dimensions ranging from sub-nm to hundreds of nm, extremely low permeability,
and strong interaction of water molecules with the structure of clay minerals and associated
exchange cations. However, the recent success of gas shale and shale oil plays has stimulated
research interest in shales/mudrocks.
1.1

Challenges in shales/mudrocks
Over the past decade, research interest in shale physical properties has increased due to

the commercial success of unconventional resources (UCR), particularly gas shale and liquid
shale oil plays. Scientic studies on imaging the porestructure in the shales/mudrocks
2

was done in order to understand how these plays are producing in spite of their nanodarcy
range permeabilities (e.g., Javadpour, 2009; Loucks et al., 2009; Milliken & Reed, 2010;
Passey et al., 2010; Schieber, 2010; Walls et al., 2011; Curtis et al., 2012; Milliken et al.,
2013). Advancement in the imaging techniques lead to the discovery of nanometersized
pores associated with the inorganic as well as organic components of the shales/mudrocks.
These modern imaging methods are invaluable to understanding shale petrology and pore
structure, however imaging results are not quantitative with regard to pore volume, and
current resolution limits can only detect pores with diameters larger than about 4 nm.
Quantication of pore structure attributes in shales/mudrocks still remains a challenge.
Reservoir properties estimation, including porosity, in eld scale are mostly done with petrophysical evalution of wireline logging data. These indirect petrophysical inversion are veried and calibrated with independent measurements obtained from core analysis. In absence
of proper petrophysical inversion techniques for shales/mudrocks, measurements on cores
becomes the only way to estimate porosity of these reservoirs. The lack of accurate and
standardized techniques to quantify pore structure attributes from both core and log measurements lead to signicant uncertainty in the economic assessment of these commercial
plays.
The current industry adopted technique of total porosity measurement in shales/mudrocks is based on the method developed in Gas Research Institute (Luel & Guidry, 1989;
Luel & Guidry, 1992; Luel et al., 1992) and is generally referred as GRI technique. Several
issues was raised on the methodology and also lack of standardized practices. Each company
has adopted dierent protocols which lead to signicant and consistent discrepancies in the
reported data between laboratories (Karastathis, 2007; Passey et al., 2010; Sondergeld et al.,
2010, Spears et al., 2011). Mercury intrusion porosity (MIP), the widely accepted technique
to measure pore-structure attributes such as pore size distribution in sandstones and carbonate, also have signicant applicability issues for shales/mudrocks (Bustin et al., 2008;
Sigal, 2009; Comisky et al., 2011; Sigal, 2012). The details about the limitations to cur-

rent porosity and pore-structure attribute measurements in shales/mudrocks are discussed

in later chapters.
1.2

Terminology
The terminologies used for both shales/mudrocks and porosity are varied and often slackly

used. There are dierent terminologies used for the same thing or property and also same
terminology is used to dene dierent things or properties, based on the application. In the
following subsections, the terminology that will be used in this thesis are explained.
1.2.1

Shales / Mudrocks

Shales and mudrocks (sometimes mudstones) are both widely used terms for ne-grained
sedimentary rocks and have being interchangeably used in the literature and by the hydrocarbon industry. Folk (1974) dened the term mudrocks as the terrigenous rocks that
contains more than 50 percent silt and/or clay, mudstones are mudrocks with equal percentage of silt and clay grains and the term shale be restricted to indurate, ssile variety of
the same. It should be noted that, the terms shale and mudrocks are used for terrigenous
rocks and not for chemically precipitated rocks (allochemical and orthochemical). Friedman
(2003) states while dening shales in his Classication of sedimentary rocks as Probably
the most satisfactory denition of shale is according to particle size. In terms of size, the
name shale is the lithied equivalent of mud. Javadpour (2009) used the same denition of
mudrocks to dene rocks with more than 50% of the grains smaller than 62.5 m in diameter.
This denition includes mudstones and shales. In this thesis, the terminology mudrock will
be used to strictly dene on the basis of particle-size. This terminology does not carry any
implication about the mineralogy and composition of the rock, i.e. whether its clay rich
or not. The term shale will only be used in case of stratigraphic nomenclature, e.g. the
Haynesville Shale, or when it is used to indicate certain kind of hydrocarbon play, e.g. oil
shale, shale oil, gas shale.

1.2.2

Organic Matter

Varied terminology is used for the organic components of the mudrocks. The term kerogen is sometimes loosely used to refer to the organic matter in the mudrocks, however, sensu
stricto denition of kerogen is based upon the insolubility to organic solvents. Similarly, bitumen is the component of organic materials that is extractable using organic solvents. Bitumen is sometimes referred to the heavy oil reservoirs as a designation for a hydrocarbon
uid with a gravity of 10 API or lower, based upon the classication of the US Department of
Energy http://www.glossary.oilfield.slb.com/en/Terms/b/bitumen.aspx, (Schlumberger Oileld Glossary, 2013). In this thesis, the term organic matter is used to refer to
the general organic components in absence of any solubility data and the term bitumen and
kerogen are only used to imply their solubility in organic solvents.
1.2.3

Porosity

Porosity is dened as the fraction of the bulk volume of the porous sample that is
occupied by pore or void space (Dullien, 1991). This denition implies that a pore is the
void space in the porous media, however, in natural reservoirs space is always occupied and
never void. According to the Schlumberger Oileld Glossary (2013) dened a pore as a
discrete void within a rock, which can contain air, water, hydrocarbons or other uids.
Therefore, the part of the rock occupied by solid material is the grain and the part that is
occupied by uid system is the pore. This seemingly straight-forward denition becomes
complicated in mudrocks because of the presence of clays and organic matter.
Clays are reactive of water and exchangeable cations present on clay mineral surfaces
and in interlayers form strongly held shells of water molecules around them. This water is
immovable for hydrocarbon production applications and therefore termed ineective porosity
for ow of hydrocarbon in the reservoirs. The terminology clay bound water (CBW) is used to
quantify the volume of water adsorbed on clay mineral surfaces, which includes the interlayer
space in the expandable clay minerals and external surfaces of crystallites in non-expandable

clay species. The term total porosity is dened as total wateraccessible porosity and eective
porosity as total porosity minus the clay-bound water.
The presence of organic matter (OM) complicates the denition of pores and porosity,
as their variable nature make it dicult to assign it to the solid framework part of the
uid pore part of the rock. These OM are mixtures of materials having variable chemical
composition, molecular structure, thermal stabilities, solubility properties and mechanical
properties. Also it is dicult to classify organic matter based on one criteria. These shale
play reservoirs are generally self-sourcing and what is generated in these rocks remain in
the rocks over the geologic history. There is a mixture of volatile hydrocarbons, liquid
hydrocarbons, not-so-liquid hydrocarbons, not-so-solid hydrocarbons to solid hydrocarbons
and there is no clear distinction between each component or phase. In that case, it has
to be operationally dened, e.g. the S1, S2, S3 components in RockEval techniques. For
the purpose of this thesis, the components of the OM that gives a peak in the S1 region of
RockEval is assigned to the pore/uid part of the rock and anything beyond that thermal
stability is assigned to the solid framework part of the rock.
Porosity is expressed in porosity unit, which is equal to the percentage of pore space in
a unit volume of rock. It is abbreviated to p.u. and lies between 0 and 100. (Schlumberger
Oileld Glossary, ). The pore volume is expressed as specic pore volume per unit mass of
sample.
1.2.4

Pore-size

There are dierent attributes to quantify porestructure and porosity is one of them.
Porosity is a singlevalue quantication (Nimmo, 2004) and does not represent the heterogeneity of these porestructures that occur in the nature. One other attribute that is used to
characterize pore-structure is the poresize distribution (PSD), which quanties the relative
volumes associated with the dierent poresizes. Pore-size does not have a precise denition due to the irregular and variable interconnected network of pores. The irregular nature
of pore-structure make dening and partitioning of the pore-space into individual pores
6

dicult. One of the important conceptualization of pore-structure often applied to soil science and earth science is that a pore-structure essentially consists of a collection of channels
whose eective width varies along its length (Nimmo, 2004). Relatively wider portions of
those channels (pore-body) are connected and separated by relatively narrow portions called
the pore-throats (Nimmo, 2004). Moreover, each measurement technique records the porestructure dierently and the raw measurement is converted into quantitative descriptions of
pore structure that usually are based on model assumptions.
The most common and applied model invoked during inverting experimental data to
PSD is the capillary bundle approximation, where the porous material is assumed to contain
aligned, separate cylindrical open-ended capillaries with dierent diameters. In real porous
media, a complex array of pore shapes between irregularly shaped particles exists. The
assumption of cylindrical pores is analogous to the assumption of the spherical model for
grain size distribution determinations and hence the pore sizes, in this thesis, will be referred
to as equivalent cylindrical pore diameter. The pore-size distribution obtained from the
assumption of capillary bundle approximation model should be treated as a one-dimensional
approximation of the real three-dimensional complex pore-structure.
In the literature, pore-size nomenclature is varied and a wide variety of terminology is
used for pores with dierent sizes (e.g. nanopore, mesopore, micropores). The International
Union of Pure and Applied Chemistry [IUPAC] (Sing et al., 1985) recommend the following
classication of pores according to their size:
1. Micropores: pores with pore-size below 2 nm
2. Mesopores: Pores with pore-size between 2 nm to 50 nm
3. Macropores: Pores with pore-size greater than 50 nm
The advantage of this classication is that it is based on the physics of nitrogen adsorption
in various pore-sizes of a porous material at -197.3 C (77.3K) and 1 atmosphere pressure.
Dierent mechanisms, such as multilayer adsorption (macropores), a combination of multi7

layer adsorption and capillary condensation (mesopores), and micropore lling (micropores)
dominate in the dierent pore size regimes. Its disadvantage is that the classication is not
intuitive; nanosized pores are termed as micropores and mesopores. Additionally, pores
larger than 50 nm are not subdivided further. They are jointly classied as macropores.
However, since the IUPAC classication (Sing et al., 1985) is a nonarbitrary classication,
the IUPAC poresize classication is used in this thesis as N2 gas adsorption technique is
used as the primary tool for PSD measurements. The IUPAC classication is more appropriate for shales since we expect shales to have a considerable volume of their pore space in
nanometersized pores. For future work, we recommend dening a poresize classication
that is more appropriate for mudrocks.
1.3

Research Objectives
The importance of quantitative pore structure characterization in understanding shale

reservoir behaviors and lack of measurement techniques are the two main motivation factors
behind this research. The research objectives that will be addressed in this thesis are as
follows:
1. Develop or adapt mudrock specific pore-structure attribute measurement
techniques: The pore structure attributes that were obtained are total porosity, specic surface area and pore-size distribution. A new total porosity measurement technique for shale gas applications are developed. Nitrogen gas adsorption technique was
used and adapted for shale applications.
2. Quantitative understanding of pore-structure attributes in mudrocks and
the controls of composition: This is achieved in two steps. The pore-structure of
compositional end members of shale system, clays and organic matter were investigated. Quantitative pore-structure attributes in shales was correlated with compositional characterization to observe any systematic trends.

1.4

Organization of Thesis
This thesis is organized in to four broad categories, divided into eight chapters. The rst

two chapter, Chapter 1 and Chapter 2, introduce the main research problem and the samples
that will be used for this research, respectively. The next two chapter are centered around
methodologies for measuring mudrock pore-structure attributes. Chapter 5 and Chapter
6 presents the detailed pore structure in mudrocks and their compositional end members.
The last two chapters summarizes the observations and its application in other studies on
mudrock characterization. The details of each individual chapters are as follows:
Chapter 2 presents the details of the sample characterization methodologies, and compositional and textural characterization of the mudrock samples used for this thesis.
Chapter 3 presents a new methodology of measuring total porosity in mudrocks using water
immersion porosimetry (WIP). The results of systematic study the WIP technique,
including evaluating the reproducibility and experimental uncertainty, on three sets
of total seventy-four shale samples, from (i) an Eastern European Silurian gas shale
formation, (ii)the Haynesville Formation of East Texas, USA, and (iii) the Niobrara
Formation of Northern Colorado, USA are presented. The WIP results are compared
the measurements by the GRI and MIP techniques and the signicant dierences are
discussed.
Chapter 4 presents the experimental procedure of the N2 gas adsorption measurement
technique, interpretation of raw data obtained and theoretical concepts for inverting
the raw data to pore structure parameters. Some issues and considerations for applying
this technique for mudrock pore structure characterizations are discussed.
Chapter 5 describes the experimental results of nitrogen gas adsorption experiment on two
important mudrock end member, clays and organic matter (OM). The implications of
the experimental ndings are discussed in light of understanding the pore structure

of the end members components of a mudrock system. The insights are later used to
understand the compositional controls on mudrock pore structure.
Chapter 6 presents three case studies on characterizing pore structure from a suite of
mudrock samples to understand the compositional controls on their pore structure.
The three case studies used (i) an assorted set of mudrock samples from dierent
locations with dierent clay contents (ii) a suite of 19 samples from a well drilled in
the Niobrara Formation of Northern Colorado, USA and (iii) a suite of 16 samples
from a well drilled in the the Haynesville Formation of East Texas, USA.
Chapter 7 deals with the applicability of porosity and pore structure measurements for
real life gasshale applications. The potential applications of the new understanding
of mudrock porosity and pore structure parameters in exploration and production
strategies of gas shale resources are demonstrated.
Chapter 8 sums up the conclusions for the dissertation. Recommendations for future work
are also summarized in this chapter.
1.5

List of Publications
Published in peer reviewed journals

1. Kuila, U. & Prasad, M. (2013). Specic surface area and poresize distribution in clays
and shales, Geophysical Prospecting, 61 (2), 341362
Submitted in peer review journal
1. Kuila, U., McCarty, D.K., Derkowski, A., Fisher, T. & Prasad, M. , Total porosity measurement in gas shales by the water immersion porosimetry (WIP) method, Submitted
to Fuel.

10

Conference Proceedings and Abstracts

1. Kuila, U. & Prasad, M. (2013). Variation of elastic moduli of clays with humidity,
AGU 2012, Poster presentation at American Geophysical Unions 45th Annual Fall
Meeting held in San Francisco, California, 37 December, 2012
2. Kuila, U., Prasad, M., Derkowski, A. & McCarty, D.K. (2012). Compositional controls on mudrock pore-Size distribution: An example from Niobrara Formation. SPE
160141-PP, Presented at presentation at the SPE Annual Technical Conference and
Exhibition held in San Antonio, Texas, USA, 8-10 October 2012.
3. Kuila, U. & Prasad, M. (2012). Pore Structure in articially compacted and naturally
compacted clays, CMS 2012. Oral presentation at 49th Annual Meeting of The Clay
Minerals Society, July 7 - July 11, 2012, Golden, CO, USA
4. Kuila, U. and Prasad, M. (2010). Pore size distribution and ultrasonic velocities of
compacted Na-montmorillonite clays. SEG Expanded Abstracts 29, 2590-2594. Oral
presentation at SEG 80th Annual Meeting, October 17 - 22, 2010 , Denver, Colorado,
USA.
In preparation
1. Kuila, U., McCarty, D.K., Fisher, T., Wigand, M., Derkowski, A., Prasad, M. & Milliken, K., Specic surface area and pore-size distribution of organic matter in mudrocks.
2. Kuila, U., Kenvin, J., McCarty, D.K., Fisher, T., & Prasad, M., Some issues on application of nitrogen gas-adsorption for mudrock characterization.

11

CHAPTER 2
SAMPLE CHARACTERIZATION AND DESCRIPTION

Tolle numerum omnibus rebus et omnia pereun

Take from all things their number and all shall perish
- Saint Isidore of Seville
Etymologies [c.600], Book III, chapter 4, quoted in E. Grant (ed.), A Source Book in
Medieval Science (1974), trans. E. Brehaut (1912), revised by E. Grant

Experimental work to characterize porestructure in mudrocks was carried out on samples from dierent locations having dierent mineralogy, clay content, organic matter (OM)
content and thermal maturity. The dierences in the composition of these mudrocks allowed
investigating of the compositional controls on the mudrock porestructure. Therefore compositional characterization becomes important since it gives quantitative information about
the sample that will be measured. The purpose of this chapter is to describe the sample
sets in detail to provide information on the compositional characteristics and also to present
the sample characterization experimental methods in detail. The sample characterization
methodology is described in section 2.1. The compositional characteristics of the mudrocks
samples used in the later experiments are presented in 2.2
2.1

Sample Characterization Methods

The samples were characterized for whole rock quantitative mineral and elemental com-

position, along with cation exchange capacity (CEC). For compositional characterization
purposes, each sample was ground to <425 m (<40 mesh) powder, homogenized, and
divided following rigorous procedures to obtain mineralogical and chemically equivalent portions (McCarty, 2002), which are then used for dierent analyses. The main sample characterization methodologies are presented as following.

12

2.1.1

Quantitative Mineral Phase (QPA) Analysis

Following the methods of rodo et al. (2001), quantitative mineral phase compositions
were obtained by inverting X-ray diraction (QXRD) data of randomly oriented powders
spiked with 10% by weight ZnO (Chevron program QUANTA, Omotoso et al. (2006)). In
this method, all dioctahedral Al-rich 2:1 clay minerals (illite, dioctahedral smectite, muscovite, mixed-layered illitesmectite) are quantied together as the illite+smectite (I+S)
group mineral (rodo et al., 2001).
Majorelement and traceelement chemical analyses of the samples were obtained by inductively coupled plasma-atomic emission spectrometry (ICPAES) from ActLabs, Ontario,
Canada (www.actlabs.com). The cation exchange capacity (CEC) was measured by the
Cohexamine technique of Orsini & Remy (1976) and Bardon et al. (1983). The CEC values
were reported on a dry sample weight basis, obtained after dehydration at 200 C (rodo
& McCarty, 2008). The measured CEC value is used to quantify the smectite-equivalent
component (% S) in the bulk rock I+S clay mineral group. This calculation was determined
by assigning the bulk rock CEC to the total illite+smectite weight fraction from QXRD
corrected for minor CEC contributions from other minerals. The whole rock CEC values
were used to estimate a virtual IS expandability of the I+S mineral group assuming a pure
illite end member CEC of 15 meq/100g (CECI ) and a pure smectite end member CEC of
100 meq/100g (CECS ) using the following relationship:
%SI -S =

measCECI -S CECI
100
CECS CECI

(2.1)

The QXRD mineralogy, the majorelement and traceelement chemistry data and CEC
data were combined and optimized , to calculate individual mineral composition and anhydrous grain density.

The

Chevron proprietary BestRockmodeling method, similar to the modeling approach presented in

rodo & Kawiak (2012), was used

13

2.1.2

Organic Matter Characterization

Total organic carbon (TOC) content and characterization (type and thermal maturity)
were obtained commercially by Weatherford Laboratories, Houston, TX, using the LECO
TOC and the RockEval II methodology. The details of the LECO TOC and RockEval
methodologies are given in Jarvie (1991). The TOC is the measure of total organic carbon
as wt.% of the rock. Conversion of TOC to OM wt.% requires information of the carbon
content in OM. The wt.% carbon within OM is highly variable ranging from 76 to 90%
depending upon maturity (Ungerer et al., 1981). In absence of this data, the OM content
(wt.%) was calculated assuming average carbon concentration of 83% by molecular weight
(Jarvie, 1991). The RockEval II S2 pyrograms were used to as an indicator of the chemical
heterogeneity of the OM.
2.1.3

Textural Characterization

Textural characterization was made on representative subsamples using eld-emission

scanning electron microscope (FESEM) imaging on polished uncovered thin sections. Uncovered polished sections were prepared using surface impregnation with a low-viscosity
medium before the nal polish to reduce mechanical damage. Gold-coated thin sections
were imaged using a JEOL JSM-7000F FESEM at a facility located in the Metallurgical
and Material Science Department at the Colorado School of Mines. Component identication and characterization were made with the aid of EDAX Genesis EDS, TSL, and EBSD
spectra.
2.1.4

Thermogravimetric Analysis

Adsorbed water content was evaluated by thermogravimetric analysis and mass spectrometry (TGA-MS). Generally, 30-50 mg of <40 mesh (0.4 mm) powdered samples were
analyzed at a heating rate of 5 C /m in a TA Q5000 IR TG instrument purged with dry
N2 at a ow rate 50 ml/min. The weighting error of the TA Q5000 IR instrument under
these measurement conditions is <0.001 mg. To record the atomic mass signals of evolved
14

gas and H2 O during heating, the tube was connected to a ThermoOnix quadropole mass
spectrometer. Reagent grade N2 was used as a carrier gas owing at 50 ml/min Derkowski
et al. (2012)
2.2

Description of Sample Sets

Mudrocks from dierent locations with dierent mineralogy, clay and OM content were

used for this thesis. Three dierent set of mudrock samples from an Eastern European gas
shale play, the Haynesville Formation and the Niobrara Formation were used. The compositional correlation and geological context of the Haynesville Formation and the Niobrara
Formation were studied in greater details. The mineralogical correlations will be used later
to understand porosity and poresize distributions associated with specic mineral assemblages. A single mudrock samples from dierent locations, previously studied by Sarker
(2010), was also analyzed for pore-structure characterization. To understand the physical
properties of the clays endmembers, source clays from the Clay Mineral Society was also
used. In the following subsections, compositional characteristics of the dierent sample sets
are described in details.
2.2.1

Source Clays

Wyoming montmorillonite (SWy2) and low-defect Georgia kaolinite (KGa1b) were obtained from the Source Clay Repository of Clay Mineral Society. These are natural samples
and so have some minor contaminants but they are dominated by the specied minerals.
These two samples were studied in asreceived powdered form to understand the porestructure of these clays in general. The as-received pure clays will be referred to by the
Clay Mineral Society naming (e.g. SWy2, KGa1b) for the rest of the thesis. Clays in
mudrocks are compacted over the geological history and the physical properties, such as
density, porosity, acoustic velocities, permeability change with depth of burial. Mechanical
compaction, which is a function of increase in overburden stress, results in a change in the
pore structure, exponential decrease in pore volume with depth and in the alignment of clays.

15

To examine the eect of compaction on pore-structure, Wyoming montmorillonite powder

(SWy2) was articially compacted using cold pressing techniques to make compacted pellets. The powdered samples were put inside a one-and-half inch cylindrical mold with a
piston. The inner surfaces of the barrels were coated with Teon tapes to prevent sticking
of charged clays with the walls during the pressing. This also helps in reducing the friction
associated while pushing the sample out of the barrel and sample recovery was made simple.
Then pressure was applied uniaxially to the mold via the piston at a very slow rate in a cold
press. The movement of the piston is controlled by a pump owing at a constant rate of
0.05 ml/min. Samples were prepared by pressing at dierent pressures (between 4000 psi to
10000 psi) to obtain pellets with dierent porosity and level of compaction. The pellets were
named using clay mineral identier and the pressure applied, e.g. sample SWy24K means
the pellet is made from SWy2 Wyoming Montmorillonite clay by applying 4000 psi of axial
load. Rectangular chips (of about 0.5 gm weight) were cut from the pellets using a diamond
saw to use for pore-size distribution measurements.
2.2.2

Eastern European Silurian gas shale play

A set of twentytwo samples from a Silurian gas shale play in Eastern Europe was studied
for this thesis. Henceforth, this sample set will be referred as SS1 (Sample Set 1). The details
of geological formation and exact location for these samples were not available. The samples
are composed mostly of I+S clay minerals, quartz, and feldspar with a low to moderate
amount of carbonate minerals (up to 20 wt.% calcite and dolomite; Figure 2.1(a), Table 2.1).
The equivalent smectite content in the I+S clay group is <5 %, indicating the clay minerals
are mostly illites. The TOC content ranges from 1.0 to 6.0 wt.% with an average value of
2.6 wt.% (n = 22). The RockEval II S1, S2, and HI values Table 2.1 and S2 pyrograms
(Figure 2.1(b)) indicate that the OM in these samples is thermally mature with essentially
inert with no present hydrocarbon generation potential from kerogen. These mudrocks are
texturally immature, poorly sorted and composed of silt- and clay-sized grains. Silt-size
quartz, pyrite, altered clay psuedomorphs and dolomite particles, are uniformly surrounded
16

by a matrix of clay-size material (Figure 2.1(c)). OM exists in lamination parallel to the

bedding plane (Figure 2.1(c)) and also dispersed within the clay matrix (Figure 2.1(d)).
Recrystallized carbonate (calcitic) cements are observed occasionally (Figure 2.1(d)).
2.2.3

Haynesville Formation

The Upper Jurassic Haynesville Formation is one of the important currently productive
gas shale play located in east Texas, USA, with an estimated total reserves of in excess of 100
trillion cubic feet (tcf), and the estimate of ultimate recovery (EUR) per well ranges between
4 and 7.5 bcf (Wang & Hammes, 2010). The Haynesville Shale is an organic, siliceous and
carbonaterich post-rift transgressive mudrock deposited during the Upper Jurassic (Kimmeridgian to the early Tithonian) in a deep partly euxinic and anoxic basin synchronous
with the opening of the Gulf of Mexico (Hammes et al., 2011). A complex depositional
environment existed in the palaeoHaynesville Basin that was partly surrounded by carbonate shelves and sedimentation by both carbonatedominated and siliciclasticdominated
inll prevailed depending upon proximity to carbonate platforms and siliciclastic source
respectively. A wide compositional variation was reported and three main lithofacies were
identied: bioturbated and laminated calcareous mudstone, silty peloidal siliceous mudstone,
and unlaminated siliceous organicrich mudstone (Hammes et al., 2011).
Thirty four of samples from the a well drilled in the Haynesville Formation in East Texas,
USA, were selected for this thesis. Henceforth, this sample set will be referred as SS2 (Sample Set 2). The samples are composed mostly of I+S clay, quartzo-feldspathic minerals,
pyrite with a variable amount of carbonate minerals (calcite, ankerite\Ca-excess dolomite
and minor amounts of siderite) up to 59 wt.% (Figure 2.2(a), Table 2.2). The I+S clay
species contains <5% up to 9 % expandable smectite layers. The TOC ranges from 0.5 to
6.3 wt.% with a mean of 2.8 wt.% (n=34). The RockEval II S1, S2 and HI values (Table 2.2)
indicate that the OM is thermally mature but, not as highly mature as SS1. The variable
RockEval II S2 pyrogram patterns for these samples (Figure 2.2(b)) indicate that the OM
is very heterogeneous and has a signicant amount of bitumen indicated by a dominant low
17

(a) Mineral Composition

(b) RockEval II S2 pyrogram

(c) FESEM thin section sample SS116

(d) FESEM thin section sample SS116

Figure 2.1: Compositional and textural characterization of the samples from Silurian gas
shale play (SS1) (a) Mineral composition obtained from QXRD; (b) RockEval II S2 pyrograms of dierent representative samples. The low values of S2 peak (y axis FID response)
and nearly at pyrogram indicates that the OM in SS1 are thermally mature with almost
negligible amount of pyrolysable kerogen. (c) & (d) Textural variations seen in backscattered electron images (COMPO)of polished thin section representative samples from SS1
(see text for detailed description) QTZ = Quartz. DOL = Dolomite. PYR = Pyrite. ANA
= Anatase. OM = Organic Matter. Alt Clay (Pseudo) = Clay mineral pseudomorph of
altered detrital grain. ALB = Albite.

18

temperature S2 peaksbetween 350 to 450 C (Clementz, 1979). Thin section studies using
FESEM imaging indicates that the majority of the siliciclastic dominated lithologies are
texturally immature where quartz and albite silt grains are scattered in a clay-size particle
dominated matrix (Figure 2.3(a) and Figure 2.3(b)). The clay matrix shows a preferred
orientation of the particles approximately parallel to the bedding direction (Figure 2.3(b)).
The OM is dispersed within the clay matrix (Figure 2.3(b)) and also as discrete particulate grains (Figure 2.3(c)). Some sample show that extensive calcite cementations forming
laminations of variably cemented zones (Figure 2.3(a) and Figure 2.3(c)). Other recrystallized mineral cementing materials include albite and quartz (Figure 2.3(c)). Figure 2.3(d)
shows an example of carbonate dominated samples showing extensive recrystallization of the
detrital carbonate grains.

(a) Mineral Composition

(b) RockEval II S2 pyrogram

Figure 2.2: Compositional characterization of the samples from Haynesville Formation (SS2)
(a) Mineral composition obtained from QXRD; (b) RockEval II S2 pyrograms of dierent
representative samples. The presence of low temperature peak indicates thermally mature
OM with high percentage of bitumen.

The total carbonate content correlates inversely with I+S clay content (Figure 2.4(a)).
This is due to varying sedimentation environments that prevailed during the Haynesville

19

QTZ
OM
(Particulate)

QTZ

OM dispersed
with clay

(a) Siliciclastic dominated lithofacies representative

sample

(c) FESEM thin section sample SS26

(b) FESEM thin section sample SS21

(d) Detrital Carbonate dominated lithofacies representative sample SS22

Figure 2.3: Textural variations seen in backscattered electron images (COMPO)of polished
thin section representative samples from the Haynesville Formation (see text for detailed
description) QTZ = Quartz. OM = Organic Matter. ALB = Albite.

20

deposition with dierent relative amounts of carbonates versus siliciclastics basin inux controlled by proximity to the carbonate platform source and siliciclastic source respectively
(Hammes et al., 2011). The OM content show a negative correlation with the I+S clay content (Figure 2.4(b)). Such correlation is due to the dilution of OM by inux of siliciclastic
sediments during deposition (Hammes et al., 2011).

(a) Calcite vs. I+S

(b) OM vs. I+S

Figure 2.4: Mineralogical associations in Haynesville samples. (a) Relationship between

calcite and I+S group clay content (wt.%). (b) Relationship between OM content and I+S
group clay content.

2.2.4

Niobrara Formation

The Niobrara sample set consisted of twenty-two samples obtained from the Berthoud
State No. 3 well (NE SW Sec 16, T4N, R69W), Berthoud Field, Larimer County, CO. Cores
from Berthoud State No. 3 well have been studied extensively (Pollastro, 1992; Longman
et al., 1998; Landon et al., 2001). Henceforth, this sample set will be referred to as SS3
(Sample Set 3). The Niobrara Formation was deposited during an Upper Cretaceous major
marine transgression along the eastern shelf of Laramidia. The depositional environment in

21

the broad, shallow, asymmetrical Western Interior Seaway resulted in nannofossil-rich carbonate sedimentation (primarily coccoliths). The dominant rock types are nely laminated,
nannofossiliferous chalk and interbedded black shaly marls, organized into two main members: the Fort Hays Limestone at the base and the overlying Smoky Hill Member (Pollastro,
1992). The Smoky Hill Member is further divided into seven alternating chalk-rich (Chalk
A, Chalk B, Chalk C and Chalk D) and marl-rich members (Marl AB, Marl BC, Marl C
D). The samples were selected to reliably represent the individual members in the Niobrara
Formation, and to evaluate preferential pore structure associations with certain constituents,
such as clay minerals and OM. The rocks are mostly composed of calcite and clay minerals(Figure 2.5(a), Table 2.3). I+S clay is the most dominant phase in the non-carbonate
portions. Total organic carbon (TOC) content varies from 0.1 to 5.3 wt.%. The OM in
Niobrara was dominantly Type II sapropelic (Landon et al., 2001). The RockEval II S2 pyrograms have a single relatively high temperature peak indicating dominance of pyrolysable
kerogen in the OM. There is no indication of the presence of bitumen and the Tmax varies
from 417 C to 443 C, with an average Tmax of 436 C ( 6 C) (Figure 2.5(b), Table 2.3).
The oil window generally corresponds to a Tmax of 432 C (Landon et al., 2001). The values
obtained in this study indicate that thermal maturity of the OM is either immature or just
at the onset of the oil window.
The total calcite mineral wt.% correlates inversely with I+S clay content (Figure 2.6(a)).
This is consistent with the reported eustatic dominated cyclic sedimentation pattern of the
Niobrara Formation that alternated between deposition of pelagic chalk units during warmer
current inux and calcareous mudrocks or shaly chalk units during cold Arctic paleo-current
ow (Pollastro, 1992; Longman et al., 1998). The Chalk D and the Fort Hays Member
from the basal units of the Niobrara Formation, have comparatively low organic content
(<1 wt.%). The organic carbon content is thought to have been controlled by a gradual
deepening of the basin and uctuating paleo water depth resulting in periods of anoxia
(Longman et al., 1998). The basal units were deposited under shallow oxidizing conditions

22

(a) Mineralogy

(b) OM characterization

Figure 2.5: Compositional characterization of the samples from Niobrara Formation SS3
(a) Mineral composition obtained from QXRD; (b) RockEval II S2 pyrograms of dierent
representative samples.

with low preservation potential for organic carbon. The OM content increases up section from
1.0 to 6.4 wt.%, where the sediment was deposited in deeper, more anoxic water (Longman
et al., 1998). In this upper organic rich section, two trends are observed with regard to the
clay content (Figure 2.6(b)). Samples with low clay content show a positive trend with OM
content, and samples with higher clay content (>15 wt.%) show a negative trend with OM
content.
2.2.5

Assorted set of mudrocks

An assorted set of individual mudrock samples, previously studied by Sarker (2010) for
elastic and ow properties, were selected. Henceforth, this sample set will be referred as ASS
(Assorted Sample Set). The samples analyzed are North Sea Shale, Pierre Shale, Mancos
B Shale, Cox Argillite, Middle Bakken Shale and Woodford Shale. Detailed mineralogical,
compositional and textural descriptions of the samples are given in Sarker (2010). The exact
samples used by Sarker (2010) were not avaliable and separate sample piece was used for
this study. Quantitative mineralogical analysis by QXRD was repeated to verify the eect
of sub-sampling and sample heterogeneity. Except for Pierre Shale and Woodford shale,

23

(a) Calcite vs. I+S

(b) I+S vs. OM content

Figure 2.6: Mineralogical associations in Niobrara samples. (a) Relationship between calcite
and I+S group clay content (wt.%). (b) Relationship between OM content and I+S group
clay content.

the mineralogy is consistent and within the limits of QXRD, as shown in Table 2.4. Along
with these samples, an illite-rich Cambrian mudrock sample from Silver Hill, Montana was
obtained as ground rock chips from the Clay Mineral Society. This particular mudrock is
used as the source for the Clay Mineral Society standard (Imt-2) for illitic clays, however
it contains signicant amount (10 wt.%) of non-clay minerals (Table 2.4). Therefore, this
sample is treated as an illite-rich mudrock, instead as pure illite clay samples.
2.3

Summary
In this chapter, the sample characterization methodologies and the compositional descrip-

tion of the sample set were presented in detail. The samples are characterized for whole rock
quantitative mineral and elemental composition, along with cation exchange capacity (CEC)
and OM characterization using LECO TOC and RockEval II. These data were combined
and optimized using the Chevron proprietary BestRock modeling method to calculate individual mineral composition and anhydrous grain density. Textural characterization was
24

done on some representative samples using FESEM imaging on polished thin section. A detailed compositional description of various mudrock samples including Haynesville, Eastern
European Silurian mudrock formations, Niobrara and others that were used for experimental
work are presented in this chapter.

25

Table 2.1: Mineral and OM content (wt.%), CEC and RockEval II parameters of samples from Eastern European Silurian
gas-shale play (SS1). The concentration values reported are wt.% normalized to 100% obtained after combining the mineral
phases wt.% (from QXRD) and OM wt.% (from LECO). Carbon wt.% in OM is assumed to be 83% (Yen & Chilingar, 1976).
ID

Qtz

SS1-1
SS1-2
SS1-3
SS1-4
SS1-5
SS1-6
SS1-7
SS1-8
SS1-9
SS1-10
SS1-11
SS1-12
SS1-13
SS1-14
SS1-15
SS1-16
SS1-17
SS1-18
SS1-19
SS1-20
SS1-21
SS1-22

32
24
26
29
29
29
28
27
33
29
24
44
33
39
44
38
43
33
46
44
31
34

KPlag
spar
0
1
2
3
2
2
2
3
3
2
4
3
2
2
3
2
3
2
2
2
3
2

6
5
7
7
6
7
6
5
8
8
5
4
5
5
5
6
6
9
4
3
8
6

Cal

Ank

6
9
3
0
0
1
0
0
2
3
0
3
4
3
1
1
1
1
0
0
1
2

4
10
7
2
4
6
6
4
7
8
0
1
3
2
1
2
2
2
2
2
3
4

Sum
Pyr
Carb

Hal

Ana

Org

Sum
Non Chl
Clay

IS

%SI -S

Sum
CEC TOC S1
Clay

3.7
5.0
2.9
4.9
0.7
1.0
1.0
3.0
1.9
2.0
3.7
7.1
1.0
1.6
2.9
4.9
2.7
2.5
2.4
3.0
1.2
1.9

0.9
0.0
0.9
0.7
1.0
0.9
0.9
0.2
0.0
2.0
0.8
0.0
0.0
0.0
0.0
0.5
0.0
1.0
0.5
0.6
1.9
1.0

0.6
0.0
0.7
0.0
0.3
0.7
0.7
0.0
0.0
0.2
0.5
0.0
0.5
0.0
0.0
0.2
0.2
0.2
0.5
0.3
0.7
0.9

1.2
2.5
1.7
2.1
1.4
2.1
1.5
2.4
1.4
1.2
4.2
7.4
3.3
6.7
7.7
6.4
4.9
4.2
1.5
1.4
1.8
2.7

55
57
50
49
44
50
45
44
55
55
42
70
52
59
64
60
63
54
59
57
51
53

8
4
8
9
10
9
8
9
6
8
9
6
4
4
3
5
5
6
9
6
8
5

37
39
42
42
47
41
48
47
39
37
48
24
44
37
32
33
32
40
32
37
42
42

7
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5

45
43
50
51
56
50
55
56
45
45
58
30
48
41
36
40
37
46
41
43
49
47

9
19
10
2
4
7
6
4
9
11
0
4
7
5
2
3
3
2
2
2
4
6

9.1
7.0
7.5
7.4
6.2
6.1
5.5
8.0
3.7
4.1
3.9
4.3
4.2
3.6
3.6
4.7
3.0
4.9
3.1
3.4
3.3
1.2

1.03
2.06
1.42
1.76
1.13
1.71
1.27
1.97
1.19
1.01
3.52
6.13
2.76
5.56
6.43
5.31
4.10
3.45
1.27
1.18
1.48
2.21

0.03
0.01
0.02
0.05
0.02
0.04
0.02
0.04
0.02
0.02
0.08
0.04
0.04
0.05
0.06
0.05
0.05
0.07
0.03
0.04
0.03
0.06

S2

S3

HI

OI

0.09
0.01
0.02
0.14
0.03
0.08
0.04
0.01
0.01
0.01
0.13
0.33
0.10
0.25
0.30
0.16
0.10
0.09
0.01
0.03
0.04
0.04

0.16
0.2
0.23
0.13
0.09
0.14
0.24
0.20
0.20
0.12
0.12
0.41
0.16
0.23
0.29
0.26
0.57
0.16
0.54
0.23
0.14
0.38

9
0
1
8
3
5
3
1
1
1
4
5
4
4
5
3
2
3
1
3
3
2

16
10
16
7
8
8
19
10
17
12
3
7
6
4
5
5
14
5
43
20
9
17

Qtz = Quartz (wt%); K-spar = K- feldspar(wt%); Plag = Plagioclase(wt%); Cal = Total Calcite and Mg-rich Calcite(wt%); Ank = Ankerite
including excess Ca-Dolomite (wt%); ; Pyr = Prite; Chl = Chlorite; I-S = Illite+Smectite group (wt%), total dioctahedral 2:1 layer clay: illite,
mixed-layer illite-smectite, smectite, and mica; CEC = Cation Exchange Capacity (meq/100 g); TOC = Total Organic Carbon(wt%); S1 = volatile
hydrocarbon (HC) content (mgHC/g of rock); S2 = pyrolysable organic matter, mgHC/ g of rock; S3= carbon dioxide contenet (mgCO2/ g of rock);
HI = Hydrogen index, S2 x 100 / TOC (mgHC/g TOC); OI = Oxygen index, S3 x 100/TOC (mgHC/g TOC)

26

Table 2.2: Mineral and OM content (wt.%), CEC and RockEval II parameters of samples from Haynesville gas-shale play
(SS2). The concentration values reported are wt.% normalized to 100% obtained after combining the wt.% of mineral phases
(from QXRD) and wt.% of OM (from LECO analysis). The conversion factor used for the conversion of TOC to wt.% OM is
83% (Yen and Chilingar, 1976)
ID

Qtz

Plag Cal

SS2-1
SS2-2
SS2-3
SS2-4
SS2-5
SS2-6
SS2-7
SS2-8
SS2-9
SS2-10
SS2-11
SS2-12
SS2-13
SS2-14
SS2-15
SS2-16
SS2-17
SS2-18
SS2-19
SS2-20

20
9
15
16
7
13
19
18
21
18
17
19
13
20
22
19
21
23
24
21

5
9
16
10
6
7
6
3
5
6
10
8
10
9
9
8
9
10
10
8

2
3
22
14
27
35
8
2
1
3
5
31
31
8
6
9
9
10
7
9

Sid
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Dol
Total
+
Pyr
Carb
Ank
2
44
1
1
32
0
1
1
1
1
0
0
1
1
0
1
0
1
1
0

4
47
23
15
59
35
9
3
2
4
5
31
32
8
6
10
10
11
8
9

2.0
0.9
2.0
5.0
1.0
3.8
2.0
2.0
3.0
2.0
2.0
2.0
7.3
1.9
2.0
1.0
1.9
1.0
2.0
1.0

Anh

Org

Sum
Non Kao Chl
Clay

IS

%SI -S

Sum
CEC
Clay

TOC

S1

S2

S3

HI

OI

1
1
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0

1.0
1.0
4.0
4.3
1.3
7.8
1.0
1.0
0.5
1.0
2.9
3.9
7.2
3.9
3.7
2.8
4.6
2.3
3.6
3.5

33
69
60
51
75
68
37
27
32
31
37
64
67
43
43
40
46
47
47
43

0
0
0
0
0
0
0
0
6
1
0
2
2
2
3
5
2
3
2
1

52
21
31
38
19
27
47
60
50
60
55
32
28
46
45
44
43
40
42
47

<5
<5
7
5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
6
8
5

67
31
40
50
25
32
63
73
68
69
63
36
33
57
57
60
54
53
53
57

10.1
4.2
7.2
8.1
0.6
5.1
8.5
10.9
9.9
11.6
10.0
4.0
3.3
9.6
9.4
7.5
8.4
8.9
10.3
10.1

0.85
0.87
3.30
3.54
1.10
6.51
0.83
0.82
0.45
0.82
2.38
3.22
5.95
3.23
3.04
2.33
3.82
1.87
2.98
2.87

0.16
0.40
2.68
1.70
0.56
4.88

0.40
0.49
1.86
1.01
0.63
3.18

0.19
0.16
0.24
0.22
0.12
0.23

47
56
56
28
57
48

22
18
7
6
11
4

15
10
9
12
5
6
16
14
12
8
9
3
2
9
9
11
8
10
9
10

Continued on next page

Qtz = Quartz (wt%); K-spar = K- feldspar(wt%); Plag = Plagioclase(wt%); Cal = Total Calcite and Mg-rich Calcite(wt%); Dol + Ank = Dolomite,
Ankerite including excess Ca-Dolomite (wt%); ; Pyr = Prite; Chl = Chlorite; I-S = Illite+Smectite group (wt%), total dioctahedral 2:1 layer clay:
illite, mixed-layer illite-smectite, smectite, and mica; CEC = Cation Exchange Capacity (meq/100 g); TOC = Total Organic Carbon(wt%) LECO; S1
= volatile hydrocarbon (HC) content (mgHC/g of rock); S2 = pyrolysable organic matter, mgHC/ g of rock; S3= carbon dioxide contenet (mgCO2/
g of rock); HI = Hydrogen index, S2 x 100 / TOC (mgHC/g TOC); OI = Oxygen index, S3 x 100/TOC (mgHC/g TOC)
TOC from Corg chemistry measurement. No Leco TOC and RockEval data available.

27

Table 2.2: Continued.

ID

Qtz

SS2-21
SS2-22
SS2-23
SS2-24
SS2-25
SS2-26
SS2-27
SS2-28
SS2-29
SS2-30
SS2-31
SS2-32
SS2-33
SS2-34

21
21
19
21
20
22
20
28
24
23
22
22
24
19

Plag Cal
8
8
9
5
6
5
5
5
5
5
5
6
6
6

6
10
7
32
25
18
27
15
22
16
30
22
22
25

Sid
0
0
0
0
0
1
0
0
0
1
0
1
1
0

Dol
Total
+
Pyr
Carb
Ank
1
1
0
0
0
0
4
1
2
0
1
0
1
0

7
10
7
32
25
19
31
17
24
16
31
23
24
25

1.9
2.9
1.9
5.0
3.8
3.8
2.8
4.8
1.9
3.8
2.9
3.8
2.9
4.8

Anh

Org

Sum
Non Kao Chl
Clay

IS

%SI -S

Sum
CEC
Clay

TOC

S1

S2

S3

HI

OI

0
0
0
0
0
0
0
0
0
0
0
0
0
0

2.8
2.7
3.4
7.6
3.6
4.4
3.4
2.7
2.7
4.1
3.6
5.0
3.5
4.6

40
45
41
71
58
53
62
57
57
52
64
59
60
59

1
3
4
0
2
2
3
0
1
3
2
1
1
2

49
47
47
29
38
41
30
38
37
42
30
36
35
36

<5
6
<5
<5
<5
<5
<5
<5
<5
7
<5
<5
9
<5

60
55
59
29
42
47
38
43
43
48
36
41
40
41

10.0
10.3
8.7
3.4
7.7
7.8
4.0
6.9
4.7
9.7
4.1
5.7
8.8
7.4

2.30
2.21
2.80
6.32
3.02
3.65
2.81
2.27
2.23
3.37
3.01
4.17
2.87
3.79

1.81
2.08
1.45
0.54
0.55
1.57
1.21
3.00
1.67
3.13

0.76
1.25
0.92
0.54
0.53
0.91
0.6
1.51
0.79
1.31

0.58
0.72
0.62
0.48
0.41
0.52
0.59
0.7
0.69
0.6

25
34
32
23
23
27
19
36
27
34

19
20
22
21
18
15
20
17
24
16

28

10
5
9
1
2
4
5
5
5
4
4
5
4
3

Table 2.3: Mineral and OM content (wt.%), CEC and RockEvalII parameters of the Niobrara (SS3) samples. The concentration
values reported are wt.% normalized to 100% obtained after combining the mineral phases wt.% (from QXRD) and OM wt.%
(from LECO). Carbon wt.% in OM is assumed to be 83% (Yen & Chilingar, 1976).
ID
SS3-1
SS3-2
SS3-3
SS3-4
SS3-5
SS3-6
SS3-7
SS3-8
SS3-9
SS3-10
SS3-11
SS3-12
SS3-13
SS3-14
SS3-15
SS3-16
SS3-17
SS3-18
SS3-19
SS3-20
SS3-21
SS3-22

Depth Litho
(ft.) Member

Qtz Plag Cal Dol

Total
Pyr
Carb

2928
2932
2945
2963
2978
2988
3017
3029
3048
3056
3068
3076
3080
3090
3103
3119
3138
3150
3161
3195
3203
3205

3
8
3
6
10
9
8
6
12
16
12
5
7
6
17
13
10
17
17
4
3
3

86
51
86
77
54
46
69
82
56
44
53
86
66
85
41
55
60
38
59
91
82
91

Chalk A
Chalk A
Chalk A
Marl A-B
Marl A-B
Marl A-B
Chalk B
Chalk B
Chalk B
Marl B-C
Marl B-C
Chalk C
Chalk C
Chalk C
Marl C-D
Marl C-D
Marl C-D
Chalk D
Chalk D
Fort Hays
Fort Hays
Fort Hays

1
3
2
2
2
2
2
2
2
3
2
1
2
2
3
2
3
5
3
2
1
1

85
50
85
76
52
44
67
81
53
41
52
85
65
84
40
54
60
37
55
90
80
90

1
1
1
1
2
2
2
1
3
3
1
0
0
0
1
0
0
1
4
1
2
1

2.2
9.4
2.1
4.2
7.8
10.3
2.0
0.9
2.1
2.1
6.3
0.5
2.1
0.9
3.0
2.9
4.1
2.0
0.5
0.0
0.6
0.0

Sum
Anh OM Non Kao IS
Clay

Sum
CEC TOC S1
%SI -S
Clay

0.0
0.0
0.0
0.2
0.2
0.2
0.4
0.0
0.2
0.1
0.2
0.2
0.4
0.0
0.3
0.1
0.4
0.0
0.4
0.0
0.5
0.2

34
>95
57
62
51
32
33
19
25
18
25
14
46
19
14
28
42
17
28
ND
32
ND

2.0
5.9
1.4
2.3
4.2
3.9
3.1
1.9
4.4
4.3
3.6
1.4
4.8
1.2
3.2
6.3
6.4
0.5
0.5
0.1
0.4
0.1

95
78
95
92
78
72
85
93
77
69
77
94
82
95
68
79
85
65
80
97
88
96

0
0
0
0
2
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0

5
22
5
8
20
28
15
7
23
29
23
6
18
5
32
21
15
35
20
3
12
4

5
22
5
8
22
28
15
7
23
31
23
6
18
5
32
21
15
35
20
3
12
4

2.4
23.5
3.5
5.9
13.3
12.1
6.9
2.3
8.6
9.2
8.5
1.7
9.9
1.7
9.2
9.2
8.2
10.9
8.0
ND
5.4
0.0

1.6
4.9
1.2
1.9
3.5
3.2
2.6
1.6
3.6
3.6
3.0
1.2
4.0
1.0
2.7
5.2
5.3
0.4
0.4
0.1
0.3
0.1

1.01
0.92
1.21
1.53
1.47
1.48
1.32
1.33
0.61
0.63
0.58
0.86
1.05
0.55
0.85
1.13
0.92
0.11
0.12
0.06
0.18
0.1

S2

S3

Tmax
HI OI
( C)

5.91
16.22
4.49
6.46
11.42
12.51
9.53
5.61
12.72
12.07
9.71
4.02
13.12
3.28
6.80
15.63
15.87
0.50
0.54
0.14
0.70
0.2

0.36
0.48
0.34
0.30
0.15
0.37
0.29
0.23
0.25
1.27
0.34
0.22
0.28
0.37
0.29
0.29
0.36
0.23
0.19
0.18
0.29
0.2

434
437
432
433
434
436
439
440
439
417
439
436
441
437
443
442
441
438
437
425
430
434

363
329
387
342
330
386
374
352
351
340
322
350
327
322
255
301
300
120
142
215
229
297

22
10
29
16
4
11
11
14
7
36
11
19
7
36
11
6
7
55
50
277
95
313

Qtz = Quartz; Plag = Plagioclase; Cal = Total Calcite and Mg-rich Calcite; Dol = Dolomite including excess Ca-Dolomite and Ankerite; Pyr = Prite;
OM = Organic matter = TOC (wt% of the rock) / 0.83; Kao= total dioctahedral 1:1 layer clay: kaolinite, dickite, nacrite; I-S = total dioctahedral 2:1
layer clay: illite, mixed-layer illite-smectite, smectite, and mica; CEC = Cation Exchange Capacity (meq/100 g); TOC = Total Organic Carbon(wt%)
LECO; S1 = volatile hydrocarbon (HC) content (mgHC/g of rock); S2 = pyrolysable organic matter, mgHC/ g of rock; S3= carbon dioxide contenet
(mgCO2/ g of rock); HI = Hydrogen index, S2 x 100 / TOC (mgHC/g TOC); OI = Oxygen index, S3 x 100/TOC (mgHC/g TOC)

29

Table 2.4: Major mineral and organic matter content of the studied samples from the assorted set of mudrocks . The concentration values reported here are weight percent normalized to 100 % obtained after combining the wt. % of mineral phases
(obtained from QXRD). Following the methods of rodo et al. (2001), quantitative mineral phase compositions were obtained
by X-ray diraction (XRD) analysis of randomly oriented powders spiked with ZnO, (QUANTA, Omotoso et al. (2006)). All
dioctahedral Al-rich 2:1 clay minerals (illite, dioctahedral smectite, muscovite, and mixed-layered illite-smectite) are quantied
together as the illite smectite group mineral (rodo et al., 2001). The Total Organic Carbon (TOC) is measured by LECO
technique. The mineralogy of samples from (Sarker, 2010) is shown for reference.
Sample
Silver Hill Cambrian Shale
Middle Bakken
Middle Bakken (Sarker, 2010)
Pierre Shale
Pierre Shale (Sarker 2010)
Cox Argillite
Cox Argillite (Sarker 2010)
Woodford Shale
Woodford Shale (Sarker 2010)
North Sea Shale
North Sea Shale (Sarker 2010)
Mancos B shale
Mancos B shale (Sarker 2010)

Qtz

KPlag Cal
spar

7
46
49
30
45
24
26
63
5
7
9
43
39

1
4
5
8
5
1
2
15
0
2
1
4
4

1
2
0
11
15
2
1
3
2
1
1
2
4

1
27
25
1
tr
24
21
0
0
0
2
7
9

Sid
0
0
0
0
0
0
1
0
0
0
1
0
1

Dol. Pyr
0
13
16
4
6
2
5
1
87
0
1
8
8

0
2
3
1
2
1
1
1
1
1
1
1
2

Hal
0
0
0
0
0
0
0
0
0
0
2
0
0

Sum
Opal Other
Non Kao Chl
C/CT min
Clay
0
0
0
0
0
0
0
0
0
17
14
0
0

0
0
0
0
0
0
0
0
3
0
0
0
0

10
94
98
56
73
54
58
83
98
29
32
65
67

0
0
0
0
0
1
1
0
0
14
15
6
6

1
0
<1
8
3
3
2
0
0
0
0
0
0

IS

Sum
Clay

TOC

89
6
2
36
24
40
39
17
2
56
53
28
27

90
6
2
44
27
44
42
17
2
71
68
35
33

N/A
N/A
0.11
N/A
18
N/A
0.52
N/A
1.07
N/A
0.2
N/A
1.36

Qtz = Quartz (wt%); Kspar = Kfeldspar(wt%); Plag = Plagioclase(wt%); Cal = Total Calcite and Mg-rich Calcite(wt%); Sid = Siderite; Dol +
Ank = Dolomite, Ankerite including excess Ca-Dolomite (wt%); ; Pyr = Prite; Chl = Chlorite; I-S = Illite+Smectite group (wt%), total dioctahedral
2:1 layer clay: illite, mixed-layer illite-smectite, smectite, and mica;TOC = Total Organic Carbon(wt%) LECO.

30

CHAPTER 3
TOTAL POROSITY MEASUREMENT IN MUDROCKS USING WATER IMMERSION
POROSIMETRY (WIP) TECHNIQUE

Accurate and minute measurement seems to the nonscientic

imagination, a less lofty and dignied work than looking for something
new. But nearly all the grandest discoveries of science have been but the
rewards of accurate measurement and patient long-continued labour in the
minute sifting of numerical results.
- William Thomson, 1st Baron Kelvin
Porosity is a most fundamental reservoir property. Energy industry requirements include
the accurate measurement of porosity in order to determine the hydrocarbon storage capacity, which is an important parameter in evaluating reservoir quality. Despite the commercial
importance of mudrock formations, porosity measurements from core samples are still challenging and the lack of an accurate measurement technique leads to a signicant uncertainty
in assessment of these plays. The current energy industry standard technique for measuring
porosity in mudrocks is based on crushed rock methodology developed by the Gas Research
Institute (GRI). The aggressive sample treatment in the GRI technique and the lack of a
standardized protocol reduce the accuracy of GRI results. An alternative method of measuring total porosity in mudrocks without using crushed rock is developed and presented
in this chapter. The chapter is organized as follows: Section 3.1 gives an introduction to
porosity measurement techniques. The potential pitfalls and limitations of current porosity
measurement techniques for mudrocks are discussed in Section 3.2. Porosity measurement
methodology using an immersion-saturation technique with deionized water is adopted and
modied for mudrock applications and details of the experimental procedures are presented
in Sections 3.3 and 3.4. The results of water immersion porosimetry (WIP) technique from
31

Eastern Europe Silurian mudrock samples (SS1), the Haynesville Formation (SS2) and
the Niobrara formations (SS3) are presented in Section 3.5. The results from the WIP
measurements are compared with other standard techniques like GRI and mercury intrusion
porosimetry (MIP). Finally, a detailed discussion providing an assessment of the advantages,
potential errors, pitfalls and reproducibility of this method are also presented in Section 3.6.
3.1

Introduction
Direct porosity measurement in mudrocks is complicated because of the ne grained

texture, the presence of extremely small pore sizes, often <10 nm, nanodarcy range permeability, and the strong interaction of water molecules with the clay mineral surfaces and
associated exchangeable cations. The current energy industry standard measurement protocol for porosity is based on a helium pycnometry technique developed by the Gas Research
Institute (GRI) (Luel & Guidry, 1989; Luel & Guidry, 1992; Luel et al., 1992). The
crushing procedure specied in the GRI technique and the lack of a standardized measurement practices signicantly reduces the accuracy of the GRI results (Karastathis, 2007;
Passey et al., 2010; Sondergeld et al., 2010, Spears et al., 2011). Comparisons of porosity measurements on mudrocks from immersion methods, helium pycnometry, and mercury
porosimetry demonstrate large inconsistencies between the methods (Dorsch et al., 1996).
In addition to basic methodology questions, signicant uncertainty exists in GRI porosity
measurements because unknown quantities of water molecules may be retained before injection of the measurement gas if temperature and humidity are not well controlled during the
experiment. Grain density and porosity values can be signicantly underestimated due to
such uncontrolled humidity conditions. Quantifying the experimental uncertainties is seldom
practiced during routine core analysis, although it is very important for mudrocks because
of their low porosity. Experimental uncertainties in porosity measurements need to be quantied because they have implications in the economic risk assessment of these commercial
plays. In addition to the accuracy of the porosimetry measurement technique, simplicity,
convenience of the methodology and reproducibility of the results must be considered to
32

make it useful for routine core measurements (Washburn, 1921b).

The objective of this study is to develop an alternative total porosity measurement
method for mudrock gas formations which avoids the inherent pitfalls associated with using
crushed rock such as with the GRI method. A combination of liquid saturation, immersion
and thermogravimetric techniques is optimized to measure the total porosity on intact mudrock samples. Deionized water is used for both the saturating and the immersing uid. A
systematic study of the WIP technique, including evaluating the reproducibility and experimental uncertainty, was conducted on several sets of samples including the Eastern European
Silurian gas shale play (SS1), the Haynesville Formation (SS2) and the Niobrara Formation
(SS3). The WIP results are compared the measurements by the GRI and MIP techniques
and the signicant dierences are discussed.
3.2

Current Porosity Measurement Techniques for Mudrocks

Common conventional reservoir core porosity measurement techniques include Boyles

Law helium porosity analysis, immersion technique using Archimedes principle and mercury injection porosimetry (MICP) porosimetry. Porosity measurements in mudrocks are
complicated because of the ne grained texture, small pore sizes and extremely low permeability. Nanodarcy permeabilities make traditional core measurement techniques extremely
dicult. To overcome these problems, Luel & Guidry (1989) suggested a crushed rock
methodology to obtain total porosity in such tight rocks. This technique has been adopted
by the upstream hydrocarbon industry for routine shale core analysis and is referred as the
GRI (Gas Research Institute) technique (Figure 3.1). In the GRI method, bulk density is
measured by mercury immersion using Archimedes Principle on a large (300 g) block of
intact untreated core material. The sample is then crushed in order to accelerate the extraction and pretreatment stages and also to facilitate complete invasion of helium in the
pore space during the subsequent grain volume measurement. The crushed sample is solvent extracted with hot toluene using the Dean Stark method, and then dried in an oven at
200 C. Grain volume of the crushed material is measured by helium pycnometry (Boyles
33

law). Pore volume is calculated using the dierence between bulk volume and grain volume
(Figure 3.1).
Signicant discrepancies exist in porosity and grain density values of comparable samples measured by dierent commercial laboratories using the GRI method (Karastathis, 2007;
Passey et al., 2010; Sondergeld et al., 2010; Spears et al., 2011). These discrepancies are
attributed to dierent and often undisclosed experimental procedures. The potential sources
of errors in the GRI method are discussed here and summarized in Figure 3.1. Inconsistent
crushing and sample handling procedures can introduce a sampling bias between the dierent stages of measurements and might explain the dierences in measured values between
laboratories. Moreover, these clay bearing rocks have high hydration potentials and abundant microporosity making them sensitive to changes in the relative humidity (RH) of the
environment. An unknown quantity of water molecules may be adsorbed or desorbed during
an uncontrolled experiment due to the hydration capacity of clays and the condensation
potential of micropore systems in moderate relative humidity. This can signicantly aect
the measured grain density (Figure 3.2) and produce even greater uncertainties in porosity
calculations. Another disadvantage of most GRI protocols is that they require crushing a
large mass of valuable core material (300 g) in order to tolerate a small mass loss during
the crushing step. Karastathis (2007) developed a modied methodology to address the
issue of mass conversion during crushing and grain volume measurement process, which improved the consistency of results while using a smaller quantity of sample (1014 g). The
reported deviation in GRI porosity measurements between dierent laboratories range from
0.5 porosity unit (p.u.) Luel et al., 1992; Karastathis, 2007; Passey et al., 2010) to 1.5
p.u. (Sondergeld et al., 2010; Spears et al., 2011).
For conventional reservoir rocks such as sandstones and limestones, the MIP technique
is widely accepted method to measure porosity and pore size distribution. However, current
instruments can only go up to 414 MPa (60,000psi), corresponding to a theoretical limit
Porosity

unit: A unit equal to the percentage of pore space in a unit volume of rock. It is abbreviated to
p.u. and lies between 0 and 100. (Schlumberger Oileld Glossary, )

34

Bulk density measured (

b)

on preserved full core (~300 g) by Mercury Immersion.

Samples crushed to powders and chips, ~150 g of crushed sample separated using splitter

Potential problems: Inconsistent crushing and sieving methods

between labs: Some use 20-35 mesh fraction and discard finer
fraction, some labs use all the crushed material.
Crushed samples weighed (ISWcs). Bulk volume calculated.

Potential problems: Loss of native saturation during crushing

technique, which could affect b. ISWCS & b are different portions of
sample, different saturations.

Fluid extraction using Dean stark. Sample weighed after extraction.

Potential problems: Dissolves solid bitumen creating porosity artifact,
artificially increases G.
Sample heated in oven up to 200 C (GRI report, 1989) & weighed after drying (FSWCS) Grain
density ( G) measured by He pycnometry.

Potential problems: Most labs only preheat to 110 C, hydroscopic

samples may adsorb moisture from air if exposed for any length of
time, may decrease G.

Total porosity calculated by difference of BV and GV.

Potential problems: Potential error in BV and GV from GD

measurement.

Fluid Saturation calculated from pore-volume and saturation calculation.

Figure 3.1: Flow chart summarizing the analytical procedure steps followed in the GRI
technique. The potential pitfalls associated with each Analytical step are highlighted in the
dark grey boxes.

35

Figure 3.2: Whole rock grain density measured by helium pycnometry of the samples, preheated at 200 C,when kept under a controlled vacuum chamber (dry) compared to when
equilibrated to room conditions. The dashed line is the tted linear equation for all data
points.
on the smallest pores that can be measured of diameter 3.6 nm. In tight mudrocks, there
is signicant pore volume that is too small to be accessed by MIP. Furthermore, at such
high pressures, blank eects from the heating and compression of mercury and the rock
structure, particle deformation, and the opening of closed pores may signicantly aect the
results (Penumadu & Dean, 2000; Bustin et al., 2008; Sigal, 2009; Comisky et al., 2011).
Sigal (2012) ported that the MIP porosity of samples from Barnett Shale is about 20%
to 50% lower than the measured helium porosity, mainly because mercury cannot access
pore volumes with throats smaller than 3.6 nm. In crushed mudrocks, the total specic
pore volume measured with MIP increases with a decreasing size of the crushed particles
(Comisky et al., 2011). The increased pore volume is related to better pore accessibility in
the smaller crushed particle size, but there also may be additional creation of pore space
during the crushing process (Comisky et al., 2011). Given the limitations associated with
crushing in the GRI and the inability of the MIP technique to access pores <3.6 nm, an

36

alternate method to measure total porosity on intact mudrock samples was developed.
3.3

Total Porosity Measurement by Liquid Saturation and Immersion Techniques

Liquid saturation and immersion techniques are standard techniques for measuring bulk

density, grain density (or specic gravity) and porosity adopted in many disciplines such
as the ceramic industry, material sciences, archaeology, mining and rock mechanics and
petroleum industry. The saturation and immersion technique determines porosity by saturating a sample with a liquid with known density, and calculating the pore volume from
the weight dierence between the fully saturated and dehydrated states. The total volume of the sample is determined using Archimedes principle. Thus, bulk density, grain
density, and porosity are determined from the same experimental sample. Liquid saturation and immersion techniques have been used for porosity measurements by the American
Petroleum Institute (API RP40), the American Society for Testing and Materials (ASTM
Standard C2000; ASTM Standard C830-00), and the International Society for Rock Mechanics (Franklin et al., 1981). The method has been employed in analysis of a wide variety
of rock types including low porosity granites (Alexander et al., 1981; Katsube & Kamineni,
1983; Melnyk & Skeet, 1986), oil sands and sandstones (Barnes, 1931; Plummer & Tapp,
1943), dolomites and limestones (Goldstrand et al., 1995), and mudrocks (Shapiro, 1975;
Brownell, 1977; Nuhfer, 1979; Howard, 1991; Katsube & Scromeda, 1991; Katsube et al.,
1992; Dorsch et al., 1996; Dorsch & Katsube, 1996; Gaucher et al., 2004). These studies
show that saturation and immersion technique is eective in measuring porosity in a wide
variety of materials when the experimental procedures are tailored for the particular type
of material. In this study, the saturation-immersion technique is modied to measure total
porosity in thermally mature mudrock lithologies such as gas shale that are characterized
by low total porosity, a signicant content of hydroscopic clay minerals, thermally mature
organic matter and a high amount of micro and mesopore volumes.

37

In the following subsections, the experimental considerations for adopting this technique
for gas-shale total porosity measurements are discussed. The two most important experimental details that needs to be considered in the choice of saturating and immersing uid
and the pretreatment conditions.
3.3.1

Choice of Saturating and Immersing Fluids

The accuracy of porosity measurements using the liquid saturation and immersion technique depends upon having a saturating and immersing uid that will eciently saturate
the entire pore network and be stable in laboratory conditions (Melnyk & Skeet, 1986). The
API Recommended Practices 40 document (API RP40) suggests using brine, light rened
oil, or a high-boiling point solvent as a saturating and immersion uid. Previous saturating
and immersion technique based porosity studies and published methodologies used various
uids that include acetylene tetrachloride (Barnes, 1931), deionized (DI) water (Washburn
& Fooyitt, 1921; Shapiro, 1975; Alexander et al., 1981; Katsube & Kamineni, 1983; Melnyk
& Skeet, 1986; Katsube & Scromeda, 1991; Katsube et al., 1992; Goldstrand et al., 1995;
Dorsch et al., 1996; Dorsch & Katsube, 1996), light hydrocarbon ((Brownell, 1977); Howard,
1991), kerosene ((Nuhfer, 1979); Gaucher et al., 2004) and other petroleum products and
paran (Washburn, 1921a; Washburn & Bunting, 1922). An ideal saturating uid should
have
1. low surface tension and a high wetting tendency,
2. low viscosity,
3. a high vapor pressure and slow evaporation rate,
4. low reactivity with the porous material,
5. stable composition and density,
6. be non-hazardous and have safe handling properties.

38

The driving forces to saturate porous media are the pressure gradient and the surface
tension. The coecient of penetration of a liquid into a capillary system should be high for
an ideal saturating uid. The coecient of penetration (Washburn, 1921a) for a wetting
uid, assuming a zero contact angle, is:
z=

(3.1)

where is the surface tension of the uid in dynes/cm, is the dynamic viscosity of the
uid in poise (dyne-s/cm2 ), and z is the coecient of penetration in cm/s. Higher penetration coecients of deionized water compared to other commonly used uids indicate faster
penetration into the micropores and capillaries (Table 3.1).
Table 3.1: Penetration coecient (Washburn, 1921a) of commonly used saturating uid for
liquid-saturation and immersion techniques, assuming the contact angle to be zero.

Liquid

Surface
Dynamic
Penetration Vapor
Tension
Viscosity
Coeecient Pressure
(dynes/cm) (centipoise) (cm/s)
(mmHg)

Acetelyene tetrachloride 31.1

Deionized Water
72.7
Kerosene
27.5
Light Hydrocarbon
n-Hexane
18.5
Octane
21.6
Decane
23.8
Isopar V
27.0

1.610
1.002
1.233

965.6
3630.9
1115.2

72.37
17.55
2.21

0.312
0.541
0.912
18.005

2964.7
1998.1
1305.5
75.0

121.41
10.43
0.96
<0.08

Deionized (DI) water has distinct advantages in saturating mudrocks because they can
have signicant clay mineral content and cation exchange capacity (CEC). Exchangeable
cations present on clay mineral surfaces and in interlayers form strongly held shells of water
molecules around them even under low humidity conditions and control the water adsorption
of the bulk sample (Drits & McCarty, 2007; rodo & McCarty, 2008; Derkowski et al.,
2012). The polar nature of water molecules has a strong anity for exchange cations and
the high specic surface area of clay minerals. Thus, saturating the pore system with water
39

molecules is ideal because water will penetrate small pores and capillaries associated with
clay mineral aggregates. DI water is stable, has a well-known specic gravity property, and
is not reactive or hazardous. Because water is used as both the saturating and immersing
uid, this procedure is referred to as the water immersion porosimetry (WIP) technique.
3.3.2

Pretreatment Conditions

All total porosity measurement techniques involve cleaning, or removing naturally occurring formation uids, before saturating the pore system with a measurable quantity of
external uid. The most used cleaning techniques to remove volatile compounds, including water and hydrocarbons, are organic solvent extraction using a Soxhlet or Dean Stark
apparatus, distillation by heating or a combination of both. The abundance of chemically
heterogeneous OM in organic rich mudrock reservoirs must be considered in the choice of
the pretreatment techniques. Solvent extraction and distillation by heating clean dierent
components of the residual oil and organic matter (Cooles et al., 1986; Wilhelms et al., 1991;
Noble et al., 1997). Distillation by heating removes volatile hydrocarbons in the C5 -
C25 composition range (Deygout, 2011), whereas solvent extraction removes the volatile hydrocarbons and the heavy-end" C12+ components, including nonvolatile asphaltenes and
high molecular weight polar nitrogen, sulfur, and oxygen (NSO) compounds (Larter, 1988).
Continuous solvent extraction under reuxing conditions (Soxhlet, Dean Stark) on crushed
rocks can dissolve the solid bitumen from the sample, which is thermally stable above 300 C
during RockEval pyrolysis and can be considered a framework component. The cleaning
technique using distillation by heating method was opted over the solvent extraction techniques because distillation does not remove the thermally stable solid bitumen out of the
rock, which is a framework grain under reservoir conditions.
The heating temperature for pretreatment must be high enough to remove volatile hydrocarbons and adsorbed water hydrating clay minerals, but not so high as to alter the solid
organic and inorganic framework. Drits & McCarty (2007), rodo & McCarty (2008) and
Derkowski et al. (2012) that smectite and expandable mixedlayered illitesmectite (IS)
40

clay structures are capable of retaining signicant electrostatic bound water as strongly held
exchange cation hydration spheres in excess of 200 C. Luel et al. (1992) and Luel &
Guidry (1992) suggested a preheating temperature of 110 C to remove the free and clay
bound water, which is similar to the recommendation of API RP40. However, the original
GRI report (Luel & Guidry, 1989) suggested a pre-heating temperature of 204 C (400 F)
to remove adsorbed water from illitic clay minerals. These authors reported an average
additional mass loss of 0.07 wt.% from a mudrock sample on heating to 204 C for 12 days
after prior heating to 110 C for 21 days. This additional 0.07 % weight loss translates to
underestimation of about 0.2 p.u. Thermogravimetrimetric (TGA) measurements on a illitic
mudrock from the Haynesville Formation suggest signicant dehydration mass loss happens
between 110 C and 200 C (Figure 3.3). Thus, a pretreatment temperature of 200 C for
12-16 hours is used for measurement of dry weight for the study samples. Unpublished data
on heat testing a wide variety of organic rich samples clearly indicates that a 200 C heating
temperature does not aect the solid OM present in the rock and will be discussed later (Dr.
Arkadiusz Derkowski and Dr. Douglas McCarty, personal communication).
3.4

Methodology
In the following subsections, the details of the WIP technique, calculations, and experi-

mental protocol will be discussed.

3.4.1

WIP Measurements Calculations

The total porosity of any sample measured by WIP (W IP ) is determined by the relationship:
W IP =

(B G )
100
(H2 O G )

(3.2)

where G is the grain density, B is the bulk density of the water-saturated sample, and
H2 O is the density of water at the measurement temperature T C. The bulk density is
determined by Archimedes Principle. Thus, by weighing a fully saturated sample in air and

41

Figure 3.3: TGA curve (black) for a representative mudrock sample from the Haynesville
Formation using a heating protocol as shown by the temperature curve (grey). Total weight
loss from start to 200 C and weight loss from 110 to 200 C are shown.
submerged in the immersion uid, bulk density (B ) can be calculated using Equation 3.3.
"

Sat.W t.Air
B =
(H2 0 air ) + air
Sat.W t. Air Sat.W t.Sub

(3.3)

where Sat.W t.Air is the weight of the saturated sample in air and Sat.W t.Sub is the weight
of the saturated sample submerged in DI water and air is the air density (0.0012 g/cm3 ).
The density of water is calculated using Equation 3.4, obtained by tting a second order
polynomial to the published water density data (Haynes et al., 2012, Chapter 6).
H2 0 = 0.0000053 T 2 + 0.0000081 T + 1.0001627

(3.4)

where T is temperature of the water during measurement. The dry grain density (G ) is
determined by weighing the completely dehydrated sample weight in air and the completely
saturated sample submerged in water:
"

DryW tAir
(H2 0 air ) + air
G =
DryW tAir Sat.W t.Sub

42

(3.5)

where DryW t.Air is the weight of completely dehydrated sample. The porosity measured
by WIP is the total water accessible porosity, including water adsorbed on clay surfaces,
which includes the interlayer space in the expandable clay minerals and external surfaces of
crystallites in nonexpandable clay species (i.e. clay-bound water). Therefore, the eective
porosity of a sample will generally be lower than the WIP porosity. Measurements of water
saturated weight in air (Sat.W t.Air ) and submerged weight in water (Sat.W t.Sub ) were repeated for at least ve times for each study sample. Averaging the repeated measurements
of Sat.W t.Air and Sat.W t.Sub allowed determination of the experimental uncertainties in
these parameters. The dry weight (DryW t.Air ) was measured only once for each sample.
However, based on repeated measurements of standard weights, an uncertainty of 0.005 g
was used for the dry weights. These uncertainties were carried through in the subsequent
calculations by propagating the error to nd the experimental uncertainty in the total porosity. The reported uncertainty is the expanded uncertainty with a coverage factor of 2, which
denes an interval having a level of condence of 95.45 percent (Taylor & Kuyatt, 1994). A
sample calculation is shown in Appendix A.
3.4.2

WIP Measurements Procedure

The following steps were followed during the WIP procedure (Figure 3.4). Henceforth,
this protocol will be referred to as the adopted protocol (AP). The experimental details and
parameters (e.g. sample size and mass, pretreatment time, saturating pressure and duration,
etc.) reported here are the optimum values obtained from this study, however protocols and
parameters can be varied depending upon the experimental requirements.
STEP 1 Sample Preparation On Day One, the samples were prepared by cutting two
to three rectangular shaped chips from core material of 35 g each and a total sample
weight of 10 g. The chips were cut using a diamond blade to smooth out irregularities
which would hold surface water outside the pore system. The samples were then heated
in a vacuum oven at 200 C overnight (1216 hours) in order to remove all volatile

43

components (water and hydrocarbons).

STEP 2 Dry Sample measurement On Day Two, the dry weight of the samples (DryW t.Air )
was measured in a moisture analyzer (Mettler Toledo HB43, readability 0.1 mg) by
heating the sample at 200 C for 15 minutes and the dry weight recorded at the end
of the heating cycle. Use of a moisture analyzer and measuring the dry weight at the
200 C condition prevents re-adsorption of moisture from the air during the dry weight
measurement. The weight of the sample was monitored at every ve minutes during
the heating cycle.
STEP 3 Saturating the samples After the dry weight measurement, the samples were
saturated under pressure in a Vinci Manual pressure saturator. The pressure saturation chamber was connected to an ultra-high vacuum pump to evacuate the air in
the sample prior to the introduction of water to ensure maximum possible saturation.
The samples were kept in vacuum of less than 1.33 Pa (10 mHg) overnight. Following
evacuation, on Day Three, DI water was introduced into the sample chamber of the
pressure saturator and a pressure of 13.7 MPa (2000 psi) was applied. The samples
were kept under this pressure for 24 hours.
STEP 4 Water Saturated Sample measurement On Day Four, the samples were
taken out of the pressure saturator and transferred in a water lled beaker to avoid
exposure to air. The chip portions were analyzed for bulk density using a conventional
highprecision jolly balance set up with a Mettler Toledo XS instrument that has
a readability of 0.01 mg. The saturated samples were weighed ve times while submerged in a beaker of DI water (Sat.W t.Sub ). The immersion DI water is mixed with
a drop of surfactant, provided by the Jolly balance instrument manufacturer, per 100
mL of water to relieve the surface tension of water eecting the submerged weight
measurement. The temperature of the immersion water was recorded. The saturated
samples were then weighed at least ve times in air (Sat.W t.Air ). The surface of the
44

chips was gently blotted with a small ne bristle paint brush on weigh paper to remove
surface water so there was no sheen left and care was taken to minimize the exposure
in air to minimize loss of water by evaporation. The samples were placed back in DI
water lled beaker in between the weighing to minimize prolonged exposure in air.

Figure 3.4: Flow chart summarizing the analytical procedure steps followed in the WIP
technique

3.4.3

Reproducibility test

A test was performed on samples from SS1 with dierent chip sizes and saturating conditions to nd the optimum parameters and to evaluate the reproducibility of the technique.
The dierent test conditions and the adopted protocol, described in Section 3.4.2 are summarized in Table 3.2. In Test 1 and Test 2, dierent samples from the same core depth were
used, however, the sample (chip) sizes were changed. Test 3 was repeated on exactly the
same samples that were used in Test 2.

45

Table 3.2: Test Conditions for the reproducibility test. AP = Adopted Protocol; Op. =
Operator
Test
#

Sample
Mass

AP

1011 g

Test1

4.55.5
g

Test2

56 g

Test3 5-F-6 g

3.4.4

Samples & Chips

Description
Samples from a
depth interval
Chip size: 5.0 g
Dierent sample
from same depth
interval Chip size :
0.5 g
Dierent sample
from same depth
interval Chip size :
2.5 g
Exactly same
samples
re-measured after
Test 2

Vacuum
system

Saturation
System

High
Vacuum

Pressure
Saturator,
2000 psi

Low
Vacuum

No
pressure,
Bench top

1&
2

Low
Vacuum

Pressure
Saturator,
2000 psi

High
Vacuum

Pressure
Saturator,
2000 psi

Op.

Additional Comments
DryW t.Air measured after oven
drying
DryW t.Air
measured after
jolly
balance
measurement
DryW t.Air
measured after
jolly
balance
measurement
Samples kept in
water saturated
for 2 weeks before analysis

Other porosity measurements

Depending on sample availability, WIP results were compared with other commonly
used mudrock measurement techniques, including the GRI and MIP techniques. Porosity
by the GRI technique was done by a commercial vendor using 300g of sample representing
stratigraphic interval of about 15 cm, in contrast to the discrete spot samples of about 10g
used for WIP.
MIP experiments were done using a Micromeritics AutoPore IV 9500 porosimeter. A
35 g sample was kept in an oven at 110 C overnight and then degassed at less than 6.6 Pa
(50 mHg) evacuation pressure for at least 30 minutes in the instrument. The samples were
not crushed but small chips were used. The bulk volume of the sample is measured after
lling the sample holder with mercury under a positive pressure of 10.34 kPa (1.5 psi). Pore
volumes are measured by intruded mercury volume at discrete pressure steps up to 413.7
MPa (60000 psi). Every pressure point was equilibrated below 0.001 L/g/s intrusion rate.

46

A conformance correction was done following the Bailey method as described in Comisky
et al. (2011)
3.5

Results
A total of 75 samples from three dierent shale plays sample sets were analyzed for total

porosity using the WIP technique: 22 samples from SS1 (Eastern European Silurian formation), 34 samples from SS2 (Haynesville Formation) and 19 samples from SS3 (Niobrara)
Formation. Detailed compositional description of the three dierent sample sets were presented in Section 2.2.2, Section 2.2.3 and Section 2.2.4, respectively. SS1 was selected based
on sample availability for reproducibility test and optimization of the experimental protocols for the WIP technique. In the following subsections, the results for the WIP studies are
presented.
3.5.1

Reproducibility Test

Despite signicant dierences in some of the experimental conditions, bulk density and
grain density measurements are indistinguishable within the uncertainty interval (Figure 3.5,
Table 3.3). Porosity calculations from these data using Equation 3.2 are also consistent with
dierences between the maximum and minimum measured total porosity ranging from 0.10
to 0.85 p.u. The average dierence in WIP total porosity measurements was 0.41 p.u. The
dierence between the measured total porosity can be accounted for by the measurement
uncertainties and the mean total porosity values are repeatable within range of experimental
uncertainty with a few exceptions. Sample SS111 had a maximum dierence of 0.85 p.u.,
and Sample SS116 a dierence of 2.03 p.u. A repeated measure ANOVA test was conducted
on the WIP data (n = 19) and indicates that there was no signicant dierence between the
dierent tests, p = 0.224 (Table 3.4).

47

Figure 3.5: Measured bulk density, grain density and total porosity values for the 22 test
samples from SS1 using WIP for adopted protocol, test 1 and 2. The errors bars indicate
expanded uncertainty of the experiment with a coverage factor of 2. The results are consistent
and repeatable within the experimental uncertainty.

48

Table 3.3: Results of the repeatability test of WIP technique. The details of the test conditions are provided in Table 3.2.
Adopted Protocol
ID
SS-1
SS-2
SS-3
SS-4
SS-5
SS-6
SS-7
SS-8
SS-9
SS-10
SS-11
SS-12
SS-13
SS-14
SS-15
SS-16
SS-17
SS-18
SS-19
SS-20
SS-21
SS-22

Bulk
Density
(g/cm3 )

Grain
Density
(g/cm3 )

Porosity
(%)

2.734(2)
2.701(3)
2.684(2)
2.698(2)
2.681(2)
2.661(4)
2.678(4)
2.659(3)
2.693(2)
2.699(4)

2.791(3)
2.797(3)
2.761(3)
2.792(3)
2.754(2)
2.741(4)
2.752(3)
2.754(2)
2.751(3)
2.764(3)

3.16(17)
5.31(23)
4.37(19)
5.25(18)
4.16(17)
4.54(31)
4.18(28)
5.40(22)
3.33(20)
3.71(26)

2.446(2)
2.596(2)
2.489(2)
2.457(2)
2.505(3)
2.532(4)
2.563(4)
2.661(2)
2.671(6)
2.657(2)
2.627(2)

2.540(2)
2.680(3)
2.575(2)
2.549(2)
2.603(3)
2.625(2)
2.662(3)
2.714(2)
2.727(3)
2.734(3)
2.709(3)

6.15(18)
5.00(20)
5.47(18)
5.93(17)
6.09(27)
5.75(26)
5.98(28)
3.08(18)
3.26(35)
4.41(18)
4.77(20)

Test 1

Test 2

Bulk
Density
(g/cm3 )

Grain
Density
(g/cm3 )

Porosity
(%)

2.734(4)
2.652(6)
2.681(5)
2.663(5)
2.677(4)
2.669(7)
2.682(5)
2.652(5)
2.690(5)
2.696(6)
2.621(5)
2.463(8)
2.603(5)
2.491(4)
2.463(4)
2.508(3)
2.540(4)
2.568(7)
2.657(5)
2.673(4)
2.652(4)
2.634(7)

2.790(6)
2.759(6)
2.769(5)
2.754(6)
2.752(6)
2.744(7)
2.749(6)
2.750(6)
2.754(6)
2.759(7)
2.706(5)
2.555(5)
2.681(6)
2.586(6)
2.557(5)
2.616(5)
2.626(6)
2.670(6)
2.712(6)
2.735(5)
2.737(6)
2.709(8)

3.11(38)
6.06(49)
4.97(40)
5.18(43)
4.28(40)
4.25(54)
3.81(46)
5.58(42)
3.63(41)
3.60(48)
5.00(41)
5.91(58)
4.66(44)
5.95(46)
6.05(41)
6.66(36)
5.23(41)
6.12(54)
3.21(43)
3.61(38)
4.90(42)
4.40(58)

49

Test 3

Bulk
Density
(g/cm3 )

Grain
Density
(g/cm3 )

Porosity
(%)

2.730(12)
2.675(14)
2.687(7)
2.648(3)
2.673(10)
2.665(7)
2.718(8)
2.652(7)
2.687(9)
2.691(9)
2.653(1)
2.539(5)
2.587(3)
2.482(6)
2.453(5)
2.489(5)
2.537(8)
2.557(10)
2.652(9)
2.675(12)

2.787(4)
2.778(4)
2.766(5)
2.739(4)
2.753(4)
2.744(4)
2.798(4)
2.751(4)
2.751(5)
2.754(4)
2.725(4)
2.638(3)
2.672(4)
2.575(3)
2.547(3)
2.619(3)
2.621(4)
2.660(4)
2.712(4)
2.725(4)

3.20(70)
5.78(84)
4.47(47)
5.24(28)
4.57(62)
4.52(48)
4.46(51)
5.67(45)
3.66(56)
3.59(54)
4.15(22)
5.98(34)
5.11(30)
5.92(40)
6.11(41)
8.06(36)
5.18(51)
6.19(61)
3.49(57)
2.93(70)

2.621(6)

2.711(3)

5.24(40)

Bulk
Density
(g/cm3 )

Grain
Density
(g/cm3 )

Porosity
(%)

2.733(4)
2.658(2)
2.672(3)
2.646(7)
2.662(6)

2.789(4)
2.774(4)
2.760(5)
2.740(4)
2.746(4)

3.13(30)
6.56(25)
5.02(28)
5.39(45)
4.85(41)

2.703(5)
2.636(7)
2.681(3)
2.686(3)
2.642(4)
2.534(4)
2.576(2)
2.475(4)
2.445(7)
2.478(5)
2.533(4)
2.562(3)
2.648(11)
2.669(3)
2.643(9)
2.618(5)

2.795(5)
2.747(4)
2.751(5)
2.757(4)
2.725(4)
2.638(3)
2.674(4)
2.581(5)
2.552(3)
2.618(3)
2.625(3)
2.666(4)
2.714(5)
2.726(4)
2.738(4)
2.714(4)

5.13(38)
6.38(47)
3.98(31)
4.04(25)
4.84(33)
6.32(30)
5.86(28)
6.69(37)
6.87(48)
8.61(35)
5.65(31)
6.22(27)
3.85(70)
3.33(29)
5.47(58)
5.57(38)

Comparisons between Test 2 and Test 3, which was a repeat of the measurement on the
same samples, shows there is consistently higher total porosity measured in the Test 3 samples, which were subjected to dehydration and saturation for the second time (Figure 3.6(a)).
This dierence in total porosity between Test 2 and Test 3 ranges from -0.07 to 0.83 p.u.
with a mean of 0.47 p.u. Note that the grain density values do not change (Figure 3.6(b)),
but there is systematic decrease in the bulk densities of the Test 3 samples compared to Test
2 (Figure 3.6(c)). This is supported by the paired ttest results (Table 3.5) on these samples
(n = 20) where bulk density values dier signicantly (p = < 0.0005), while the grain density
variation is not signicant (p = 0.558). The dierences observed in the WIP values may be
due to micro fractures or swelling from repeated sample saturation and dehydration, because
this would not aect the grain density, but would decrease the bulk density of the samples.
3.5.2

Water Immersion Porosity Measurement Results

The bulk density values from SS1 (n = 22) range from 2.446 to 2.734 g/cm3 , with a mean
value of 2.623 (0.088 [2]) g/cm3 . Note that the measured bulk density is with full DI
water saturation and does not represent the in-situ bulk density. The grain density values
of SS1 range from 2.540 to 2.797 with a mean of 2.704 ( 0.081 [2]) g/cm3 . The total
porosity calculated from SS1 ranges from 3.08 to 6.16 p.u. with a mean of 4.73 p.u. (1.02
p.u. [2]). In samples with low OM content, the total porosity increases with increasing clay
content (Figure 3.7(a)). The samples with the highest OM content have the highest total
porosity (Figure 3.7(a) and Figure 3.7(b)). The total porosity is inversely correlated with
total carbonate mineral content (Figure 3.7(c)).
The bulk density values in the Haynesville samples (SS2, n = 34) range from 2.432 to
2.727 g/cm3 with a mean of 2.543 (0.069 [2]) g/cm3 . The grain density values range from
2.527 to 2.845 g/cm3 with a mean of 2.694 (0.070 [2]) g/cm3 . The total porosity values
range from 3.45 to 12.68 p.u. with a mean of 8.90 p.u. (1.90 p.u. [2]). There is an inverse
correlation between the total porosity and carbonate mineral content (Figure 3.8(a) and
Figure 3.8(c)). Total clay content (Figure 3.8(a)) shows in general a positive correlation with
50

(a) WIP

(b) Grain Density

(c) Bulk Density

Figure 3.6: Comparison of measured (a) total porosity (b) grain density and (c) bulk density
on exact same samples between Test 2 and Test 3. Repeated measurement on the same
samples for the second time (Test 3) gives higher total porosity values compared to that
measured for the rst time (Test 2). The dierence in the measured WIP is due to possible
cracking of the samples by repeated sample saturation and dehydration which decreases the
bulk density but the grain density remain unaected.

51

Table 3.4: Results of repeated-measure ANOVA test comparing porosity values for the same
samples under dierent test condition (AP: Adopted protocol, Test 1 and Test 2, Table 3.2).
Sample SS1-11 and SS1-21 are excluded because of missing data, SS16 omitted as outlier.
Mauchly s Test of Sphericity indicated that the assumption of sphericity had not been violated, 2 (2) = 1.133, p =0.568, and, therefore, repeatedmeasure ANOVA with sphericity
assumption was used. There was no signicant eect of dierent test condition (Table 3.2)
on the measured porosity, F(2, 36) = 1.560, p = 0.224
Descriptive Statistics
Mean
WIP_AP 4.6737
WIP_Test1 4.7163
WIP_Test2 4.8058

SD
1.0225
1.0441
1.0417

N
19
19
19

Mauchly s Test of Sphericity

Measure:

WIP

Within
Subjects
Eect
Test

.936

Measure:

WIP

Mauchlys W

Approx.
Chidf
Square
1.133 2

Sig.
.568

Epsilon
Greenhouse- Huynh- LowerGeisser
Feldt
bound
.939
1.000
.500

Tests of Within-Subjects Effects

Source

Type
III
SS

df

Mean
Square

Test
Conditions

Sphericity
Assumed

.173

.086

Error
(Test)

Sphericity
Assumed

1.993

36

.055

52

Sig.

1.560

.224

Noncent. Obs.
Partial
PaPower
Eta
rame- =
Squared
ter
0.05
.080

3.120

.309

Table 3.5: Results of the paired-samples t-test conducted to compare the bulk density (BD),
grain density (GD) and porosity (WIP) values of the samples under test condition Test 2
and Test 3 (Table Table 3.2). Sample SS16 and SS121 are excluded because of missing
data.
Paired Samples Statistics
Std.
Error
Mean

Mean

Std.
Dev.

Pair 1

BD_Test2 2.6203
BD_Test3 2.6129

20
20

0.0831
0.0825

0.0186
0.0185

Pair 2

GD_Test2 2.7041
GD_Test3 2.7046

20
20

0.0723
0.0700

0.0162
0.0157

Pair 3

WIP_Test2 4.9500
WIP_Test3 5.4145

20
20

1.2577
1.3532

0.2812
0.3026

Paired Samples Test

Paired Dierences
95% C.I. of the Di
Mean

Std.
Dev.

Std.
Error
Mean

Lower

Upper

Sig.
df (2
tailed)

Pair
1

BD_Test2

0.0075
BD_Test3

0.0060

0.0013

0.0047

0.0103

5.557 19 0.000

Pair
2

GD_Test2

0.0005
GD_Test3

0.0038

0.0008

0.0022

0.0013

0.596 19 0.558

Pair
3

WIP_Test2

0.4645
WIP_Test3

0.2520

0.0564

0.5825

0.3465

8.242 19 0.000

53

(a) WIP vs. Total Clay Content

(b) WIP vs. OM

(c) WIP vs. Total Carbonate Content

Figure 3.7: Compositional controls on total porosity of mudrocks from SS1 (Eastern European Silurian mudrocks). (a) Total porosity as a function of total clay content (wt.%).
The samples are color-coded by their OM content (wt.%). The dotted line indicates general
trend and not a linear t. (b) Total porosity as function of OM content (wt.%), samples
color coded by their total clay content. (c) Total porosity as a function of total carbonate
content (wt.%)

54

the WIP porosity, however, the high clay content samples with low carbonate content show
constant range of WIP values. OM content (Figure 3.8(b)) does not show any independent
control on the total porosity in the Haynesville samples.
The Niobrara samples of SS3 have a bulk density range from 2.414 to 2.624 g/cm3 with
a mean of 2.500 (0.053 [2]) g/cm3 . The grain density ranges from 2.486 to 2.761 g/cm3
with a mean of 2.654 (0.070 [2]) g/cm3 . The total porosity in SS3 ranges from 4.81 to
12.74 %. There is no obvious correlation with the mineralogy and total porosity in SS3. All
the measured total porosity data for the three sample sets are provided in Appendix B.
3.5.3

Comparison of WIP with GRI measurements

To compare with commonly used methodologies, total porosity was obtained commercially by the GRI technique for nine SS1 samples, and 24 SS2 samples. Note that the sample
size used in GRI technique is 100300 g from a 15 cm stratigraphic interval, compared to
a 2-3 cm-thick WIP sample of 10 g taken from the same the 15 cm interval.
The total porosity and grain density obtained for SS1 by the WIP method are consistently higher than those for the comparable GRI samples except for one high OM sample
(Figure 3.9(a) and Figure 3.9(b)). Two distinct trends can be identied in the SS2 data. In
general, samples with lower OM content have slightly higher WIP porosity values compared
to GRI porosity, while samples with high OM have higher GRI porosities (Figure 3.10(a)).
The same is true of the grain density values(Figure 3.10(b)). Samples with low organic content have nearly the same grain densities obtained by both methods, but the samples with
highest OM have distinctly higher grain densities from the GRI technique.
3.5.4

Comparison of WIP with MIP measurements

Porosities using the MIP technique were obtained for the 22 samples in SS1, and from 13
samples from SS2. Comparison of measured porosity and grain density data obtained for the
same sampling interval shows that MIP porosity and grain densities are consistently lower
relative to WIP porosity for all comparable samples (Figure 3.11(a) and Figure 3.11(b)).

55

(a) WIP vs. Total Clay Content

(b) WIP vs. OM

(c) WIP vs. Total Carbonate Content

Figure 3.8: Compositional controls on total porosity of Haynesville samples SS2. (a) Total
porosity as a function of total clay content (wt.%). The samples are color-coded by their
total carbonate content (wt.%). (b) Total porosity as function of OM content (wt.%), samples color coded by their total carbonate content. (c) Total porosity as a function of total
carbonate content (wt.%).

56

(a) Comparison of Total Porosity

(b) Comparison of Grain Density

Figure 3.9: Comparison of (a) measured total porosity and (b) measured grain density of
the samples from SS1 by WIP and GRI techniques

(a) Comparison of Total Porosity

(b) Comparison of Grain Density

Figure 3.10: Comparison of (a) measured total porosity and (b) measured grain density of
the samples from SS2 by WIP and GRI techniques

57

The dierence in porosity values ranges from a minimum of 2.7 p.u. to a maximum of 4.9
p.u. for SS1 and from 2.4 p.u. to 7.1 p.u. for SS2.

(a) Comparison of Total Porosity

(b) Comparison of Grain Density

Figure 3.11: Comparison of (a) measured total porosity and (b) measured grain density of
the samples from SS1 (circles) and SS2 (triangles) by WIP and MIP techniques.

3.6

Discussion
The purpose of this study was to evaluate and optimize the WIP method for measur-

ing total porosity in thermally mature mudrock lithologies. Signicant advantages of the
WIP methodology over the GRI and MIP methods have been found. In contrast to the GRI
method, (a) all required measurements are conducted on the same portions of sample, (b) no
crushing is required and there is no preferential removal of material by sieving, and (c) only
1015 g of material are used. Moreover, gravimetric measurements are more direct and precise compared to gas based volumetric measurement techniques such as helium pycnometry.
Gas based volumetric measurement are prone to leaks and temperature uctuations in the
system and generally requires frequent accurate calibration of the reference volume. Uncontrolled exposure to external environments during GRI dehydrated condition measurements

58

can introduce signicant errors in grain density measurements (Figure 3.2). The water activity (or RH) environment during the saturation (100% water activity) and dehydration (0%
water activity) weight measurements are well controlled in the WIP methodology. Moreover,
the protocol included repeated measurements of each weight for quantication of the experimental uncertainty of each total porosity measurement. The WIP experimental procedure
is simple, quick to implement even with repeated measurements, which makes it applicable
for routine core measurements.
Reproducibility tests demonstrate the WIP technique is precise and reproducible. Optimization of the experimental methodology have allowed reduction of the absolute experimental uncertainty to be 0.20.3 p.u. at the combined expanded uncertainty with a coverage
factor of 2, signifying 95.45% condence interval in the total porosity calculation (Figure 3.5,
Table 3.3). An additional internal reproducibility test, to test the operation bias in the measurements, (Dr. Douglas K. McCarty and Dr. Timothy Fisher, personal communication)
was performed on ve randomly selected mudrock samples and measured for WIP by three
separate operators. The last step of the WIP measurement is weighing the saturated rock
chips submerged in water and in air. The saturated weight in air measurement requires
removal o surface water. Prolonged exposure to room conditions and aggressive brushing
could create the potential for operator bias in the measured weights. The operator biases in
the total porosity measurements are within the experimental uncertainties of the measurement and there no discernible trend in the data that can be attributed to a specic operator
(Figure 3.12). The average dierence in WIP measurements was 0.21 p.u. with a maximum
disparity of 0.5 p.u. (Figure 3.12). The reported standard error for GRI porosity measurements range from 0.5 p.u. (Karastathis, 2007; Luel et al., 1992; Passey et al., 2010) to 1.5
p.u. (Sondergeld et al., 2010; Spears et al., 2011). A part of the total error is due to lack of
standardized operating procedures for GRI measurements among commercial laboratories.
The accuracy of the WIP technique depends on the complete saturation of the samples.
The high penetration coecient of DI water (Table 3.1) makes it the best choice of satu-

59

Figure 3.12: Measured WIP data for each of ve randomly selected samples with propagated
expanded uncertainty with coverage factor of 2. The colors are operator-specic:Op.1 = blue,
Op.2 = green, Op.3 = peach. (Dr. Douglas K. McCarty and Dr. Timothy Fisher, personal
communication)
rating uid to achieve complete saturation, assuming that water is the wetting phase with
a zero contact angle. There is concern that water will not completely penetrate the organic
pore system, and there is uncertainty about the wetting behavior of kerogen. Passey et al.
(2010) speculated that the organic pore may be hydrocarbon wet, while Siskin et al. (1991),
Schimmelmann et al. (1999), and citations within, reported that at low temperatures, water hydrogen exchanges with organic hydrogen, most of which is bound to NSOfunctional
groups in OM.
Because of the abundant polar functional groups in kerogen at all levels of maturity, along
with its open and porous structure, there is no reason to expect that water would not ll its
pore system. Kerogen extracted from a SS2 sample using the method of Ibrahimov & Bissada
(2010) (courtesy: Dr. Adry Bissada, University of Houston), initially equilibrated at 75%
RH, desorbed 2.3 wt.% adsorbed water upon heating to 105 C and the readsorbed during re
equilibration to laboratory condition (room temperature 25 C and RH 50%) ( Figure 3.13),
60

which demonstrates that there are strong polar interactions with water molecules and the
kerogen functional groups and that this material is far from hydrophobic. Recent molecular
simulation studies of water in the pores of organic matter suggested pore lling by capillary
condensation occurs in pores as small as 1.2 nm (Hu et al., 2013).

Figure 3.13: Thermogravimetric analysis (TGA) data of isolated kerogen from Haynesville
samples (SS2). The kerogen was equilibrated at 75% RH before analysis. The sample is
heated to 105 C from room temperature at 5 Cmin1 rate under owing nitrogen gas and
then kept for 26 hours under laboratory conditions (50% RH).
Another possible problem with WIP is that there could be swelling of the bulk rock due
to clay mineral expandability or simple mechanical expansion during the saturation process.
Since gas shale formations are generally at a high thermal maturity and diagenesis level,
detrital smectite is usually transformed to illite (e.g., rodo et al., 2009). This, along with
a high level of compaction and cementation, minimizes the chances of swelling, which was
not observed during the experiments. The potential for expansion and incomplete saturation
will cause grain density to be underestimated compared to the actual values. Therefore, WIP
grain density can never be too high. The only scenario where WIP will over predict the grain
density is if the sample contained a signicant quantity of a water soluble naturally occurring
61

mineral such as halite, which could dissolve during saturation. Creation of microcracks
during sample handling is a possibility which may result in higher than actual total porosity,
but this would not aect the grain densities (Test 2 and Test 3, Figure 3.6). Even with the
creation of microcracks, the average deviation in total porosity from Test 2 and Test is about
0.47 p.u. with a maximum of 0.83 p.u.
To assess the accuracy of the WIP method, the results were compared with those obtained
by other measurement techniques (GRI and MIP). Note that the GRI and MIP techniques
have their own limitations and these should be considered while making the comparisons.
Comparison between WIP and MIP results indicate that MIP systematically underestimates
the porosity and grain density of the samples studied because MIP is limited to samples with
pore throat sizes greater than 3.6 nm. Most gas shale type rocks have pore throats sizes
<3.6 nm and the complete pore system cannot be accessed by MIP (Sigal, 2012). MIP
measurements performed using plugs and chips also indicate there is incomplete mercury
intrusion (Comisky et al., 2011; Sigal, 2012). Application of MIP on crushed powder tends
to give systematically higher porosity than that measured on plugs and chips because of
improved pore accessibility (Comisky et al., 2011).
Comparison of GRI and WIP total porosity indicates that WIP porosity often has higher
values than GRI porosity for low organic content rocks (<3.0 wt.%), but an opposite trend
in high organic content samples. The WIP grain density for low-organic content rocks
from SS1 is higher than GRI grain density by 0.05 g/cm3 (Figure 3.9(b)). These samples
have a trace of halite (<2 wt.%) which cannot explain the 0.05 g/cm3 dierence in grain
density. This dierence is probably due to the uncertainty in GRI measurements from an
unknown quantity of water molecules which may not be removed or were adsorbed before
injection of the helium measurement gas. In SS2, there is a good agreement between WIP
and GRI grain densities for low organic content samples (Figure 3.10(b)). However, in the
high OM samples, the GRI technique systematically measured higher grain density. The
most likely cause, however, is that the solvent extraction treatment used in the Dean Stark

62

procedure not only removes volatile hydrocarbons (liquid oil) and but also a portion of solid
bitumen that forms the rock matrix. This increases total porosity and skews the bitumen
containing samples to higher grain densities. Previous studies comparing thermal distillation
(pretreatment by heating) and solvent extraction suggest that these two techniques extract
dierent fractions of organic material (Cooles et al., 1986; Wilhelms et al., 1991; Noble et al.,
1997). Comparisons between RockEval pyrograms of pre and postDean stark extraction
portions from the same samples indicate that there is a signicant loss of the S2 component
(Figure 3.14). The dierence in the S2 pyrogram patterns at temperatures > 400 C after
Dean Stark indicates that solid bitumen is dissolved in the process. Based on the proportion
of S2 recorded in the natural sample and the samples after Dean Stark procedure, the
OM fraction (solid bitumen) removed could account for 0.5 p.u. of extra total porosity,
assuming a bitumen density is estimated at 1.1 g/cm3 . Compared to Dean Stark solvent
extraction, the thermal distillation by preheating at 200 C process does not alter the solid
OM in rock, based on consistent RockEval II S2 and S3 values before and after preheating
treatment on thirty samples from Baltic Basin (Figure 3.15, Dr. Arkadiusz Derkowski and
Dr. Douglas McCarty, personal communication). Removal of this low density bitumen
results in systematically higher measured grain densities in Dean Stark extracted samples,
as in GRI technique, compared to thermally treated samples at 200 C such as in the WIP
procedure. Other possible causes may be:
1. There could be incomplete saturation of the high organic content rocks if water did
not access all the organic pores, resulting in an under-prediction of WIP grain densities. However, the water adsorption experiment (Figure 3.13) indicates that achieving
complete saturation in organic pores will not be a problem. Comparing WIP results
with other hydrocarbon based saturating uids such as kerosene and Isopar could
provide additional useful information .
2. The crushing of samples and subsequent sieving in the GRI procedure may selectively
remove the lighter, lower density, and ner-grained material thus introducing a bias
63

toward the higher density components. The organic materials in the mudrocks tend
to grind into smaller grains during crushing and may be discarded during the sieving. Additional experiments investigating whether sampling bias caused by the sieving
methodology or not are recommended.

Figure 3.14: Comparison of RockEvalII S2 pyrogram between a natural mudrock and its
equivalent Dean Stark extracted portion. The Dean Stark extraction method aects the
organic matter and removes the solid bitumen (low temperature peak/shoulder) and some
parts of pyrolysable kerogen compared to the natural rock.

3.7

Summary
A modied methodology using a DI water immersion porosimetry (WIP) technique was

adapted to measure total porosity in mudrocks. The results suggest that:

1. Measured total porosity, grain density, and bulk density obtained by the WIP technique
are precise and reproducible.
64

(a) Comparison of RockEval II S2

(b) Comparison of RockEval II S3

Figure 3.15: Comparison of RockEval II (a) S2 and (b) S3 parameters measured before and
after preheating at 200 C on thirty samples from Baltic Basin. (Dr. Arkadiusz Derkowski
and Dr. Douglas McCarty, personal communication)

2. The absolute experimental uncertainty in the WIP porosity measurement is about

0.20.3 p.u. This is lower than the reported 0.5 p.u. uncertainty in total porosity
measurement by the industry adopted GRI technique.
3. Potential limitations include possible incomplete saturation, water soluble mineral dissolution, and sample swelling. The possible inuence of these factors on the measured
total porosity of the samples is considered small.
4. Comparison of results from WIP and MIP measurements indicates that MIP systematically underestimates the porosity and grain density, because a signicant pore volume
in these mudrocks have minimum access pore throat diameters below the MIP measurement limit of 3.6 nm.
5. The Dean Stark pretreatment in the GRI method dissolves solid bitumen and can
create porosity artefacts and increase the grain density compared to WIP.

65

6. Despite possible limitations, the WIP technique is a more reliable measurement for
mudrock total porosity than the GRI and MIP methods. The simplicity of WIP
methodology makes it especially attractive for routine mudrock core total porosity
measurements.

66

CHAPTER 4
N2 GAS ADSORPTION TECHNIQUE FOR MEASURING PORESTRUCTURE
PARAMETERS IN MUDROCKS.

We never really see time. We see only clocks. If you say this object
moves, what you really mean is that this object is here when the hand of
your clock is here, and so on. We say we measure time with clocks, but
we see only the hands of the clocks, not time itself. And the hands of a
clock are a physical variable like any other. So in a sense we cheat
because what we really observe are physical variables as a function of
other physical variables, but we represent that as if everything is evolving
in time
Carlo Rovelli

4.1

Introduction
The pore-structure of a porous media inuences most behavior of porous media, for ex-

ample, elastic and mechanical behavior, movement and ow of uids etc. Understanding and
quantication of pore-structure parameters is crucial in modeling behavior of porous media.
In the previous chapter, measurement of an important pore-structure parameter namely, total porosity for mudrock lithologies was discussed. However, total porosity is a single-value
quantication (Nimmo, 2004) and does not reveal any insight in the textural attributes
and morphology of the pore-structure. The poresize distribution (PSD) of a porous media
quanties the relative pore volumes associated with the dierent poresizes. Understanding
of their poresize distributions is critical in estimating transport and storage properties of
unconventional mudrock reservoirs. Mudrocks exhibit multiple scale porestructures that
are more complex than that of conventional reservoir rocks. Advanced imaging techniques

67

reveal a nanometer scale pore structure in the clay and OM components within such mudrocks (for example, Javadpour, 2009; Loucks et al., 2009; Passey et al., 2010; Milliken &
Reed, 2010; Curtis et al., 2011; Ambrose et al., 2012; Curtis et al., 2012; Milliken et al.,
2013). With such small pores, the assumption of standard continuum approach falls apart
and at those nanometer scales, the physics of transport mechanism change.
Another important pore-structure parameter describing the porous media is surface area
and is expressed as surface area per unit mass of sample or specic surface area. The
external sample specic surface areas are negligible and insignicant in comparison to that
on the internal pore wall (Gregg & Sing, 1983). Smaller pores have higher specic surface
areas and nanoporous materials typically possess large specic surface areas,for example,
Vycor glass 7930, an open-cell porous glass with average pore diameter of about 4 7 nm,
has an specic surface area of 250 m2 /g (Levitz et al., 1991). Natural gas is adsorbed on
these internal surfaces contributing to large gas storage capacity of these mudrocks in spite
of their low porosity.
Most of the pore space characterization studies on mudrocks are based on imaging techniques, which are qualitative visual descriptions of mudrock porosity. These imaging techniques cannot resolve pores smaller than 5 nm and thus cannot see a signicant portion of
the pore structure (Chalmers et al., 2012). Furthermore, the scale of observation is generally limited to about 1500 m2 unless multiple images are taken into account and thus the
imaged pore structure may not be representative of the rock due to millimeter scale stratigraphic heterogeneity (Chalmers et al., 2012). However, by combining multiple images, a
larger area can be studied. Imaging techniques are useful for qualitative visual porosity and
fabric description, but additional eort is required to provide indirect quantication of of
total porosity and pore-size distribution (for example Curtis et al., 2012; Milliken et al.,
2013).
PSD and specic surface area can be quantied using dierent techniques; the two most
popular and widely used methods are mercury intrusion porosimetry and N2 gas adsorp-

68

tion. Historically, mercury intrusion porosimetry (MIP) has been the preferred technique
for pore-size analysis of macroporous conventional reservoirs. The limitations of MIP technique for nanoporous mudrock porestructure characterization was discussed in Section 3.2,
paragraph 3 (page 34). Gas adsorption techniques are widely used in the soil, chemical,
ceramic, and pharmaceutical industries to characterize the pore structures and surface areas of powders and ne particles. Subcritical N2 gas adsorption techniques can be used to
investigate ne pores in the range of 1.7200 nm that are prevalent in mudrocks and coals.
To date, this method has rarely been used in the petroleum industry to characterize rock
pore structure. Ross & Bustin (2009) studied the pore structure of DevonianMississippian
and Jurassic mudrocks from western Canadian sedimentary basins using both subcritical gas
adsorption and MIP. Adesida (2011), Chalmers et al. (2012), Clarkson et al. (2012a,b) and
Clarkson et al. (2013) demonstrated usefulness of gas adsorption techniques to evaluate the
quantitative porestructure parameters of mudrocks from dierent gas shale plays in North
America. Application of gas adsorption methods in the characterization of unconventional
reservoirs will undoubtedly become more widespread in the energy industry. In this thesis,
the N2 gas adsorption is used as the principle investigative tool for PSD and specic surface
area quantication and the results N2 gas adsorption are compared with conventional MIP
techniques or some samples. In this chapter, the experimental procedure of N2 gas adsorption measurement techniques, interpretation of raw isotherm data and theoretical concepts
for inverting the raw data to porestructure parameters are discussed in details.
4.2

Theory of N2 Gas Adsorption Technique

Adsorption can be dened as the tendency of one component of the system to have a

higher concentration at the interface than it has in either of the adjacent bulk phases" (Barnes
& Gentle, 2005). In the case of solid-uid system, there is a increase in the density of the
uid happens near the soliduid interface. Depending upon whether the force of interaction
between the uid molecules and the surface layer is chemical or physical; adsorption can be
of two types: physisorption or chemisorption.
69

Physisorption experiments are done with dierent gases and at dierent pressure-temperature
conditions. However, lower temperatures and higher pressures will make adsorption more
prominent and easier. In low pressure adsorption experiments the temperaturepressure
regime is below the critical point of the uid used. These experiments yield valuable information about the textural properties of porous material, such as surface area and porestructure. Since the gas is below its critical point, capillary condensation becomes important
in these experiments which give us the information of poresizes. N2 (at 77K) is the most
commonly used gas for surface area and mesopore characterization, however, alternative
gases can also be used, such as krypton (at 77K), argon (at 87K), carbon dioxide (at 273 K).
Adsorption studies conducted at higher pressures, where the gas is in supercritical phase,
give information about the nature of surface and its adsorption potential of gases. These
experiments have applications in natural gas production from shale reservoirs, fuel cells
storage, gas purication and separation process, adsorption heat pumps, thermotransformers,
refrigerating instruments etc. These experiments provide information about the surface area
and surface properties but generally do not give information about the pore-sizes as the gas
never condenses during the experiments.
In this thesis, subcritical N2 gas adsorption at liquid nitrogen (LN2 ) temperature and subatmospheric pressure is used for characterizing surface area and pore-structure of mudrocks.
The following sections will discuss the measurement procedure, qualitative interpretation
of raw experimental data and reduction of data to pore-structure parameters like specic
surface area and PSD.
4.2.1

Experimental Procedure

Subcritical N2 gas adsorption experiments are performed using various procedures. In

this study, the amount of gas adsorbed is determined by a discontinuous static volumetric
method. The experiments were conducted on a Micromeritics ASAP 2020 instrument.
Sample preparation was carried out in accordance with I.S.O 9277:2010(E). One to three
grams of sample was degassed at 200 C for 24 hours under vacuum (<10 m Hg) prior to
70

the analysis. The sample is kept under heat and vacuum until the sample outgassing rate is
< 0.005 Torr/min over a 15 minute interval (usually 1624 hours). The heating temperature
is consistent with pre-treatment conditions used in the WIP experiment (Section 3.3.2). This
ensures removal of any adsorbed claybound and capillary water, while avoiding irreversible
damage to their mineral and solid organic matter structure and sample texture. After the
degas procedure, the tube and sample are weighed to determine the analysis weight of the
sample. The sample tube is then placed on the analysis port of the instrument and the
adsorption/desorption isotherms are collected.
In the analysis port, at rst the free space (dead volume of the tube) is measured volumetrically using helium (reagent grade 99.99% pure) before measurement of the adsorption
isotherm. The sample tube, after evacuating the helium, is kept in a cryogenic liquid nitrogen (LN2 ) dewar and dosed with a known amount of N2 at a series of precisely controlled
pressures. The molar quantity of nitrogen dosed to the sample is calculated from the pressure
and temperature measurement, using the real gas equation of state, in a pre-calibrated manifold volume. The gas is allowed to equilibrate with the sample while pressure is monitored
continuously. When the pressure change per equilibration time interval (rst derivative) is
less than 0.01% of the average pressure during the interval, the equilibrium pressure (P)
and temperature are recorded. The molar quantity of gas adsorbed is calculated by mass
balance between the dosed amount and the molar quantity of gas in the tube after equilibration. The saturation vapor pressure (P 0 ) of N2 at LN2 temperature is determined every
two hours during the experiment using a vapor pressure thermometer. The quantity of adsorbed gas on the solid surface is measured at discrete pressure (P) steps over the relative
equilibrium pressure (P/P 0 ) range of 0.0075 to 0.995 at constant temperature. The experiment systematically increases pressure up to the condensation pressure (adsorption branch)
followed by reduction of pressure from P 0 (desorption branch) and the data are reported as
the adsorption isotherm: quantity of gas adsorbed per mass expressed as moles or volume
in cm3 /g (S.T.P.) as a function of relative equilibrium pressure (P/P 0 ). Additional details

71

regarding analytical procedures are found in Webb & Orr (1997).

4.2.2

Qualitative Interpretation of N2 Isotherm Profile

The shape of the isotherm and its hysteresis pattern provide useful information about the
physisorption mechanism, the solid and gas interactions, and can be used to qualitatively
predict the types of pores present in the adsorbent. IUPAC (Sing et al., 1985) classied
the adsorption isotherms into six types (Type I to VI), along with four hysteresis pattern
types (H1 to H4). The dierent hysteresis patterns H1 to H4 are characteristic of dierent
mesopore shapes . A detailed description of the IUPAC isotherm classication is presented
in Sing et al. (1985) and Rouquerol et al. (1998). Three isotherm types, especially important
for unconventional gas and oil formations (Figure 4.1) are described here (adapted from Sing
et al. (1985) and Rouquerol et al. (1998)).
Type I isotherm A purely microporous adsorbent will exhibit a concave-shaped isotherm
with very high adsorption at low relative pressure (P/P 0 <0.01) before it reaches a
plateau (Figure 4.1(a)).The narrow micropores result in overlapping adsorption potential elds of opposite walls. This overlapping potential energy creates an enhanced
absorbent-adsorbate interaction resulting in high adsorption uptake at low relative
pressures. This phenomenon is called micropore lling. The limiting uptake (plateau)
depends on the cumulative accessible micropore volume present in the sample.
Type II isotherm A non-porous or a macroporous material will exhibit a completely reversible isotherm (Figure 4.1(b)) where the adsorption and desorption follow exactly
the same path. These isotherms initially have a concave shape at lower relative pressures (P/P 0 < 0.2), followed by a linear region and nally a convex shape near to the
P/P 0 axis at higher relative pressure (P/P 0 > 0.4). This isotherm prole represents
a monolayermultilayer adsorption mechanism of the gas on the open and stable solid
surface. The concave part at low relative pressure indicates a relatively high heat of
adsorption resulting in molecules adsorbed into a layer one molecule thick (i.e., mono72

layer coverage). The sharp inection point at the transition between the concave and
the linear part of the isotherm is considered to represent the completion of monolayer
coverage and beginning of multilayer coverage. The amount adsorbed at this point is
directly related to the surface area of the solid.
Type IV isotherm A dominantly mesoporous material will have a characteristic hysteresis
loop (Figure 4.1(c), which is associated with capillary condensation and evaporation
taking place in mesopores. At lower relative pressure (P/P 0 < 0.4) the isotherm prole
is similar to the Type II isotherms indicating a monolayer to multilayer adsorption
mechanism on the walls of meso and macropores. At higher relative pressures,
the gas condenses in the mesopores to bulk liquid at pressures below the saturation
vapor pressure with the formation of a gasliquid meniscus (capillary condensation).
The isotherms prole exhibit a plateau with reduced adsorption at high P/P 0 . The
limiting adsorption plateau at high P/P 0 0 indicates complete lling of mesopores and
subsequent adsorption on the external surfaces.

(a) Type I isotherms

(b) Type II isotherm

(c) Type IVIsotherM

Figure 4.1: Typical isotherm shape exhibited by (a) purely microporous material (Type I
isotherm prole) (b) non-porous and macroporous material (Type II isotherm prole) and (c)
purely mesoporous materials (Type IV isotherm prole). Modied from Sing et al. (1985).

The hysteresis loop in the Type IV isotherms is characteristic of mesopores. There are
multiple models to explain the hysteresis phenomena in adsorption isotherms (Naumov,
73

2009). Nonconnecting cylindrical mesopores can show hysteresis due to dierences in the
meniscus shapes during the capillary condensation (adsorption) and evaporation (desorption)
process. Capillary condensation is preceded by a metastable uid state with a cylindrical
meniscus while capillary evaporation occurs with a hemispherical meniscus resulting into
lower equilibrium pressures P of the phase transition for the same pore size (Groen et al.,
2003). Such phenomenon generate hysteresis prole in mesoporous materials with non
connecting cylindrical pore systems like mesoporous silica SBA15 (Naumov, 2009). Other
mechanisms that contribute to hysteresis include pore network and pore connectivity eects
(Conner et al., 1986; Mason, 1982; Naumov, 2009). In such interconnected porenetworks
system, the path dependence is due to varying sequence in which the probe molecule encounters the pores during the adsorption and desorption process, resulting in hysteresis.
The characteristic H1 to H4 hysteresis patterns (Figure 4.2) result from dierent mesopore
shapes (Sing et al., 1985). The H1 hysteresis loop (Figure 4.2(a)) types are mostly seen in
materials with a narrow distribution of cylindrical or tubular pores. The H2 (Figure 4.2(b))
types indicate a complex, interconnected pore structure with narrow pore openings. The H3
hysteresis loop (Figure 4.2(c))types are mostly found in materials with platy particles having
slit-shaped pores. The H4 hysteresis loop (Figure 4.2(d))types also result from slit-shaped
pores, but are associated with micropores in general.
A number of isotherms do not readily t into any of the above classications. These
include isotherms with characteristics of more than one Type, such as the mixed Type I/IV
and mixed Type II/IV. There is a special type of isotherm shape, classied as Type IIB
(Rouquerol et al., 1998), which is characteristic of mudrocks (Figure 4.3). The adsorption
branch of the isotherm has a general shape like Type II isotherms, but with a distinct H3
type hysteresis loop. As mentioned above, hysteresis indicates the presence of mesopores.
These materials are not purely mesoporous as there is no indication of the completion of
mesopore lling that would result in a plateau at higher relative pressures as in a typical Type
IV isotherm. A signicant volume of macropores in these materials results in the absence of

74

(a) H1

(b) H2

(c) H3

(d) H4

Figure 4.2: The four hysteresis shapes of adsorption isotherm usually found by subcritical
N2 adsorption. Modied from Sing et al. (1985).

the isotherm plateau at high P/P 0 relative pressures as seen in Type IV isotherms.
Another important feature observed in many hysteresis patterns is the forced closure of
the desorption branch where the isotherm closes at P/P 0 relative pressures around 0.41
0.48 for N2 isotherms (Figure 4.3). This phenomenon is attributed to a process called the
Tensile Strength Eect (TSE) (Gregg & Sing, 1983). Disappearance of the hysteresis is
due to the instability of the hemispherical meniscus during capillary evaporation in pores
with diameters smaller than approximately 4 nm (Groen et al., 2003). In these pores, the
surface tension forces are stronger than the tensile strength of the liquid causing the meniscus
collapse which leads to a spontaneous evaporation of the bulk liquid phase. The presence
of forced closure in the isotherm shape may indicate a signicant volume of pores with
diameters smaller than 4 nm.
4.2.3

Specific Surface Area Determination

In mesopores and macropores, monolayer-multilayer adsorption onto the pore walls occurs
initially in the approximate P/P 0 range of 0.050.30. Generally a sharp bend knee is observed
which indicates the completion of single molecular layer (monolayer) coverage of adsorbate
molecules on the available pore surface. The specic surface area of the material is directly

75

Figure 4.3: N2 adsorption isotherms at 77K of a representative mudrock sample from Haynesville (Sample SS21). Inset gure is a detailed view showing the isotherm forced closure
in the relative pressure (P/P 0 ) range 0.350.55 due to tensile strength eect.
related to the monolayer capacity, i.e., the amount (moles or cm3 S.T.P.) of N2 required
for monolayer coverage of the total surface area, including external surface, and meso and
macropore surface . The monolayer capacity (Vm ) is converted into specic surface area
by multiplying it with the average molecular cross-sectional area(0.162 nm2 , as suggested by
I.S.O 9277:2010(E)) of N2 at -197.3 C and the Avogadro constant (number of molecules in
one mole, 6.023E+23).
Vm am L 1018
SSA =
22414

(4.1)

where Vm is the monolayer volume expressed in cm3 (STP)/g , am is the cross sectional area
of the nitrogen molecule in nm2 , L is Avogadros number and 22414 cm3 /g is the molar
volume of one mole of an ideal gas at Standard Temperature and Pressure (S.T.P). and SSA
is the specic surface area in m2 /g.
The most common way to obtain the monolayer capacity is to invert the adsorption
isotherm using the BET (named after initial of the proponents of the theory Brunauer, Emmett, and Teller) theory. Brunauer et al. (1938) extended the Langmuir model of monolayer

76

adsorption to multilayer adsorption and developed the following model with some assumptions about the nature of interaction between solid and gas :
P/P0
1
C 1
=
+
(P/P0 )
V (1 P/P0 ) Vm C Vm C

(4.2)

where V is quantity adsorbed in cm3 (STP)/g and C is a constant related to the heat of
adsorption of the gas on the solid. The adsorption isotherm when plotted in BET transform
P/P0
C 1
plot of
versus P/P 0 gives a straight line with slope s =
and intercept
V (1 P/P0 )
Vm C
1
1
. Solving these two equations will give us the monolayer capacity Vm =
and
i=
Vm C
s+i
s
the BET constant C = + 1.
i
Brunauer et al. (1938) showed that the transform is linear with a positive intercept over
the approximate P/P 0 range of 0.05 to 0.30 for a variety of dierent materials. However,
this range is debated and depends on the type of material (Gregg & Sing, 1983; Rouquerol
et al., 1998). BET theory works best for nonporous, mesoporous, and macroporous materials
and is theoretically not applicable for microporous material (with < 2 nm pores) because of
the occurrence of micropore lling instead of multilayer adsorption. However it is dicult
to precisely distinguish between the regions of micropore lling and monolayermultilayer
adsorption.
A modied BET approach was suggested by Rouquerol et al. (2007) to evaluate surface
areas for materials containing micropores. They suggested an objective way to nd the linear
tting P/P 0 range using the following to criteria:
P/P0
1. The tted linear trend should have a positive intercept in the yaxis
V (1 P/P0 )
of BET transform plot. A negative intercept will result in a negative value of C which
!

is non-physical.
2. The tted P/P 0 range should show a continuously increasing trend in the transform
plot of V (P0 P) vs. P/P 0 (also known as Rouquerol transform plot, Figure 4.4) and
the upper limit of the tting range should be selected at local maxima of the V (P0 P)
vs. P/P 0 plot.
77

The lower limit of the tting range is selected to maximize the R2 coecient of t
(minimum accepted R2 0.9999). Another check of consistency, recommended by Rouquerol
et al. (2007), states that the BET monolayer capacity Vm should lie in between quantity
adsorbed (V ) values at the limits of tted relative pressure (Figure 4.4). However, the
authors pointed out these values should be treated as equivalent specic surface area, not as
the absolute true surface area, because the theoretical basis of BET theory and monolayer
capacity does not honour nitrogen adsorption behavior in micropores. The advantage of this
method is that reduces any subjectivity in the assessment of the tting range of the BET
plot and gives an objective way to determine the tting range of BET transform plot rather
than using same relative pressure range for all materials. The modied BET technique has
been adopted in I.S.O 9277:2010(E) and was followed in this study. It is done using the
Micromeritics MicroActive DataMasterv5.00 software from Micromeritics. The dierence
in measured specic surface area using the modied Rouquerol approach and traditional
tting range of 0.05-0.30 P/P 0 was most prominent for micropore rich mudrocks and varied
by a maximum of 8 relative percent (absolute 2 m2 /g) for the studied samples. The
equivalent specic surface area obtained will be henceforth referred as eSSA.
4.2.4

Total Specific Micropore Volume Determination: tplot Technique

tplot (Lippens & De Boer, 1965) is a widely used technique to estimate the specic
micropore volume and the open surface area, which is the surface area from mesopores,
macropores, and external surface area. In a tplot, the adsorbed N2 volume (V a) is plotted
against statistical thickness (t) of the adsorbed layer of N2 . The thickness (t) depends on the
relative pressure (P/P 0 ). The thickness curve should be ideally obtained from N2 adsorption
on a reference solid with the same chemical and surface properties but without pores, which
is dicult to obtain for mudrocks, because of the compositional heterogeneity and presence
of clays and OM. Several thickness equations were proposed in the literature as representative universal thickness curve of which the most generally applicable and most frequently
employed are those of HarkinsJura, Halsey and Broekho-de Boer; each developed for a
78

Figure 4.4: Modied BET technique following Rouquerol et al. (2007) applied on isotherm
data from sample SS21 (Figure 4.3). The red dots and the dashed lines shows the P/P 0
tting range and the tted data points for the BET linear trend, respectively.

79

specic type of material using N2 as the adsorptive and the constant values are empirical
(see summary in Webb & Orr, 1997). The other commonly used thickness equations are
the Kruk-Jaroniec-Sayari and Carbon STSA equation. The choice of the universal tcurve
for mudrocks applications, used in this thesis, is discussed later in Section 4.3.2. Figure 4.5
shows idealized examples of tplot for a microporous solid and for a sample free of micropores (Webb & Orr, 1997). If the V a vs. t plot yields a straight line that passes through the
origin, then the sample is considered to be free of micropores. tplot of microporous material
shows a straight line at medium t and a concavedown curve at low t. Extrapolation of the
linear region to the V a axis gives the specic micropore volume and the slope of the linear t
gives the open surface area. At higher t, convex-up deviation from the linear trend indicates
capillary condensation in mesopores.
4.2.5

Total Specific Pore Volume Determination

The total specic pore volume for N2 gas adsorption technique was obtained from general principle of Gurevich rule (Gregg & Sing, 1983; Rouquerol et al., 1998). The specic
pore volume is obtained by liquid molar volume adsorbed at a predetermined P/P 0 . The
liquid molar volume adsorbed at the relative pressure (P/P 0 ) of 0.990 (either measured or
interpolated) is calculated by converting the adsorbed gas volume to liquid molar volume
using N2 density of 0.808 g/mL at liquid nitrogen temperature (77K). A relative pressure
of 0.99 corresponds to a pore-size (diameter) of 193.5 nm and hence the total specic pore
volume indicated is the specic pore volume of pores smaller than 193.5 nm diameter.
4.2.6

PoreSize Distribution Analysis

In a N2 adsorption experiment, capillary condensation occurs after multilayer adsorption

from a vapor where the pore spaces are lled with bulk liquid separated from the gas phase by
a meniscus. The gas in the pore condenses at pressures below the saturating vapor pressure
of the gas (P 0 ). The gas phase in pores with dierent diameters will condense at dierent
pressures and thus provide quantitative information about the relative volume of dierent

80

Figure 4.5: Schematic example of tplot for two samples with and without micropores (modied from Webb & Orr, 1997). An example of tplot obtained from isotherm data of sample
SS21 are chosen as an example for microporous material. A sample free of micropores
(referred as No micropores) will have a linear trend that passes through the origin. Microporous samples shows a concave-down curve at low thickness (t) values indicating micropore
lling, followed by a straight linear part indicating multilayer adsorption. The intercept on
the yaxis gives the specic micropore volume. The deviation at higher t is indicative of
capillary condensation.

81

poresizes in the sample.

The poresize distribution is obtained by application of the BJH (Barret, Joyner, and
Halenda) technique, assuming cylindrical, non-connecting pore geometry. The Kelvin equation describes the eect of surface curvature of the liquidvapor meniscus on the vapor
pressure and relates the pore diameter with the relative pressure (P/P 0 ):
RT ln (P/P0 ) =

2VL
rm

(4.3)

where P is the vapor pressure over curved liquidvapor meniscus surface with an eective
radius of rm , VL is the molar volume of the liquid absorptive, is the liquid surface tension.
P 0 is also the saturation vapor pressure as it corresponds to the vapor pressure where rm =
, R is the universal gas constant and T is the temperature. The main assumption of
this equation is VL is independent of the pressure, i.e. that the liquid is incompressible.
Equation 4.3 suggests that the vapor pressure P over a concave meniscus will be less than
the saturation pressure P 0 . The capillary condensation of a vapor to a liquid should occur
in a pore at some pressure P determined by the value of rm for that pore and will be
always less that the saturation pressure (provided the meniscus is concave). The smaller
the pore, the lower the condensation pressure. For our application of N2 adsorption at LN2
temperature , the following values are used for surface tension and molar volume of liquid
nitrogen (Gregg & Sing, 1983): = 8.85dynecm1 ; VL = 34.71cm3 mol1 . Substituting these
values in Equation 4.3, the equation reduces to:
1
2 (8.85 dyne cm1 ) (34.71 cm3 mol1 )
9.55339 108
=

cm
(8.314 107 dyne cm K1 mol1 ) (77.35 K) ln P/P0
ln P/P0
0.955339
nm
(4.4)
rm =
ln P/P0

rm =

where rm is the mean radius of curvature of the meniscus. Note that this rm is not equal to
the pore radius. During a physisorption experiment, capillary condensation takes place after
the pore walls are coated with multiple layers of adsorbed gas molecules. The thickness of
the adsorbed layer, t, will depend upon P/P 0 . Thus, the capillary condensation does not
82

occur against solid pore walls and rm ,the mean radius of curvature of the meniscus (or the
Kelvin radius), has to be correlated with the pore radius rp (Figure 4.6). The conversion of
rm to rp involves some assumptions, particularly about the pore-shape and the contact angle
, between the capillary condensate (bulk liquid) and the adsorbed layer. If the thickness
of the adsorbed layer is t and the contact angle between the adsorbed lm and the bulk
condensate has a nite value , then the relationship becomes:
rp = rm cos + t

(4.5)

Figure 4.6: Schematic representation of the relationship between Kelvin radius, rm , the core
radius rk and the pore radius rp of a cylindrical pore with a hemispherical meniscus.
The contact angle between a liquid and a solid pore wall is assumed to be equal to
zero for application purpose. However, this assumption is highly debated (see review in
Gregg & Sing, 1983). can have any value between 0 to 90 . In the physisorption process,
the contact is between the adsorbed lm and the bulk condensate. The adsorption eect
is limited to a distance of 34 molecular diameters from the surface of pore wall. Beyond
that distance the adsorbed uid has properties of bulk liquid. So at higher pressure (near
saturation pressure), where the thickness of adsorbed layer is large and the adsorbed lm
has bulk-liquid like properties, the contact angle between the adsorbed layer and the bulk
83

condensate will become zero. At lower P/P 0 and smaller pores, thinner lm thickness will
exhibit dierent structure than bulk liquid due to extended inuence of the potential of the
solid surface and would therefore have a nite angle of contact with it. Lowell & Shields
(1991) advocated that for cos (for 20 ) 0.94, the error introduced by this assumption
of zero contact angles will be small and negligible when compared to errors introduced by
assuming simple pore shape models.
Barrett et al. (1951) suggested a computational method of applying the Kelvins equation and accounting for thinning of adsorbed layer to get the poresize distribution from
adsorption isotherm data and is referred as the BJH method named after the initial of the
proponents (Barrett, Joyner and Halenda). Generally one of the universal thickness curve,
discussed in Section 4.2.4, are used unless an experimental t-curve is generated for the samples. The eect of using dierent t-curve on inverted PSD is discussed later (Section 4.3.2).
Even though, theoretically the BJH method is a desorptive method (it works from higher
P/P 0 to lower P/P 0 to model imaginary emptying of condensed adsorptive), it can be
used for both the adsorption isotherm and the desorption curve. However, in the case of a
broad distribution of pore sizes the desorption branch is particularly sensitive to the tensile
strength eect (TSE) phenomenon of the adsorbed phase shown in Figure 4.3. PSD obtained
from desorption curve would be limited to 45 nm, because TSE phenomenon limits the
applicability of Kelvins equation beyond a critical diameter of 4 nm (Groen et al., 2003).
The limit of pore diameter that can be evaluated using BJH theory from the adsorption
branch is 1.7 nm because the Kelvins equation becomes invalid in micropores. The choice
of adsorption branch and desorption for PSD inversion was discussed later in Section 4.4.
The reduction of data using BJH theory was done using the Micromeritics MicroActive
DataMasterv5.00 software.
Theoretically there is no upper limit to the measurable pore diameter using capillary
condensation techniques. However the upper limit of pore-size measurement of about 200
nm is due to the practical experimental constraints of maintaining very high P/P 0 > 0.99

84

throughout the system (Gregg & Sing, 1983). Gregg & Sing (1983) showed that a P/P 0
corresponding to a pore-sizes of 2 micron and 10 micron are 0.9990 and 0.9998 respectively
and such P/P 0 variations are dicult to monitor and thus pores in in this size range that
is become practically impossible to dierentiate. Moreover, maintaining a precise relative
pressure so close to unity is impractical and extremely sensitive to slight changes in temperatures. Hence, for all practical purpose, the upper limit of reliable measurement using N2
adsorption technique is around P/P 0 of 0.99 which corresponds to a pore size of 193.5 nm.
The Kelvins equation and BJH theory are classical thermodynamic approaches and is
based on simplied assumptions. Over the last two decades, dierent molecular simulation
approaches like Grand Canonical Monte Carlo simulation (GCMC) and molecular dynamics
are used to model the distribution of gas molecules in a system in equilibrium based on the
microscopic properties of the system, such as the uid-uid and uid-solid interaction energy
parameters, the pore size, the pore geometry, as a function of temperature. Density functional theory (DFT) is a practical alternative to these techniques as it is computationally
less intensive compared to the previous methods. DFT mathematical formalism is presented
in Webb & Orr (1997) and the citations within. DFT modeling results in a realistic equilibrium density distribution for the conned uid as a function of temperature and pressure
(Figure 4.7(a)). Therefore, a model isotherm (quantity adsorbed versus pressure) can be
generated while accounting for solid and gas interactions along with geometrical conguration of pore walls (Figure 4.7(b)). These models are calibrated against real isotherm data
of non-porous material and real behavior of uids. Several such models accounting for interactions for a particular pair of solidgas in particular nite pore geometry are proposed,
for example N2 on carbon surface with cylindrical pore, N2 on oxide surface with cylindrical
pores, N2 on pillared clays, N2 on carbon slit pore. The implications of using one model
over the other from PSD inversion of mudrock samples will be discussed later.
For pore size analysis using DFT models, the experimental isotherm is assumed to be
a combination of isotherm behavior in each pore with dierent sizes (Figure 4.7(c)). Each

85

(a)

(b)

(c)

Figure 4.7: Schematic diagram of constructing a DFT kernel (images from MicroActive
DataMaster v5.00 software manual). An example model of Ar on Argon on Carbon at
87.3 K is chosen for illustration. (a) DFT approach used to model the realistic equilibrium
density distribution for the conned uid, Ar on carbon surface at P/P 0 =0.5 and T =
87.3K; (b) The model isotherm (quantity adsorbed versus pressure) generated for a slit pore
of 4 nm; (c) Kernel q(p, H) or collection of model isotherms with dierent pore sizes (H).
This kernel q(p, H) is used to deconvolve the experimental isotherm to obtain distribution
function of pore size H.

pore size present contributes to the total isotherm in proportional to the fraction of the total
area of the sample that it represents. Mathematically, this relation is expressed by
Q(p) =

dH q(p, H) f (H)

(4.6)

where Q(p) is the experimental isotherm , q(p, H) is the quantity adsorbed per unit area
at pressure p in an ideal pore of size H, and f (H) is the total area of pores of size H in
the sample. The q(p, H) is numerically obtained from the assumed DFT model and can
be thought of a kernel or collection of several modeled individual pore isotherms. The distribution f (H) is numerically solved using a deconvolution algorithm. The Micromeritics
MicroActive DataMaster v5.00 software was used for the DFT analysis. Overtting of
the experimental data by the deconvolution process in case of noisy data or wrong model
is generally controlled by the regularization parameter (or the smoothing parameter). The
regularization parameter is chosen by optimizing the distribution roughness factor and the
RMS error of t. MicroActive DataMaster v5.00 software manual suggest that little sensitivity of distribution and residuals obtained with the value of the regularization parameter

86

indicates proper choice of model to represent data and high sensitivity indicate inadequate
or noisy data or a poor choice of model.
Considerations on Application of N2 Gas Adsorption Techniques for Mudrock Characterization

4.3

There are dierent methodologies and algorithms to invert the isotherm data into pore
structure attribute information, of those the most commonly used methods were discussed
in the previous section Section 4.2). Also dierent studies have used dierent measurement
protocols for mudrock application, such as use of crushed sample (Adesida, 2011) and intact
core samples (Clarkson et al., 2012b). This have raised the following questions concerning
the best practices of N2 gas adsorption techniques for mudrocks applications:
Should the analysis be done in intact core samples or on crushed powders?
What are the best inversion techniques for pore-structure parameters from isotherm
data: BJH or DFT?
What is the best and correct way of representation of pore-size distribution data?
How reproducible and repeatable the measurements are?
In this following subsections, these topics will be discussed with a view to standardize
the technique for mudrocks applications.
4.3.1

Sample Preparation: Intact or Crushed Samples?

As discussed previously, all direct porosimetry techniques involve penetrating a sample

with a measurable quantity of an introduced gas or liquid. The nanodarcy range permeability of mudrock formations makes the diusion and/or penetration and equilibration of the
probe molecules (for example, N2 ) impossible or impractical for large volume samples. Most
core measurements of mudrocks are made on crushed rock samples following the procedure
established by the Gas Research Institute (Luel & Guidry, 1989). The reason behind using
crushed powder is that the path length for the gas to access the entire pore structure is
87

decreased, thereby, equilibrium can be achieved within reasonable time. This is particularly
important for N2 gas adsorption because the low temperature (LN2 temperature, 197 C)
used decreases the gas diusion rate through the pores. Clarkson et al. (2012b) have used intact coreplugs, instead of crushing; however, proper equilibration of N2 molecules to record
entire pore-structure remains a concern. Adesida (2011) and Comisky et al. (2011) studied
the eect of crushing on pore-structure parameters measured by N2 gas adsorption and MIP
techniques, respectively. They found the total specic pore volume measured increases with
decreasing sample particle size. The increased pore volume is related to better pore accessibility at smaller grain sizes, however, creation of articial pore space by microfracturing
during the crushing process cannot be ruled out (Comisky et al., 2011). Both studies sieved
the crushed mudrock sample into dierent sizes which may lead to a sampling bias. The
dierent mechanical properties of the constituent mineral grains of mudrocks results in differential tendency to grind into smaller sizes, and may result in a bias in the composition and
mineralogy of each separated fraction. However, this is a speculation and needs additional
measurements and observations to estimate and quantify.
The eect of sample size on N2 gas adsorption analysis on mudrock samples was investigated. Figure 4.8 show examples of poor isotherm data obtainedN2 adsorption during two
dierent analysis runs on the chip samples from Eastern European Silurian mudrock sample.
The samples were prepared by cutting rectangular shaped chips (23 cm wide, 0.5-1 g
each) from core material with a total sample weight of 14 g. The analysis on the chip
samples took unusually long ( 34 days) and the isotherm data do not match with the
IUPAC suggested isotherm type proles. The middle part of the adsorption branch and the
entire desorption branch is too at, indicating a severe diusion problem. Using higher mass
resulted in lower specic adsorbed volume. The diculty of N2 molecules to access into and
out of the porestructure prevented complete equilibration. This experimental result suggest that reliable data for gas adsorption cannot be obtained using intact chips for mudrock
samples.

88

Figure 4.8: Experimental results of N2 gas adsorption runs on chips of a mudrock sample
from Eastern European Silurian gas shale play. Note the raw isotherm data with at
adsorption and desorption branches indicating signicant problem of N2 to diuse into the
pore structure.

89

Therefore, the crushed samples were used for N2 gas adsorption analysis on mudrock
samples in this thesis. McCarty (2002) found that samples ground with minimal energy to
pass a < 420 m (40 mesh) sieve that are then homogenized and divided with a louvered
splitter produce mineralogically and chemically equivalent sample portions for compositional
and adsorption analyses. This sample preparation protocol was used. Note that the samples are not sieved through < 420 m mesh seive, instead the sample was crushed until the
entire mass passes through the sieve to prevent potential sample biasing due to sieving. To
investigate whether complete equilibration can be achieved using < 420 m crushed samples,
a test was performed to examine the eect of increasing equilibration time during isotherm
measurement. Results from two dierent Haynesville samples: one with a large volume of
ne pores between 2 and 5 nm (SS21, see composition in Table 2.2) and the other a sample that has an abundant volume of pores >10 nm (SS25, see composition in Table 2.2),
are shown for comparison. The equilibration check algorithm in the Micromeritics ASAP
2020 instrument uses a SaviskityGolay convolution method to compute weighted average and rst derivative of pressure over 10 windows of user-dened equilibration interval
(MicroActive DataMaster v5.00 software manual). The equilibration criteria used is that
the rst derivative should be <0.01% of the average pressure over 10 equilibration interval
windows. The samples were degassed and then a sequence of adsorption measurement between P/P 0 range of 0.0075 to 0.30 was made with varying equilibration intervals, keeping
all other analysis parameters constant. The samples were kept on the analysis port between
successive tests to avoid reweighing and redegassing. There is no systematic variation in
measured eSSAN 2 with dierent equilibration interval suggesting that complete equilibration
is achieved when handground < 420 m powder are used and the small relative variation of
2% and 6% in measured eSSAN 2 for SS21 and SS25, respectively (Table 4.1) are due to
the measurement precision.
However, crushed samples will articially create external surface area and will show higher
specic surface areas than their equivalent intact material, depending upon the particle size

90

Table 4.1: eSSAN 2 for sample SS21 and SS25 with dierent equilibration interval. C =
BET constant; Vm = monolayer capacity
Equilibration
Interval
SS2-1
10 sec
20 sec
30 sec
45 sec
SS2-5
10 sec
20 sec
20 sec (Run II)
30 sec
45 sec
100 sec

eSSAN 2
m2 /g

Vm
STP

Fitting)
Range (P/P 0 )

BET fitting
R2

22.85
23.29
23.01
23.11

160.7
155.1
163.1
164.8

5.249
5.35
5.286
5.309

0.0500.201
0.0500.200
0.0490.201
0.0500.201

0.999994
0.999996
0.999993
0.999994

4.11
4.24
4.19
4.25
4.36
4.32

109.3
99.5
102.2
106.2
100.8
102.8

0.945
0.974
0.962
0.976
1.001
0.992

0.0400.250
0.0500.250
0.0400.250
0.0500.226
0.0500.250
0.0500.225

0.999978
0.999984
0.999983
0.999991
0.999984
0.999989

distribution of the crushed powder. To understand the eect of crushing, a particle size was
analyzed by laser prole size analysis (LPSA) technique of the asreceived powdered source
clay SWy2 and three dierent < 420 m crushed mudrocks suspended in isopropyl alcohol.
Assuming spherical particles with grain density of 2.85 g/cm3 for dehydrated smectites, the
calculated external specic surface area is equal to 0.065 m2 /g Figure 4.9), which is an
order of magnitude lower than the measured eSSA (31.13 m2 /g) obtained for the sample.
Similar external specic surface areas between 0.053 to 0.070 m2 /g were obtained for the
dierent mudrock crushed samples (Figure 4.9), indicating crushing the samples will not
signicantly add to measured eSSA of samples. Also the median particle size of the crushed
powders is more than 200 times greater than the maximum pore-size of 200 nm measurable
by nitrogen gas adsorption technique. Hence it can be safely assumed that the grinding and
crushing process does not aect the measured porestructure parameters within the range
of investigation for N2 gas adsorption analysis.

91

Figure 4.9: Incremental volume percent particle size distribution of asreceived SWy2 and
hand-ground mudrock < 424m powders measured by laser prole size analysis in suspension
in isopropyl alcohol (black curve with lled circles). The cumulative external specic surface
area (blue line) of the powder is calculated assuming spherical grains. Note the reported
particle size distribution is that of a crushed powder and not the actual grain size distribution
of the sample.

92

4.3.2

Choice of Thickness Equation for t-plot and BJH PSD Inversion

One of the key assumptions used for inverting the isotherm data to obtain micropore
volume and PSD by t-plot and BJH technique, respectively, is the thickness equation that
quanties the thickness of the adsorbed layer on the pore surface as a function of relative
pressure (P/P 0 ). As discussed in Section 4.2.4, in a t-plot, the quantity adsorbed (Va) is
plotted against statistical thickness (t) of the adsorbed layer of N2 given by the thickness
equation. In the BJH algorithm, the thickness equation is used to account for the adsorbed
multilayer on the pore surface prior to commencement of capillary condensation in the pores
(Section 4.2.6). In absence of an experimentally obtained thickness equation applicable for
mudrocks, one of the several universal thickness equation are generally used. t-plot (Figure 4.10) and BJH inversion of adsorption isotherm data using dierent thickness equations,
keeping all other parameters constant, for two samples (SS21 and SS25) are compared in
Figure 4.11. The inverted micropore volume for SS21 varied from 0.006 to 0.012 with a
coecient of variability of 0.249 while the inverted micropore volume for SS2-5 is extremely
low and show variation beyond three signicant digits (Table 4.2). Note that the transform
plots (Figure 4.10) using Halsey and Broekho-de Boer does not show the expected convexup deviation from the linear trend at higher t, which is indicative of capillary condensation
in mesopores.
Signicant dierences in the calculated dierential pore volume exist for the same isotherm
experiment depending on which thickness equation is used (Figure 4.11). In each case, the
dominant pore modes remain the same regardless of the thickness equation used but the
dierential pore volume for each pore-size diers by up to 40%, particularly at the smaller
size range. The mismatch in the dierential pore volume is not signicant for pore diameters > 10 nm. The BJH computation method starts from higher pressures, hence larger
pore sizes, to lower pressures (smaller pore sizes) in the later steps. Due to the method of
calculation, errors become cumulative towards the smaller pore sizes and the dierence is
more pronounced.
93

Figure 4.10: t-plot transform plots of sample SS21(solid plus symbols) and SS25 (open circles) using dierent thickness equations. The best-t straight line shows for the extrapolation
of the linear region on the y-axis, which gives the specic micropore volume.

94

Table 4.2: t-plot inversion results of sample SS21 and SS25 using dierent thickness equations. KJS = KrukJaroniecSayari; HJ = HarkinsJura; BdB = Breokho-de Boer.
Thickness
Equation

Micropore
Vol.(cm3 /g)

Inter
cept

External
eSSA (m2 /g)

Slope

Fitting
Range (nm)

R2

0.008
0.009
0.010
0.012
0.006

5.193
6.367
6.605
7.853
3.702

9.31
8.95
8.26
6.40
14.58

6.02
5.78
5.34
4.14
9.43

0.891.39
0.981.59
0.831.23
1.011.58
0.620.77

0.999831
0.999988
0.999971
0.999782
0.999987

0.000
0.000
0.000
0.000
0.000

-0.091
-0.186
0.303
0.260
0.231

3.52
4.28
3.37
3.45
3.54

2.28
2.77
2.18
2.23
2.29

0.611.01
0.630.95
0.610.88
0.620.98
0.410.63

0.999977
0.999989
0.999991
0.999996
0.999965

SS2-1
KJS
Halsey
HJ
BdB
CarbonSTSA
SS2-5
KJS
Halsey
HJ
BdB
CarbonSTSA

Figure 4.11: PSD obtained from inversion of N2 gas adsorption data for two samples SS21
and SS25 using dierent thickness equation in the BJH method. The pore-size distribution
data shows the similar peaks but the absolute volume diers depending upon the thickness
equation used.

95

Since the physical measurement of the true thickness equations is dicult to obtain
experimentally for compositionally heterogeneous mudrocks, for a comparative purposes the
choice of any given thickness equation is immaterial as long as it is kept consistent across
a sample set. In this study, the Harkins-Jura thickness equation was selected based on
consistent t-plot transform curve-shape and closest (0,0) intercept in t-plots for majority
of samples with low eSSA samples that likely contain no micropore. The Harkins-Jura
thickness equation was also used for the multilayer adsorbed volume in the BJH inversion
to be consistent with t-plot analysis. No other corrections , for example Fass correction and
Kruk-Jaroneic-Sayari correction, were applied. However, it should be noted that the real pore
structures are much more complex than the models can account for and the inversion of the
isotherm is based on theoretical grounds that includes simplifying assumptions (Rouquerol
et al., 1994). The absolute or true value of any pore attribute cannot be measured and
each experimental technique and theoretical inversion gives a unique value (Rouquerol et al.,
1994). The results should be used in a semiquantitative way to understand the dierence
in pore structure between samples of interest. The reported pore sizes should be interpreted
as an equivalent cylindrical poresize distribution and should not be considered absolute.
4.3.3

PSD Inversion Technique: BJH or DFT?

The two most common inversion technique for pore-size distribution analysis from N2
isotherm data includes BJH and DFT techniques. While BJH uses a classical thermodynamic
approach, DFT method uses statistical thermodynamic model isotherms to t experimental
isotherm data to obtain PSD. The DFT technique has been preferred by researchers for
mudrock pore-structure characterization. Adesida (2011) and Clarkson et al. (2012b) used
DFT model of N2 adsorption on carbon slit 77K model as the kernel for PSD inversion.
There are other DFT kernel for N2 adsorption on dierent other surface types and dierent
pore geometries. The inverted PSD is sensitive to the choice of DFT kernels and results in
noisy and inconsistent PSD data (Figure 4.12, left) despite each of them producing a good
t for the experimental isotherm data points (Figure 4.12, right). Thus, a good t of the
96

model and experimental isotherm data is not an indicator of the correct choice of model and
PSD of the sample (for example, as used in Clarkson et al., 2012b). At present, DFT models
are not adequate to model the solid-gas interaction in real mudrocks mainly because of the
compositional heterogeneity and dierences in surface energies of their constituent silicate
minerals and organic matter. While the DFT method is more rigorous and has potential for
mudrock applications in the future, the BJH method based on the Kelvin model is suggested
for the mesopores PSD inversion of isotherm data for mudrock applications until appropriate
DFT kernel are developed for complex mineralogy.

Figure 4.12: (Left) Comparison of PSD results obtained by using dierent DFT kernels with
N2 as adsorptive for inversion of the adsorption data for sample SS21. The regularization
parameter (Reg) is optimized with the RMS error t and is reported in the legend. The
inverted results are sensitive to the choice of methods. (Right) Model t to the experimental
adsorption data (plus symbols). The models show reasonable match with the experimental
data but the inverted PSD results are noisy and inconsistent with respect to each other.
Inf. = innite; As = aspect ratio; Cyl = cylindrical pore; SWNT = single wall nanotube;
MWNT = multi-wall nanotube.

97

4.3.4

Graphical Representation of PSD Data

The PSD can be represented in dierent ways each having its own advantages and disadvantages. The most commonly used representations are ,
1. cumulative specic pore volume (ner or larger) versus pore size;
2. incremental specic pore volume versus pore size;
3. dierential pore volume (dV /dD) versus pore size and
4. log dierential pore volume (dV /d log D) versus pore size.
Both the N2 gas adsorption and MIP inversion techniques use a stepwise calculation
of the incremental specic pore volume (i.e. normalized per unit weight of sample) V
at each pressure interval which is then related to the particular pore size interval. The
total specic pore volume measured is the sum of all specic pore volume increments. The
partial sums of the specic pore volume up to particular pore size Di plotted versus the
pore size Di gives the cumulative specic pore volume curve. The cumulative curve can be
represented in two ways, either larger or smaller than the current pore size, depending upon
the direction of summation. The xaxis of these cumulative distribution plots are mostly
represented in the log10 scale because this permits convenient presentation of a range of
diameters covering several orders of magnitude as found in geologic materials. The yaxis
is plotted to an arithmetic scale. It should be noted that this transformation of the x
axis is only for visual representation, not a real coordinate transformation in the sense of
Di log10 Di and consequently Vcum (Di ) Vcum (log10 Di ) (Meyer & Klobes, 1999). The
cumulative distribution is the most unbiased way of presenting PSD data as the there is
no subsequent mathematical transformation of the data prior to plotting (Diamond, 1970).
However, the cumulative distribution does not explicitly highlight the modal pore size.
In the incremental specic pore volume distribution representation, the absolute incremental specic pore volume Vi is plotted versus the mid-point of the limits of the related
98

pore sizes interval. Sometimes they are represented as a bar graph plot with width of the
each bar indicating the pore-size increment and the height equal to the absolute incremental
specic pore volume for that particular increment. The advantage of such representation is
that the distribution shape shows the modal pore sizes explicitly. The disadvantage is that
the distribution shape is dependent upon experimental point spacing. An non uniform point
spacing along the pore-size axis can skew the representation by assigning higher incremental
pore volumes to larger intervals (Meyer & Klobes, 1999).
The dependence of the distribution shape on the irregular point spacing is eliminated by
normalizing the absolute incremental specic pore volume Vi with the interval size Di
(Meyer & Klobes, 1999). The log dierential distribution (dV /d log D) uses the dierence
in log of the upper and lower pore sizes of the interval ( log Di ). Both distributions are
plotted with the dierential value along yaxis versus the midpoint of the related pore-size
interval. For bimodal or multimodal pore-size distribution, these two distributions have a
dierent shape and gives a dierent visual impression of the PSD of the investigated sample
(for example, Clarkson et al., 2012a, 2013).
Comparison of PSD prole for these two representations for SS21 from N2 gas adsorption and Berea sandstone from mercury porosimetry (Figure 4.13 and Figure 4.14) suggest
that the dV /dD distribution function shape visually overemphasizes the smaller pores, even
though, mathematically both of them contains the same information. A direct comparison
of these two distributions is incorrect as the units of these two distribution functions are
cm3 /g
and it indicates the pore
dierent (Meyer & Klobes, 1999). The unit of dV /dD is
nm
volume density. A graphically correct way of presenting poresize distribution function is
such that the peak area under the curve between any two pore-sizes should be proportional
to the partial specic pore volume for the specic pore-size interval. Hence the distribution
dV /dD should only be used with linear abscissa (poresize) and linear ordinate (dV /dD)
axis, as presented in Figure 4.13(c) and Figure 4.13(d). The peak areas on the dV /dD plot
with logarithmically compressed pore-diameter abscissa (Figure 4.13(a) and Figure 4.13(b))

99

are not proportional to total volumes (Meyer & Klobes, 1999).

Since a logarithmic abscissa is preferred to represent the poresize scale for geologic
materials, dV /d log D distribution representation is recommended for such samples. This is
because the peak area under the dV /d log D curve between any two pore-sizes in the log scale
is proportional to the partial specic pore volume for the specic pore-size interval as shown
in the following integration equation.
Z D
2
D1

Z D
Z D
2
2
dV
dV
D

dD =
d log D = dV = V |D2
1
d log D
D1 d log D
D1

(4.7)

where D is equal to logarithm of pore diameter D.It should be noted that the parameter
dV /d log D is related to log10 D and hence the correct abscissa of representation is a linear
axis of log10 D (Meyer & Klobes, 1999). The unit of dV /d log D (cm3 /g) is similar to the
unit of specic pore volume; however these two quantities should not be confused. Note
Figure 4.14, the values of dV /d log D are higher than the measured total specic pore volumes
of the sample. The correct partial volumes should be obtained using a numerical integration
(area under the curve)of the curve with respect to the linear log10 D axis.
4.3.5

Reproducibility of N2 Gas Adsorption Data

Reproducibility test was performed on ve dierent samples from SS2 (SS21 to SS25)
using the same instrument (Micromeritics ASAP 2020) in two dierent laboratories located in Golden, Colorado and Houston, Texas. Both the instruments were calibrated by
measurement on for certied standards. The atmospheric pressure at these two places is different which resulted in variation of measured saturation pressure of N2 under experimental
conditions between 603626 torr in Golden compared to 757766 torr in Houston. Dierent
masses are used from the exact same split were used (3 g used in Houston compared to
5 g in Golden). The other major dierence in the analysis conditions is the sample tube
valves used, a TransSealvalve that keeps the sample under vacuum after degas was used
in the Golden laboratory whereas a SealFritvalve that need a positive backll gas (N2 ) to
prevent contamination of sample by moisture between degas and analysis. The comparison
100

(a)

(b)

(c)

(d)

Figure 4.13: dV /dD representation of PSD for two samples Berea sandstone (blue) obtained by MIP and mudrock SS22 obtained by N2 gas adsorption (red). Note that dV /dD
distribution function shape visually overemphasized the smaller pores. (a) and (b) dV /dD
representation with logarithmically compressed pore-size abscissa; (c) and (d) dV /dD representation with linear pore-size abscissa.

101

(a)

(b)

Figure 4.14: dV /d log D representation of PSD for two samples Berea sandstone (blue)
obtained by MIP and mudrock SS22 obtained by N2 gas adsorption (red). Total specic
pore volume measured for Berea and SS22 are 0.0894 cm3 /g and 0.0206 cm3 /g, respectively,
and the dV /d log D values are higher than the total pore volume. dV /d log D is related to
linear log10 D values scale (abscissa at the top of the plots) and correct partial pore volume
for each poresize range can be obtained by calculating the peak area with respect to linear
log10 D scale.

102

of raw adsorption data suggest that the measurements in Golden laboratory reported systematically higher amount of adsorbed gas compared to the Houston laboratory. Hence the
reported eSSAN 2 , total specic pore volume by Gurevich rule and PSD are systematically
higher when measured in Golden compared to than measured in Houston. The variation
in measured eSSAN 2 ranges from 1.4% to 17.9% while the variation in the measured total
specic pore volume ranges from 5.4% to 9.1% (Table 4.3). A reproducibility study on a certied mesoporous material CRM BAM-PM-103 between 25 laboratories reported a variation
of 9.2% (mean 156.0 m2 /g) in specic surface area using BET Method and variation of 6.4%
(mean 0.250 cm3 /g) in specic pore volume by BJH method (Klobes et al., 2006). The PSD
(Figure 4.15) obtained show similar distribution shape and modal pore-size values, however
the porevolume for all pore-sizes are consistently higher by about 5 -10% when measured
in Golden compared to when measured in Houston. The reason for this consistent dierence
in the measurements at the two locations is not clear. The results of the reproducibility
test suggest that the measurement uncertainty of the pore structure attributes in general is
about 515 relative %.
Table 4.3: Results of reproducibility test for samples ve samples between two laboratories
located in Golden, CO and Houston, TX
Total Specific Pore Volume
(cm3 /g)
Golden
Houston

eSSAN 2
Sample
(m2 /g)
Golden Houston
SS2-1
SS2-2
SS2-3
SS2-4
SS2-5

26.28
9.73
8.99
18.65
4.41

23.29
9.26
8.87
15.59
3.87

0.0309
0.0208
0.0273
0.0333
0.0140

a Single

0.0287
0.0197
0.0257
0.0304
0.0131

point adsorption specic pore volume of pores with less than 149.1 nm diameter, calculated by
Gurevich rule. Quantity adsorbed intrapolated at P/Po = 0.987.

103

Figure 4.15: Comparison of PSD for dierent samples measured at two dierent labs, Golden
CO (lighter shade) and Houston TX (darker shade). (Left) Red: SS21; Blue: SS22; Black:
SS23. (Right) Green: SS24; Magenta: SS25.
4.4

Comparison of N2 gas adsorption and MIP techniques

N2 gas adsorption techniques can be used to characterize materials dominated by micro

and mesopores (2-50 nm). A limitation of this technique is that it fails to measure pores
with diameter >200 nm. The intrusion technique using a non-wetting liquid (a contact angle
greater than 90 with a given solid) is better suited for measurement of larger pores. Since
the meniscus is convex, the nonwetting liquid has to be forced by pressure to intrude into
the empty pores and the applied pressure is related to the poresize by Washburn equation
(Webb & Orr, 1997). Mercury is preferred choice as the nonwetting liquid because of its
large contact angle with most kinds of solid surfaces, low vapor pressure, and chemically inert
to most solids. This technique is most frequently used for pore-structure characterization of
conventional reservior rocks in earth science and petroleum engineering. Serious limitations
on the applicability of this technique for mudrock porestructure characterization have been

104

reported as highlighted in Section 3.2, (page 34). Despite these limitations, the ability
of MIP to record macropores with orders of magnitude makes it a useful tool to provide
complimentary information of pores structure along with N2 gas adsorption (Gregg & Sing,
1983; Bustin et al., 2008). Joyner et al. (1951), and Caro & Freeman (1961) showed similarity
of PSD from MIP and N2 gas adsorption in dierent materials such as charcoal, porous
alumina, and phosphate fertilizers. Other studies also used the combination of these two
methods to characterize pore structures of carbon (Dubinin, 1966), clays and soils (Sills
et al., 1973; Echeverra et al., 1999), cements and concrete (Diamond, 1971), porous silica
(Brown & Lard, 1974), and mudrocks (Bustin et al., 2008; Chalmers et al., 2012; Clarkson
et al., 2012a; Clarkson et al., 2013).
One of the most striking feature of both N2 gas adsorption and MIP results in heterogeneous porous structure is that they show a strong hysteresis pattern. The porenetworks
and interconnectivity of the pores aects the experimental results for both gas adsorption
and MIP and the hysteresis is due to varying sequence in which the probe molecule encounters the pores. The hysteresis observed in gas adsorption and MIP may be attributed to
various other physical phenomena discussed in the literature; however the interconnected
porestructure has been used most often to explain the scanning curves inside the hysteresis
loop observed in both these techniques (Mason, 1982). The cylindrical capillary bundle pore
model approximation used to inversion of both data to PSD does not account for network
connectivity of the pore. There will be a dierence in the PSD obtained based on the branch
of experimental data used. This leads to question that which branch of the hysteresis loop
to appropriate for PSD analysis.
A comparison of PSD results on clays and mudrocks from N2 gas adsorption and MIP
techniques will be presented in the following sections and the equivalence and disparity
between each measurement will be evaluated. The phenomena of adsorptiondesorption and
extrusion-intrusion hysteresis patterns in gas adsorption and mercury porosimetry will
be discussed.

105

4.4.1

Theory of MIP

Mercury is a strongly nonwetting uid and can only penetrate into the poresystem
when external pressure is applied. MIP utilizes this nonwetting property of the uid to
determine the porestructure of the materials. The Washburns equation relates the pressure
(P ) required to penetrate a pore capillary with the contact angle of mercury () against the
pore surface, the gas/liquid surface tension () and the pore diameter Dm , and is given by
P =

4 cos
Dm

(4.8)

Equation 4.8 suggests that with increasing pressure, the mercury will penetrate pores with
smaller diameters. In a typical MIP experiment, a sample is placed in the bulb of previously
volume calibrated penetrometer(a special sample holder for such experiment) and lled with
mercury. Pressure is gradually increased up to about 414 MPa (60,000 psi) and change of
volume of mercury is constantly recorded and is reported as the intrusion curve (Figure 4.16).
The total pore volume is calculated by the volume of mercury intruded at 414 MPa (60000
psi). The extrusion curve (Figure 4.16) is generated by recording the volume change as the
pressure is progressively decreased. One important feature to be noticed in Figure 4.16 is
that all the intruded mercury has not extruded as the pressure is reduced back to ambient
and some mercury has become entrapped inside the sample. This phenomenon is discussed
in the later sections. The pressure data is converted into equivalent pore-sizes using the
Washburn equation (Equation 4.8).
A key uncertainty involved in inverting for PSD from MIP data using Washburns equation is the value of the contact angle, commonly assumed to be 140 . However, several
studies have demonstrated a wide range of contact angles, even between mercury and similar surfaces (Diamond, 1970; Gregg & Sing, 1983; Groen et al., 2002). Klobes et al. (2006)
reported relative deviation of about 35% in the interpreted pore diameter with changing the
contact angle from 140 to 120 . Figure 4.17 shows the change in the inverted data with dif-

106

Figure 4.16: Raw experimental data obtained by MIP in a representative mudrock showing
show strong hysteresis pattern between the intrusion and extrusion curves.
ferent contact angle assumption for two dierent samples, Berea sandstone and Haynesville
Shale. Sills et al. (1973) demonstrated using a value of contact angle of 140 for clay rich
samples; MIP pore-size distribution plot is displaced by approximately 20% compared to
N2 PSD and a value of 153 for mercury contact angle made the plots overlapping. In this
study the results were analyzed using both the advancing and retreating contact angles be
140 , and the surface tension was taken as 0.485 N/m(485 dyne/cm).
Pore-structure characterization in mudrocks are hindered by the extremely low permeability of shales. Comisky et al. (2011) found that mercury intrusion results in mudrocks
show a strong dependence on the sample particle size; crushed sample with smaller size
record higher porosity compared to plug samples due to the higher pore-accessibility in the
crushed samples. However, using a crushed powder will have higher conformance related
errors at low intrusion pressures that need to be corrected. Comisky et al. (2011) presented
the Bailey methodology to account for the conformance and porevolume compressibility
corrections. The Bailey methodology is adopted in this thesis for correcting the MIP data
for mudrocks.

107

Figure 4.17: : Deviation in the inverted pore size distribution as a function of the mercury
contact angle
Another important source of error in MIP data is the apparent volume change caused
by the compressibility of the system (apparatus and mercury) and bulk samples particularly
at higher pressures. Sigal (2009) suggested a methodology for blank correction using blank
runs on nonporous quartz crystal standards; however he also pointed out that methodology
developed could not be extended to mudrocks due to their heterogeneous composition. The
eect of blank correction can be minimized by using a higher sample mass leaving a smaller
dead-space in the sample holder. In this study, no separate blank correction was applied to
the results but higher sample mass was used to minimize blank correction eects.
4.4.2

Phenomenological comparison between MIP and N2 gas adsorption

Both methods can only record interconnected pores which have some communication with
the exterior of a porous body and therefore, data obtained must always be viewed as the pore
size distribution of the connected pores. Both methods mentioned above are based on the
phenomenon of surface tension or the surface free energy of a uid (liquid or gas) in a conned
capillary or pore-space, however, the uid properties of mercury and N2 are dierent. These
uids sequentially measure the interconnected porespace network in a dierent mechanism.
108

Mason (1982) and Conner et al. (1986) explained in details the phenomena of adsorption
desorption process of nitrogen gas adsorption and intrusionextrusion process of mercury
technique. Conner et al. (1986) demonstrated the equivalence between mercury intrusion
and nitrogen desorption and between mercury extrusion and nitrogen adsorption. In the
following, the key ndings from Mason (1982) and Conner et al. (1986) highlighting the
equivalence and dierence in these two techniques are summarized.
During N2 adsorption process, at lower relative pressure (P/P 0 <0.4) the uid molecules
adsorb on the pore surface forming a monolayermultilayer of adsorbed gas molecules. At
higher relative pressure (P/P 0 >0.4), the gas condenses in the pores to a bulk liquid phase by
capillary condensation. The gas in the pores condenses to the bulk liquid phase at pressures
below the saturated vapor pressure of the gas (P 0 ). The two phases are separated by a
concave meniscus. Kelvins equation relates the vapor pressure with the surface curvature
of the liquidgas meniscus, which is dependent on the pore diameter. Thus, gas in pores
with dierent diameters will condense at dierent pressures giving information about the
pore-sizes present in the sample. During the adsorption process, the pores ll sequentially
from smaller to larger sizes. All of the interconnected pores are accessed by the gas phase
prior to the commencement of capillary condensation and the condensation in any given
pore is independent of network eect. On the other hand, larger pores will be sequentially
emptied rst by capillary evaporation during the desorption process, however, a larger pore
connected by smaller pores cannot be emptied until the smaller pore size empties. Therefore,
the desorption process will be controlled by sizes of constrictions in the pore-structure (porethroats). The pore-size distribution shape obtained from the nitrogen adsorption will be
dominated by the relative wider portions of the porestructures (pore-body). During the
desorption process, the pore volumes will be skewed toward the smaller pore sizes and hence
the distribution shape will be more dominated by porethroats.
In mercury porosimetry technique, the larger pores will ll rst at lower pressures during
the intrusion process and during the extrusion process, the smallest pore will empty rst.

109

Both processes are sequential pore access mechanism and therefore are subject to the network
eects. However, the network eects are dierent in intrusion versus extrusion processes.
More details of the network eects were presented in Conner et al. (1986). Another important
feature of mercury extrusion process is that some mercury is trapped and retained in the pore
structure (Figure 4.16). This is related to the pore structure heterogeneity with mercury
snaping-o in narrow pore constrictions connecting larger porebody elements (Rigby &
Edler, 2002). The mercury intrusion process and N2 desorption process are analogous and
similar that in both processes show similar network eect and sequential pore emptying
where the emptying of larger pores are controlled by the emptying of the smaller pore
constrictions. The poresize distributions are dominated by the pore-throats. N2 adsorption
and mercury extrusion are similar in the way that both of them measures the smaller pores
rst. However, the nitrogen adsorption process is not aected by the network and retention
eects as in mercury extrusion process. In the following study, the similarity in N2 gas
adsorption and MIP measured data for mudrock samples was tested. The main advantage
will be extend the ranges of the PSD by combining these techniques and gaining an insight
about the interconnected pore network structure.
4.4.3

Methods and Materials

MIP data was obtained on pure SWy2 clay samples, SWy2 compacted pellets, and
fteen SS2 samples (SS21 to SS25, SS225 to SS234, Table 2.2) using Micromeritics
AutoPore IV 9500 . The pressure range applied was 0.14420 MPa corresponding to an
equivalent pore-size range of 3.6 nm to 104 nm. The samples were degassed in vacuum at 50
mHg pressure for at least 30 minutes. 1015 g of smaller chip samples were used for SS2
samples. Smaller (25 g) samples were used for the clay samples. Every pressure point was
equilibrated below 0.001 L/g/s intrusion rate. More data points were collected during the
intrusion branch with sparse data points along the extrusion curve for most of the sample
based on the conventional application protocols. The conformance correction of the MIP
data for the SS2 samples was performed using the Bailey methodology. A separate blank
110

correction was not applied, however, a higher sample mass to ll most of the penetrometer
bulb was used to minimize blank correction eects.
The MIP data on mudrock sample set ASS were previously reported by Sarker (2010).
The analysis was done commercially on whole core samples and the reported MIP data are
conformancecorrected, however, exact methodology of the conformance correction was not
known (Dr. Rituparna Sarker, personal communication).
The total pore volume from N2 gas adsorption was obtained by application of Gurevich
rule, as explained in Section 4.2.5. PSD inversion was done using the BJH method assuming
Harkins-Jura thickness equation for multilayer adsorption. The PSD from the adsorption
branch was limited to 1.7 nm as the assumptions of Kelvins equation becomes invalid beyond
that poresize. The PSD from the desorption branch was limited to 5 nm because of the
TSE eect.
4.4.4

Results

Total Pore Volumes A comparison of total pore volume data measured from N2 gas
adsorption and MIP techniques on dierent samples shows two dierent trends (Figure 4.18).
The total pore volumes measured by the N2 gas adsorption technique for the Haynesville
samples (SS2) are higher than those measured by MIP (Figure 4.18(a)), whereas for other
mudrocks from ASS and pure SWy-2 clay samples show higher total pore volume measured
by MIP (Figure 4.18(b)).
PSD A comparison of PSD obtained from inverting the adsorption and desorption
branch for N2 isotherm for asreceived source clay SWy2 samples shows signicant difference in the two distributions (Figure 4.19(a)). The PSD from the desorption branch was
limited to 5 nm as the isotherm data below P/P 0 of 0.47 were not included in the inversion.
The PSD from the adsorption branch tends to be broader with modes at higher pore-sizes
compared to the PSD from the desorption branch. This dierence is also observed were
observed between the inverted PSD obtained from extrusion and intrusion curves of MIP
111

(a) The Haynesville Formation

(b) Pure Clays and other mudrocks

Figure 4.18: Comparison of total specic pore volume measured by N2 gas adsorption and
MIP techniques on (a) the Haynesville samples (SS2), and (b) on pure SWy-2 clays and
other mudrocks. The solid line is the line of equal values.

data (Figure 4.19(b)). The PSD from the extrusion curves tend to be broader with higher
modal poresizes compared to the PSD obtained from the intrusion curve. Another marked
dierence between the two PSD is that the extrusion curve PSD shows lower pore volumes
and the modal peak at higher poresize is missing.
Comparing the PSD obtained from inverting dierent branches between N2 isotherm
data and MIP data indicates that the PSD from desorption and intrusion branches have
similar shapes (Figure 4.19). Similarly, the PSD from adsorption and extrusion branches
are closer in shape and distribution. Detailed comparisons between PSD from desorption
and intrusion branches for the pure clay samples indicate similar modal pore sizes and poresize ranges, however, the MIP data shows higher pore volume and therefore slightly sharper
prole (Figure 4.20(a)). The adsorption branch PSD have similar modal pore-sizes as the
extrusion branch but have higher pore volume (Figure 4.20(b)). Similar dierence in PSD
was obtained for the mudrocks Figure 4.21. The reasons for these similarities and dierence
will be discussed in the following section.

112

(a) N2 gas adsorption

(b) MIP

Figure 4.19: Comparison of PSD of as-received SWy-2 clays obtained by inverting the
dierence branches of (a) N2 isotherm data and (b) MIP data. Note the yaxis is broken for
MIP data plot.

(a) PSD from Desorption and Intrusion branch

(b) PSD from Adsorption and Extrusion branch

Figure 4.20: Comparison of PSD of as-received SWy-2 clays and the compacted pellets (a)
obtained from N2 desorption and MIP intrustion data and (b) obtained from N2 adsorption
and MIP extrustion data. Note the change in scales of the pore-size axis in (a) and (b)

113

Figure 4.21: Comparison of PSD obtained from N2 adsorption (gray, N2 desorption (blue)
and MIP intrusion (red) branch from four mudrock formations.

114

4.4.5

Discussion

The higher measured pore volume in the Haynesville samples by N2 gas adsorption techniques (Figure 4.18(a)) indicates that these samples have a signicant volume of small pores
with <3.6 nm poresize, which cannot be accessed by MIP technique. In addition, MIP has
limited poreaccessibility in such mudrock chip sampless due to the low permeability of the
mudrocks and high viscosity of mercury. For the pure clay and other ASS mudrock samples,
the higher measured total pore volume obtained from MIP indicates that these samples have
signicant volume of larger macropores > 200 nm, the practical limit of measurable poresize
by P/P 0 gas adsorption technique. By addressing the high pressure compression correction
and the pore accessibility issues in MIP applications for mudrocks, a combination of MIP
and N2 gas adsorption can be a powerful tool to provide information for the entire poresize
range.
The hysteresis observed in both the experimental techniques suggest there is a strong path
dependence on how the uid molecules access the pore structure. Inverting any experimental
data to obtain PSD involves several simplifying theoretical assumptions, the most commonly
used assumption is the cylindrical capillary bundle approximation. Such a theory does not
honor the real complexity of interconnected pore network system. Therefore, such inversions
provide only equivalent characteristics values, depending upon the experiment and theory
(Rouquerol et al., 1994). The results should be treated as one dimensional approximation
of the real pore structure.
The PSD from N2 adsorption and MIP extrusion branches have broader proles with
higher modal pore-sizes as compared to PSD obtained from N2 desorption and MIP intrusion branches. The dierence in the distribution proles is due to the fact the adsorption and
extrusion mechanism is more controlled by the relative wider portions of the pore structure,
the porebody (Mason, 1982; (Conner et al., 1986)) and therefore they provide an approximation of pore-body size distribution. Similarly the desorption and intrusion processes are
essentially a porethroat controlled. The dierence in the modal size between porebody and
115

porethroat ranges between 1020 nm in natural mudrocks (Figure 4.21) and from 7080 nm
in pure clays (Figure 4.20). The dierence in shapes of porebody and porethroat size distributions are also illustrated by Dullien & Dhawan (1974) who reported dierent pore-size
distribution of same sandstone sample using optical microscope stereological image analysis
on a two-dimensional slice and mercury intrusion measurements(Figure 4.22). The image
analysis resulted in the wide and at distribution function compared to a narrow, sharp distribution resulting from mercury porosimetry. The modal size is larger in the image-analysis
distribution compared to mercury intrusion. The dierence in the pore-size distribution was
explained in terms of porebody measurement which is revealed in quantitative photomicrography measurement compared to porethroat measurement in mercury intrusion. Similar
dierence in PSD was reported by Dewers et al. (2012) in their study of the Haynesville
Formation using FIBSEM and MIP techniques.

Figure 4.22: Comparison of PSD from Clear Creek Sandstone measured using MIP and
optical microscope stereological image analysis techniques (from Dullien & Dhawan, 1974).

116

It is demonstrated in this study that dierent hysteresis branches provides dierent PSD
results. For combined N2 gas adsorption and MIP measurements of pore structure, the correspondence of the mechanism along the hystersis branches should be honored and comparison
of N2 adsorption PSD and MIP intrusion PSD are not recommended (for example, as used
in Clarkson et al., 2012b, Schmitt et al., 2013). The dierence in PSD data have implications and application for modeling transport behavior and elastic behavior of mudrocks.
The understanding of the dierence between porebody size distribution and porethroat size
distribution can aid in modeling transport properties and elastic properties. The porethroat
size distribution correlates very well with concept of hydraulic radius and controls the capillary pressure behavior and Poisuelles ow through the porous media. However, the Knudsen
diusivity ow regime, where the gas molecules bounces between the pore walls, should be
more controlled by the pore-body size distribution. The elastic behavior of porous media
is controlled by the aspect ratio of the pores, however, there is no robust measurement of
aspect ratio. The ratio of the pore-throat sizes and the pore-body size can be used as an
approximation of the aspect ratio of the pores.
4.5

Future Recommendations
A systematic study by crushing the samples with dierent grain sizes, with sieving and

without sieving, should be carried to understand the eect of crushing and sieving in sampling
bias and porestructure characterization for mudrocks. Development of a mudrockspecic
thickness equation by N2 adsorption analysis on non-porous samples with representative
chemical composition similar to mudrocks is recommended. Poresize distribution inversion
using a DFT deconvolution method using classical thermodynamic kernels, like Kelvins
equation, to model kernel functions is recommended. In BJH algorithm, the errors become
cumulative towards the smaller pore sizes and such approach utilizing the deconvolution
tting can distribute the errors for the entire pore-size range. Classical thermodynamic
theory based DFT kernel is better suited for mudrock application compared to currently
available other DFT kernels.
117

Application of dierent porenetwork models during inverting the N2 isotherm and MIP
data (for example Conner et al., 1986, and citations within) to explain the hysteresis can
provide important information such as pore connectivity and pore aspect ratio,which are
valuable attributes for ow and elastic modeling. N2 gas adsorption and MIP techniques are
generally utilized separately for pore structure characterization and using it in conjunction
will give additional insights about the pore structure. Give the complexity of pore-structures
in mudrocks, combined approaches such as simultaneous inversion of N2 isotherm and MIP
data (for example, Murray et al., 1999), and integrated experimental methods (for example
Rigby & Fletcher, 2004; Rigby et al., 2004) should be developed.
4.6

Summary
In this chapter, detailed descriptions about the experimental procedure and common

inversion methods to obtain porestructure parameters from N2 gas adsorption data were
presented. Application issues for adapting this technique for mudrock pore-structure characterization, for example, sample preparation technique, choice of inversion techniques of the
raw isotherm data, representation of pore-size distribution (PSD) data and reproducibility
of the measurement data, were discussed. A detailed comparison between the N2 gas adsorption and MIP techniques were also presented. The following summarize the conclusions
based on the observations of gas adsorption results on mudrocks:
1. Due to slow diusivity and incomplete equilibrium issues during N2 gas adsorption
analysis in mudrocks, handground <420 m mesh powder are recommended. Crushing
to <420 m mesh does not potentially damage the porestructure in the range of
investigation for N2 gas adsorption of up to 200 nm.
2. The eSSA, total pore volume and PSD results obtained on handground <420 m mesh
powder of mudrock samples are consistent and reproducible within 515 relative %.
3. The tplot and BJH inverted results depend upon the choice of thickness equation.
The choice of thickness equation should be kept consistent across a sample set or com118

parison. Furthermore, the results should be treated as equivalent characteristic values

and used in a semiquantitative way to understand the dierence in pore structure
between samples of interest.
4. BJH inversion is preferred over DFT inversion technique.
5. Linear dV /d log D ordinate vs. logarithmic poresize abscissa plots are recommended
for PSD data representation
6. Combined N2 gas adsorption and MIP techniques are useful as they provide complementary information about the pore-structure.
7. The commonly used capillary bundle approximation does not account for the interconnectedness of real pore structure. The PSD results obtained from N2 adsorption and
MIP extrusion branch give an approximation of the pore-body size distribution whereas
the N2 desorption and MIP intrusion branch gives a porethroat size distribution.
Most of the recent pore space characterization studies on mudrocks have centered around
using imaging techniques, which are limited to provide visually detailed yet qualitative descriptions of mudrock porosity and cannot provide detailed porestructure parameter quantication. To address this issue, N2 gas adsorption technique has been used to understand
the porestructures in mudrocks and its end members such as clay minerals and organic
matter, which were described in the next chapters. Only the adsorption branch of the N2
isotherm data were inverted for this thesis.

119

CHAPTER 5
PORESTRUCTURE IN MUDROCK END MEMBER COMPONENTS: CLAYS AND
ORGANIC MATTER

Divide each diculty into as many parts as is feasible and necessary to

resolve it.
Ren Descartes
The applicability and justication of using the N2 gas adsorption technique to characterize the mudrock porestructure was demonstrated in Chapter 4. The results and insights
gained about the mudrock pore structure obtained by this technique will be presented in the
following two chapters. The presence of micropore and mesopore networks either exclusively
within the organic matter or as pore systems in the inorganic clay components is widely
debated. Previous imaging studies have indicated that the majority of nanometer scale pore
structure observed in mudrocks are associated with the clay and organic matter (OM) components (e.g., Javadpour, 2009; Loucks et al., 2009; Milliken & Reed, 2010; Passey et al.,
2010; Schieber, 2010; Walls et al., 2011; Curtis et al., 2012; Milliken et al., 2013). Typical
pores associated with clay minerals includes elongate to angular pores with dimensions ranging from 30 nm to 2 m, often occuring within or at the edge of the clay aggregates or against
rigid silt grains (Milliken & Reed, 2010; Schieber, 2010). OM porosity occurs as irregular,
bubble-like, circular to elliptical pores with diameters as low as 5 nm (minimum resolution
of current imaging techniques). The relative importance of organic hosted porosity as compared to clay hosted porosity and its contribution to total mudrock porosity is also highly
debated (Schieber, 2010; Modica & Lapierre, 2012; Curtis et al., 2012). For example, Curtis
et al. (2012) reported that OM porosity is dominant in Barnett, Kimmeridgian, Woodford
and Horn River shales, whereas porosity in the Haynesville Shale is dominated by the clay
microstructure. Due to the dierence in surface energy, dierentiating and quantifying the

120

pore volume in OM and in clay aggregates and other inorganic components is critical for
modeling gas adsorption and desorption processes in mudrocks with signicant production
implications in these self-sourcing reservoirs.
Natural mudrocks are heterogeneous with wide composition range and therefore, to understand the compositional controls on mudrock pore structure, it becomes important to
characterize the pore-structure associated with the pure compositional end member components. In this chapter, the porestructure associated with the two important compositional
end member components of a mudrock system: clay mineral, particularly illite+smectite
(I+S) group, and organic matter (OM), are discussed in Section 5.1 and Section 5.2, respectively. Pure standard clay mineral samples were used. A combination of dierent sample
selection and experimental treatment procedures were used to understand the OM pore
structure, in absence of readily available OM samples.
5.1

Pore Structure in Clay Minerals

The pore-structure of pure clays have been studied in details using several measurement

techniques, such as transmission electron microscopy (e.g., Kim et al., 1995; Jullien et al.,
2005), scanning electron microscopy; gas adsorption techniques (Quirk & Aylmore, 1971;
Cases et al., 1992; Rutherford et al. (1997);Neaman et al., 2003) and MIP techniques (Diamond, 1970; Ahmed et al., 1974). The pore volumes, specic surface areas and poresizes of
clays, particularly illite+smectite group (I+S), depends mainly on the origin of the sample,
layer charge or cation exchange capacity, the nature of the exchanged cation and the dehydration conditions during sample preparation. The most distinctive characteristic of I+S
group clay minerals microstructure is their multiscale stacked fabric. Several individual elementary units comprising of aluminosilicate tetrahedral-octahedral-tetrahedral (TOT) layers
with interlayer cation of unit cell size of 915 , 0.91.5 nm form mesoscopic tactoids or
quasicrystals. Slitshaped or wedged micropores are present along on the broken edges or
overlapping regions of the turbostratically stacked (disorder stacking of layer with dierent
rotations with respect to the cyrstallogarphic caxis) elementary TOT units within each
121

tactoid (Quirk & Aylmore, 1971; Cases et al., 1992; Neaman et al., 2003). Rutherford et al.
(1997) used with N2 and neohexane adsorption techniques on pure homoionic smectites to
study the micropores in them and found that micropore volume range from 14% to 66% of
the total measured pore volume, depending upon the size of the exchangeable cations in the
TOT interlayers. The tactoids are stacked to form aggregates which results in intertactoid
medium mesopores (Neaman et al., 2003; Jullien et al., 2005, Moyano et al., 2012). Stacking
of aggregates with preferred alignment form interaggregate pores, which are usually micron
sized. Neaman et al. (2003) reported that compaction of pure smectites results in reduction
of pore volume corresponds to intertactoid and inter-aggregate pores. Diamond (1970) reported presence of micronsized pores in clays using the MIP technique. Ahmed et al. (1974)
reported dierent eect of compaction behavior in poresize distribution of clays depending
upon initial moisture content of samples.
Poresize distribution results obtained from N2 gas adsorption and mercury intrusion
studies on pure clays are presented in Section 5.1.1. MIP techniques were used to compare
the results obtained from N2 gas adsorption. The poresize distribution results obtained
from N2 gas adsorption and MIP are discussed to understand the multiscale pore-structure
in clays and its evolution with compaction.
5.1.1

PSD Results on Pure Clays and Compacted Pellets

Two source clays, Wyoming montmorillonite (SWy-2) and low-defect Georgia kaolinite
(KGa-1b), from the Clay Mineral Society were used for this study. Wyoming montmorillonite
powder (SWy-2) was articially compacted to study the eect of compaction on clay pore
structures. The details of the sample description and preparation of compacted pellets were
presented in Section 2.2.1.
Source Clays The N2 gas adsorption isotherm of as-received powders of Wyoming
Montmorillonite SWy2 and Kaolinite KGa1b show marked dierence in shapes (Figure 5.1). Uncompacted kaolinite shows almost reversible Type II nitrogen isotherm. The
122

isotherm shape indicates that kaolinite is dominantly macroporous and micromesopores are
absent. The hysteresis is extremely narrow and may indicate presence of ne macropores.
Low adsorbed volume at P/P 0 <0.01 indicates that kaolinite has negligible or nonexistent
micropores.
The Wyoming Montmorillonite SWy2 asreceived shows Type IIB isotherm pattern
with a signicant H3 type hysteresis pattern. Such aisotherm shape indicates that the
material contains both mesopores, which are responsible for the H3 type hysteresis , and
macropores, which results in an absence of the plateau and having steep slopes in the relative
pressure range of 0.98 - 1.00 (Figure 5.1). The H3 hysteresis pattern indicates presence of
slitlike pores. The presence of the forced closure (Tensile Strength Eect) of the desorption
branch at 0.41 < P/P 0 < 0.47 indicates that SWy-2 has a signicant pore volume with
diameter <4 nm. The volume uptake at P/P 0 near 1.0 indicates the total porosity of
materials up to 200 nm pore-size range. Montmorillonite powders have larger pore-volumes
< 200 nm pore-size range compared to the kaolinite powders. Furthermore, they have more
micropores than kaolinite, as indicated by a larger adsorbed volume at P/P 0 <0.01.
As-received kaolinite KGa1B have similar modied BET eSSA and t-plot open specic
surface area (Table 5.1), whereas SWy-2 shows signicant dierence between the two surface
areas. The dierence in these values can be correlated to surface area associated with
micropores (pore less than 2 nm diameter) since the open specic surface area obtained from
t-plot typically provides the the surface area from mesopores, macropores, and external
surface area.
Measured MIP pore volume are higher than the measured N2 gas adsorption pore volume
(Table 5.2). This dierence can be attributed to presence of macropores not measured by the
N2 gas adsorption method. Kaolinite show insignicant tplot micropore volume, whereas
micropores in SWy-2 constitutes about 7% of the measured N2 gas adsorption total pore
volume. The PSD for SWy-2 is bimodal with a major peak at 6080 nm and a minor
but prominent peak around 3 nm. KGa-1b have mostly a unimodal distribution with peak

123

Figure 5.1: N2 gas adsorption isotherms of as-received powders of Wyoming Montmorillonite SWy-2 and Kaolinite KGa-1b. The isotherm shapes indicate the kaolinite powders are
mostly macroporous, with negligible or non-existent micropores and mesopores. Montmorillonite powders shows presence of signicant volumes micropore, mesopores and macropores.
(See text for detailed description).

Figure 5.2: PSD of as-received source clay powders: Montmorillonite (SWy2) and Kaolinite (KGa1b) obtained from inverting N2 gas adsorption isotherm. SWy2 have bimodal
distribution with a major peak between 6080 nm with a minor peak around 3 nm. KGa1b
shows unimodal pore distribution with peak around 100 m.
124

Table 5.1: Modied BET eSSA,open specic surface area obtained from t-plot and micropore
specic surface area of the samples based on N2 adsorption isotherms. The uncertainty
estimate is only calculated from the mist of the regression t to the actual data point.
Other experimental uncertainties are not evaluated.
Sample
ID

Modied BET eSSA

(m2 /g)

KGa-1b
SWy-2
Mont-4K
Mont-6K
Mont-8K
Mont-10K

t-plot Open SSA

(m2 /g)

% SSA by
micropores (%)

As-received Source Clays

11.49 0.02
11.40 0.04
31.13 0.16
19.68 0.24
Pellets made from SWy-2
31.81 0.16
18.73 0.34
30.26 0.14
17.71 0.22
31.55 0.15
18.50 0.40
29.75 0.13
18.26 0.41

0.78%
36.78%
41.12%
41.47%
41.36%
38.62%

Table 5.2: Total pore volume of the pure clay samples measured by N2 gas adsorption and
MIP. Micropore volume obtained by using tplot technique on N2 isotherm data.
Sample
ID

N2 Total Pore
Volume (cm3 /g)

KGa-1b
SWy-2

0.049
0.068

Mont-4K
Mont-6K
Mont-8K
Mont-10K

0.063
0.060
0.062
0.056

MIP Total Pore t-plot Micropore

Volume(cm3 /g) Volume (cm3 /g)
As-received Source Clays

<0.0005
0.517
0.005
Pellets made from SWy-2
0.206
0.006

0.005
0.170
0.006
0.129
0.005

125

Micropore Volume
(% N2 pore volume)
0.00%
7.35%
9.52%
8.33%
9.68%
8.93%

around 100 nm (Figure 5.2).

Compacted Clay Pellets The compacted pellets does not show a denitive change
in eSSA with increasing compacting load (Figure 5.3(a)). The total pore volume, measured
by both N2 gas adsorption and MIP, decreases with compaction (Figure 5.3(b)). This agrees
with the observed compaction curves where porosity of mudrocks decreases with depth of
burial. The specic micropore volume is almost constant for all the montmorillonite clay
samples, both the compacted pellets and the as-received powder (Table 5.2).

(a) eSSA

(b) Total Pore Volume (cm3 /g

Figure 5.3: (a) Specic surface area obtained by modied BET technique on N2 adsorption
isotherm; (b) Total pore-volume, obtained from N2 gas adsorption (black) and MIP (blue)
techniques, on pure SWy-2 samples, as a function of with increasing axial load. eSSA does not
show a denitive trend with compaction. The total pore volume decreases with compaction.

The compacted pellets show similar bimodal PSD, obtained from N2 gas adsorption
technique, as the powder SWy2 (Figure 5.4(a)). The pore volume around 3 nm peak does
not change with the applied axial load . There is a signicant decrease in pore volume in the
compacted pellets around the 60 nm modal peak with the greatest decrease seen between
Swy2 powder and pellets compacted at 4000 psi (Mont-4K). With further compaction there
is a slight decrease in porosity in 60 nm range. The distribution patterns are consistent for all
126

the compacted pellets. PSDs obtained from MIP technique on the compacted pellets show
a dominant bimodal PSD with modes at 20 nm and around 10002000 nm (Figure 5.4(b)).
The total pore volumes associated with each modal peak and the modal poresizes decrease
as the compaction load increases; the pore volumes of uncompacted as-received powders are
the highest, followed by Mont-4K and Mont-8K and lowest in Mont-10K. The primary mode
poresize decreases from 10 m (104 nm) in asreceived powders to 2 m (2000 nm) in
Mont-4k to 1 m (1000) nm in Mont-10K. The secondary mode pore-size also shifts towards
lower poresizes as compaction load is increased.

(a) PSD from N2 gas adsorption

(b) PSD from MIP

Figure 5.4: (a) PSD of as-received SWy-2, compacted clay pellets Mont4K, Mont8K
and Mont10K obtained from N2 gas adsorption technique, (b) PSD of as-received SWy-2,
compacted clay pellets Mont4K, Mont8K and Mont10K obtained from MIP technique.

5.1.2

Discussion

The measured specic surface area of the source clays SWy-2 and KGa-1b (31.13 and
11.49 m2 /g, respectively) is in good agreement with the reported values by the Clay Mineral Society (31.82 and 10.4 m2 /g, respectively) (http://www.clays.org/SOURCE%20CLAYS/
SCdata.html). The presence of micropores agrees well with the reported micropore volume
in homoionic pure Wyoming Montmorillonite (Rutherford et al., 1997) and the reported
127

micropore volumes, between 0.0040.008 cm3 /g for dierent exchangeable cations, are in
the range of our measured micropore volumes for natural untreated montmorillonite (0.006
cm3 /g).
The PSD obtained from N2 gas adsorption and MIP techniques indicates multiple scale
of poresizes associated with smectitic clays, in accordance with the reported multiscale
structure in clays (Quirk & Aylmore, 1971; Cases et al., 1992; Neaman et al., 2003; Jullien
et al., 2005; Moyano et al., 2012). The measurements show that such building blocks are
associated with specic pore sizes. Based in our PSD results, we identify the following porosity associating with the dierent multiscale structure of clay minerals and their aggregates
(Figure 5.5):
Intratactoid porosity (fine mesopores < 4nm): Pores with diameter 3 nm are correlated with the pore spaces within tactoids formed by turbostractic stacking of clay
elementary units.
Intertactoid or intra-aggregate porosity (large mesopores and fine macropores
4 nm to 100 nm): pores with diameter around 50-100 nm account for the porosity
within aggregates formed by stacking of the tactoids. The MIP PSD suggests a pore
diameter of 20-25 nm. suggesting that the pore-throats associated with these pores
have a modal size of about 2025 nm.
Interaggregate porosity (micron sized): the mode around 2000 nm obtained from the
MIP data can be correlated with pores between the aggregates of clays
The total pore volume decreases with compaction load (Figure 5.3(b)) indicating pore
collapse by increasing overburden stress Compaction has the biggest eect on the inter
aggregate macropores. Better alignment of clay particles and compaction of pore with applied
stress decreases the interaggregate pores and also shifts their modal pore-size to lower values.
The intra-aggregate porosity is less aected from compaction, while intratactoid porosity
(pores with about 3 nm diameters) does not change with compaction. The intratactoid
128

Fundamental
I-S structure

Aggregates

Tachoids or Quasicrystals

Mode B

Mode A

dV/dlog(D) Pore Volume (cm/g)

Courtesy: Dr. Douglas

K. McCarty

Neaman et al. (2002)

Mode C
Moyano et al. (2012)

C
B

Figure 5.5: Comparison of the PSD obtained from N2 gas adsorption and MIP for pure
SWy-2 clays and compacted pellets. The three modes observed in these distribution can
be correlated with multi-scale porosity associated with clay structures especially montmorillonite. Mode A corresponds to intra-tactoid pores, Mode B corresponds to intertactoid or
intra-aggregate mesopores and Mode C corresponds to macropores between aggregates. (See
text for detailed description).
129

ne mesopores (< 4 nm) are shielded from compaction because of their small poresizes and
strong molecular forces that prevails in these pores. Equivalence can be drawn from studies
of collapse strength of cylindrical pipes with constant pipe wall thicknesses, which suggest
that the elastic buckling collapse pressure of such pipes is inversely proportional to the cube
of their diameter (Sharp, 1981). The smaller diameter of the pores increases their collapse
strength which shields them from compaction. These pores are in size range where molecular
forces dominate which also helps withstand overburden pressures. The eSSA of the samples
does not show any denitive change with compaction (Figure 5.3(a)) because the volume
of smaller pores, which have higher surface area per unit volume, remains unchanged with
compaction.
5.2

Pore Structure in Organic Matter

One of the important characterization parameters in evaluating self-sourcing reservoirs

such as gas shale plays is the total organic carbon (TOC), which is an indicator of the amount
of organic matter (OM) present in these mudrock reservoirs. The amount of OM present
is directly correlated with several rock properties, including total porosity (Passey et al.,
2010; Milliken et al., 2012), microporosity Chalmers & Bustin, 2007; Ross & Bustin, 2009),
total gas content (Strapo et al., 2010), methane sorption capacity (Lu et al., 1995; Cui
et al., 2009; Zhang et al., 2012), mechanical properties (Aoudia et al., 2010; Zargari et al., in
press), and the bulk rock elastic properties (Vernik & Liu, 1997; Prasad, 2001; Prasad et al.,
2011). Knowledge of the OM properties is critical to predicting the physical properties of
the bulk mudrock. Changes in OM properties, especially pore structure, during the thermal
evolution in sedimentary basins are not clearly understood. Recent studies using imaging
techniques have reported the presence of nanometer scale pore networks in the OM present
in thermally mature mudrocks (Walls et al., 2011; Curtis et al., 2012; Milliken et al., 2012;
Milliken et al., 2013, and citations within). The origin and evolution of OM pore structures
with thermal maturity is debated and is the focus of recent and on-going research (Schieber,
2010; Curtis et al., 2012; Loucks et al., 2012; Modica & Lapierre, 2012; Milliken et al., 2013,
130

Zargari et al., in press).

Several studies have investigated the pore structure of immature OM by using the subcritical N2 gas adsorption technique on both thermally immature oil shale lithology and
modern ocean bottom sediments. No evidence of presence of primary open porosity was
found within the immature OM (Tisot, 1962; Weiler & Mills, 1965; Schrodt & Ocampo,
1984; Mayer, 1994, 1999; Mayer et al., 2004; Han et al., 2006; Han et al., 2010b; Bai et al.,
2012. Laboratory pyrolysis experiments on immature OM rich oil shale samples indicated
that the development of an open pore structure only occurs at elevated temperatures above
400 C (Schrodt & Ocampo, 1984; Han et al., 2010b; Bai et al., 2012; Elbaharia, 2012; Tiwari et al., 2013). These studies conrms that welldeveloped primary porosity is absent in
OM prior to or with sediment deposition; and that OM porosity is secondary and develops
during the late diagenesis processes. Imaging studies on natural mudrocks with dierent
thermal maturities indicates that porosity development in OM starts when the thermal maturity of the rock moves from the oil to the gas window (Schieber, 2010; Curtis et al., 2011;
Curtis et al., 2012). Other studies report development of porosity in OM at the beginning
of oil window (Loucks et al., 2012; Modica & Lapierre, 2012). Zargari et al. (in press) have
postulated that visual techniques are inadequate to image lled pores that have low optical
or density contrast with their surrounding material. However, there is no observed denite
correlation between the abundance and morphology (shape and size) of OM porosity with
increasing thermal maturity (Curtis et al., 2012; Milliken et al., 2013).
Most studies on OM hosted pore structure of thermally mature mudrocks are based on
imaging techniques that provide only a qualitative and visual description. These studies
are valuable in understanding organic rich mudrock petrology, however, additional eort is
required for quantication of total porosity and pore-size distribution (e.g., Curtis et al., 2012;
Milliken et al., 2013). Other methods are needed to understand the quantitative relationships
and contributions of OM to the total porosity and the PSD evolution in sedimentary basins.

131

N2 gas adsorption technique was used to obtain quantitative pore structure parameters,
including the specic surface area and PSD of OM with dierent thermal maturity. To
accomplish this, natural mudrock samples with varying thermal maturity and bulk composition were subjected to a variety of treatments including selectively removing or isolating the
organic components. Comparing PSDs before and after removing OM allows estimates of
the relative proportions of pore volumes between the organic and inorganic components for
a specic size range in these mudrocks. The details of the samples and sample preparations
are discussed in Section 5.2.1 and Section 5.2.2, respectively. The pore structure results
obtained N2 gas adsorption are presented in Section 5.2.3 and the insights obtained from
the experimental ndings to understanding the porestructure of OM with dierent thermal
maturity are discussed in Section 5.2.4.
5.2.1

Samples

Seven OM rich samples from ve geographic locations were chosen for detailed analysis
based on contrasting levels of thermal maturity and OM type. The samples were analyzed
to determine composition and the nature of the pore structure. The details of the samples
are listed below
Natural Solid Hydrocarbons: Two samples from Utah Gilsonite veins and from Gulf of
Mexico (GOM) solid bitumen veins were analyzed. The Utah Gilsonite occur naturally
in parallel, near-vertical veins in the Eocene Uinta Formation of the Uinta Basin of
Utah and Colorado (Monson & Parnell, 1992). The sample for this study was obtained
from Bonanza mines, Utah, mined by American Gilsonite Co. The proposed model of
origin of these gilsonite veins in the Uinta Fm. suggests an initial creation of large scale
fracture systems by elevated pore pressures generated by hydrocarbon generation in the
underlying Green River Fm., followed by injection and solidication of expelled liquid
hydrocarbon from Green River into these fracture systems. Solidication of this liquid
hydrocarbon to gilsonite was caused by subsequent degassing of the volatile components

132

from the liquid hydrocarbons due to pressure release, leaving behind the heavier residue
(Monson & Parnell, 1992). The degassing of the volatiles was substantiated by presence
of numerous small bubble shaped pores revealed by SEM analysis ((Monson & Parnell,
1992). Thermal maturity studies suggest that the gilsonite fall into the thermally
immature category (Anders et al., 1992; Monson & Parnell, 1992). The second solid
bitumen sample from the Gulf of Mexico is also postulated to have formed in GOM
are through similar processes (Han et al., 2010a) as the gilsonite.
Oil Shale Samples: Two oil-shale samples from Uinta Basin, one natural and another
cooked at 350 C (Elbaharia, 2012), were used to understand the eect of temperature
(or maturity) in evolution of OM pore structure. The natural sample, Mahogany Fm.
collected at White River Oil Shale Mines in Utah, is composed mostly of carbonate
minerals and immature Type I OM (22 wt.% TOC) with minor amount of clays (Elbaharia, 2012) . An aliquot of the natural sample was heated to 350 C at ambient
pressure conditions, and hereafter, will be referred as the cooked Mahogany formation sample. Pore-structure characterization was also performed on isolated OM from
Green River Fm.
Thermally mature mudrocks: Four samples from two dierent thermally mature gas
shale plays: a single sample SS115 (Table 2.1) from Silurian gas-shale in Eastern
Europe and three samples SS2A, SS2B and SS2C from the Jurassic Haynesville gasshale in East Texas, USA were used to study the pore-structure in thermally mature
OM. The composition of the natural samples obtained from QXRD and LECO TOC are
shown in Table 5.3. The dominant constituent of these samples are clay minerals and
will inuence the pore-structure obtained from these natural mudrocks. To understand
the pore-structure of the OM part only in these natural mudrocks, two dierent sample
preparation techniques were used, which are discussed in the following subsection.

133

Table 5.3: Mineral and OM content (vol.%) and RockEval II parameters of the natural samples used in OM pore-structure
characterization study. The concentration values reported are volume.% normalized to 100% obtained after combining the wt.%
of mineral phases (from QXRD) and wt.% of OM (from LECO analysis) and converting them to vol. %. The conversion factor
used for the conversion of TOC to wt.% OM is 83% (Yen & Chilingar, 1976). The density of OM assumed to be 1.25 g/cm3 .
The data for Mahogany Shale is obtained from (Elbaharia, 2012).
Samples

Total
K
Plag Cal Dol
Qtz
Carb
spar

Mahogany Fm. 12
Eastern European Shale Gas
3
5
SS115 42
Haynesville Fm.
0
5
SS2A 28
0
7
SS2B 13
0
6
SS2C 26

%S
Pyr AnhyOrg Kao Chl I+S
in
I-S

24

31

0.7

44.0

1.5

15.6

29

10
33
12

0
0
0

11
33
13

1.0
1.9
1.0

0
1
0

10.0
16.0
11.6

0
0
2

2
5
3

43
24
38

134

SUM LECO
S1
CLAY TOC

S2

S3

22.00

>225

<5

32

6.43

0.06 0.30 0.29

<5
<5
<5

45
29
43

3.21
6.51
3.77

1.47 1.57 0.39

2.32 1.43 0.54
0.10 0.65 0.15

5.2.2

Sample Preparation

Two dierent sample preparation techniques, separation of OM and removal of OM are

used to understand the OMhosted porestructure in natural mudrocks. The details of the
procedure are discussed below:
Separation of OM The OM separation was done using two dierent protocols. The
separation of OM from samples SS115 and SS2A (Table 5.3) was done by commercial
kerogen isolation process by GeoMark Research, Ltd. The eectiveness of the commercial
kerogen isolation process was assessed with bulk powder Xray diraction analysis (XRD)
(Figure 5.6). An example bulk powder XRD pattern from the commercially isolated OM
samples shows abundant mineral phases indicating that it is not pure OM but concentrated
OM (Figure 5.6(a)). Hereafter, the samples from the commercial isolation process will be
referred to concentrated OM.
Similar observation about the inadequacy of commercial isolation process to yield pure
OM was made by Ibrahimov & Bissada (2010) who suggested an alternative fully automated,
conservative, closedsystem methodology under an inert atmosphere for kerogen isolation.
Kerogen isolation using the Ibrahimov & Bissada (2010) methodology on a dierent sample
from Haynesville (SS2B) was done by Dr. Adry Bissada. The XRD pattern of the separated
OM from the Haynesville shale (SS2B) conrms the almost complete removal of mineral
phases with traces of pyrite and rutile (Figure 5.6(b)). Hereafter, this pure separated OM
obtained by the Ibrahimov & Bissada (2010) technique will be referred to isolated OM.
Three isolated OM samples were studied: two samples of isolated OM from the Green River
Formation (courtesy Dr. Dehua Han, from University of Houston) and isolated OM from
SS2B.
Removal of OM by bleaching The acid digestion treatment of dissolving mineral
matter during the kerogen isolation process does not chemically alter OM structure, nor does

135

(a) Concentrated OM

(b) Isolated OM SS2B

Figure 5.6: Bulk XRD pattern of (a)Concentrated OM obtained from commercial kerogen
isolation (b) Isolated OM from sample SS22 using method of Ibrahimov & Bissada (2010).
Note the suggests abundance of inorganic mineral phases in concentrated OM. The broad
peak on the XRD pattern of isolated OM represents the amorphous OM background and
lack of peaks indicate the purity of the OM. Q = Quartz; Cal = Calcite; Py = Pyrite; Pyh
= Pyrrhotite (FeS) Rut = Rutile.

136

it generate new organic compounds (Vandenbroucke & Largeau, 2007; Ibrahimov & Bissada,
2010). However, repeated ushing of chemicals during the process may potentially damage
the porestructure in these mechanically soft OM. An alternative indirect method of removing the OM from the natural sample and comparing changes in porestructure parameters
between natural gas shale samples and upon OM removal was used for OM pore-structure
characterization. Similar methodology of removing OM for porestructure characterization
was previously applied for immature oil shales Tisot (1962) and for recent ocean bottom sediments and soils (Weiler & Mills, 1965; Mayer, 1994, 1999; Mayer et al., 2004). The thermally
mature gasshale samples (SS115, SS2A, SS2B and SS2C, Table 5.3) were crushed in
a mortar to pass through a <420 m (<40 mesh) sieve. A louvered laboratory splitter was
used to split each of the samples into two mineralogically and chemically equivalent portions
(McCarty, 2002). Removal of OM was done on one split using standard organic removal
protocol for soil mineralogical analysis (Anderson, 1963; Jackson & Barak, 2005). OM was
removed by oxidation: the sample was treated with a 6% solution of sodium hypochlorite
(NaOCl) solution adjusted to a pH of 8.59 at 60-70 C, followed by washing with sodium
chloride solution. The eectiveness and the impact of bleaching on the composition of the
samples were evaluated by RockEval II and XRD analysis of the samples before and after
bleaching. The LECO TOC (Table 5.4) results show that bleaching removes the majority of
OM from the samples. However the RockEval II S2 pyrograms indicate some high molecular number bituminous organic matter remains in the bleached samples (Figure 5.7(a) and
Figure 5.7(b)). The XRD patterns of the natural and the bleach sample (Figure 5.8) suggest
that the bleaching process also oxidizes and removes pyrite from the samples. All other
inorganic minerals remain unaected by the bleaching process.
5.2.3

Results

Pore structure parameters, such as eSSA and PSD were obtained using the subcritical
N2 gas adsorption experiments. The details about experimental procedure and inversion of
raw data were presented previously in Section 4.2. A degassing temperature of 100 C used
137

Table 5.4: RockEval II data of the Haynesville Formation samples before and after removal
of OM by bleaching.
Sample LECO TOC

S1

S2

S3

Haynesville
SS2A Natural
SS2A Bleached

3.04
0.18

1.47 1.57 0.39

0.12 0.43 0.63

SS2B Natural
SS2B Bleached

5.54
0.48

2.32 1.43 0.54

0.11 0.40 1.24

(a) SS2A

(b) SS2B

Figure 5.7: Comparison of RockEval II S2 pyrograms of natural and bleached sample of (a)
SS2A, (b) SS2B. The bleaching removes majority of OM from the samples but some high
molecular number bituminous organic matter remains in the bleached samples.

138

XRD Data Before and After Bleach Treatment

Q

Sample: SS1-15

Q
Q
Q

Clay

Q
Q

Bleached

Haynesville Sample
Q
SS2-C
Q

Q
Q
Q

Py

Py
Py
Cal

Q
Cal

Clay

Intensity

Untreated

Cal

Py

Py

Py

Clay

Bleached
Clay

Intensity

Untreated

2 ()

Figure 5.8: Bulk XRD patterns of untreated natural (blue) and bleached (red) sample from
(top) SS115 and (bottom) SS2C. Note the lack of pyrite peaks in the bleached samples.
Q = Quartz. Cal = Calcite Py = Pyrite.

139

for separated OM and thermally immature samples while 200 C used for clay rich thermally
mature shales samples. Generally 23 g of sample mass was used, except for the separated
OM samples where sample mass were limited to availability (0.1-0.5 g).
Thermally Immature Samples The naturally occurring solid bitumen from Utah
Gilsonite (Figure 5.9(a)) and GOM bitumen (Figure 5.9(b)) showed extremely low quantities
of adsorbed N2 , within the detection limit of the instrument, indicating a nonporous nature
of the samples. The reversible isotherm prole for the natural and cooked Mahogany shale
samples represents a monolayer-multilayer adsorption mechanism of the gas on the open
and stable solid surface and indicates that the samples are dominantly macroporous or nonporous (Figure 5.9(c)). The isolated Green River OM samples show similar isotherm proles
but with hysteresis at higher relative pressures (P/P 0 >0.75) indicating presence of larger
mesopores and ner macropores in these samples (Figure 5.9(d)). The extremely low eSSA,
total tplot micropore volume and total pore volume (Table 3) and PSD (Figure 5.10) of the
Mahogany shales indicate absence of micropores and mesopores in those samples. Cooking
the Mahogany shales to articially mature the OM did not show any development of porosity
in pores below 200 nm (Figure 5.10). The isolated Green River OM shows higher eSSA and
total pore volume compared to the natural Mahogany shales but the t-plot suggests absence
of any micropores. Comparison of PSD of the isolated OM from Green River with the
natural organic rich Mahogany shales indicate greater volume of mesopores in the isolated
OM compared to their natural equivalent (Figure 5.10).
Thermally Mature Samples The isotherm proles (Figure 5.11) of concentrated OM
from SS115 and SS2A and isolated OM from SS2B suggest that the samples are dominated by mesopores (indicated by the hysteresis pattern) and macropores (indicated by the
absence of plateau at higher P/P 0 >0.95). The concentrated OM isotherm does not show
strong forced closure of the desorption branch due to the Tensile Strength Eect (Groen
et al., 2003). The isolated OM from SS2B do show a prominent forced closure of the
140

(a) UTAH Gilsonite

(b) GOM Bitumen

(c) Mahogany Fm. Shales

(d) Green River Isolated OM

Figure 5.9: N2 adsorption isotherm (at LN2 temperature) for the immature OM samples.
(a) Utah Gilsonite, (b) Gulf of Mexico (GOM) Bitumen, (c) Mahogany Fm. shale natural
and Mahogany shale cooked at 350 C, (d) Green River (GR) isolated OM.

141

Figure 5.10: PSD of the immature OM samples.

desorption branch, indicating the absence and presence of signicant amount of pore volume
with diameter <4 nm, respectively. The concentrated OM samples from samples SS115 and
SS2A have eSSA of 10.58 m2 /g and 5.51 m2 /g respectively. They show similar PSD patterns (Figure 5.12(a)) as compared to isolated Green River OM , however the concentrated
OM samples shows presence of pores between 510 nm size which are practically absent in
Green River OM. The tplot (Table 5.5) results also suggest insignicant micropore (<2
nm) volume in the concentrated OM samples. The SS2B isolated OM sample has a very
high eSSA of 126.61 m2 /g and very high total pore volume of 0.365 cm3 /g compared to
the concentrated OM samples. It also has signicant amount of micropores and the tplot
micropore volume is 0.006 cm3 /g. The PSD shows similar distribution shape as the other
concentrated and isolated OM samples but the pore volumes are signicantly larger for all
pore sizes (Figure 5.12(b)).

142

Figure 5.11: (Left)N2 gas adsorption isotherm (at LN2 temperature) for the concentrated
and isolated OM. (Right) Zoomed image showing detailed isotherms for concentrated and
isolated OM from Green River (GR), SS11 and SS21.

(a) PSD of concentrated OM from SS115 and SS2A

(b) PSD of isolated OM from SS2B

Figure 5.12: (a) PSD of the concentrated OM from SS115 and SS2A. PSD from isolated
Green River OM is shown for comparison. (b) PSD of the isolated OM from SS2B.

143

Table 5.5: Pore structure parameters obtained by gas adsorption technique on samples used for OM pore structure characterization experiments.
Natural
eSSA
BET N 2

m2 /g

tplot
Micropore
Volume
cm3 /g

Natural Solid Bitumen

Utah gilsonite
0.10
<0.0005
GOM Bitumen
0.27
<0.0005
Mahogany Fm. Shale, Uinta Basin
Natural
0.43
<0.0005

Cooked at 350 C
0.46
<0.0005
Isolated OM, Green River
GR I

GR II

Eastern Europe Silurian Shale

SS115
21.16
0.007
Haynesville
SS2A
15.45
0.003
SS2B
26.22
0.007
SS2C
18.65
0.003

OM separates
Total
Pore
Volume
cm3 /g

BET N 2

Bleach Test
Total
Pore
Volume
cm3 /g

BET N 2

m2 /g

tplot
Micropore
Volume
cm3 /g

m2 /g

tplot
Micropore
Volume
cm3 /g

Total
Pore
Volume
cm3 /g

<0.0005
0.001

0.002
0.002

5.62
4.52

0.001
<0.0005

0.026
0.030

0.027

10.58

<0.0005

0.061

10.25

0.001

0.037

0.035
0.032
0.036

5.51
126.61

<0.0005
0.006

0.035
0.365

16.30
9.54
14.07

<0.0005
<0.0005
<0.0005

0.062
0.037
0.047

eSSA

144

eSSA

The removal of OM matter in samples SS115 (Figure 5.13(a)), SS2A (Figure 5.13(b)),
SS2B (Figure 5.13(c)) and SS2C (Figure 5.13(d)) by treatment with buered NaOCl solution resulted in decrease in tplot micropore volume (Table 5.5) but there is an increase
in total pore volume. The eSSA in general, decreases with removal of OM except for sample
SS215. Comparison of PSD for the samples before (untreated samples) and after removal of
OM by bleaching (Figure 8) show reduction of pore volume network below 3.0 nm diameter
and increase in pore volume for higher pore diameters. All the bleached samples show a
characteristic PSD bimodal shape with a prominent peak around 3 nm pore size.
The PSD obtained from the separated OM are not consistent with the PSD results obtained from its natural equivalent. The immature Green River isolated OM shows presence
of pores below 50 nm whereas the OM host mudrock from Mahogany shales do not have any
porosity below 50 nm (Figure 5.10). The concentrated OM from SS115 and SS2A have insignicant pore volume below 5 nm, but the results from the natural samples, before and after
bleaching indicate presence of pores below 5 nm (Figure 5.13(a),Figure 5.13(b)). The isolated
OM from SS2B have exceptionally high volumes of micropores and ne mesopores (Figure 5.13(c)) compared to the pore volumes measured in the natural sample (Figure 5.13(c)).
5.2.4

Discussion

The naturally occurring solid bitumen from Utah Gilsonite veins and from GOM do not
show any open porosity below 200 nm pore diameters. The extremely low porosity and
adsorption of these samples are consistent with the glassy quenched nature of their origin.
Monson & Parnell (1992) reported solidication of expelled liquid hydrocarbon as the origin
of Utah Gilsonite. They also reported the presence of gas escape pores with dimensions
from about 1 m to >1 mm diameter, which are outside the measurement range of N2 gas
adsorption technique. The GOM bitumen is low maturity organic matter mobilized from
deeper source rocks that does not occur as pore uid in any host sedimentary rocks but as
discrete softer formation in the sedimentary basin (Han et al., 2010a). Overburden loading
and creeping over geologic time might have destroyed any pore-structure in these viscoelastic
145

(a) SS11

(b) SS21

(c) SS22

(d) SS23

Figure 5.13: Comparison of PSD between the natural sample and its bleached equivalent for
(a) sample SS115 (b) sample SS2A (c) sample SS2B (d) sample SS2C. PSD from the
OM separates from these samples (blue) are shown for comparison.

146

bitumen bodies, similar to that of salts.

The OM rich Mahogany shales from Uinta Basin show similar absence of open micropore and mesopores (Figure 5.10) and extremely low surface area and pore volume. The
measured low eSSA (0.46 m2 /g) and total pore volume (0.003 cm3 /g) of Mahogany Fm.
shales is consistent with the reported specic surface area (0.58-0.79 m2 /g) and total pore
volume (0.0005 cm3 /g) from a previous study (Tisot, 1962). Cooking of the samples at
350 C does not created any new signicant pore volume below 200 nm size range. Elbaharia
(2012) reported development of porosity with pore sizes of about 270 nm from FESEM study
during heating experiments of these samples. The reported poresizes are outside the detection limit of the nitrogen gas adsorption. Mercury intrusion porosimetry should be used
to record any changes in the macropores size. Previous studies on physicochemical changes
of oil shale organic matter from Green River suggest that with increasing thermal maturity
(timetemperature integral) the original OM decomposes in successive stages producing rst
bitumen, then immiscible oil, followed by gas and pyrobitumen (Lewan, 1994). Schnackenberg & Prien (1953) and Johnson et al. (1975) reported an intermediate stage where the
kerogen softens to a rubber like material rubberoid, showing elastic and swelling properties, prior to development of bitumen. Porestructure studies on oilshales during retorting
have reported an initial decrease or constant specic surface area and total pore volume of
oil shales on heating up to 400 C followed by increase in specic surface area and total pore
volume beyond 400 C (Schrodt & Ocampo, 1984; Han et al., 2006; Han et al., 2010b; Bai
et al., 2012). This is attributed to the pore blockage by the remobilized bitumen upon cooling
into the pores (Schrodt & Ocampo, 1984; Han et al., 2006). No change in porosity and eSSA
in the cooked Mahogany samples may be due to the similar pore blockage of the mesopores
and micropores by the condensed remobilized bitumen. Zargari et al., in press reported
similar evidences on retorted Bakken samples, where the bitumen exhibited ow textures
and condensation in the conduits and microcracks. Using XRay CT imaging, Elbaharia
(2012) showed healing of a natural fracture in an oil shale after heating, which supports the

147

solidifying and hardening behavior of bitumen after being melted and remobilized at high
temperature. Development of open pore systems happens at higher temperature (Tiwari
et al., 2013) with successive decomposition of bitumen, explaining the reported increase of
specic surface area and total pore volume beyond 400 C (Schrodt & Ocampo, 1984; Han
et al., 2006; Han et al., 2010b; Bai et al., 2012). A combination of nitrogen gas adsorption
techniques and mercury intrusion on a systematic pyrolyzed set of samples is recommended
to fully understand the evolution of porespace with maturation of organic matter.
Comparison of poresize distributions between the bleached samples and its natural
equivalent shows a reduction of pore volume at sizes below 3.0 nm and increase in pore
volume for larger pore sizes. The bimodal PSD shape of the bleached samples with a modal
peak around 3.0 nm is similar to the characteristic signature of clay microstructure (Section 5.1.2). The increase of pore volume at higher pore diameters is due to removal of
porelling OM from the clay mesopores. Previous studies on removal of OM from immature oil shales (Tisot, 1962) and ocean oor sediments and soils (Weiler & Mills, 1965;
Mayer, 1999; Mayer et al., 2004) have reported increase in specic surface area and pore
volumes of the samples after organic removal. The increase in pore volume is attributed to
the removal of OM lling the pores of the inorganic matrix. The reduction of pore volume
at lower pore diameters after bleaching indicate that the OM have ne micropore mesopore
network within their matrix which are destroyed with removal of OM. An example ionmilled
FESEM image (Figure 5.14) from a Eastern Europe Silurian shale sample from a dierent
well location shows similar textural arrangement where the OM hosting a sponge like pore
system completely lling the space formed by the arrangement of the phyllosilicate grains.
Similar textural relation between the clay microstructure and OM for other mudrocks was
reported by Schieber (2010), Milliken et al. (2013) and Zargari et al., in press. Removal of
OM will destroy the smaller pores hosted within the OM but will simultaneously open the
mesopores with the clay microstructure. The change in eSSA will depend upon the relative
proportions of the pores in the OM matter that are destroyed compared to the pores in clays

148

that are created due to OM removal. Comparison of total pore volumes with poresize less
than 5 nm (t-plot micropore volume plus BJH pore volume between 2 to 5 nm) for the natural samples containing both clay-hosted and OM hosted pores and the bleached samples with
only clay pores suggests that OM-hosted pores account for 26-77% of the total pore volume
with pore-size less than 5 nm (Table 5.6). However, the OM in natural rocks blocks some
of the clay ne mesopores which are otherwise open in bleached sample, thereby increasing
the relative proportion of OM hosted pore volume.

Figure 5.14: FESEM image of ion milled sample from Eastern European Silurian Gas Shale
showing microstructural feature where OM hosting a sponge like pore system lls the intra
aggregate clay porosity. Image Courtesy: Dr. Kitty Milliken, Bureau of Economic Geology,
Jackson School of Geosciences, University of Texas at Austin, Austin, Texas

149

Table 5.6: Relative proportion of pores with < 5 nm pore size in OM compared to clays, obtained from comparison of tplot
micropore volume and BJHHJ ne mesopore (25 nm)volume between natural samples and bleached samples. The bleached
sample with only have clay pores while the natural samples will have both the clay pore and OM pores.
Natural
Sample

tplot
Micropore
Volume
cm3 /g

Bleached

BJHHJ
Total Pore
Pore Volume
Volume
25 nm
<5 nm
cm3 /g
cm3 /g

Eastern European Shale Gas

SS115
0.007
0.004
Haynesville
SS2A
0.003
0.004
SS2B
0.007
0.005
SS2C
0.003
0.005

tplot
Micropore
Volume
cm3 /g

Proportion of

BJHHJ
Total Pore
Pore Volume
Volume
25 nm
<5 nm
cm3 /g
cm3 /g

pore volume<5 nm
hosted in
Clay
OM
%
%

0.011

0.001

0.003

0.003

34

66

0.007
0.012
0.008

0.000
0.000
0.000

0.005
0.003
0.004

0.005
0.003
0.004

74
23
55

26
77
45

150

Pore-structure parameters obtained from separated OM cannot be reliably used to understand OM hosted pore-structures in natural mudrocks. PSD obtained the separated OM
are not consistent with the PSD of its natural equivalent. The kerogen isolation process
does not alter the OM chemical structure (Vandenbroucke & Largeau, 2007; Ibrahimov &
Bissada, 2010) but the rigorous chemical treatment involving multiple ushing with acids
physically damages pore-structure of the mechanically soft OM. The OM removal experiments using bleaching procedure provides important information about OM hosted pore
structure in natural mudrocks, however the bleaching does not remove all of the OM present
(Figure 5.7, Table 5.4). An alternate OM removal procedure using low temperature plasma
ashing technique (Mayer, 1994; Kale et al., 2010) is recommended, since this process does
not required repeated ushing of the sample with chemicals.
5.3

Conclusion
A detailed porestructure analysis of the two most important mudrock endmember

components, I+S clay minerals and organic matter, using N2 gas adsorption technique were
presented. Experiments were done on pure source clay minerals from Clay Mineral Society
and on pellets compacted with dierent uniaxial load from the as-received clay powders.
Porestructure in OM end member with varying maturity was studied using separate sample
treatment procedures, like OM isolation and OM removal by bleaching, on dierent thermally
mature mudrock samples. The following summarizes the observations:
1. Montmorillonite has micropores and ne mesopores with a characteristic pore-size of
3 nm that is missing in kaolinite.
2. Illite+smectite group of clay minerals show multiple scales of porestructures:
Intratactoid pores (3 nm) formed by stacking of elementary unit cells in tactoids.
This 3 nm pores cannot be detected by most frequently used current imaging techniques used for mudrock pore characterization because it is below the resolution
oof the techniques.
151

Intraaggregate pores (20-100 nm) formed by stacking of tactoids in an aggregate,

Interaggregate larger macropores (>1000 nm) between the aggregates.
3. Compaction of montmorillonite shows
Decrease in pore volume and reduction of pore-size in the interaggregate macropores
Some reduction in pore volume of the intraaggregate mesopores and ne macropores
No change of intratactoid micropore and ne mesopores.
4. Naturally occurring thermally immature solid bitumen samples does not show any pore
structure below 200 nm indicating their glassy nature.
5. Immature Type I OM rich Mahogany shale has negligible surface areas and pore volumes below 200 nm. Articial thermal treatment of the sample at 350 C still does
show not any evidence of increased porosity.
6. Removal of OM from thermally mature shales showed signicant reduction of pore
volume network below a diameter of 10 nm. This reduction of pore volume is indication
of pores associated with thermally mature OM.
7. The relative abundance of micro and ne mesopores in natural mudrocks is controlled
by both, the clay and the OM contents. Some of the OM in natural mudrocks occurs
as poreblocking material within the clay mesopore structures.
8. Separated OM from mature gas shale cannot be reliably used for OM porestructure
characterization due to the potential textural damage during kerogen isolation process.
It is concluded that the clay minerals and the OM, depending upon their thermal maturity, can host micropore and mesopore network system within them. In order to fully

152

understand the controls these end members have on natural mudrock porestructure, a systematic study correlating the porestructure with detailed quantitative mineralogical and
OM characterization on a natural mudrocks is necessary. This follows in Chapter 6.

153

CHAPTER 6
PORESTRUCTURE IN NATURAL MUDROCKS

Possibly many may think that the deposition and consolidation of

negrained mud must be a very simple matter, and the results of little
interest. However, when carefully studied experimentally it is soon found
to be so complex a question, and the results dependent on so many
variable conditions, that one might feel inclined to abandon the inquiry,
were it not that so much of the history of our rocks appears to be written
in this language.
Henry Clifton Sorby
The presence of micropore and mesopore pore networks within both organic and inorganic components of a mudrock, particularly clay minerals were documented in the previous
chapter. However, a systematic study comprehending the control of composition of the rock
on the pore-size distribution is critical in order to fully understand the mudrock pore networks. To achieve this, three dierent case studies correlating the porestructure results
obtained by N2 gas adsorption analysis with the detailed quantitative mineralogical and
OM characterization of a suite of well characterized mudrocks samples are presented in this
chapter. The three case studies include
1. an assorted set of mudrock samples from dierent locations and composition (Section 6.1),
2. a suite of 20 samples from a well drilled in the Niobrara Formation (Section 6.2), and
3. a suite of 16 samples from a well drilled in the Haynesville Formation (Section 6.3).

154

6.1

Case Study I: Assorted set of Mudrocks

Mudrock samples from sample set ASS, collected from dierent locations with dierent

mineralogy and clay content, were studied to study the eect of clay on poresize distribution
of mudrocks. The details of the samples are given in Section 2.2.5 (Page 23) and the
mineralogy of samples is presented in Table 2.4, Page 30.
6.1.1

Results

N2 Isotherm Profiles The natural mudrocks from sample set ASS show a wide variety
of isotherm shapes (Figure 6.1). The Pierre Shale and the Cox Argillite show Type IIB
isotherm shapes similar to SWy2. The Silver Hill Cambrian Shale shows a close to Type IV
isotherm shape without a steep slopes in the P/P 0 range of 0.98 - 1.00 and H4 hysteresis
pattern indicating that this sample has micropores and mesopores and no signicant amount
of macropores. The Mancos B Shale, the Woodford Shale and the Middle Bakken show
Type II isotherms with very low amount of gas adsorbed compared to other shales and lower
hysteresis indicating a predominantly macroporous nature of the samples. The North Sea
Shale shows extremely low adsorbed volume at low and intermediate relative pressures but
high adsorbed volume and a prominent hysteresis at higher relative pressure. The North Sea
Shale has no micropore and ne mesopores, but has a signicant volume of larger mesopores
and macropores.
SSA and Pore Volumes The North Sea Shale, the Mancos B shale and Middle Bakken
show similar SSA obtained by modied BET inversion and tplot inversion on the isotherm
data (Table 6.1). The Silver Hill Cambrian Shale, Pierre Shale, Woodford Shale and Cox
Argillite show signicant dierence between these two SSA. The dierence in these values
is correlated to surface area associated with micropores since tplot typically provides the
open SSA from mesopores, macropores, and external specic surface area. The Middle
Bakken, the North Sea Shale and the Mancos Shale show no or negligible tplot micropore
volumes (Table 6.1). Silver Hill Cambrian Shale shows highest proportion of micropores
155

Figure 6.1: N2 adsorption-desorption isotherms obtained for the assorted set of mudrocks
samples. Based on the isotherm shapes, the nature of the pore-structure for the dierent
samples is inferred (See text for detailed description).

156

constituting about 17% of the total N2 adsorption measured pore volume.

Table 6.1: Pore-structure parameters for the assorted set of natural shale samples, obtained
using N2 gas adsorption techniques. eSSA is obtained from modied BET technique. Open
SSA and micropore volume are obtained using tplot inversion. The uncertainty estimate
is only calculated from the mist of the regression t to the actual data point. Other
experimental uncertainties are not evaluated.
Sample
ID
Silver Hill Cambrian Shale
Middle Bakken Shale
Pierre Shale
Cox Argillite
Woodford Shale
North Sea Shale
Mancos B Shale

eSSA

Open SSA

Total PV

(m2 /g)

(m2 /g)

(cm3 /g)

Micropore
vol.
(cm3 /g)

31.57(18)
0.98(01)
13.93(02)
18.09(03)
1.93(01)
9.60(03)
2.82(03)

20.26(44)
1.02(00)
9.17(02)
10.34(04)
1.61(02)
10.11(03)
2.82(08)

0.030
0.006
0.040
0.036
0.006
0.062
0.014

0.005
<0.0005
0.002
0.004
<0.0005
<0.0005
<0.0005

Micropore contribution
% eSSA

% total PV

35.83
0.00
34.17
42.84
16.58
0.00
0.00

16.67
0.00
6.05
10.76
0.00
0.00
0.00

PSD Figure 6.2 show the PSD of all mudrock samples. The Silver Hill Cambrian shale
is dominated by ne mesopores with modal diameter around 23 nm. The Pierre Shale, the
Cox Argillite and the Mancos B shale show similar bimodal PSD like the montmorillonite
with a very prominent peak at 3 nm. The North Sea Shale has the highest porosity and it
does not show any ne mesopores. The dominant pores are large mesopores and macropores
with a modal peak at 20 nm. Middle Bakken and Woodford shale have very low porosity in
1.7200 nm range and are dominated by macropores.
6.1.2

Discussion

All samples with signicant amount of I+S group of clays show prominent peak around 3
nm in PSD (Figure 6.3), similar to the SWy2 powders and pellets. As indicated previously in
Section 4.3.4, the peak area under the curve between any two pore diameters of dV/d(logD)
plots is equal to the partial porosity for that particular pore diameter range. The pore
volume of the 23 nm pores in the natural mudrocks are higher in samples with higher I+S
group clay content (Figure 6.3). SEM image of the Cox Argillite (Sarker, 2010) reveals

157

Figure 6.2: PSD of the natural mudrock samples obtained from N2 gas adsorption technique.
Note the yaxis scale for the North Sea Shale is dierent from other plots.

158

presence of clay dominated continuous matrix with distributed angular detrital grains, such
as dolomite and quartz (Figure 6.4(a)). The continuous clay matrix controls the main pore
size distribution of the mudrocks. The only exception to the trend is the North Sea Shale
sample , with about 14% Opal C/CT, that shows small amount of 3 nm pore volume in spite
having 56% I+S group clays. The dominance of 1020 nm pores in this sample (Figure 6.1)
correlate well with reported presence of pore of 10 nm diameters in opals (Feoktistov et al.,
2001). Figure 6.4(b) shows a SEM image from the Middle Bakken sample demonstrating the
extensive calcite cementation in the sample destroying porosity. Hence the Middle Bakken
shale show very low porosity with mostly large mesopores and macropores.

Figure 6.3: PSD of the natural mudrocks up to 10 nm pore diameter, colorcoded by clay
content with hotter colors indicating higher clay content. The legend shows the I+S group
clay content in wt.%. The area under the curve with a particular size range is proportional
to the absolute pore volume for that size range. Samples with lowest clay content (blue)
do not have a 3 nm pore-body distribution peak and the 3 nm pores become increasingly
abundant with increasing clay content. The North Sea Shale sample does not show any 3
nm porosity in spite having 56% of I+S clays.
The compaction experiments on dry powders, presented in Section 5.1, showed presence
of incompressible 3 nm pore-size intratactoids that remains unaected by compaction
process. However, articial compaction of dry clays is very dierent from the natural com-

159

(a) Cox Argillite

(b) Middle Bakken

Figure 6.4: (a) SEM backscatter image of the Cox Argillite (Sarker, 2010). The rock shows
clay dominated continuous microstructure with distributed angular detrital grains, such as
dolomite and quartz. (b) SEM backscatter thin section image from the Middle Bakken
sample (Sarker, 2010). The microstructure in this sample is controlled by extensive calcite
cementation destroying porosity.

paction process that occurs in sedimentary basins. Natural mudrocks contain water while
the articially compacted clay samples in this study did not contain water. In spite of the
dierence in compacting environments, the PSD of naturally compacted clay rich mudrocks
show similar porestructure as obtained from the dry compaction experiment in laboratory
(Figure 5.4(a)). Figure 6.5 shows a plot of BJH cumulative pore volume between 2 to 5 nm
poresize obtained from gasadsorption as a function of the I+S clay content (wt.%) for all
the natural mudrock samples from ASS except North Sea Shale. The linear t extrapolated
to 100% I+S clay content gives a value of 0.0075 cm3 /g for 25 nm pore-volume of clays.
The extrapolated value from natural shales coincides with the BJH cumulative pore volume
between 2 to 5 nm poresize obtained for the pure clays samples. The abundance of the 3
nm pore-sizes correlates with the I+S clay content in the natural mudrock samples. This
indicates that the 3 nm intratactoids pores are also shielded from natural compaction
process.

160

Figure 6.5: BJH cumulative pore volumes between 2 to 5 nm diameter of the natural mudrock
samples except for the North Sea Shale (lled circles) are plotted as a function of their I+S
clay content. A linear trend with zero intercept (equation y = bx) is tted to the data
points. Note that the result from the pure montmorillonite (open triangles) is not used in
the regression and is shown only for reference.

161

6.2

Case Study II: The Niobrara Formation

The Niobrara Formation was selected for this study based on (1) sample availability, (2)

available characterization on the same section in previous studies (Pollastro, 1992; Longman
et al., 1998; Landon et al., 2001), and (3) wide compositional variations within one same
well. The details of the regional geology and sample compositional characterization were
presented in Section 2.2.4.
6.2.1

Results

The N2 gas adsorption isotherm proles for the Niobrara samples show Type IIB isotherm
prole with a well dened H3 type hysteresis indicating presence of mesopores and macropores (Figure 6.6). The H3 type hysteresis loop indicates there is a signicant volume of
slit shaped mesopores. The isotherm proles also show the forced closure of the desorption
branch around 0.41 < P/P 0 < 0.49 due to the tensile strength eect. A TSE parameter
(Figure 6.7(a)) was dened to quantify the extent of this forced closure by the gradient of
desorption branch between P/P 0 0.410.49:
T SE parameter =

QAP/P0 0.49 QAP/P0 0.41

0.49 0.41

(6.1)

where QAP/P0 0.49 and QAP/P0 0.41 are the quantity of adsorbed gas at P/P 0 of 0.49 and
0.41 during desorption, respectively. The TSE parameter increases with increasing I+S clay
content in the Niobrara samples (Figure 6.7(b)). This observation leads to a qualitative
conclusion that samples with higher clay content have a signicantly larger amount of ne
mesopores <4 nm.
eSSA (m2 /g) for the Niobrara samples are summarized in Table 6.2 as per recommendations of I.S.O 9277:2010(E). Chalk D samples have signicantly higher eSSAs (1014 m2 /g)
compared to the rest of the samples, while the Fort Hays Member samples have the lowest
eSSAs (1.952.1 m2 /g) (Table 6.2). The samples with low clay content (<15 wt.% I+S clay
content) have similar eSSAs around 23 m2 /g, and do not show a any strong positive correlation with clay content (Figure 6.8). In contrast, the high clay content samples (>15 wt. %)
162

Figure 6.6: Typical N2 gas adsorption isotherm prole for the Niobrara Formation (sample
Niobrara Chalk B 3048 ft. sample SS39 Table 2.3). Inset gure is a detailed view showing
the isotherm forced closure in the relative pressure (P/P 0 ) range 0.35-0.55 due to tensile
strength eect.

(a) TSE parameter determination

(b) TSE parameter vs. I+S content

Figure 6.7: (a) Detailed view of N2 adsorption isotherm for a typical Niobrara sample with
emphasis on the forced closure behavior of the desorption branch. The yaxis is the quantity
adsorbed in cm3 /g STP (QA). The TSA parameter is dened to quantify the extent of the
forced closure, as the volume desorbed per unit pressure between the P/P 0 of 0.410.49. (b)
The TSA parameter increases with increasing I+S clay contents of the sample. This indicates
samples with higher clay content have a signicantly higher amount of ne mesopores <4
nm.

163

have a clear positive correlation between eSSA and clay content (Figure 6.8), however, the
Chalk D samples, with low OM content (<0.5 wt.%), show higher eSSA (9.9114.29 m2 /g)
compared to its higher OM content equivalent.
Table 6.2: eSSA of the Niobrara samples reported as per recommendations of I.S.O
9277:2010(E). The uncertainty estimate is only calculated from the mist of the linear BET
equation to the actual data point. Other experimental uncertainties are not evaluated. C =
BET Constant; Qm = Monolayer capacity.
Sample
ID

Depth
(ft.)

Litho
Members

Sample
Mass (g)

eSSA
(m2 /g)

SS31
SS32
SS34
SS35
SS36
SS38
SS39
SS310
SS311
SS312
SS313
SS314
SS315
SS316
SS317
SS318
SS319
SS320
SS322

2928
2932
2963
2978
2988
3029
3048
3056
3068
3076
3080
3090
3103
3119
3138
3150
3161
3195
3205

Chalk A
Chalk A
Marl A-B
Marl A-B
Marl A-B
Chalk B
Chalk B
Marl B-C
Marl B-C
Chalk C
Chalk C
Chalk C
Marl C-D
Marl C-D
Marl C-D
Chalk D
Chalk D
Fort Hays
Fort Hays

2.9216
2.8453
2.8095
2.7999
2.8908
2.9318
2.8889
2.7941
3.087
2.8167
2.8934
3.0057
2.8298
1.9236
2.3887
2.0024
2.8615
3.0066
2.6186

2.77(1)
3.73(1)
2.79(1)
2.63(1)
3.93(1)
2.71(1)
4.36(1)
3.97(0)
3.93(1)
2.37(1)
3.17(1)
2.74(1)
6.13(1)
2.42(0)
2.40(1)
14.29(2)
9.91(1)
1.95(1)
2.21(1)

46.2
45.0
39.0
41.5
44.7
44.3
51.5
49.2
52.6
30.7
40.6
50.5
48.2
37.9
32.4
109.1
98.3
69.8
70.3

Qm
(cm3 /g)

BET tting
R2

0.64
0.86
0.64
0.60
0.90
0.62
1.00
0.91
0.91
0.54
0.73
0.63
1.41
0.56
0.55
3.28
2.28
0.45
0.51

0.999921
0.999993
0.999971
0.999960
0.999985
0.999930
0.999963
0.999967
0.999994
0.999879
0.999993
0.999933
0.999977
0.999991
0.999973
0.999989
0.999990
0.999996
0.999969

P/P 0 tting range

Low
High
0.040
0.090
0.059
0.060
0.090
0.042
0.080
0.080
0.100
0.040
0.110
0.040
0.100
0.090
0.110
0.069
0.069
0.100
0.060

0.301
0.301
0.276
0.301
0.300
0.301
0.340
0.341
0.300
0.341
0.300
0.301
0.340
0.301
0.300
0.251
0.276
0.300
0.340

The general PSD of the Niobrara samples varies between unimodal to bimodal distributions (Figure 6.9). The Chalk D samples has a prominent bimodal PSD with a major peak
between 60100 nm and a minor peak around 23 nm, similar to the PSD observed in pure
clays (Section 5.1.1) and some of the other natural mudrocks (Section 6.1.1). The lower clay
content samples have closer to an unimodal PSD with 3 nm distribution peak becoming less
prominent (Figure 6.9). A detailed observation of the PSD up to 10 nm poresize reveals
that the presence of 3 nm distribution peak in the high clay content samples (Figure 6.10).
The 3 nm pore volumes increase with increasing clay content for both low OM content (Figure 6.10(a)) and high OM content samples (Figure 6.10(b)). The low clay high OM content
164

Figure 6.8: eSSA of the Niobrara samples as a function of their illite+smectite group clay
content. Each color indicates separate lithostratigraphic units.
samples (Figure 6.10(b)) do not have any 3 nm diameter pores, whereas the samples with
similar clay content but low OM (Figure 6.10(a)) show minor but prominent 3 nm peak in
PSD. This indicates that the OM content inuences the presence of small 3 nm pores in
these samples.
6.2.2

Discussion

The results demonstrate the inuences of bulk rock composition, especially I+S clay content, and OM content and maturity, on the pore-structure parameters in this suite of the
Niobrara Formation samples. As discussed previously in Section 2.2.4, the bulk rock mineralogy and OM content show denitive general trends consistent with the geological depositional
environment, however, there is no consistent relation or grouping for the dierent lithostratigraphic members. (Figure 2.6(a) and Figure 2.6(b). There is also no noticeable dierence
in the ranges of mineral contents and OM content between the individual lithostratigraphic
members (e.g., between the chalk and marl members). This is because the formation is
highly laminated and the scale of lithostratigraphic nomenclature does not explains the core

165

Figure 6.9: PSD of all the Niobrara samples; color coded based on their clay content where
red indicates highest clay content. Chalk D shows characteristic bimodal distribution with
peaks around 3 nm and 60-100 nm. Other samples show unimodal to bimodal PSD.

(a) Low OM (<1 wt.% OM) content samples

(b) High OM (>1 wt.% OM) content samples

Figure 6.10: PSD of the Niobrara samples up to 10 nm poresize with (a) low OM (<1 wt.%
OM) content, and (b) high OM (>1 wt.% OM) content. Hotter colors indicate higher clay
contents.

166

scale variations. Due to that lack of lithostratigraphic and mineralogical relationship in core
scale, the samples are not grouped based on their lithostratigraphic nomenclature. Instead,
for the purpose of correlating the porestructure attributes and bulk composition, the samples were divided into four groups based on their (i) OM content, with 1 wt.% OM as the
separator between low and high OM content samples, and (ii) I+S clay content, with 15
wt.% I+S as the separator between low and high clay content samples (see Figure 2.6(b)).
eSSA Figure 6.11 summarizes the relationship of eSSA with four dierent rock qroups.
The samples with I+S clay contents below 15% have almost uniform and relatively low eSSAs
and show insignicant correlations between I+S clay content, OM content and eSSA. In these
samples, the clay content is below a critical concentration to form a continuous matrix and
hence the bulk rock porestructure is controlled by nonclay constituents. Above 15% I+S
clay content, there is a strong positive correlation between I+S clay content, OM content
and eSSA. In these samples, the clay form the continuous matrix in the rock framework,
which controls the bulk rock pore structure resulting in an increase in eSSA with increasing
clay content. Also there is a clear separation in the SSA I+S content correlation based on
organic richness: for the same I+S content, high OM content corresponds with lower SSA;
absence of organic matter leads to an increase in SSA by about 150% (Figure 6.11). The
eSSA is inversely correlated with OM content for the high clay and high OM content samples
(Figure 6.12) , however, this relationship is only the manifestation of the inverse relationship
between clay and OM (Figure 2.6(b)). It can be concluded that that the OM does not add
to the total eSSA, but its presence decreases the eSSA in the rocks.
PSD The qualitative interpretation of isotherm proles obtained for the Niobrara samples indicates that they contain a signicant volume of both meso- and macropores. The
forced closure behavior of the desorption branch indicates the presence of ne mesopores
< 4 nm pore size in the high clay content samples. The 3 nm modal peak in PSD obtained
from the quantitative inversion of the adsorption branch is consistent with the qualitative
167

Figure 6.11: Comparison between eSSA and I+S clay content for the Niobrara samples
based on groups with low OM and high clay content, high organic and high clay content,
low organic and low clay content, and high organic and high clay content

Figure 6.12: Relationship between eSSA and OM content for the high OM content (>1
wt.%) samples. The labels indicate the I+S clay content (wt.%).

168

isotherm interpretation. This observation and the relationship with mineralogy mentioned
above indicate that the 3 nm modal peaks in PSD are real and not artifacts of the inversion
algorithm.
The pore volume associated with the 3 nm modal peak increases with increasing I+S
clay content in the low OM content rocks (Figure 6.13(a) and Figure 6.13(c)). The samples
with high OM content and high clay content show similar correlation between 3 nm pore
volume with I+S clay content (Figure 6.13(b)), however, they have lower pore volume and
diused 3 nm modal peak compared to low OM content samples with equivalent clay contents. The samples with low clay content and high OM content have low and constant pore
volumes in the 1.7 to 5 nm interval and the characteristic 3 nm modal peak in PSD is absent
(Figure 6.13(d)).
The presence of a 3 nm modal peak in PSD is a characteristic signature of the clay
microtexture. The similarity of overall PSD proles with the clay PSD indicates that the I+S
clay content is the primary control on bulk rock porestructure in the Niobrara Formation
lithologies. It can be also inferred that the OM in the Niobrara Formation does not have any
associated mesoporosity below 10 nm. The OM appears to occupy the 3 nm mesopores that
are associated with clay aggregates, based on the observation of diuse 3 nm modal peaks
and smaller 3nm pore volumes in high OM content samples compared to low OM content
samples with comparable clay contents (Figure 6.13).
Implications These observations and conclusions have implications in understanding
the OM porestructure in the Niobrara Formation and role of thermal maturity in porestructure development. Recent FESEM and STEM studies on gas shales clearly demonstrate
the presence of pores smaller than 10 nm in OM. OM pore systems are related to thermal
maturity, however, the maturity at which it starts developing is highly debated. Some studies
suggest the porosity in OM begins to develop as the temperature approaches the dry gas
window (Passey et al., 2010; Schieber, 2010; Curtis et al., 2011; Curtis et al., 2012); while

169

(a) Low OM, high I+S clay content samples

(b) High OM, high I+S clay content samples

(c) Low OM, low I+S clay content samples

(d) High OM, low I+S clay content samples

Figure 6.13: Relationship of small mesopore distribution (1.7-10 nm) for the Niobrara samples based on groups with (a) low OM and high clay content, (b) high OM and high clay
content, (c) low OM and low clay content, and (d) high OM and high clay content.

170

others suggest it starts at the onset of oil window (Loucks et al., 2012; Modica & Lapierre,
2012). The RockEval II parameters (Table 2.3) and S2 pyrograms (Figure 2.5(b)) from these
samples and previous studies (Landon et al., 2001) indicate that the thermal maturity of
these the Niobrara sample from this particular well is just at the onset of oil generation
window. The eSSA and PSD results conrms the absence of open welldeveloped micro
and ner mesopore pore system with the OM matrix. The thermal maturity of these samples
was not enough to create open porosity within the OM. In addition to the absence of the
inherent pore-structure within them, the OM acts as pore-blocking material and lls the
clay intratactoid mesoporosity. This is consistent with the previous reported occurrence
of OM occupying pore space with clay microstructure observed in recent marine sediments
(Weiler & Mills, 1965; Mayer, 1999; Mayer et al., 2004), immature Green River Formation
oilshales ((Tisot, 1962)), Bakken Shales (Zargari et al., in press) and thermally mature
mudrocks (Schieber, 2010; Milliken et al., 2013) and also in accordance to the observations
from the OM removal study presented in Section 5.2.
6.3

Case Study III: The Haynesville Formation

The Haynesville Formation is compositional very complex and composed both siliciclastic

and carbonateddominated lithologies with thermally mature OM. Diagenetic alteration like
calcite cementation and recrystalization will aect the connected porestructure of the rocks.
Previous studies on pore-structure characterization on the Haynesville Formation using imaging techniques suggested that inorganicassociated porosity is the dominant porestructure
in these rocks (Curtis et al., 2010; Milner et al., 2010; Dewers et al., 2012). Dewers et al.
(2012) did not observe no distinct pore structure with organic phases, however, there limit
of resolution was 10 nm, which might miss the nanoscale OM porosity. Total porosity measurements does not show any positive correlation with OM content (Figure 3.8(b)), however,
strong evidence of presence OMhosted porosity with <5 nm poresize was obtained on
the Haynesville rocks by the organic removal test by bleaching (presented in Section 5.2).
The objective of this case study is to investigate whether there is any systematic correlation
171

between pore-structure and composition. 16 aliquots of samples SS21 to SS26, SS225

to SS234, from the Haynesville sample set (SS2), described in details in Section 2.2.3,
were investigated using N2 gas adsorption technique. The samples were selected based on
availability.
6.3.1

Results

The isotherm proles of the Haynesville samples shows a Type IIB shape with H3 type
hysteresis loops, consistent with isotherm proles observed in mudrocks. This isotherm
prole indicates dominance of both mesopores and macropores in these rocks. Isotherm
proles of some samples show signicant adsorption at very low P/P 0 of < 0.01, indicating
presence of micropores in them. The presence of ne mesopores with poresize <4 nm in
these samples were indicated by the forced closure of the desorption branch of the isotherm.
The porestructure parameters obtained from inversion of isotherm data are tabulated
in Table 6.3. The eSSA shows a general positive correlation with the I+S clay content,
however with signicant scatter in the data (Figure 6.14(a)). The eSSA shows a negative
correlation of the total carbonate content (Figure 6.14(b)) and no systematic correlation
with OM content (Figure 6.14(c)). Similar correlations with composition were observed
with tplot micropore volume and the total pore volume up to 200 nm.
All the samples show a general PSD with a signicant volume of ne mesopores and
larger macropores. The PSD in the ne mesopore range (< 10 nm) of some samples show a
very prominent 3 nm modal peak as observed in pure clay samples (Figure 6.15(a)). In some
high OM content samples, the ne mesopore does not show the characteristic 3 nm modal
peaks and have an uniform distribution of pore volume up to 10 nm (Figure 6.15(b)). The
total pore volume between 25 nm pore-size shows a general positive correlation with I+S
clay content (Figure 6.15(c))but the scatter in the data and the uniform PSD in the ne
mesopore range indicates that there are ne mesopore associated with other component in
the rock.

172

(a) eSSA vs I+S

(b) eSSA vs Total Carbonate

(c) textiteSSA vs OM content

Figure 6.14: Variation in eSSA of the Haynesville Formation samples with (a) I+S group
clay content, (b) total carbonate mineral content, and (c) OM content.

173

(a) Characteristics Clay 3 nm modal PSD peak

(b) Uniform PSD

(c) 25 nm pore volume vs. I+S clay content

Figure 6.15: Variation observed in PSD in the ne mesopore range < 10 nm for the Haynesville Formation samples. Some samples show (a) very prominent 3 nm modal peak characteristics of I+S microstructure, while (b) high OM content samples show a very uniform
distribution of pore volumes. (b) BJH cumulative pore volumes between 2 to 5 nm diameter
of the Haynesville samples are plotted as function of their I+S clay content.

174

Table 6.3: Porestructure parameters of the Haynesville Formation mudrock samples obtained from N2 gas adsorption technique.

6.3.2

Sample
ID

eSSA
(m2 /g)

tplot Micropore
Volume (cm3 /g)

BJH cumulative 25 nm
Pore Volume (cm3 /g)

Total Pore
Volume (cm3 /g)

SS2-1
SS2-2
SS2-3
SS2-4
SS2-5
SS2-6
SS2-25
SS2-26
SS2-27
SS2-28
SS2-29
SS2-30
SS2-31
SS2-32
SS2-33
SS2-34

26.16
9.73
8.93
18.61
4.37
12.68
12.55
12.33
10.53
9.59
11.85
12.05
13.60
6.96
11.67
10.00

0.010
0.003
0.001
0.004
0.000
0.002
0.002
0.002
0.001
0.001
0.002
0.002
0.003
0.000
0.002
0.001

0.008
0.003
0.002
0.005
0.001
0.003
0.004
0.003
0.003
0.003
0.004
0.003
0.004
0.003
0.004
0.003

0.032
0.022
0.029
0.034
0.015
0.025
0.033
0.033
0.030
0.030
0.034
0.032
0.032
0.027
0.032
0.030

Discussion

The Haynesville Formation is geologically very complex with a wide range of compositional variation. All the porestructure attributes obtained from N2 gas adsorption technique
show a positive correlation with I+S clay content of the samples. This is consistent with the
reported dominance of clay-associated porosity in these rocks (Curtis et al., 2010; Milner
et al., 2010; Dewers et al., 2012).
In addition to the clay microstructure, the bulk rock porestructure might be also aected
by the
1. presence of thermally mature OM
2. presence of extensive carbonated cements in the siliciclastic dominated lithofacies
3. presence of recrystallized carbonate dominated lithofacies

175

There is no reported evidence of porous OM in the Haynesville Formation, however, the

removal of OM by bleaching indicated presence of OM hosted micro and ne mesopores in
the Haynesville rocks. The ne mesopore range PSD in the high OM content samples show
uniform distribution compared to characteristic 3 nm modal peak PSD as observed in clay
dominated microstructures. Comparing the trend between pore volume of 25 nm pore
sizes for the Haynesville samples as a function of I+S clay content with the similar trend
observed in the case of mudrocks from sample set ASS (Case StudyI, Figure 6.5), indicates
that there are more ne mesopore volumes in then Haynesville samples which is probably
due to the contribution of OM hosted mesopores (Figure 6.16(a)). All these evidences
indicate presence of OM hosted porosity in thermally mature Haynesville samples. However,
there is no systematic correlation between the ne mesopore volumes with the OM content
of the rock (Figure 6.16(b)). Similar lack of systematic trend between organic porosity
and pore size with OM content was reported in SEM observations of the Woodford Shale
(Curtis et al., 2012) and the Marcellus Shale (Curtis et al., 2011). Milliken et al. (2013)
observed a stronger correlation between TOC and OM hosted porosity in the Marcellus
Shale, but only towards the lower end of the TOC range. The lack of systematic correlation
with OM content is probably due to the heterogeneity within organic matter itself (Curtis
et al., 2012; Milliken et al., 2013). OM in mudrocks are highly heterogeneous composed of
dierent component with dierent composition such as various type of macerals, bitumen,
kerogen, pyrobitumen, char and not all the components show development of porosity with
thermal maturity (Schieber, 2010; Curtis et al., 2012; Milliken et al., 2013). A complete
understanding of the controls on OM-hosted porosity requires further comprehensive detailed
studies combining SEM observations and quantitative characterization of the pore-structure
parameters with OM geochemical /petrographical characterization on dierent sample sets
with contrasting thermal maturation history and possibly, experimental work on articial
maturation of immature kerogen.

176

(a)

(b)

Figure 6.16: (a) Comparison of 25 nm pore volume trend with I+S clay content in the
Haynesville samples (lled circles) with the trend observed in Case Study I, Figure 6.5.
The numbers indicate the OM content (wt.%) in the samples. The results from the pure
montmorillonite (open triangles) are shown for reference. (b) BJH cumulative pore volumes
between 2 to 5 nm diameter of the Haynesville samples are plotted as function of their OM
content.

The Haynesville Formation has a signicant volume of carbonate minerals. Presence of

detrial calcitic bioclasts and secondary calcitic cements have been reported (Hammes et al.,
2011). The negative correlation of the porestructure attributes with total carbonate mineral
content (Figure 6.14(b)) is probably due to the (i) imprint of the mineralogical negative
correlation between I+S clay content and carbonate content (Figure 2.4(a)) or (ii) porosity
destruction by cementation and recrystallization or (iii) combination of both. To understand
whether calcitic cementation has any eect on porestructure and poreconnectivity, PSD
of two samples SS21 with 4 wt.% total carbonate mineral content and SS26 with 35 wt.%
total carbonate content were compared. FESEMEDS analysis of thin section of SS26
(Figure 2.3(c)) suggest that the carbonate minerals are secondary cementing material that
forms highly localized laminations of extensive cemented zones. To assess the eect of this
localized extensive cemented zones, a sample (SS2A) taken within 0.5 ft depth interval of
SS26 were also analyzed.
177

A comparison of PSD (Figure 6.17) of these three samples indicate that the cemented
sample have lower pore volume in the ne mesopore range. Repeat analyses on the sample
(SS2A) taken within 0.5 ft depth interval of SS26 suggest that extensive localized cementation resulted in reduction of connected porosity in the mesopore range (Figure 6.17).
Combined thermogravimetric mass spectroscopy (TGA-MS) analysis of a similar sample
(Figure 6.18) showed evolved methane gas at temperatures higher than OM and calcite decomposition temperature, which indicates that there are closed porosity occluded by calcite
cement and that methane can only escape once the calcite cement is decomposed. These
observations concluded that extensive secondary calcite cementation in the Haynesville Formation will aect the connectivity of the bulk rock pore system and resulted in presence of
closed ineective porosity.

Figure 6.17: Eect of cementation on PSD of the Haynesville samples.

6.4

Summary
In this chapter, three dierent porestructure characterization case studies on a suite of

mudrock samples were presented. The main conclusion from these three case studies are
summarized as follows:

178

Figure 6.18: Combined thermogravimetric and mass spectrometry (TGA-MS) analysis of a

highly cemented Haynesville sample. The MS data shows evolved CH4 (mass 16) signal at
high temperatures beyond the decompostion temperature of OM and calcite.
1. The clay microstructure is the most dominant controls the small scale pore features
in natural mudrocks. The eSSA and presence of micropores/ne mesopores in any
natural mudrocks is correlated with the I+S group clay content in the rock.
2. In naturally compacted mudrocks, the 3 nm intratactoids associated with I+S clay
were shielded from natural compaction eect.
3. The presence of micro- and mesoporosity within OM is related to its thermal maturity.
OM in the studied Niobrara mudrocks, which is at the onset of oil window in thermal
maturity scale, does not show evidence of associated ne mesopores and micropores.
The thermally mature (gas window) OM in the Haynesville show evidence of presence
of micropores and ne mesopores.
4. The immature OM in the Niobrara rocks lls the mesopores within the clay microstructure.
5. The micro and mesopore volume in the Haynesville Formation does not show any direct
correlation with the OM content.

179

These understandings about the presence of micropore and mesopore pore networks
within organic matter and the inorganic components has important implications for gas
storage and production capacity in unconventional gas shale reservoirs. Insights about mudrock porosity and pore structure can help in addressing other challenges related to mudrocks
applications, some of which are demonstrated in the next Chapter.

180

CHAPTER 7
APPLICATIONS

Knowing is not enough; we must apply.

- Johann Wolfgang von Goethe
The purpose of this chapter is to summarize the learning from the previous chapters
about the mudrock porestructures and demonstrate their use in addressing some challenges
in mudrock characterization and modeling. The chapter is divided into dierent sections
based on the applications.
7.1

Gas shale Exploration Strategies

One of the biggest challenge in gas shale exploration strategy is to locate the sweet spot.

There is not clear understanding about what constitutes a sweet spot, primarily because there
is a lack of understanding of ow process and lack of extensive suite of measurements by
wireline tools that works for mudrocks. However, the composition characterizing wireline
tools, like Gamma Ray and Elemental Capture Spectroscopy (ECS) tools, works well in
mudrocks and therefore linking pore-structure understandings with composition can help in
making exploration strategies. In the following subsection, dierent demonstrations about
how correlating the pore-structure attributes with composition can aid in making gas shale
exploration strategies.
7.1.1

Predicting total porosity

Figure 7.1 summarizes the relationship between measured WIP porosity and compositional attributes, like total clay content, OM content and total carbonate content, for the
three suites of mudrocks measured (SS1, SS2 and SS3). Geologically each of these mudrocks
are dierent and therefore they show dierent correlation with measured total porosity. In

181

the Eastern European gas shale play mudrocks and in the Haynesville Formation mudrocks,
total porosity show a general positive correlation with total clay content (Figure 7.1(a) and
Figure 7.1(d)). This can be attributed to the clay forms a continuous matrix in these mudrocks and therefore, exerting a primary control on the macroscopic pore structure. Apart
from this primary control, other factors can also inuence the total porosity. In general, the
presence of carbonate minerals have a negative trend with total porosity is both SS1 (Figure 7.1(c)) and SS2 (Figure 7.1(f)). In SS1, the carbonate mineral content is not correlated
with to the clay content (Figure 7.1(c)) and therefore, it is not the mineralogical imprint of
the clay correlation. The carbonate minerals are therefore present as cementing material
destroying the total porosity. In the Haynesville samples (SS2), the negative correlation
between total carbonate and total porosity may be due the mineralogical correlation between clays and carbonate (Figure 7.1(f)) or the porosity destruction by cementation and
recrystallization of carbonate grains or combination of both. The Haynesville Formation
is a complex mudrock system with both siliceous and carbonate-dominated lithofacies and
the carbonate minerals are present both as cementing material in siliceous mudrocks and
recrystallized detrial grains in the carbonate dominated mudrocks.
Total porosity has a good positive correlation with OM content in SS1 (Figure 7.1(b)).
The Eastern European mudrocks are thermally mature with no pyrolysable kerogen. The
positive correlation with between OM content and porosity indicates that the OM hosted a
pore-network system in them. However, thermally mature OM in the Haynesville does not
how any correlation with total porosity (Figure 7.1(e)), although N2 gas adsorption studies
on those samples before and after removal of OM clearly indicate presence of OM hosted
micro and mesopore system. The lack of correlation between total porosity and OM content
is probably due to the heterogeneity of the organic matter. Comparison of RockEval II S2
pyrograms between SS1 and SS2 (Figure 2.1(b) vs. Figure 2.2(b)) indicate the heterogeneous
nature of OM in SS2 with a signicant volume of bituminous compounds in some samples,
whereas the OM in SS1 is very homogeneous. Dierent components of OM in the Haynesville

182

Formation might have dierent pore structure and dierent history of thermal evolution.
In contrast to the Eastern European Silurian mudrocks and the Haynesville mudrocks,
there is no denite mineralogical correlation observed in the Niobrara samples for the total
porosity (Figure 7.1(g), Figure 7.1(h) and Figure 7.1(i)). However, the ne meso-pore system show a good correlation with the total clay content and OM content, as presented in
Section 6.2 (Figure 6.13). The ne scale (micro and meso) pore structure attributes, e.g.
the SSA, show positive correlation with beyond 15 wt.% I+S clay content, indicating a continuous continuous matrix beyond that critical concentration (Figure 6.11). The presence of
OM also inuences the ne mesopore scale pore system as it acts like a pore lling material
with the clay mesopores and does not host any porosity within them. The Niobrara is a
pelagic carbonate system and the macroscale pore structure might be inuenced by diagenetic imprints like mechnical and chemical compaction and carbonate dissolution. The ne
scale microstructure is mainly controlled by presence of I+S clay content.
A better understanding of the correlations between pore-structure attritibutes and composition of mudrocks in a geological context is critical to develop exploration strategies.
Porosity is one of the major factor for sweet spot characterization . Wireline tools for
porosity measurement does not work well in mudrock lithologies. In absence of reliable
porosity estimates from wireline measurements, correlation between core porosity measurements with composition in a geological and lithofacies context can be extended to derive
porosity map using the geological and lithofacies mapping. The sample selection for this
study was based on composition range in order to represent the wide range of composition. However, as this study has demonstrated, detailed lithofacies characterization are also
required to make predictive porosity maps based on the composition logs.
7.1.2

Predicting Adsorbed Gas Content

A signicant component of gas-in-place of the gas shale reservoirs is present as adsorbed

gas phase attached to the pore surface (Ambrose et al., 2012). The Langmuir model indicates
that the total amount of adsorbed gas is directly proportional to the pressure and isotherm
183

(a) SS1 vs. Clay

(b) SS1 vs. OM

(c) SS1 vs. Carb

(d) SS2 vs. Clay

(e) SS2 vs. OM

(f) SS2 vs. Carb

(g) SS3 vs. Clay

(h) SS3 vs. OM

(i) SS3 vs. Carb

Figure 7.1: Compositional correlation between total porosity measured by WIP technique
for mudrocks from (a), Eastern European gas shale formation (SS1), (d), the Haynesville
Formation (SS2), and the Niobrara Formation (SS3). The dashed line are not best t linear
trendlines, but are shown to illstrate general correlation.

184

prole is predicted by two constants, the Langmuir volume and the Langmuir pressure. The
Langmuir volume is the amount of adsorbed gas at innite pressure, and it represent the
monolayer capacity, i.e., amount of gas required to cover the available surface area with
a single layer of adsorbed gas molecules. All other factors remaining constant, higher the
specic surface area , greater will be the monolayer capacity and adsorbed gas content
in the reservoir should have a positive correlation with specic surface area. Therefore,
specic surface area is another important pore structure attribute for gas shale exploration
strategies and sweet spot characterization. Specic surface area is essentially an micro
and mesoporestructure attribute; pores >50 nm have insignicant contribution to the SSA
of the porous media.
Pore structure information obtained from N2 gas adsorption techniques on mudrocks in
this study suggest that micropores and ne mesopores contribute to the total pore volume in
minor proportion but most pore surface area is associated with the them. The pore-structure
characterization on mudrocks in this study suggest that the porosity associated with I+S
clays microstructure is the most dominant feature of ne scale pore structure of mudrocks.
The specic surface area and micropore- ne mesopores volumes up to 5 nm are positively
correlated with the I+S clay content of the mudrocks (Figure 6.5, Figure 6.8, Figure 6.13
and Figure 6.14(a))
The presence of OM inuences the ne scale microstructure of the mudrocks. Part of
the OM in mudrocks occupies open clay mesopore pore-network system. This was found
true for all mudrocks irrespective of their maturity and has been reported, for example,
in recent marine sediments, immature oil shales, mudrocks in oil window maturity such as
Bakken Shale and thermally mature mudrocks. The N2 gas adsorption experimental results on the Niobrara formation and the organic removal experiments also conrms this
observation Figure 6.13. However, this OM can host a micro- and mesopore system within
themselves depending upon maturity. The immature OM from the Mahogany Formation
and the Niobrara Formation lacks internal porosity. The OM removal study on thermally

185

mature Eastern European Silurian and the Haynesville Formation mudrocks indicate indicating presence of micropores and mesopores within the OM and therefore, in gas shale plays,
the OM contributes to the the SSA and micropore- ne mesopores volumes.
This indication of presence of micropores and mesopores within the OM for thermally
mature gas shales has widespread implications in their gas storage capacity (gas-in-place
calculations) and production potential. Partitioning the micropore and mesopore volumes
between OM and I+S clay components is critical due to the dierence in surface energy
between the two types of components. The second constant of the Langmuir model, the
Langmuir pressure, is dependent upon the entropy of adsorption and is related to the pore
surface energy. Methane, for example, has a greater anity for OM than for I+S and
other inorganic surfaces. Supercritical adsorption of methane (at 25 C) on the natural
and OM removed samples from a Eastern European gas shale mudrock (SS115) and a
Haynesville mudrock (SS2C) indicate that methane adsorption is strongly controlled by the
presence of OM (Figure 7.2). The OM removed samples show very low adsorption indicating
majority of the adsorbed methane in the natural mudrocks are associated with OM pores.
The results of the OM removal experimental study, presented in Section 5.2, dierentiating
quantify the contribution of OM and clay to the total eSSA and pore volume of rock. Such
quantication are needed to model the adsorption behavior for gas shale resource assessments
and production planning.
7.1.3

Completion Strategy

An important sweet spot factor that is often considered during gas shale exploration
is the frackability or the brittleness of the rock that determines the success of a fracturing
treatment. In general, carbonate minerals are considered to be relatively brittle and less
elastic than clay and quartz. Therefore, presence of carbonate minerals is desirable for
successful hydraulic fracturing.
The carbonate mineral content in the Eastern European Silurian mudrocks and in the
Haynesville Formation correlates negatively with their total porosity. This indicate that the
186

(a) SS115

(b) SS2C

Figure 7.2: Comparison of supercrtical methane adsorption at 25 C between the natural

and OM removed sample for (a) SS115 and (b)SS2C.

carbonate minerals are mostly diagenetically precipitated cement and thereby reducing the
total porosity of the rock. Extensive recrystallization of detrital carbonate grains might also
reduce the porosity of the carbonate dominated mudrocks. The pore-structure characterization and TGA-MS study on a carbonate rich Haynesville sample show that presence of
secondary calcite reduces the pore connectivity and total porosity in these mudrocks (Figure 6.17 and Figure 6.18). Thus while completions in carbonate dominated section might be
successful, the restricted pore connectivity will reduce the matrix permeability around the
fracture and in turn limit the ow of gas into the fracture system and aect the productivity
of the well in long term.
Such understanding of pore-structure attributes can help in designing the exploration and
completion targets for mudrock reservoirs. In carbonate dominated mudrock section matrix
acidizing techniques post fracturing might increase the pore-connectivity and permeability
of the rock matrix around the fracture. A core scale experimental study investigating the
change in pore-structure before and after acid ushing in a high carbonate content mudrock

187

samples is recommended to understand the applicability of eld scale acidizing treatments

for gas shale applications.
7.2

Implications in Rock Physics Modeling

Claywater composites are used as the building unit for rock-physics modeling of shales

(Hornby et al., 1994; Arne Johansen et al., 2004). However, most of the application of these
models to natural shales concentrate on the orientation distribution function of macroscopic
clay aggregates. They fail to incorporate the multiscale porosity associated with clay microstructure and their evolution under increasing stress. The mismatch between measured
and predicted elastic stiness coecients of shales by over a factor of two, lies in this inadequate microstructural representations in the models (Militzer et al., 2011). Quantitative
data on pore-size distribution of clays in pure form and in shales are critical inputs for the
rock-physics models of mudrocks.
The pore structure characterization of pure clays and natural mudrocks suggest presence
of multiscale porosity associated with I+S group of clays. The smallest scale of porosity in
these clays shows a characteristic 3 nm modal peak in PSD inverted from N2 isotherm and
remains unaected by the compaction process. The presence of the 3-nm clay porosity, which
is shielded from compaction, has implications in rock-physic modeling of shales. The endmember building block of many mudrock rock physics models is the clay water composite.
Hornby et al. (1994) used a combined self-consistent approximation (SCA) and dierential
eective medium (DEM) approach to model clay-water composite, where a biconnected
medium of clay-water medium is rst modeled using the SCA. The SCA model is later
adjusted to the target porosity using a DEM model. However, the target porosity of this
claywater composite is not exactly known. The assessment of incompressible 3 nm sized
pores associated with I+S group of clays provides an important building block for their
mineral modulus. The tactoids or the quasicrystal can be imagined as the basic elastic
end member of shale rock. Extrapolation of the tted linear trend (with zero intercept) of
pore volume between 2 to 5 nm pore size with I+S clay content in natural mudrocks to
188

100% clay content coincides with the pore-volume obtained for the pure clays (Figure 6.5).
A specic pore volume of 0.0075 cm3 /g between 2 to 5 nm for 100% clays will translate
into a porosity value of 2.1% assuming an anhydrous grain density of 2.85 g/cm3 for I+S
clays. Therefore the building block for I+S mineral modulus can be thought of clay tactoids
with 2.1% porosity. Neaman et al. (2003) explained the possible arrangement of these pores
in tactoids as slit-shaped porosity at the egdes of turbostatically arranged unit cells or as
rectangle-shaped porosity formed by initial arrangement of the interlayers. Using a Reuss
average between the dry clay modulus of 55 GPa from Katahara and water (2.2 GPa), we
derived a bulk modulus of the smectite clay-water composite as 36 GPa. This should form
a fundamental mineral end-member for I+S clay. The presence of intraaggregate pores
formed by the arrangement of tactoids within a single aggregate of clay will also reduce the
modulus the claywater composite.
The conclusion from pore-size distribution studies on organic rich mudrocks indicate that
some of the OM can occupy these 3 nm intra-tactoid porosity. Therefore, for gas shale mudrock rock physics modeling, the basic building block should be made of organo-clay packages
where OM occupies the wedge shaped pores within the clay framework, similar to the textural arrangement observed in an ion-milled organic rich mudrock from Eastern European
gas shale play (Figure 7.3(a)). Note that ne scale micro and meso pore system in the OM
depends on its thermal maturity. The thermally mature (gas window) OM in the Haynesville
Formation has open porosity within them but the thermally immature (onset of oil window)
OM from the Niobrara Formation do not. The eect of OM and its thermal maturity on
the micro and ne mesoscale microstructure of mudrocks are summarized in a total specic
ne mesopore (25 nm) volume versus I+S clay content plot template (Figure 7.3(b)). The
specic ne mesoscale pore volume increases with increasing I+S clay content in the rock
along the line joining the origin with pore volume at 100% I+S clay content. This is the
most dominant microstructural feature in all mudrocks. Presence of non porous immature
OM will reduce this ne mesoscale pore volume and thermally mature OM can contribute

189

to the ne mesoscale pore volume.

7.3

Implications in Gas-Flow Modeling

Gas ow in these mudrock resrvoirs occurs mainly through the interconnected fracture

network system. This fracture network system is postulated to be recharged by gas owing
in the micro and mesopores dominated mudrock matrix. Darcy model of permeability and
ow that is applied in conventional plays cannot be directly applied to shale reservoirs due to
the presence of micro and mesopores. Gas production from these resources is much greater
than anticipated and cannot be explained by conventional models. Gas ow in the mudrock
matrix is expected to be a combination of Knudsen diusion and slip ow in micro and
ne mesopores and Darcy-like ow in larger macropores pores. Modeling this ow requires
knowledge about the poresize and relative pore volume associated with each poresize.
Flow of uids in rocks is generally modeled using Darcys equation, which describes
advective viscous uid ow. Flow of gases in tubes (including capillaries) is described by
HagenPoiseuille equations with no-slip boundary conditions. The noslip boundary condition indicates viscous bonding of uids to the wall and is modeled by assuming the particle
velocity to be zero at the wall of the pipe (or pore). The characteristic length is dimension
that denes the scale of a physical system, and for porous media, characteristic length most
commonly used is the pore-diameter. As the characteristic length of the physical system
decreases, the assumption of standard continuum approach falls apart. The dimensionless
Knudsen number (Kn = /Rh , where is the mean free path of the gas, i.e. the average
distance between two consecutive molecular collisions, and Rh is the characteristic pore diameter, is used to determine the degree of appropriateness of applying continuum approach.
For Kn <0.01, the mean free path of the gas molecules is negligible compared to the characteristic dimension of the ow geometry (i.e. Rh parameter), the continuum hypothesis
of uid mechanics generally holds true. Rarefaction eects become important as Knudsen
number increase and consequently the pressure drop and mass ow rate cannot be predicted
from the continuum model of uid ow (Hagen-Poiseuilles equation or Darcy-like ow). At
190

(a) Organoclay building block for gas shale elastic modeling

(b) Eect of OM on mesopore volumes

Figure 7.3: (a) FESEM image of ion milled sample from Eastern European Silurian Gas
Shale showing microstructural organo-clay packages (Image taken by Dr. Kitty Milliken).
(b) Eect of OM and its thermal maturity on the ne mesopore structure in mudrocks.

191

Kn >10 the gas molecules collide with the ow boundaries more often than inter-molecule
collisions. Thus the molecules move independently of each other and in this condition the gas
composition have no importance. This ow regime is known as Knudsen diusion or free
molecule ow. The intermediate region in between 0.01 < Kn < 10 cannot be considered
neither as a continuum ow nor a freemolecule ow but has been modeled as a combination
of Knudsen ow and Poiseuille ow. Klinkenberg (1941) gave a rst order approximation by
suggesting simple addition of Knudsen and Poiseuille ow constants. Schoeld et al. (1990)
used a twoparameter semi-empirical equation combining the two ow constants to explain
experimental data of gas ow through microporous membrane. Karniadakis et al. (2005)
suggested a unied equation to model the entire ow regime where the addition of Poisuelle
ow constant and Knudsen ow constant multiplied by factors to account for rarefaction of
gas and slip coecient. Javadpour (2009) suggested similar approach of addition of Knudsen
ow constant and Poiseuille ow constant multiplied by some factor to account for factors
like wall smoothness. All of these studies show that the gas ow in mudrocks can be modeled
as weighted contribution between pure Knudsen ow and pure Poiseuille ow.
The ow constant can be obtained for Knudsen ow and Hagen-Poiseulle ow (Schoeld
et al., 1990), which are as follows:
2 r 8RT
KK =
3 M


0.5

1 1 MP
KP = r 2
8 RT

M
RT

(7.1)
(7.2)

where where KK is the ow constant for Knudsen ow, KP is the ow constant for Poiseulle
ow, r is the radius of the capillary tube, is the total porosity, is the tortuousity, is
the viscosity of gas, P is the average pressure, T is the temperature, R is the universal gas
constant amd M is the molecular weight of the gas. It should be noted that the Knudsen
ow constant KK is independent of pressure while the Poiseulle ow constant is dependent
on pressure. Comparing the ow constants KK and KP at dierent pore radius r will give
us an idea about the relative importance of these ow regimes and their contribution to total
192

ux.
Figure 7.4 shows the ow constants as a function of pore radius at dierent pore pressures.
The porosity and the tortuosity values used 0.08 and 3.00, respectively, which are typical
values for mudrocks. At low pore pressure (14.7 psi) Knudsen ow will dominate over
Poiseuille ow for poreradius below 300 nm. Since KK is independent of pore pressure
and KP is directly proportional to pore pressure, at higher pore pressures, Poiseuille ow
dominates over Knudsen ow. The pore size at which KP dominates over KK moves towards
smaller pores as pore pressure is increased. The ratio of KK to KP for a 3 nm radius pore
capillary decreases from 102.3 to 0.3 as pore pressure increases from 14.7 psi to 5000 psi.

(a) 14.7 psi pore pressure

(b) 100 psi pore pressure

(c) 1000 psi pore pressure

(d) 5000 psi pore pressure

Figure 7.4: Flow constants as a function of pore-radius at dierent pressures

193

The above plots suggest that ow in larger pore sizes generally follow Darcy ow but
contribution from Knudsen ow will be important in ows through the small meso and
micropores. The PSD obtained from N2 gas adsorption techniques can be used to model the
contribution of Knudsen ow and Poiseuille ow for shale systems. The pore-size-distribution
data are modeled assuming the pores to be a bunch of straight capillary tubes and the
Knudsen ow constant and the Poiseuille ow constant are calculated using the following
formula:
KK =

KP =

0.5

ri 2

i 1 M P
RT

i 8RT
ri
3 M

X2

X1
r

M
RT

(7.3)
(7.4)

where i is the partial porosity associated with the pore-radius ri . The PSD obtained from
BJH inversion is used to get the distribution of porosity. The specic pore volumes obtained
from N2 gas adsorption for the Haynesville samples were converted to partial porosity distribution data using the grain density obtained from WIP measurements. Figure 7.5 shows the
contribution of Knudsen ow with respect to the total ow as a function of their measured
WIP total porosity. A general trend is observed with increasing contribution of Knudsen
ow to the total ow with decreasing total porosity.
The liquid-permeability of the rock can be modeled from the pore-size distribution from
the following equation:
k=

X1
r

i
8
ri

(7.5)

where k is the permeability of the rock in m2 and the pore-radius ri is in m. The permeability
of the mudrocks from assorted set was calculated and compared with previously measured
liquid permeability (Sarker, 2010). The modeled permeability show an overall general trend
with the measured permeability (Figure 7.6), however, the modeled values are two order
of magnitude higher than the measured permeability. The general correlation suggest the

194

Figure 7.5: The modeled contribution of Knudsen diusive ow to the total ow as a function
of measured total porosity in the Haynesville samples.
control of PSD on permeability behavior. However, any conclusion from these modeled
values needs to be applied with caution as the simplistic models are used. The purpose of
this exercise is only to demonstrate the applicability of PSD information in ow models and
can be used in more rigorous permeability and simulation models to predict ow behavior
through mudrocks.
7.4

Implications in Petrophysical Evaluation: Density of OM

One of the major problem in petrophysical evaluation of mudrock reservoirs is the range

of variability in reported grain densities of OM. Most reported grain densities for OM in
mudrocks range between 0.95 and 1.725 g/cm3 (Nwachukwu & Barker, 1985). Density
gradient centrifugation (DGC) of separated OM from mudrocks suggest high heterogeneity
in composition and density of dierent constitutive macerals, even within a single mudrock
(e.g., Robl et al., 1987; Robl & Taulbee, 1995). Robl & Taulbee (1995) reported a grain
densities of 1.01 to 1.40 g/cm3 for dierent macerals with the OM separated from a single
Cleveland Shale sample. Current petrophysical practices uses total organic carbon (TOC)
as the only characterization attribute for OM. TOC quanties on the total wt.% of organic
195

(a) Linear Plot

(b) Log-Log plot

Figure 7.6: Comparison of modeled liquid permeability from Poiseuilles equation with reported measured permeability data(Sarker, 2010) on mudrocks in (a) linear scale, (b) logarithmic scale. The solid line indicate the tted linear trend. Note the linear trend show a
curve in log-log plots. The Mancos B data (red) is not included in the t.
carbon in the rock and since all petrophysical inversions are based on volume percentages,
the TOC need to be converted into vol.% OM. This conversion requires knowledge about two
parameters; (i) the wt.% of carbon within OM is required convert TOC to wt.% OM in the
rock, and (ii) the grain density of OM to convert the wt.% OM to volume.% OM. Because
of the dierence in the densities between the OM and the mineral constituents of the rock,
OM volume percent and also the whole rock grain density is sensitive to the assumed OM
grain density (Figure 7.7).
Direct information on the OM grain density can be obtained by kerogen isolation and
DGC studies which are tedious, expensive and time consuming. An indirect method of computing OM grain density can be used using the combination of precise bulk rock grain density
measurement by WIP technique, TOC measurement and quantitative mineral phase analysis
(QPA) on suite of sample from the same mudrock formation. The bulk QXRD mineralogy,
elemental composition and cation exchange capacity (CEC) were combined and optimized
using the Chevron proprietary BestRock modeling method, similar to the modeling ap196

Figure 7.7: Whole rock grain density sensitivity analysis due to the variation in assumed
OM density variation.
proached presented in rodo & Kawiak (2012), to calculate individual mineral composition
for each of the three dierent samples set, which were also measured for bulk rock grain
density by WIP technique. This technique provides a robust calculation of the individual
and cumulative mineral grain densities for the each individual sample sets (Table 7.1). The
individual anhydrous grain densities obtained can be combined using the QXRD wt.% to
obtain the grain density of the inorganic mineral fraction of the mudrock. The OM grain
density of the each individual sample can be calculated using the equation 7.7.
1
rock

OM =

wt.%min wt.%OM
+
min
OM

1
rock

wt.%OM
100 wt.%OM

min

(7.6)

(7.7)

wt.%OM is obtained from TOC measurements and converted assuming a 83% carbon
content in OM. The wt.% carbon within OM is also variable ranging from 76 to 90% de-

197

Table 7.1: BestROCK derived grain densities of indiviual mineral for each sample set
Sample Set 1
Mineral
Quartz
K-feldspar
Plagioclase
Calcite
Ank.
Pyrite
Halite
Anatase
Clay
Minerals
Chlorite
I+S c

BestRock
Grain
Density
2.651
2.570
2.633
2.722
2.979
5.050
2.164
3.906

3.143
2.811

Sample Set 2
BestRock
Grain
Density

Mineral
Quartz
Plagioclase
Calcite
Ank. a
Pyrite
Barite
Anatase
Fluorapatite
Clay
Minerals
Kaob
Chlorite
I+S

a Ank.

2.651
2.619
2.716
2.983
5.013
4.479
3.906
3.129
2.622
2.995
2.854

Sample Set 3
Mineral

BestRock
Grain
Density

Quartz
Plagioclase
Calcite
Ank.
Pyrite
Anhydrite

2.664
2.689
2.717
3.008
5.024
3.030

Clay
Minerals
Kao

2.622

I+S

2.773

= Ankerite or excess-Ca dolomite

= total dioctahedral 1:1 layer clay: kaolinite, dickite, nacrite, haloisite
c I+S = total dioctahedral 2:1 layer clay: illite, mixed-layer illite-smectite, smectite, and possibly mica
b Kao

198

pending upon maturity (Ungerer et al., 1981), and have some sensitivity in the calculated
OM density. However, to prevent the system of equations to be underdetermined, the conversion factor is kept constant for all the sample sets. rock is the anhydrous grain density of
the bulk rock measured by WIP technique. min is the grain density of the inorganic mineral
fraction of the rock. OM grain density for all the samples, except for low OM content (<0.5
wt.%) in the SS3, are calculated.
The calculated OM grain density values obtained for SS1, SS2 and SS3 are 1.369 g/cm3
(0.250 g/cm3 [2], n = 22), 1.211 g/cm3 (0.284 g/cm3 [2], n = 34) and 1.290 g/cm3
(0.242 g/cm3 [2], n = 12, respectively (Figure 7.8). Note that for SS3, only the high
OM content (<0.5 wt.%) sample were used for OM grain density inversion and the two
outliers are excluded. The scatter in the data is probably due to a combination of the errors
associated with the each measurements and the heterogeneity in the grain densities of OM
within the same formation.
The higher OM grain density for the Eastern European gas shale samples (SS1) can
be explained by their high thermal maturity with essentially no prysolysable kerogen left
(Figure 2.1(b)). The OM in the Haynesville Formation (SS2) is less mature than SS1 with
relatively higher HI and OI values (Table 2.2) and contains a high amount of bitumen
(Figure 2.2(b)). The higher standard deviations in the calculated OM grain densities for the
SS2 samples because of the heterogeneity and the variable amounts of bitumen and kerogen
in these samples. The heterogeneity of the OM in SS2 is expressed by the variable RockEval
S2 region of the pyrograms (Figure 2.2(b)). The thermal maturity of the OM in the Niobrara
Formation from SS3 is either immature or just at the onset of the oil window, based on the
Tmax and HI values. The S2 pyrograms have a relatively high temperature kerogen peak
( 430 C) with no indication of bitumen (Figure 2.5(b)). The OM grain density of 1.29
g/cm3 is within the reported range of OM grain density from Kimmeridgian shale that has
the same maturity level (Okiongbo et al., 2005).

199

(a) Eastern European gas shales SS1

(b) Haynesville SS2

(c) Niobrara SS3

Figure 7.8: Histogram of calculated OM densities, using Equation 7.7, for (a) the Eastern
European gas shale samples SS1, (b) the Haynesville Formation, and (c) the Niobrara Formation. The red bar in the histogram on OM density from the Niobrara are treated as
outliers.

7.5

Summary
Applications of experimental results of total porosity measurements and pore-structure

parameter analysis using N2 gas adsorption techniques in dierent other mudrock characterization application and modeling studies were presented in this Chapter. Insights obtained
by correlating compositional attributes with pore-structure attributes for mudrocks is useful
for making gas shale exploration and completion strategies and aid in nding the sweet
spot for drilling wells. The understanding about the I+S clay and OM microstructure helps
to model the fundamental building blocks in elastic and rockphysics modeling of shales.
The use of experimental results on PSD of mudrocks in modeling of gas ow through these
reservoirs were demonstrated. Finally, combination of bulk rock grain density measurements
and quantitative mineral phase analysis were used to obtain the grain density of OM, which
is a fundamental input required for petrophysical evaluation of these mudrocks.

200

CHAPTER 8
CONCLUSION AND RECOMMENDATIONS

A conclusion is the place where you got tired thinking.

- Martin H. Fischer
The two main objectives of this thesis were to develop methodologies for measuring pore
structure attributes, such as total porosity, specic surface area and pore-size distribution, in
mudrocks and understand the controls of composition on the same. Two dierent methodologies were adapted and modied to measure dierent pore-structure attributes of mudrocks.
The measured pore structure attributes were correlated with compositional characterization
to get insights about mudrock pore structure. The applications of the experimental results
and the insights obtained from this study in other elds of mudrock characterization and
modeling were demonstrated. A detailed summary of the results and conclusions is given in
section 8.1, while recommendations for future work are given in section 8.2.
8.1

Conclusions
This section is divided into three subsections based on the broad research topics: pore-

structure measurement methodologies, pore-structure in mudrocks and compositional controls and application of the experimental results and pore-structure information for other
eld of studies.
8.1.1

Mudrock Pore-Structure Measurement Methodologies

Water immersion porosity (WIP) technique was developed for total porosity measurements in mudrocks without using crushed rock methodology. N2 gas adsorption was adapted
and modied for mudrock applications. The major conclusions on the experimental techniques are as follows:

201

 The WIP technique developed in this thesis allows measurements of mudrock total
porosity with quantied experimental uncertainties. Salient features of this techniques
are:
1. It uses intact samples instead of crushed rocks.
2. The protocol is simple and fast.
3. Measured total porosity, grain density and bulk density are reproducible
4. The absolute experimental uncertainty is about 0.20.3 p.u.
N2 gas adsorption technique is recommended to characterize ne scale (up to 200 nm
poresize) pore-structure parameters, such as specic surface area, micropore volume
and poresize distribution (PSD), in mudrocks.
1. Crushed handground <420 m (<40 mesh) powder allows faster diusion and
complete equilibrium of the gas during analysis.
2. Crushing does not damage the porestructure in the range of investigation of up
to 200 nm.
3. The measured porestructure parameters are consistent and reproducible within
515 relative %.
4. Inversion of N2 isotherm data is nonunique and and depends on the choice of
inversion algorithms. The inverted pore-structure results should be treated as
equivalent characteristic values, rather than true value.
5. The BJH inversion algorithm is recommended over DFT inversion techniques.
The thickness equation for BJH and tplot techniques should be kept consistent
across datasets for comparative purposes.
8.1.2

Mudrock Pore Structure

The mudrock pore struture was studied at by rst investigating the porestructure of
the compositional end members: clays and organic matter (OM). The insights from the end
202

member pore-structure is then used to understand compositional eects on pore-structure

of natural mudrocks. There are compositional controls on mudrock pore structure:
Illite+smectite group of clay minerals show three scales of porestructures associated
with dierent scale of stacked clay microstructure:
1. intra-tactoid pores with shows a characteristics 3 nm modal peak in N2 adsorption
PSD;
2. intertactoid or intraaggregate pores have diameter around 50-100 nm and a
pore-throat diameter 20-25 nm; and
3. interaggregate pores are micron sized.
Immature OM does not have any signicant open ne scale (up to 200 nm) pores.
Thermally mature (gas window) OM has signicant pore volumes with poresizes <5
nm. Part of the OM in natural mudrocks exists within the clay mesopore space.
Clay microstructure is the fundamental control in nescale pore structure attributes
such as specic surface area, micropore volume and ne mesopore pore volume (25
nm) of all mudrocks.
1. There is a direct correlation between these ne scale pore structure attributes
with I+S group clay content.
2. The PSD of natural mudrocks show the characteristics 3 nm modal peak indicative
of the intra-tactoid clay pore space.
3. These clay mesopores are also shielded from the eect of compaction.
The thermally immature (onset of oil window) OM from the Niobrara Formation samples lacks open micropore and ne mesopores. The thermally mature (gas window)
OM from the Eastern European Silurian mudrocks and the Haynesville Formation
mudrocks does show evidence of open micropores and ne mesopores.
203

8.1.3

Application of the Experimental Results

The experimental results and the insights obtained about the compositional controls on
mudrock porestructure are useful for dierent other mudrock characterization and modeling
applications, which are summarized as follows
The understanding of compositional controls on macroscopic and ne scale pore structure attributes can be used in developing gas shale exploration strategies. The compositional controls on macroscopic pore structure attributes depend upon the geological
setting of the specic mudrocks formations.
The distribution of the specic surface area between the OM and I+S clay components
can be used to predict the adsorbed gas content inside the gas shale reservoirs.
The mechanically incompressible 3 nm pores associated with I+S clay tactoids can be
used as the building block for mineral modulus of I+S clays in rockphysics modeling
of mudrocks. The OM residing within the clay mesopores in organic rich mudrocks
can be used as the fundamental organo-clay building block for organic rich mudrock
elastic modeling.
The PSD data obtained from the N2 gas adsorption experiments aid in predicting the
dominant gas ow mechanism with these mudrocks.
The grain density measurements from WIP technique and detailed quantitative mineral
phase modeling can provide indirect estimates of grain density of the OM matter. The
results of this study suggest a mean OM grain density of 1.369, 1.211 and 1.290 g/cm3
for the Eastern European gas shale, the Haynesville and the Niobrara mudrock samples,
respectively.
8.2

Recommendations for Future Work

The work performed in this thesis has answered some questions on mudrock pore structure

and its quantication techniques, and more importantly, posed more new ones, opening
204

new avenues for research in multiple disciplines involved in mudrock characterization. The
following recommendations are made based on the insights gained and experiences during
this research:
The biggest challenge faced during the course of the research work is the lack of consistent terminology and standardized measurement protocols applicable for mudrock
applications. Some of the terminology and denitions developed for conventional hydrocarbon system and have serious limitations to address the complexity of mudrock
systems. Sondergeld et al. (2010) advocated developing standards specic to mudrock
applications. The complexity of the mudrock system makes it challenging to have distinct clear cut denitions. In that case, for mudrocks atleast, an operationally dened
standardized terminology is recommended. With the growing and continuing interest
in mudrock plays, this is a high time for both the industry and the academia to build
a standard for mudrocks.
The sample sets used in this study did not cover the full thermal maturity spectrum, especially from the wet gas/condensate window. WIP measurements and pore structure
characterization studied from suite of wet gas/condensate window mudrock samples
are recommended.
Additional work should be carried out with other high wetting liquids such as kerosene
and Isopar using the general liquid immersion and saturation techniques. The comparison of results with WIP measurements will give insights about the accuracy of
WIP techniques and can potentially provide other information about pore structure.
Multi technique pore structure characterization is recommended. CO2 adsorption at
0 C is recommended to obtain the micropore size distribution of the samples. Integrated N2 and MIP experimental techniques and simultaneous inversion of N2 isotherm
and MIP data using porenetwork models are recommended.

205

 The N2 gas adsorption data can also be inverted to characterize the entropy of adsorptions and surface energies of the solids hosting the pore. This parameter along
with the total surface area controls supercritical adsorption behavior. Correlating the
parameters from subcritical N2 and CH4 adsorption studies at LN2 temperature, with
supercritical CH4 results will be useful. The advantage of this study is that the subcritical adsorptions experiments are easier to perform and can be used routinely for
adsorbed gas content characterization and modeling.

206

REFERENCES CITED

Adesida, A. (2011). Pore Size Distribution of Barnett Shale Using Nitrogen Adsorption Data.
Masters thesis.
Ahmed, S., Lovell, C. W., & Diamond, S. (1974). Pore Sizes and Strength of a Compacted
Clay : Technical Paper.. Technical Report Publication FHWA/IN/JHRP-74/02 Joint
Highway Research Project, Indiana Department of Transportation and Purdue University
West Lafayette, Indiana, USA.
Alexander, J., Hall, D. H., & Storey, B. C. (1981). Porosity measurements of crystalline rocks
by laboratory and geophysical methods. Institute of Geological Sciences, Report EPNU ,
(pp. 8110).
Ambrose, R., Hartman, R., Diaz-Campos, M., Akkutlu, I. Y., & Sondergeld, C. (2012).
Shale Gas-in-Place calculations part I: New pore-scale considerations. SPE Journal, 17 ,
219229.
Anders, D. E., Palacas, J. G., & Johnson, R. C. (1992). Thermal maturity of rocks and
hydrocarbon deposits, Uinta Basin, Utah. In T. D. Fouch, V. F. Nuccio, & T. C. Chidsey
(Eds.), Hydrocarbon and Mineral Resources of the Uinta Basin, Utah and Colorado, Utah
Geological Association Guidebook (Vol. 20) (pp. 5376).
Anderson, J. (1963). An improved pretreatment for mineralogical analysis of samples containing organic matter. Clays and Clay Minerals, 10 , 380388.
Aoudia, K., Miskimins, J., Harris, N., & Mnich, C. (2010). Statistical analysis of the eects
of mineralogy on rock mechanical properties of the woodford shale and the associated
impacts for hydraulic fracture treatment design, 2010, ARMA 10-303. In 44th US Rock
Mechanics Symposium, Salt Lake City, UT . Salt Lake City, UT,usa.
API RP40 (1998). Recommended practices for core analysis.
Arne Johansen, T., Ole Ruud, B., & Jakobsen, M. (2004). Eect of grain scale alignment on
seismic anisotropy and reectivity of shales. Geophysical Prospecting, 52 , 133149.
ASAP 2020, Accelerated Surface Area and Porosimetry System Operators Manual. V4.00
(2011). Asap 2020, accelerated surface area and porosimetry system operators manual.
v4.00.

207

ASTM Standard C2000 (2010). Standard test methods for apparent porosity, water absorption, apparent specic gravity, and bulk density of burned refractory brick and shape
for boling water.
ASTM Standard C830-00 (2011). Standard test methods for apparent porosity, liquid absorption, apparent specic gravity, and bulk density of refractory shapes by vacuum pressure.
Bai, J., Wang, Q., & Jiao, G. (2012). Study on the pore structure of oil shale during
low-temperature pyrolysis. Energy Procedia, 17, Part B, 1689 1696.
Bardon, C., Bieber, M., Cuiec, L., Jacquin, C., Courbot, A., Deneuville, G., Simon, J.,
Voirin, J., Espy., M., Nectoux, A., & Pellerin, A. (1983). Recommandations pour la
dtermination exprimentale de la capacit dchange de cations des milieux argileux.
Revue de Institut Francais Du Ptrole, 38 , 621626.
Barnes, G., & Gentle, I. (2005). Interfacial science. Oxford University Press Oxford.
Barnes, K. B. (1931). A Method for Determining the Eective Porosity of a Reservoir-rock.
School of Mineral Industries, State College, Pennsylvania.
Barrett, E. P., Joyner, L. G., & Halenda, P. P. (1951). The determination of pore volume and
area distributions in porous substances. i. computations from nitrogen isotherms. Journal
of the American Chemical society, 73 , 373380.
Brown, S. M., & Lard, E. W. (1974). A comparison of nitrogen and mercury pore size
distributions of silicas of varying pore volume. Powder Technology, 9 , 187190.
Brownell, W. (1977). Procedures for the characterization of devonian shales. In Proc. First
Eastern Gas Shales Symp. U.S. Dept. Energy, N.T.I.S., MERC/SP-77-5 (pp. 659666).
U.S. Dept. of Energy, Technical Information Center.
Brunauer, S., Emmett, P. H., & Teller, E. (1938). Adsorption of gases in multimolecular
layers. Journal of the American Chemical Society, 60 , 309319.
Bustin, R., Bustin, A., Cui, A., Ross, D., & Murthy Pathi, V. (2008). Impact of shale
properties on pore structure and storage characteristics. In SPE Shale Gas Production
Conference.
Caro, J. H., & Freeman, H. P. (1961). Physical structure of fertilizer materials, pore structure
of phosphate rock and triple superphosphate. Journal of Agricultural and Food Chemistry,
9 , 182186.

208

Cases, J. M., Berend, I., Besson, G., Francois, M., Uriot, J. P., Thomas, F., & Poirier, J. E.
(1992). Mechanism of adsorption and desorption of water vapor by homoionic montmorillonite. 1. the sodium-exchanged form. Langmuir, 8 , 27302739.
Chalmers, G. R., Bustin, R. M., & Power, I. M. (2012). Characterization of gas shale pore
systems by porosimetry, pycnometry, surface area, and eld emission scanning electron
microscopy/transmission electron microscopy image analyses: Examples from the barnett,
woodford, haynesville, marcellus, and doig units. AAPG bulletin, 96 , 10991119.
Chalmers, G. R. L., & Bustin, R. M. (2007). The organic matter distribution and methane
capacity of the lower cretaceous strata of northeastern british columbia, canada. International Journal of Coal Geology, 70 , 223239.
Clarkson, C., Freeman, M., He, L., Agamalian, M., Melnichenko, Y., Mastalerz, M., Bustin,
R., Radliski, A., & Blach, T. (2012a). Characterization of tight gas reservoir pore structure using USANS/SANS and gas adsorption analysis. Fuel, 95 , 371 385.
Clarkson, C., Solano, N., Bustin, R., Bustin, A., Chalmers, G., He, L., Melnichenko, Y.,
Radliski, A., & Blach, T. (2013). Pore structure characterization of North American
shale gas reservoirs using USANS/SANS, gas adsorption, and mercury intrusion. Fuel,
103 , 606 616.
Clarkson, C., Wood, J., Burgis, S., Aquino, S., & Freeman, M. (2012b). Nanopore-structure
analysis and permeability predictions for a tight gas siltstone reservoir by use of lowpressure adsorption and mercury-intrusion techniques. SPE Reservoir Evaluation & Engineering, 15 , 648661.
Clementz, D. M. (1979). Eect of oil and bitumen saturation on source-rock pyrolysis:
Geologic notes. AAPG Bulletin, 63 , 22272232.
Comisky, J., Santiago, M., McCollom, B., Buddhala, A., & Newsham, K. (2011). Sample size eects on the application of mercury injection capillary pressure for determining
the storage capacity of tight gas and oil shales. In Canadian Unconventional Resources
Conference. 15-17 November 2011, Alberta, Canada. SPE 149432-MS.
Conner, W., Cevallos-Candau, J., Weist, E., Pajares, J., Mendioroz, S., & Cortes, A. (1986).
Characterization of pore structure: porosimetry and sorption. Langmuir, 2 , 151154.
Cooles, G. P., Mackenzie, A. S., & Quigley, T. M. (1986). Calculation of petroleum masses
generated and expelled from source rocks. Organic Geochemistry, 10 , 235245.
Cui, X., Bustin, A., & Bustin, R. M. (2009). Measurements of gas permeability and diusivity
of tight reservoir rocks: dierent approaches and their applications. Geouids, 9 , 208223.

209

Curtis, M., Ambrose, R., Sondergeld, C., & Rai, C. (2011). Investigation of the relationship
between organic porosity and thermal maturity in the Marcellus Shale. In North American
Unconventional Gas Conference and Exhibition. 14-16 June 2011, The Woodlands, Texas,
USA. SPE 144370-MS.
Curtis, M., Ambrose, R., Sondergeld, C., & Sondergeld, C. (2010). Structural characterization of gas shales on the microand nanoscales. In Canadian Unconventional Resources
and International Petroleum Conference. 19-21 October 2010, Calgary, Alberta, Canada.
SPE 137693-MS.
Curtis, M. E., Cardott, B. J., Sondergeld, C. H., & Rai, C. S. (2012). Development of organic
porosity in the woodford shale with increasing thermal maturity. International Journal of
Coal Geology, 103 , 26 31.
Derkowski, A., Drits, V. A., & McCarty, D. K. (2012). Rehydration of dehydrated
dehydroxylated smectite in a low water vapor environment. American Mineralogist, 97 ,
110127.
Dewers, T. A., Heath, J., Ewy, R., & Duranti, L. (2012). Threedimensional pore networks
and transport properties of a shale gas formation determined from focused ion beam serial
imaging. International Journal of Oil, Gas and Coal Technology, 5 , 229248.
Deygout, F. (2011). Volatile emissions from hot bitumen storage tanks. Environmental
Progress & Sustainable Energy, 30 , 102112.
Diamond, S. (1970). Pore size distributions in clays. Clays and clay minerals, 18 , 723.
Diamond, S. (1971). A critical comparison of mercury porosimetry and capillary condensation pore size distributions of portland cement pastes. Cement and concrete research, 1 ,
531545.
Dorsch, J., & Katsube, T. J. (1996). Eective porosity and pore-throat sizes of mudrock
saprolite from the Nolichucky shale within Bear Creek Valley on the Oak Ridge Reservation: Implications for contaminant transport and retardation through matrix diusion,
No. ORNL/GWPO - 025 . Technical Report Oak Ridge National Lab., Environmental
Sciences Div., TN (United States).
Dorsch, J., Katsube, T. J., Sanford, W. E., Dugan, B. E., & Tourkow, L. M. (1996). Effective porosity and pore-throat sizes of Conasauga Group mudrock: Application, test and
evaluation of petrophysical techniques, No. ORNL/GWPO021 . Technical Report Oak
Ridge National Lab., Environmental Sciences Div., TN (United States).

210

Drits, V. A., & McCarty, D. K. (2007). The nature of structure-bonded H2 O in illite and leucophyllite from dehydration and dehydroxylation experiments. Clays and Clay Minerals,
55 , 4558.
Dubinin, M. (1966). Porous structure and adsorption properties of active carbons. In P. L. W.
Jr. (Ed.), Chemistry and Physics of Carbon chapter 2. (pp. 51 120). New York, U.S.A.:
Marcel Dekker, Inc. volume 2.
Dullien, F., & Dhawan, G. (1974). Characterization of pore structure by a combination of
quantitative photomicrography and mercury porosimetry. Journal of Colloid and Interface
Science, 47 , 337 349.
Dullien, F. A. (1991). Porous media: uid transport and pore structure. (2nd ed.). Academic
press.
Echeverra, J. C., Morera, M. T., Mazkiarn, C., & Garrido, J. J. (1999). Characterization
of the porous structure of soils: adsorption of nitrogen (77 k) and carbon dioxide (273 k),
and mercury porosimetry. European Journal of Soil Science, 50 , 497503.
Elbaharia, M. (2012). High Temperature Characterization of Oil Shales: The Green River
Formation. Masters thesis Colorado School of Mines.
Feoktistov, N., Golubev, V., Hutchison, J., Kurduykov, D., Pevtsov, A., Ratnikov, V.,
Sloan, J., & Sorokin, L. (2001). Tem and hrem study of 3d silicon and platinum nanoscale
assemblies in dielectric opal matrix. Semiconductor science and technology, 16 , 955.
Folk, R. L. (1974). Petrology of sedimentary rocks.. Hemphill Publishing Company.
Franklin, J., Vogler, U., Szlavin, J., Edmond, J., & Bieniawski, Z. (1981). Suggested methods
for determining water content, porosity, density, absorption and related properties and
swelling and slake-durability index properties. In E. Brown (Ed.), Rock Characterization
Testing and Monitoring, ISRM Suggested Methods (pp. 7994). Oxford: Pergamon Press.
Friedman, G. (2003). Classication of sedimentary rocks. In G. V. Middleton (Ed.), Encyclopedia of sediments and sedimentary rocks Kluwer Academic encyclopedia of earth sciences
series. Dordrecht, The Netherlands ; Boston: Kluwer Academic Publishers.
Gaucher, E., Robelin, C., Matray, J. M., Ngrel, G., Gros, Y., Heitz, J. F., Vinsot, A.,
Rebours, H., Cassagnabre, A., & Bouchet, A. (2004). ANDRA underground research
laboratory: interpretation of the mineralogical and geochemical data acquired in the
Callovian-Oxfordian formation by investigative drilling. Physics and Chemistry of
the Earth, Parts A/B/C , 29 , 5577.

211

Goldstrand, P., Menefee, L., & Dreier, R. (1995). Porosity development in the Copper Ridge
Dolomite and Maynardville Limestone, Bear Creek Valley and Chestnut Ridge, Tennessee.
Technical Report Oak Ridge Y-12 Plant, TN (United States); Nevada Univ., Reno, NV
(United States). Dept. of Geology; Appalachian State Univ., Boone, NC (United States).
Dept. of Geology; Oak Ridge National Lab., TN (United States).
Gregg, S. J., & Sing, K. S. W. (1983). Adsorption, Surface Area, and Porosity. (2nd ed.).
New York: Academic Press.
Groen, J., Peer, L., & Prez-Ramrez, J. (2002). Incorporation of appropriate contact
angles in textural characterization by mercury porosimetry. In J. R. F. Rodriguez-Reinoso,
B. McEnaney, & K. Unger (Eds.), Characterization of Porous Solids VI Proceedings of the
6th International Symposium on the Characterization of Porous Solids (COPS-VI) (pp.
91 98). Elsevier volume 144 of Studies in Surface Science and Catalysis.
Groen, J. C., Peer, L. A. A., & Prez-Ramrez, J. (2003). Pore size determination in
modied micro-and mesoporous materials. pitfalls and limitations in gas adsorption data
analysis. Microporous and Mesoporous Materials, 60 , 117.
Hammes, U., Hamlin, H. S., & Ewing, T. E. (2011). Geologic analysis of the upper jurassic
haynesville shale in east texas and west louisiana. AAPG bulletin, 95 , 16431666.
Han, G., Osmond, J., & Zambonini, M. (2010a). A USD 100 million Rock: Bitumen in
the Deepwater Gulf of Mexico. SPE Drilling & Completion, 25 , 290299.
Han, X., Jiang, X., Yan, J., & Liu, J. (2010b). Eects of retorting factors on combustion
properties of shale char. 2. pore structure. Energy & Fuels, 25 , 97102.
Han, X., Jiang, X., Yu, L., & Cui, Z. (2006). Change of pore structure of oil shale particles
during combustion. part 1. evolution mechanism. Energy & fuels, 20 , 24082412.
Haynes, W. M., Lide, D. R., & Bruno, T. (2012). CRC Handbook of Chemistry and Physics
20122013 . (93rd ed.). CRC Press.
Hornby, B., Schwartz, L., & Hudson, J. (1994). Anisotropic eectivemedium modeling
of the elastic properties of shales. GEOPHYSICS , 59 , 15701583.
Howard, J. J. (1991). Porosimetry measurement of shale fabric and its relationship to illite/smectite diagenesis. Clays and Clay Minerals, 39 , 355361.
Hu, Y., Devegowda, D., Striolo, A., Ho, T. A., Phan, A., Civan, F., & Sigal, R. (2013). A
pore scale study decribing the dynamics of slickwater distribution in shale gas formations
following hydralic fracturing. In Unconventional Resources Conference USA. 10-12 April
2013, The Woodlands, Texas, USA. SPE 164552-MS.

212

Ibrahimov, R. A., & Bissada, K. (2010). Comparative analysis and geological signicance of
kerogen isolated using open-system (palynological) versus chemically and volumetrically
conservative closed-system methods. Organic Geochemistry, 41 , 800811.
I.S.O 9277:2010(E) (2010). Determination of the specic surface area of solids by gas adsorption - BET method.
Jackson, M., & Barak, P. (2005). Soil Chemical Analysis: Advanced Course. Parallel Press,
University of Wisconsin-Madison Libraries.
Jarvie, D. M. (1991). Total organic carbon (TOC) analysis. In R. Merrill (Ed.), Treatise
of Petroleum Geology: Handbook of Petroleum Geology, Source and Migration Processes
and Evaluation Techniques chapter 11. (pp. 113118). Tulsa: American Association of
Petroleum Geologists.
Javadpour, F. (2009). Nanopores and apparent permeability of gas ow in mudrocks (shales
and siltstone). Journal of Canadian Petroleum Technology, 48 , 1621.
Johnson, W. F., Walton, D. K., Keller, H. H., & Couch, E. J. (1975). In situ retorting of
oil shale rubble: a model of heat transfer and product formation in oil shale particles. Q.
Colo. Sch. Mines;(United States), 70 .
Joyner, L. G., Barrett, E. P., & Skold, R. (1951). The determination of pore volume and area
distributions in porous substances. ii. comparison between nitrogen isotherm and mercury
porosimeter methods. Journal of the American Chemical Society, 73 , 31553158.
Jullien, M., Raynal, J., Kohler, ., & Bildstein, O. (2005). Physicochemical reactivity in
clay-rich materials: tools for safety assessment. Oil & gas science and technology, 60 ,
107120.
Kale, S., Rai, C., & Sondergeld, C. (2010). Rock typing in gas shales. In SPE Annual Technical Conference and Exhibition. 20-22 September 2010, Florence, Italy. SPE 134539-MS.
Karastathis, A. (2007). Petrophysical measurements on Tight Gas Shales. Masters thesis
University of Oklahoma.
Karniadakis, G., Beskok, A., & Aluru, N. (2005). Microows and Nanoows: Fundamentals
and Simulation (Interdisciplinary Applied Mathematics). (1st ed.). Springer.
Katahara, K. W. (). Clay mineral elastic properties. In SEG Technical Program Expanded
Abstracts 1996 chapter 448. (pp. 16911694).
Katsube, T. J., & Kamineni, D. C. (1983). Eect of alteration on pore structure of crystalline
rocks; core samples from Atikokan, Ontario. The Canadian Mineralogist, 21 , 637646.

213

Katsube, T. J., & Scromeda, N. (1991). Eective porosity measuring procedure for low
porosity rocks. Geological Survey of Canada, Paper, 91-1E, 291297.
Katsube, T. J., Scromeda, N., & Williamson, M. (1992). Eective porosity of tight shales
from the Venture Gas Field, Oshore Nova Scotia. Geological Survey of Canada, Paper,
92-1D, 111119.
Kim, J.-W., Peacor, D. R., Tessier, D., & Elsass, F. (1995). A technique for maintaining
texture and permanent expansion of smectite interlayers for tem observations. Clays and
Clay Minerals, 43 , 5157.
Klinkenberg, L. (1941). The permeability of porous media to liquids and gases. Drilling and
production practice, .
Klobes, P., Meyer, K., & Munro, R. G. (2006). Porosity and specic surface area measurements for solid materials. US Department of Commerce, Technology Administration,
National Institute of Standards and Technology.
Landon, S. M., Longman, M. W., & Luneau, B. A. (2001). Hydrocarbon source rock potential
of the Upper Cretaceous Niobrara Formation, Western Interior Seaway of the Rocky
Mountain region. The Mountain Geologist, .
Larter, S. (1988). Some pragmatic perspectives in source rock geochemistry. Marine and
Petroleum Geology, 5 , 194204.
Levitz, P., Ehret, G., Sinha, S., & Drake, J. (1991). Porous vycor glass: The microstructure as probed by electron microscopy, direct energy transfer, small-angle scattering, and
molecular adsorption. The Journal of chemical physics, 95 , 61516161.
Lewan, M. D. (1994). Assessing natural oil expulsion from source rocks by laboratory pyrolysis. In L. Magoon, & W. Dow (Eds.), The Petroleum SystemFrom Source to Trap,
Association of Petroleum Geologists Memoir chapter 11. (pp. 201210). volume 60.
Lippens, B., & De Boer, J. (1965). Studies on pore systems in catalysts: V. the t method.
Journal of Catalysis, 4 , 319323.
Longman, M. W., Luneau, B. A., & Landon, S. M. (1998). Nature and distribution of
Niobrara lithologies in the Cretaceous Western Interior Seaway of the Rocky Mountain
region. The Mountain Geologist, .
Loucks, R. G., Reed, R. M., Ruppel, S. C., & Hammes, U. (2012). Spectrum of pore types and
networks in mudrocks and a descriptive classication for matrix-related mudrock pores.
AAPG bulletin, 96 , 10711098.

214

Loucks, R. G., Reed, R. M., Ruppel, S. C., & Jarvie, D. M. (2009). Morphology, genesis, and
distribution of nanometer-scale pores in siliceous mudstones of the Mississippian Barnett
Shale. Journal of Sedimentary Research, 79 , 848861.
Lowell, S., & Shields, J. E. (1991). Powder surface area and porosity volume 2. Springer.
Lu, X.-C., Li, F.-C., & Watson, A. T. (1995). Adsorption measurements in devonian shales.
Fuel, 74 , 599603.
Luel, D., & Guidry, F. (1989). Core analysis results Comprehensive Study Wells Devonian
Shales: Topical Report July 1989 . Technical Report Restech Houston, Inc.
Luel, D. L., & Guidry, F. K. (1992). New core analysis methods for measuring reservoir
rock properties of Devonian shale. Journal of Petroleum Technology, 44 , 11841190.
Luel, D. L., Guidry, F. K., & Curtis, J. B. (1992). Evaluation of Devonian shale with new
core and log analysis methods. Journal of Petroleum Technology, 44 , 11921197.
Mason, G. (1982). The eect of pore space connectivity on the hysteresis of capillary condensation in adsorptiondesorption isotherms. Journal of Colloid and Interface Science,
88 , 3646.
Mayer, L. M. (1994). Surface area control of organic carbon accumulation in continental
shelf sediments. Geochimica et Cosmochimica Acta, 58 , 12711284.
Mayer, L. M. (1999). Extent of coverage of mineral surfaces by organic matter in marine
sediments. Geochimica et Cosmochimica Acta, 63 , 207215.
Mayer, L. M., Schick, L. L., Hardy, K. R., Wagai, R., & McCarthy, J. (2004). Organic
matter in small mesopores in sediments and soils. Geochimica et Cosmochimica Acta, 68 ,
38633872.
McCarty, D. K. (2002). Quantitative mineral analysis of clay bearing mixtures: The Reynolds
Cup contest. IUCr CPD Newsletter, 27 , 1216.
Melnyk, T. W., & Skeet, A. M. M. (1986). An improved technique for the determination of
rock porosity. Canadian Journal of Earth Sciences, 23 , 10681074.
Meyer, K., & Klobes, P. (1999). Comparison between dierent presentations of pore size
distribution in porous materials. Fresenius journal of analytical chemistry, 363 , 174178.
MicroActive DataMaster v5.00 software manual (2011). Micromeritics microactive datamaster v5.00 software manual.

215

Militzer, B., Wenk, H.-R., Stackhouse, S., & Stixrude, L. (2011). First-principles calculation
of the elastic moduli of sheet silicates and their application to shale anisotropy. American
Mineralogist, 96 , 125137.
Milliken, K. L., Esch, W. L., Reed, R. M., & Zhang, T. (2012). Grain assemblages and
strong diagenetic overprinting in siliceous mudrocks, barnett shale (mississippian), fort
worth basin, texas. AAPG bulletin, 96 , 15531578.
Milliken, K. L., & Reed, R. M. (2010). Multiple causes of diagenetic fabric anisotropy in
weakly consolidated mud, nankai accretionary prism, IODP Expedition 316. Journal of
Structural Geology, 32 , 18871898.
Milliken, K. L., Rudnicki, M., Awwiller, D. N., & Zhang, T. (2013). Organic matterhosted
pore system, Marcellus Formation (Devonian), Pennsylvania. AAPG Bulletin, 97 , 177
200.
Milner, M., McLin, R., & Petriello, J. (2010). Imaging texture and porosity in mudstones and
shales: comparison of secondary and ion-milled backscatter sem methods. In Canadian
Unconventional Resources and International Petroleum Conference. 19-21 October 2010,
Calgary, Alberta, Canada. SPE 138975-MS.
Modica, C. J., & Lapierre, S. G. (2012). Estimation of kerogen porosity in source rocks as
a function of thermal transformation: Example from the mowry shale in the powder river
basin of wyoming. AAPG bulletin, 96 , 87108.
Monson, B., & Parnell, J. (1992). The origin of gilsonite vein deposits in the Uinta Basin,
Utah. In T. Fouch, V. Nuccio, & T. Chidsey (Eds.), Hydrocarbon and Mineral Resources
of the Uinta Basin, Utah and Colorado, Utah Geological Association Guidebook (Vol. 20)
(p. 257270).
Moyano, B., Spikes, K., Johansen, T., & Mondol, N. (2012). Modeling compaction eects
on the elastic properties of clay-water composites. GEOPHYSICS , 77 , D171D183.
Murray, K. L., Seaton, N. A., & Day, M. A. (1999). Use of mercury intrusion data, combined with nitrogen adsorption measurements, as a probe of pore network connectivity.
Langmuir, 15 , 81558160.
Naumov, S. (2009). Hysteresis Phenomena in Mesoporous Materials. Ph.D. thesis Universitt
Leipzig.
Neaman, A., Pelletier, M., & Villieras, F. (2003). The eects of exchanged cation, compression, heating and hydration on textural properties of bulk bentonite and its corresponding
puried montmorillonite. Applied Clay Science, 22 , 153 168.

216

Nimmo, J. R. (2004). Porosity and pore size distribution. Encyclopedia of Soils in the
Environment., (pp. 295303).
Noble, R. A., Kaldi, J. G., & Atkinson, C. D. (1997). Oil saturation in shales: applications in
seal evaluation. In R. Surdam (Ed.), Seals, traps, and the petroleum system: AAPG Memoir 67 chapter 2. (pp. 1330). Tulsa, Oklahoma, U.S.A. 74101: The American Association
of Petroleum Geologists.
Nuhfer, E. B. (1979). Determination of density and porosity. In E. Nuhfer, J. Florence,
J. Clagett, J. Renton, & R. Romanosky (Eds.), Procedures for petrophysical, mineralogical
and geochemical characterization of ne-grained clastic rocks and sediments chapter 4.
(pp. 1124). Morgantown, West Virginia, U.S.A: United States Department of Energy.
Nwachukwu, J. I., & Barker, C. (1985). Variations in kerogen densities of sediments from
the orinoco delta, venezuela. Chemical Geology, 51 , 193198.
Okiongbo, K. S., Aplin, A. C., & Larter, S. R. (2005). Changes in type II kerogen density as
a function of maturity: Evidence from the Kimmeridge Clay Formation. Energy & fuels,
19 , 24952499.
Omotoso, O., McCarty, D. K., Hillier, S., & Kleeberg, R. (2006). Some successful approaches
to quantitative mineral analysis as revealed by the 3rd reynolds cup contest. Clays and
Clay Minerals, 54 , 748760.
Orsini, L., & Remy, J. (1976). Using chloride cobaltihexamine for the simultaneous determination of exchange capacity and exchangeable bases of soils science. Bull. AFES , 4 ,
269279.
Passey, Q., Bohacs, K., Esch, W., Klimentidis, R., & Sinha, S. (2010). From oil-prone
source rock to gas-producing shale reservoir-geologic and petrophysical characterization of
unconventional shale gas reservoirs. In Proc. International Oil and Gas Conference and
Exhibition, 8-10 June 2010, Beijing, China, SPE 131350-MS. Beijing, China.
Penumadu, D., & Dean, J. (2000). Compressibility eect in evaluating the pore-size distribution of kaolin clay using mercury intrusion porosimetry. Canadian Geotechnical Journal,
37 , 393405.
Plummer, F. B., & Tapp, P. F. (1943). Technique of testing large cores of oil sands. AAPG
Bulletin, 27 , 6484.
Pollastro, R. M. (1992). Natural fractures, composition, cyclicity, and diagenesis of the Upper
Cretaceous Niobrara Formation, Berthoud eld, Colorado. chapter Geological Studies
Relevant to Horizontal Drilling: Examples from Western North America. (pp. 243255).
Rocky Mountain Association of Geologists.

217

Prasad, M. (2001). Mapping impedance microstructures in rocks with acoustic microscopy.

The Leading Edge, 20 , 172179.
Prasad, M., Mba, K., Sadler, T., & Batzle, M. (2011). Maturity and impedance analysis of
organic-rich shales. SPE Reservoir Evaluation & Engineering, 14 , 533543.
Quirk, J., & Aylmore, L. (1971). Domains and quasi-crystalline regions in clay systems. Soil
Science Society of America Journal, 35 , 652654.
Rigby, S. P., & Edler, K. J. (2002). The inuence of mercury contact angle, surface tension,
and retraction mechanism on the interpretation of mercury porosimetry data. Journal of
Colloid and Interface Science, 250 , 175 190.
Rigby, S. P., & Fletcher, R. S. (2004). Interfacing mercury porosimetry with nitrogen sorption. Particle & Particle Systems Characterization, 21 , 138148.
Rigby, S. P., Fletcher, R. S., & Riley, S. N. (2004). Characterisation of porous solids using
integrated nitrogen sorption and mercury porosimetry. Chemical Engineering Science, 59 ,
41 51.
Robl, T., Taulbee, D., Barron, L., & Jones, W. (1987). Petrologic chemistry of a devonian
type ii kerogen. Energy & fuels, 1 , 507513.
Robl, T. L., & Taulbee, D. N. (1995). Demineralization and kerogen maceral separation
and chemistry. In Composition, Geochemistry and Conversion of Oil Shales (pp. 3550).
Springer.
Ross, D. J. K., & Bustin, R. M. (2009). The importance of shale composition and pore
structure upon gas storage potential of shale gas reservoirs. Marine and Petroleum Geology,
26 , 916927.
Rouquerol, J., Avnir, D., Everett, D. H., Fairbridge, C., Haynes, M., Pernicone, N., Ramsay,
J. D. F., Sing, K. S. W., & Unger, K. K. (1994). Guidelines for the characterization of
porous solids. Studies in surface science and catalysis, 87 , 19.
Rouquerol, J., Llewellyn, P., & Rouquerol, F. (2007). Is the bet equation applicable to
microporous adsorbents? Studies in surface science and catalysis, 160 , 4956.
Rouquerol, J., Rouquerol, F., & Sing, K. S. W. (1998). Absorption by powders and porous
solids. Academic press.
Rutherford, D. W., Chiou, C. T., & Eberl, D. D. (1997). Eects of exchanged cation on the
microporosity of montmorillonite. Clays and Clay Minerals, 45 , 534543.

218

Sarker, R. (2010). Elastic and ow properties of shales. Ph.D. thesis Colorado School of
Mines.
Schieber, J. (2010). Common themes in the formation and preservation of intrinsic porosity
in shales and mudstones-illustrated with examples across the phanerozoic. In SPE Unconventional Gas Conference, 23-25 February 2010, Pittsburgh, Pennsylvania, USA, SPE
132370-MS. Pittsburgh, Pennsylvania, USA.
Schieber, J., & Zimmerle, W. (1998). The history and promise of shale research. In
J. Schieber, W. Zimmerle, & P. Sethi (Eds.), Shales and Mudstones: Basin studies, sedimentology, and paleontology (pp. 110). Stuttgart: Schweizerbartsche Verlagsbuchhandlung volume 1.
Schimmelmann, A., Lewan, M. D., & Wintsch, R. P. (1999). D/H isotope ratios of kerogen,
bitumen, oil, and water in hydrous pyrolysis of source rocks containing kerogen types i, ii,
iis, and iii. Geochimica et Cosmochimica Acta, 63 , 37513766.
Schlumberger Oileld Glossary (2013). http://www.glossary.oilfield.slb.com/.
Schmitt, M., Fernandes, C. P., da Cunha Neto, J. A., Wolf, F. G., & dos Santos, V. S. (2013).
Characterization of pore systems in seal rocks using nitrogen gas adsorption combined with
mercury injection capillary pressure techniques. Marine and Petroleum Geology, 39 , 138
149.
Schnackenberg, W. D., & Prien, C. H. (1953). Eect of solvent properties in thermal decomposition of oil shale kerogen. Industrial & Engineering Chemistry, 45 , 313322.
Schoeld, R. W., Fane, A. G., & Fell, C. J. D. (1990). Gas and vapour transport through
microporous membranes. i. knudsen-poiseuille transition. Journal of Membrane Science,
53 , 159 171.
Schrodt, J. T., & Ocampo, A. (1984). Variations in the pore structure of oil shales during
retorting and combustion. Fuel, 63 , 15231527.
Shapiro, L. (1975). Rapid analysis of silicate, carbonate and phosphate rocks - revised edition
volume 1401 of USGS Bulletin. US Government Printing Oce.
Sharp, A. G. (1981). Design curves for oceanographic pressure-resistant housings. Technical
Report Technical Memorandum 3-81 Woods Hole Oceanographic Institution Woods Hole,
Massachusetts, USA.
Sigal, R. F. (2009). A methodology for blank and conformance corrections for high pressure
mercury porosimetry. Measurement Science and Technology, 20 , 045108.

219

Sigal, R. F. (2012). Mercury capillary pressure measurements on Barnett core. http:

//shale.ou.edu/Content/Member%20Area/Papers/CapillaryPressure.pdf, Last Accessed on March 21, 2013. Submitted in Petrophysics.
Sills, I. D., Aylmore, L. A. G., & Quirk, J. P. (1973). A comparison between mercury
injection and nitrogen sorption as methods of determining pore size distributions. Soil
Science Society of America Journal, 37 , 535537.
Sing, K., Everett, D., Haul, R., Moscou, L., Peirotti, R., Rouquerol, J. et al. (1985). IUPAC
commission on colloid and surface chemistry including catalysis. Pure Appl. Chem, 57 ,
603619.
Siskin, M., Katritzky, A. R. et al. (1991). Reactivity of organic compounds in hot water:
Geochemical and technological implications. Science(Washington), 254 , 231237.
Sondergeld, C., Newsham, K., Comisky, J., Rice, M., & Rai, C. (2010). Petrophysical considerations in evaluating and producing shale gas resources. In Proc. SPE Unconventional
Gas Conference, 23-25 February 2010, Pittsburgh, Pennsylvania, USA, SPE 131768MS. Pittsburgh, Pennsylvania, USA.
Spears, R. W., Dudus, D., Foulds, A., Passey, Q., Sinha, S., & Esch, W. L. (2011). Shale
gas core analysis: Strategies for normalizing between laboratories and a clear need for
standard materials. In Proc. 52nd SPWLA Annual Logging Symposium, Colorado Springs,
CO. Colorado Springs, CO,USA.
rodo, J., Clauer, N., Hu, W., Dudek, T., & Bana, M. (2009). K-ar dating of the lower
palaeozoic k-bentonites from the baltic basin and the baltic shield: implications for the
role of temperature and time in the illitization of smectite. Clay minerals, 44 , 361387.
rodo, J., Drits, V. A., McCarty, D. K., Hsieh, J. C., & Eberl, D. D. (2001). Quantitative
X-ray diraction analysis of clay-bearing rocks from random preparations. Clays and Clay
Minerals, 49 , 514528.
rodo, J., & Kawiak, T. (2012). Mineral compositional trends and their correlations with
petrophysical and well-logging parameters revealed by QUANTA + BESTMIN analysis:
Miocene of the Carpathian Foredeep, Poland. Clays and Clay Minerals, 60 , 6375.
rodo, J., & McCarty, D. K. (2008). Surface area and layer charge of smectite from CEC
and EGME/H2 O-retention measurements. Clays and Clay Minerals, 56 , 155174.
Strapo, D., Mastalerz, M., Schimmelmann, A., Drobniak, A., & Hasenmueller, N. R.
(2010). Geochemical constraints on the origin and volume of gas in the new albany shale
(devonianmississippian), eastern illinois basin. AAPG bulletin, 94 , 17131740.

220

Taylor, B. N., & Kuyatt, C. E. (1994). Guidelines for evaluating and expressing the uncertainty of NIST measurement results. Natl. Inst. Stand. Technol. Tech. Note, 1297 .
Tisot, P. R. (1962). Properties of green river oil shale determined from nitrogen adsorption
and desorption isotherms. Journal of Chemical and Engineering Data, 7 , 405410.
Tiwari, P., Deo, M., Lin, C., & Miller, J. (2013). Characterization of oil shale pore structure
before and after pyrolysis by using Xray micro CT. Fuel, .
Ungerer, P., Behar, E., & Discamps, D. (1981). Tentative calculation of the overall volume
expansion of organic matter during hydrocarbon genesis from geochemistry data. implications for primary migration. In M. Bjory (Ed.), Advances in organic geochemistry (pp.
129135). volume 10.
Vandenbroucke, M., & Largeau, C. (2007). Kerogen origin, evolution and structure. Organic
Geochemistry, 38 , 719 833.
Vernik, L., & Liu, X. (1997). Velocity anisotropy in shales: A petrophysical study. GEOPHYSICS , 62 , 521532.
Walls, J. D., Diaz, E., Derzhi, N., Grader, A., Dvorkin, J., Arredondo, S., Carpio, G., &
Sinclair, S. W. (2011). Eagle ford shale reservoir properties from digital rock physics. First
Break, 29 , 97101.
Wang, F. P., & Hammes, U. (2010). Eects of petrophysical factors on Haynesville uid ow
and production. World Oil, 231 , 6366.
Washburn, E. W. (1921a). The dynamics of capillary ow. Physical review, 17 , 273.
Washburn, E. W. (1921b). Porosity. I. purpose of the investigation. II. porosity and the
mechanism of absorption. Journal of the American Ceramic Society, 4 , 916922.
Washburn, E. W., & Bunting, E. N. (1922). Porosity: V. recommended procedures for
determining porosity by methods of absorption. Journal of the American Ceramic Society,
5 , 4856.
Washburn, E. W., & Fooyitt, F. F. (1921). Porosity: III. water as an absorption liquid.
Journal of the American Ceramic Society, 4 , 961982.
Webb, P. A., & Orr, C. (1997). Analytical methods in ne particle technology volume 229.
Micromeritics Norcross, GA.
Weiler, R., & Mills, A. (1965). Surface properties and pore structure of marine sediments.
Deep Sea Research and Oceanographic Abstracts, 12 , 511 529.

221

Wilhelms, A., Larter, S. R., & Leythaeuser, D. (1991). Inuence of bitumen-2 on rock-eval
pyrolysis. Organic Geochemistry, 17 , 351 354.
Yen, T., & Chilingar, G. V. (1976). Chapter 1 introduction to oil shales. In T. F. Yen, &
G. V. Chilingarian (Eds.), Oil Shale (pp. 1 12). Elsevier volume 5 of Developments in
Petroleum Science.
Zargari, S., Prasad, M., Mba, K. C., & Mattson, E. D. (). Organic maturity, elastic properties
and textural characteristics of self resourcing reservoirs. Submitted to Geophysics.
Zhang, T., Ellis, G. S., Ruppel, S. C., Milliken, K., & Yang, R. (2012). Eect of organicmatter type and thermal maturity on methane adsorption in shale-gas systems. Organic
Geochemistry, 47 , 120 131.

222

APPENDIX A - SAMPLE CALCULATION WIP

A example sample calculation for porosity based on WIP measurements for sample SS11
Table A.1 is shown here.
Table A.1: WIP measurement data for sample SS11
Reading
1
2
3
4
5

Sat.W t.Air
(g)
10.2530
10.2521
10.2512
10.2543
10.2548

Sat.W t.Sub
(g)
6.5138
6.5147
6.5156
6.5145
6.5146

DryW t.Air
(g)
10.135

Mean
Uncertainty

10.2531
0.0030

6.5146
0.0013

10.135
0.005

Uncertainty = 2 * measured sample standard deviation

Uncertainty based on based on repeated weight measurements on standard weights

Water Temperature (T) = 22.4 C

A.1

Calculation:

Water Density
H2 0 = 0.0000053 T 2 + 0.0000081 T + 1.0001627
= 0.0000053 (22.4)2 + 0.0000081 (22.4) + 1.0001627 = 0.9977 g/cm3
Bulk density
"

Sat.W t.Air
B =
(H2 0 air ) + air
Sat.W t. Air Sat.W t.Sub
223

10.2531
(0.9977 0.0012) + 0.0012
10.2531 6.5146


= 2.734 g/cm3
Uncertainty in bulk density

U (B ) =

v
u
u
t

(B )
(Sat.W t.Air )

v
u 

u
u (

)
air
u
H2 0
(Sat.W t.
u

u
t

U (Sat.W t.Air ) +

(B )
(Sat.W t.Sub )

!2

U (Sat.W t.Sub )2

2
Sat.W t. Air
U (Sat.W t.Air )2
(Sat.W t. Air Sat.W t.Sub )2

2
Sat.W t. Air
(H2 0 air )
U (Sat.W t.Sub )2
(Sat.W t. Air Sat.W t.Sub )2

v
u
u
u
u
u
u
t

!2

1
Air Sat.W t.Sub )

(H2 0 air )

+ (H2 0 air )

Sat.W t. Sub
(Sat.W t. Air Sat.W t.Sub )2

2

Sat.W t. Air
(Sat.W t. Air Sat.W t.Sub )2

2

U (Sat.W t.Air )2
U (Sat.W t.Sub )2

v
u 

2
u
6.5146
u (0.9977 0.0012)
0.00302
u
(10.25316.5146)2
u 
2

u
t
10.2531
2

+ (0.9977 0.0012)

(10.25316.5146)2

= 0.002 g/cm3

0.0013

Grain Density

"

DryW tAir
G =
(H2 0 air ) + air
DryW tAir Sat.W t.Sub
=

10.135
(0.9977 0.0012) + 0.0012
10.135 6.5146


= 2.791 g/cm3

Uncertainty in grain density

U (G )

224

v
u
u
t

(G )
(DryW tAir )

v
u
u
u
u
u
u
t

!2

U (DryW tAir ) +

(H2 0 air )

+ (H2 0 air )

v
u
u
u
u
u
u
t

(bulk )
(Sat.W t.Sub )

Sat.W t. Sub
(DryW tAir Sat.W t.Sub )2

2

DryW tAir
(DryW tAir Sat.W t.Sub )2

(0.9977 0.0012)

+ (0.9977 0.0012)

2

6.5146
(10.1356.5146)2

10.135
(10.1356.5146)2

!2

U (Sat.W t.Sub )2

U (DryW tAir )2
U (Sat.W t.Sub )2

2

2

0.0052
0.00132

= 0.003 g/cm3

Water Immersion Porosity (WIP)

W IP =

(2.734 2.791)
(B G )
100 =
100 = 3.18
(H2 0 G )
(0.9977 2.791)

Uncertainty in WIP
U (W IP ) =

v"
u
u
t 100

v"
u
u
t 100

v
u
u
t

W IP
B

1
(H2 0 G )

1
(0.9977 2.791)

#2

#2

!2

U (B ) +
2

"

U (B ) + 100

"

0.0022 + 100
= 0.17

225

W IP
G

!2

U (G )2

(B H2 0 )
2

(H2 0 G )

#2

U (G )2

(2.734 0.9977)
(0.9977 2.791)2

#2

0.0032

APPENDIX B - MEASURED POROSITY DATASET

226

Table B.1: Measured bulk density, grain density and porosity of all the samples by dierent
techniques.
WIP

GRI

B
g/cm3

G
g/cm3

SS1-1
SS1-2
SS1-3
SS1-4
SS1-5
SS1-6
SS1-7
SS1-8
SS1-9
SS1-10
SS1-11
SS1-12
SS1-13
SS1-14
SS1-15
SS1-16
SS1-17
SS1-18
SS1-19
SS1-20
SS1-21
SS1-22

2.734(2)
2.701(3)
2.684(2)
2.698(2)
2.681(2)
2.661(4)
2.678(4)
2.659(3)
2.693(2)
2.699(4)

2.790(3)
2.796(3)
2.760(3)
2.791(3)
2.753(2)
2.739(4)
2.751(3)
2.753(2)
2.750(3)
2.763(3)

3.11(17)
5.26(23)
4.32(19)
5.20(18)
4.10(17)
4.47(31)
4.14(28)
5.35(22)
3.28(20)
3.66(26)

2.446(2)
2.596(2)
2.489(2)
2.457(2)
2.505(3)
2.532(4)
2.563(4)
2.661(2)
2.671(6)
2.657(2)
2.627(2)

2.540(2)
2.679(3)
2.574(2)
2.548(2)
2.602(3)
2.624(2)
2.662(3)
2.713(2)
2.726(3)
2.733(2)
2.708(3)

6.10(18)
4.94(20)
5.43(19)
5.89(17)
6.05(27)
5.70(26)
5.93(28)
3.03(18)
3.21(36)
4.36(18)
4.72(20)

SS2-1
SS2-2
SS2-3
SS2-4
SS2-5
SS2-6
SS2-7
SS2-8
SS2-9
SS2-10
SS2-11
SS2-12
SS2-13
SS2-14
SS2-15
SS2-16
SS2-17
SS2-18
SS2-19
SS2-20
SS2-21

2.627(4)
2.727(3)
2.496(4)
2.468(2)
2.698(6)
2.472(5)
2.606(2)
2.622(13)
2.648(10)
2.616(16)
2.523(7)
2.495(4)
2.474(3)
2.494(5)
2.432(4)
2.534(8)
2.480(8)
2.554(13)
2.496(7)
2.524(10)
2.530(12)

2.799(5)
2.845(5)
2.603(4)
2.629(5)
2.755(6)
2.527(5)
2.758(5)
2.804(12)
2.810(12)
2.798(18)
2.715(8)
2.593(5)
2.557(5)
2.653(5)
2.658(5)
2.683(9)
2.620(8)
2.704(12)
2.651(8)
2.683(11)
2.702(11)

9.58(34)
6.36(31)
6.67(36)
9.86(28)
3.25(46)
3.60(49)
8.69(29)
10.09(94)
8.95(83)
10.14(125)
11.21(58)
6.12(39)
5.31(36)
9.65(41)
13.60(38)
8.86(69)
8.64(66)
8.79(102)
9.36(62)
9.41(83)
10.08(90)

ID

B
G
g/cm3 g/cm3
2.664 2.723
2.682 2.703
2.679 2.694

2.694 2.710
2.435 2.530

2.486 2.706
2.652 2.661
2.670 2.687
2.270 2.686

MIP

B
G
g/cm3 g/cm3

G
g/cm3

2.740
2.728
2.706
2.705
2.701
2.703
2.698
2.689
2.730
2.727
2.666
2.505
2.625
2.522
2.509
2.546
2.569
2.600
2.679
2.696
2.684
2.660

0.20
0.59
0.39
0.61
0.35
0.69
0.55
0.69
0.43
0.59
1.05
1.34
0.80
1.28
1.43
1.47
1.20
1.13
0.38
0.35
0.57
0.58

2.775
2.767
2.763
2.762
2.744
2.729
2.745
2.748
2.743
2.749
2.701
2.619
2.677
2.582
2.558
2.620
2.635
2.660
2.734
2.739
2.716
2.702

2.711 2.815
2.507 2.622
2.537 2.696

3.42
3.71
4.72

2.734
2.712
2.695
2.688
1.09 2.691
2.684
2.683
2.671
0.98 2.718
4.04 2.711
2.638
2.472
2.604
2.489
2.473
8.49 2.508
2.539
2.571
0.72 2.669
0.67 2.687
15.93 2.669
2.645
2.59
1.19

2.626
2.694
2.452
2.430
2.678
2.438
2.609

2.805
2.845
2.636
2.444
2.774
2.579
2.758

7.92
5.93
7.62
1.75
3.79
6.23
7.67

2.653
2.633
2.487
2.538
2.460
2.458
2.489
2.497
2.468
2.510
2.482
2.484
2.502

2.822 8.30
2.802 8.25
2.726 10.47
2.682 6.35
2.703 9.79
2.684 9.73
2.694 8.69
2.693 8.38
2.666 8.51
2.696 7.85
2.678 8.27
2.688 8.77
2.688 8.12

BestROCK

2.804
2.845
2.653
2.699
2.788
2.549
2.797
2.812
2.813
2.796
2.721
2.641
2.595
2.672
2.676
2.698
2.644
2.709
2.674
2.679
2.717

Continued on next page

227

Table B.1: Continued.

WIP

GRI

MIP

ID

B
g/cm3

G
g/cm3

B
G
g/cm3 g/cm3

SS2-22
SS2-23
SS2-24
SS2-25
SS2-26
SS2-27
SS2-28
SS2-29
SS2-30
SS2-31
SS2-32
SS2-33
SS2-34

2.550(10)
2.525(8)
2.476(3)
2.504(5)
2.518(2)
2.538(5)
2.607(4)
2.561(4)
2.494(7)
2.528(1)
2.502(3)
2.531(4)
2.488(5)

2.713(11)
2.691(9)
2.540(4)
2.662(3)
2.665(3)
2.677(3)
2.754(3)
2.701(3)
2.667(3)
2.673(2)
2.663(3)
2.691(2)
2.656(3)

9.56(85)
9.82(69)
4.13(32)
9.50(34)
8.80(20)
8.30(31)
8.34(27)
8.22(29)
10.37(44)
8.67(14)
9.64(23)
9.44(27)
10.10(32)

2.529 2.709
2.480 2.697
2.449 2.644

7.87
9.05
8.75

SS3-1
SS3-2
SS3-3
SS3-4
SS3-5
SS3-6
SS3-7
SS3-8
SS3-9
SS3-10
SS3-11
SS3-12
SS3-13
SS3-14
SS3-15
SS3-16
SS3-17
SS3-18
SS3-19
SS3-20
SS3-21
SS3-22

2.471(4)

2.685(6)

12.69(41)

2.471(4)
2.456(5)

2.581(5)
2.580(6)

6.92(40)
7.85(49)

2.541(5)
2.488(3)
2.499(8)
2.502(4)
2.505(7)
2.497(4)
2.481(5)
2.527(5)
2.502(6)
2.474(4)
2.476(7)

2.650(6)
2.663(5)
2.614(9)
2.617(5)
2.686(8)
2.698(6)
2.591(6)
2.692(6)
2.696(8)
2.571(6)
2.588(7)

6.61(48)
10.51(34)
7.11(71)
7.11(38)
10.75(62)
11.81(41)
6.87(49)
9.72(41)
11.40(57)
6.17(45)
7.03(61)

2.597(12)
2.624(4)
2.562(6)
2.599(4)

2.740(10)
2.715(6)
2.719(7)
2.718(6)

8.22(87)
5.27(38)
9.10(47)
6.89(38)

B
G
g/cm3 g/cm3

2.461
2.427
2.496
2.596
2.527
2.474
2.506
2.449
2.513
2.491

2.600
2.560
2.625
2.723
2.616
2.559
2.629
2.546
2.649
2.586

BestROCK

G
g/cm3

5.32
5.19
4.91
4.65
3.42
3.31
4.67
3.80
5.13
3.68

2.715
2.689
2.556
2.679
2.666
2.679
2.729
2.695
2.675
2.665
2.644
2.680
2.661
2.682
2.658
2.699
2.698
2.691
2.737
2.653
2.668
2.622
2.624
2.689
2.678
2.602
2.686
2.664
2.570
2.578
2.743
2.718
2.712
2.722
2.715

228

APPENDIX C - STANDARD OPERATING PROCEDURES

The standard operation procedures for the experimental techniques are presented here.
C.1

Sample Homogenization and Splitting

Rigorous sample homogenization and splitting techniques should be adopted to prevent sampling bias

between dierent measurements. The procedure suggested in McCarty (2002) is adopted for this purpose.

Equipments
1. 40 mesh sieve
2. Porcelain mortar and pestle
3. Whirl pak bags
4. Safety gloves
5. Safety glasses
6. Isopropyl alcohol
7. Kimwipe
8. Rietype sample splitter

Procedure
1. Clean the porcelain mortar and pestle and the Rietype sample splitter with Isopropyl Alcohol. Air
dry afterwards
2. Use 56 g of samples at a time.
3. Crush the samples in the mortar using minimal energy. The samples should be pounded from grinded.
4. Sieve the crushed powder through the 40 mesh sieve and put the portion of the samples that didnot
pass back to mortar.
5. Continuing crushing and sieving, until the entire sample mass pass through the sieves
6. DONOT DISCARD ANY SAMPLE, ENTIRE SAMPLE MASS SHOULD PASS THROUGH
40 mesh sieve
7. Split the crushed sample in the Rie splitter
8. Store the crushed sample splits in whirl pak bags and label

229

C.2

Water Immersion Porosimetry

Equipments
1. Pressure Saturator
2. Rock Cutting Diamond Saw
3. Oven that heats to 200 C
4. Mettler Toledo Halogen Moisture Analyzer
5. Mettler Balance on top of Marble Table
6. Jolly Balance Apparatus for Mettler Balance with hanging basket and thermometer
7. Deionized (DI) Water
8. Surfactant
9. Filter Paper
10. Aluminum Trays
11. Glass Beakers
12. Small Artist Paint Brush
13. Tweezers
14. Safety gloves
15. Safety glasses
16. Kimwipe
17. Hot oven gloves

Procedure Day 1 Sample Preparation Step

1. Cut 3-4 rectangular to cubic chips from core samples of 3-5 g each, for a total sample weight of 10
g using diamond saw. DONOT USE WATER, cut dry. Flat, smooth sides of blocky shaped chips
are optimum (for subsequent brushing of water from surfaces while weighing.), Avoid wedges or acute
angle edges or any irregular shape which may allow for sample loss by breakage upon handling.
2. Sample faces may be polished using diamond blade to smooth out irregularities
3. transfer WIP chips to Al trays (Label with red sharpie because it stays on best in oven)

230

4. Desiccate WIP portion of samples in Al trays in 200C vacuum furnace overnight for 12-16 hours.
Day 2 Dry Sample Measurement
1. Calibrate Mettler Toledo Halogen Moisture Analyzer instrument ahead of time using 10 g standard
(half an hour required)
2. Same metal tray may be used for all samples, remember to tare between each sample.
3. Allow sample to cool briey after run is complete before storing in a whirl pak plastic bag
4. Allow sucient time for the instrument to cool down before starting the next measurement (<50 C).
5. Record the weight of each sample at the end of the analysis as the Dry weight
Day 3-4 Pressure Saturation Step
1. Vacuum overnight 12-16 hours the sample in the pressure saturator
2. Use DI water mixed with surfactant as saturant. 10 drops surfactant/1 L water.
3. Pressurize to 2,000 psi
4. Saturate for 24 hours.
Day 5 MEASUREMENT STEP
1. Remove samples from saturation system
2. Keep samples in beaker with DI water/surfactant mix
3. Set up jolly balance apparatus, with hanging basket and thermometer
4. Make 600 mL of DI water with 6 drops of wetting agent (surfactant to relieve surface tension)
5. Fill glass beaker with surfactant water: use brush to pop bubbles, dont allow bubbles to touch
basket, make sure basket does not touch thermometer.
6. Using tweezers, individually place each chip on the hanging basket and allow equilibration with new
water before taking rst wet weight (i.e. wait until weight stabilizes).
7. Take dry weight by transferring chips with basket to weigh paper, and blotting chips on weigh paper
with small brush. Brush surfaces of each chip quickly with paint brush until matte color of black (not
shiny, not dry gray color) Donot allow chips to be exposed to air for very long
8. Use tweezers to transfer blotted chips to top of tarred basket in scale (portion out of the water, above
beaker)

231

9. Immediately transfer chips back into beaker and re-equilibrate chips in water before taking next wet
weight
10. Take 6 wet weights and 6 dry weights for each sample. Take water temperature for each wet weight
11. Change water between samples if it is dirty, note if any pieces of chips that are lost during transfer,
brushing or in beaker.
DATA PROCESSING
1. Input data into WIP macro.
2. Use the Moisture analyzer drift (from calibration of the 10g standard) to adjust the dry weight of the
sample.
C.3

Nitrogen Gas Adsorption Experiments using Micromeritics ASAP 2020

Modied after the ASAP 2020 Manual (ASAP 2020, Accelerated Surface Area and Porosimetry System

Operators Manual. V4.00). Contribution from Lemuel Godinez. Major parts of the following text are
taken from ASAP 2020, Accelerated Surface Area and Porosimetry System Operators Manual. V4.00
directly.

Getting accustomed with the set-up A sample tube set consists of the following parts:
1. Sample tube
2. Either: Stopper, Seal frit or TranSeal. In this study, we used mostly TranSeal
3. Filler rod
ASAP 2020
1. Analysis Port
2. Degas Port
3. Dewars
4. Connections of sample tube: O-rings, ferrule and connector nuts
5. Software.

Checks
1. Sample log

232

2. Cleaning tubes apparatus: Ultrasonic units, brush, oven

3. Check whether there is enough liquid nitrogen or nor before starting analysis. Get it from Physics
department.

Cleaning and Labeling Sample Tubes

1. Check the reservoir of the ultrasonic cleaning unit to make sure it is clean.
2. Use 5 grams of Alconox per 500 mL of warm water; ll the reservoir of the ultrasonic unit upto the
labeled mark. Make sure the detergent is dissolved before placing the sample tubes in the water.
3. Fill the sample tubes with warm water and place them in the reservoir of the ultrasonic cleaning unit.
DONOT PUT THE FILLER RODS IN THE ULTRASONIC UNIT. Turn on the ultrasonic cleaning
unit for fteen minutes.
4. Using rubber gloves, remove the sample tubes from the reservoir.
5. Clean the interior of the sample tubes with the brush supplied with the ASAP 2020.
6. Rinse the sample tubes and ller rods thoroughly with hot water. Then rinse them with isopropyl
alcohol or acetone using a waste container to collect used solvent.
7. Dry the interior of the sample tubes and ller rods by blowing compressed air.
8. Stand the sample tubes on the sample tube rack in inverted position for hr to drip the water out.
Then put in oven for 1 hr.
9. Remove the sample tubes from the oven and allow them to cool.
10. Use latex gloves to handle the sample tubes and ller rod [NOT BARE HANDS]. Use microber cloth
or lint-free cloth to wipe the outside of the sample tube and ller rods.
11. Place a ller rod in each sample tube by holding the sample tube horizontally and sliding the ller rod
into the sample tube slowly. DONOT DROP THE FILLER ROD IN THE SAMPLE TUBE WHILE
HOLDING IT VERTICALLY. THIS MIGHT BREAK THE SAMPLE TUBE.
12. Label the sample tube and seal frit or Transeal for identication. USE THE SAME SET OF FILLER
ROD, SAMPLE TUBE AND SEALFRIT EVERYTIME
13. Degas the empty tube using the following instructions

233

Degassing the empty sample tube

1. While holding the degas port plug, remove the connector nut and plug from the degas port by turning
the connector nut counterclockwise
2. Installing the TranSeal in the Sample Tube: (i) Place the Sample into the Sample Tube; (ii) Insert
the TranSeal into the Sample tube as illustrated below. An O-ring provides sealing with the inside
of the sample tube. This O-ring can be shifted into any one of three grooves to provide a tight seal
regardless of small variations in the inside diameter of the sample tube. Be sure the external O-ring,
other components and heating mantle are arranged as illustrated.
3. Place a heating mantle over the bulb of the sample tube and secure the mantle in place with a mantle
clip.
4. Installing the Sample Tube with TranSeal on Degas port: (i) For initial sample preparation, push the
sample tube rmly into the attachment port; (ii) Securely tighten the retaining nut. Be sure to use
the low-prole retaining nut included with the TranSeal. The sealing plunger of the TranSeal should
clearly protrude into the sample tube, showing the TranSeal is open. If it is not, loosen the retaining
nut and push harder on the sample tube and then retighten the retaining nut.
5. Make sure both the heating mantle thermocouple plugs and the power plug are inserted completely
in appropriate connector on the analyzer
6. Fill half of the cold trap Dewar with liquid nitrogen. Hang the cold trap Dewar around the cold trap
port.
7. Place the insulator/stopper over the Dewar opening
8. Select Start Degas from the Unit menu in ASAP 2020 Software; the Start Degas dialog is displayed.
9. Click Browse to the right of the Sample eld to choose your sample le. (Choose 1 if using the left
port or 2 if using the right port). Select the Cleaning the tube le.
10. Click Start to begin the degassing operation.
11. Observe the status bar of the degassing operation to determine when degassing is complete. After
EXACTLY 2 hrs, click the Check button to get the outgassing rate. Wait 15 mins and note the
outgassing rate. Click Continue once is done

234

Loading sample and weighing sample mass Avoid touching the sample with your ngers
because oils may be transferred to the sample and can alter results or create degassing problems. Do not
touch the sample or ller rod with bare hands while performing the next steps. Doing so could aect the
accuracy of results.
1. Write the Sample Tube number on the Sample Data Worksheet. (See appendix)
2. Calibrate the balance.
3. Place the sample weighing support on the balance. Tare the balance and allow it to stabilize at zero
(0). Check for the * mark in the balance.
4. Place the sample tube set (sample tube, seal frit, and ller rod) on the sample support. Record the
stabilized weight on the Sample Data Worksheet as Mass of empty sample tube (ATLEAST 3 TIMES,
tare it every time). Remove the sample support and sample tube set from the balance.
5. Clean, dry sample tubes are essential for accurate results. Use latex gloves to handle the sample tubes
and ller rod [NOT BARE HANDS]. Use microber cloth or lint-free cloth to clean the outside of the
sample tube and ller rods. Take the seal frit and ller rod out to load the sample.
6. Place a sample container onto the balance. Tare the balance and allow it to stabilize at zero (0).
Slowly add the sample to the sample container. The user will specify the weight of the sample to be
used. Record the weight.
7. Use the hollow tube to load the sample. CARE SHOULD BE TAKEN TO MAKE SURE NO
SAMPLES ARE ATTACHED TO THE NECK OF THE SAMPLE TUBE. AVOID TOUCHING
THE SAMPLE WITH YOUR FINGERS BECAUSE OILS MAY BE TRANSFERRED TO THE
SAMPLE AND CAN ALTER RESULTS OR CREATE DEGASSING PROBLEMS. MAKE SURE
THE NECK OF THE SAMPLE TUBE IS CLEAN.
8. Replace the ller rod and insert the TransSeal. Use same technique to load the ller rod (Horizontally
not vertically).
9. Place the sample weighing support on the balance. Tare the balance and allow it to stabilize at zero
(0). Check for the * mark in the balance.
10. Weigh the sample tube set containing the sample and record the weight on the Sample Data Worksheet
as Mass of sample tube plus sample (Before Degas). WEIGH ATLEAST 3 TIMES.
11. Subtract the Mass of empty sample tube from the Mass of sample tube plus sample; record this value
as the Mass of sample.

235

Creating Sample Information Files in ASAP 2020 software

1. Select File > Open > Sample information; the Open Sample Information File dialog is displayed.
2. Accept the next sequenced le number
3. Click OK, then Yes to create the le; the Sample Information dialog is displayed.
4. Enter the details of sample in the SAMPLE eld
5. Enter the mass of the sample. Donot put anything in the density.
6. Using the down arrows to the right of each parameter eld, select an appropriate le. The appropriate
les details will be supplied by the submitter.
7. Click Save, then Close.

Degassing the Sample Check the previous section Degassing the empty sample tube
1. Attach the sample tube set to the degas port. (USE THE SAME PROCEDURE AS DESCRIBED
IN SECTION Degassing the empty sample tube).
2. Fill the cold trap Dewar
3. Select Start Degas from the Unit menu in ASAP 2020 Software; the Start Degas dialog is displayed
4. Click Browse to the right of the Sample eld to choose your sample le. (Choose 1 if using the left
port or 2 if using the right port). Make sure the Degas Condition eld reads the appropriate degas
le.
5. Click Start to begin the degassing operation.
6. Observe the status bar of the degassing operation to determine when degassing is complete. After
EXACTLY 12 (or 24 hrs) hrs, click the Check button to get the outgassing rate. Wait 15 mins and
note the outgassing rate. Click Continue once is done. Check the outgassing rate once just before
the end of degas. Observe the status bar of the degassing operation to determine when degassing is
complete.
7. After degassing has completed, transfer the sample tube to the analysis port to start the analysis
(next section).

236

Transferring the Degassed Sample to the Analysis Port The sample tube must be
removed from the degas port, weighed and then installed onto the analysis port for analysis.
1. Carefully remove the heating mantle clip and the heating mantle from the sample tube and allow the
sample tube to cool to room temperature (approximately fteen minutes).
2. Removing the Sample tube with TranSeal from the Preparation Station.To retain the vacuum about
the sample, loosen the retaining nut just enough for the sample tube to be pulled down until the
TranSeal plunger is seen to close. Further loosen the retaining nut and remove the sample tube
3. Weigh the sample tube set (ATLEAST 3 TIMES). Enter the weight on the Sample Data Worksheet
as Mass of sample tube plus sample (After Degas).
4. Subtract the Mass of empty sample tube (Before Degas) from the Mass of sample tube plus sample
(After Degas) to determine the mass of the sample. Record this value as the Mass of sample (After
Degas).
5. Slide an isothermal jacket down over the sample tube stem until it touches the sample tube bulb.
6. Installing the Sample tube with evacuated sample on the Analyzer: (i) Push the sample tube into
the analysis port only enough for an external seal to be made when the retaining nut is tightened
securely (ii) Activated the evacuation system of the analyzer and remove the ambient air that was
trapped in the attachment port before the sample tube was installed (iii) When a sucient vacuum is
achieved, loosen the retaining nut only enough for the sample tube to be pushed into the attachment
port suciently to open the TranSeal (iv) Securely tighten the retaining nut. The plunger of the
TranSeal should be clearly visible extending into the sample tube (refer to the previous illustration).
7. Place the sample tube Dewar cover over the sample tube stem just above the isothermal jacket as
shown in the following illustration.

Filling Analysis Dewar

1. Fill the analysis Dewar with the analysis bath uid to about 5 cm (2 inches) from the top.
2. Incorrect uid levels can lead to measurement errors. Do not overll the Dewar.
3. Check the analysis bath uid level with the dipstick as shown below
4. Insert the analysis Dewar onto the elevator as shown in the following illustration.
5. Place a insulator over the open Dewar until you are ready to start your analysis; this helps to minimize
ice accumulation.

237

6. When you are ready to start the analysis, remove the insulator.

Performing an Analysis
1. Select Unit > Start Analysis; the Analysis dialog is displayed with the Start Analysis dialog positioned
on top.
2. Choose a le for your analysis and click OK; the Analysis dialog containing the parameters of the
selected le is displayed. CHANGE THE SAMPLE MASS (ENTER THE SAMPLE MASS AFTER
DEGAS)
3. Verify parameters and make any changes you feel necessary.
4. Click Start to begin the analysis. As data are collected, the graph will be drawn in the window. Click
Next to perform another analysis; the rst view of the Analysis dialog is displayed.

238