Porosity and permeability of Green River oil shale and their changes during retorting

Fuel 203 (2017) 208–213

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Porosity and permeability of Green River oil shale and their changes during retorting Alan K. Burnham ⇑ Consultant to Total Exploration and Production, United States Energy Resources Engineering, 367 Panama St., Stanford University, Stanford, CA 94305, United States

h i g h l i g h t s

g r a p h i c a l a b s t r a c t


Estimate compaction and porosity

1.E+02

Porosity

0.51 0.42

1.E+01

0.28

1.E+00

Permeability, mD

during diagenesis as a function of organic content.
Calculate porosity as a function of oil shale grade and kerogen conversion during in-situ retorting.
Estimate permeability from porosity using simple engineering correlations.

0.19 0.11

1.E-01 5 10 15 25 35 45 55 65

1.E-02

1.E-03

1.E-04 0.0

0.2

0.4 0.6 0.8 Fraction retorted

1.0

Permeability calculated from porosity using a modified Kozeny-Carman relation as a function of oil shale grade and fraction retorted.

a r t i c l e

i n f o

Article history: Received 2 March 2017 Received in revised form 19 April 2017 Accepted 26 April 2017

Keywords: Green River oil shale Porosity Permeability Retorting Compaction

a b s t r a c t Oil shales are organic-rich mudstones that generally have little porosity and permeability until kerogen is transformed into oil and gas. A simple mathematical model is reported for how porosity and permeability values for the Green River Formation change during retorting under confinement. Unlike when retorted unconstrained, during which numerous fractures occur due to the limited tensile strength of retorted oil shale and the permeability increases from micro or nano-Darcy levels to Darcy levels, fracture permeability is minor when constrained by lithostatic loads typical of in-situ retorting, so permeabilities increase only to the milli-Darcy level. The permeability increase is related to an increase in both porosity and pore diameter, and measured permeabilities are consistent with measurements and calculations of those properties and inter-relationships developed for naturally matured petroleum source rocks. Ó 2017 Published by Elsevier Ltd.

1. Introduction Although oil shale retorting has been a source of small amounts of shale oil for centuries and is important in certain localities, global interest in oil shale as a potential source of shale oil waxes and

⇑ Address: Energy Resources Engineering, 367 Panama St., Stanford University, Stanford, CA 94305, United States. E-mail address: [email protected] http://dx.doi.org/10.1016/j.fuel.2017.04.119 0016-2361/Ó 2017 Published by Elsevier Ltd.

wanes every few decades as the price and perceived supply issues for conventional crude oil rise and fall. More recently, production of natural petroleum (oil and gas) from mature source rocks and adjacent or interbedded fine-grained yet more permeable layers has greatly increased the knowledge of porosity and permeability of organic-rich fine-grained rocks. The combination of historical and recent information gathered for oil shale processing and for production of tight oil and shale gas purposes provides some more general insights that can be useful for both applications, although

209

A.K. Burnham / Fuel 203 (2017) 208–213

2. Porosity versus grade Many years ago, Tisot [1] reported porosities for several samples of Green River oil shale from the Mahogany Zone. Porosity was determined by comparison of bulk and grain densities for cores, although the procedure for measuring grain density was not described in enough detail to be sure all porosity was accessed. The porosity was about 10% for lean oil shale with 1 wt% Total Organic Carbon (TOC) and dropped linearly to zero for organic content greater than 6 wt% TOC. However, subsequent helium pycnometry measurements at Lawrence Livermore National Laboratory (LLNL) on a variety of samples from the Piceance and Uinta Basins indicated that porosity of 1–5% still existed for samples with 7 wt% total organic carbon (TOC) [2]. So Tisot’s data should not be considered definitive. Rock porosity can also be determined from well logs. Smith et al. [3] estimated porosity over 50-ft intervals by comparing grain density measurements with neutron log densities. Values from 1 to 9% were obtained, with the higher values attributed to dissolution of nahcolite nodules. More recently, characterization of a wellbore near the center of the Piceance basin by Schlumberger Combinable Magnetic Resonance (CMR) provided a more direct measurement of the porosity via free water content over a 1500-ft thickness of the Green River Formation [4]. Results for a few selected intervals are shown in Fig. 1. The wt% TOC is approximately half the grade in gal/ton. The middle portion of the formation (including the leached zone) is not shown, because the porosity has major contributions due to dissolved nahcolite nodules. Using the USGS L-R nomenclature [5], the intervals are separated into two groups, which have zero-organic-matter intercepts of about 16 and 26%, respectively. The porosity is not simply related to either depth or mineralogy. The R0-R2 zones are typi-

0.3 BH1 850-950 ft (R8) BH1 950-1150 ft (MZ) BH1 2136-2250 ft (R0-L0)

0.2

0.1

Porosity

the primary intended application for this work is to develop a new algorithm for permeability as a function of grade and extent of kerogen decomposition for modeling fluid flow during in-situ oil shale processing. As in any field, measurement methodologies improve over time. Consequently, historical information must be critically evaluated. However, measurements from the 1960s and 1970s are still among the best available for some conditions. The current paper attempts to combine the best of the old literature with more recent measurements to draw a more comprehensive picture of how porosity and permeability evolve over the transformation of kerogen under lithostatic load typical of in-situ retorting, which can be approximated as constant volume. This information is used to develop and validate a new, simple algorithm for how porosity and permeability evolve during retorting at constant volume. More general relationships for volume versus mechanical load during retorting are still in the development stage. The approach developed here combines three simple aspects to calculate permeability as a function of grade and extent of retorting. First, an empirical relationship is developed to account for how the greater ductility of kerogen affects initial porosity as a function of kerogen content. Second, a correlation is developed between total porosity and the matrix permeability of both raw and retorted shales as expected by Kozeny-Carman and similar relationships. Third, it is shown that retorting under lithostatic load corresponding to a few hundred meters overburden yields porosities as a function of kerogen conversion roughly equal to those calculated at constant volume. These three relationships are used to create an algorithm for and a plot of permeability as a function of conversion for various oil shale grades, which is needed to model the dissipation of pore pressure generated within the formation during in-situ oil shale retorting.

0.0 0

10

20

30

40

50

60

70

0.3 BH1 1890-2135 ft (R1-R2) BH1 1151-1347 ft (L5-R6)

0.2

0.1

0.0 0

10

20

30

40

50

60

70

Grade, gal/ton Fig. 1. Porosity versus grade for five intervals in the Green River Formation in the Piceance Basin.

cally 5–15% carbonate and 20–40% illite by weight, but the R0 and L0 zones have noticeably lower porosity. The upper zones are typically 20–40% carbonate and less than 15% illite [6], and the porosity trends are almost inversely related to depth, which may be related to nahcolite deposition and dissolution. For the purposes here, the effect of organic matter on porosity is qualitatively the same for all depth and mineralogy variations. The porosity of shale versus depth is often described by Athy’s law, or more rigorously, as a function of effective stress via Terzaghi’s principle [7]. Given that immature kerogen, particularly Type I, is softer than mineral grains, it is plausible that compaction might be greater for shale with more organic content, although variability with mineralogy also occurs. Young’s modulus and compressive strength calculated from the sonic log (Schlumberger SonicScanner, chirp sampling 300 Hz to 8 kHz) using correlations of Horsrud [8] are shown in Fig. 2 for the R0-R2 interval of the Garden Gulch Member. Similar values for Young’s modulus were provided in the logging report using shear and compressive wave velocities in classical elastic wave propagation equations. These properties are for a wellbore temperature of 40–45 °C, based on temperature logs of the measurement interval, which softens the organic matter much more than the inorganic crystals. Both modulus and strength decrease as oil shale grade increases, which is consistent with both the literature [9] and nanoindentation studies showing that the inorganic crystals are stiffer and stronger than kerogen [10]. The grade dependence is more pronounced for the carbonate-rich Piceance Creek Member than for the clay-rich Garden Gulch Member. Typical mineral moduli at zero porosity are 30–50 GPa, which indicates the effect of porosity on the sonic log moduli at low grade. The high-grade modulus limit of 0.8 GPa, for which kerogen is the continuous phase, is the same as high-density polyethylene at 40 °C [11] and less than polystyrene, polycarbonate, polyethylene terephthalate,

210

A.K. Burnham / Fuel 203 (2017) 208–213

with coal being the large-scale example, are obviously loadbearing, as are organic-rich lamina in lacustrine oil shales. The depth (or effective stress) and organic content (either kerogen volume fraction or wt% TOC) effects can be combined into a single empirical formula

0.8+1.3*exp(-gpt/12)

2.0

0.076(304.8/DTCO)3.23

1.5

u ¼ u0 ead=ð1k Þ b

1.0 0.5

Uniaxial compressive strength, MPa

0.0 20

0

10

20

30

40

50

ð1Þ

where u is porosity, u0 is porosity at zero depth, d is depth, k is kerogen volume fraction, and a and b are empirical constants. As an example, this equation is fitted visually to two intervals for the Garden Gulch Member of the Green River Formation. The parameters are not tightly constrained by the data, but good fits shown in Fig. 3 are obtained for parameters listed in Table 1. Similar results could be derived for other intervals in Fig. 1 if desired. Although the depth does not change during in-situ retorting, this more general formulation would be useful for basin modeling applications.

6.5+9.5*exp(-gpt/12)

16

3. Permeability of oil shale 0.77(304.8/DTCO)2.93

12 8 4 0 0

10

20

30

40

50

Grade, gal/ton Fig. 2. Mechanical properties of illitic oil shale from the Garden Gulch Member of the Green River Formation calculated from sonic log transit times using correlations from Horsrud [8]. DTCO is the compressive wave travel time (ls/ft), and the dotted line is a correlation with grade in gal/ton (gpt).

polymethylmethacrylate, and Nylon at room temperature [12]. The compressive yield strength is also lower than most common roomtemperature polymers, with polyethylene being the closest [13]. In other words, kerogen is softer than most common synthetic polymers. Burnham [14] and White et al. [15] report macroscopic Young’s moduli measurements (static values from stress-strain measurements) for cores from the same well and interval as a function of temperature and grade. The core properties are visco-elasticplastic, so slightly different values were obtained for the loading and unloading portions of the hysteresis curves. In addition, there is some uncertainty in the grade of the cores. Nevertheless, it is certain that the Young’s modulus for oil shale grade >42 gal/ton, where kerogen is the continuous phase on average, is 1–2 GPa at 40–50 °C. Similarly, 10–20 gal/ton oil shale had a modulus of about 5 GPa at that temperature. These values agree well with those determined from the sonic log. The nanoindentation work of Eliyahu et al. [10] yield somewhat higher modulus values for kerogen, but those measurements are affected by uncertainties in Poisson’s ratio and finite deformation, which are significant for visco-elastic-plastic materials. Even so, their kerogen values are substantially smaller than for the mineral components, a result which is qualitatively consistent with Fig. 2, even though the low-grade, mineral-rich modulus in Fig. 2 is greatly affected by porosity. Whether kerogen is primarily load bearing or pore filling has been debated extensively in the literature, with a likelihood that it intermediate between those extremes in a manner than relates to the geometric dispersion of the organic matter. Organic layers,

Although there has been a resurgence of interest in the permeability of organic-rich shales due to the recent increase in oil and gas production from source rocks, permeability measurements of Green River oil shale date back many years [1,16–18]. These early measurements report negligible permeability for raw shale and focused on the porosity and permeability after kerogen conversion, combustion, and carbonate decomposition. Later, Sandvik and Mercer [19] quantified permeability for three samples with 15% TOC and found pico to nano Darcy values. Most studies of retorted and burned shale were conducted under unconfined conditions, where fracture permeability dominates the permeability increase for oil shale grade >13.5 gal/ton [1]. Without fractures, permeability of typical retorted shale is less than 1 mD. Recent X-ray tomographic studies by Kobchenko et al.

0.30 R1 Zone R1 Zone calc R0-L0 zone R0-L0 zone calc

0.25 0.20

Porosity

Young's modulus, GPa

2.5

0.15 0.10 0.05 0.00 0.0

0.2

0.4

0.6

Kerogen volume fraction Fig. 3. Comparison of measured dependence of porosity on organic content with that calculated from Eq. (1).

Table 1 Parameters for visual fits of Eq. (1) to two intervals in the Garden Gulch Member of the Green River Formation. Interval

Depth, m

u0

a

b

R0-L0 R1

651–686 614–651

0.5 0.6

0.0016 0.0011

0.7 0.5

211

A.K. Burnham / Fuel 203 (2017) 208–213

Thomas 1966 6.9 MPa AMSO cores 16.4 MPa AMSO cores 0.12 MPa Retorted LETC Raw LETC Retorted (calc) Raw (calc)

Porosity

0.6

0.4

swelling?

compaction

Thomas estimates 15% compaction for 70 gpt

0.2 Line from CMR log

0 0

20

40

60

80

Grade, gal/ton Fig. 5. Porosity for raw and retorted Green River oil shale calculated from the free volume created by kerogen conversion [1,17] AMSO is American Shale Oil (unpublished data) and LETC is Laramie Energy Technology Center. CMR is Combinable Magnetic Resonance from Schlumberger and is presumably a more accurate measure of porosity in a tight rock.

60

R0-R3 50

Grade, gal/ton

[20], Saif et al. [21], and Tiwari et al. [22] provided an incremental understanding of this fracture formation, and Tiwari et al. estimate permeabilities of hundreds of Darcies for rich oil shale. Other studies [23,24] examined permeability of rubble columns under load to determine whether compaction of rubble chimneys would shut off permeability as the shale became plastic during kerogen transformation, but those are only indirectly relevant to this work. The early study most relevant to this work is Thomas [17] who measured the porosity and permeability for cores retorted under triaxial confinement from 0.7 to 17 MPa. No apparent cracking was observed for these confinement pressures. There was an inverse relationship between confinement pressure and permeability, with a roughly constant permeability of 10 mD parallel to the bedding plane for oil shale grades greater than 35 gal/ton and confinement pressures greater than 5 MPa. Recently, Kibodeaux [25] greatly expanded the published information on porosity and permeability evolution during in-situ confined retorting using cores from inside and outside the retorted region of a field experiment. Those data are shown in Fig. 4 along with data from Thomas, which follow the same trend. The scatter in the data is typical for variability of tight rocks. Three different calculated curves are also shown, including a modified KozenyCarman relationship [30u2.5/(1  u)2], the Ergun equation [110u3/(1  u)], and a cubic power law [200u3]. There is no difference in the ability of any of these relationships to fit the data well within the measurement variability. All work better than the loglinear relationship of Li et al. [26]. The dashed line represents their base case, and they varied the slope and intercept, but their functional form simply cannot reproduce the correct relationship. Shen [27] used a similar function but did not disclose parameters. Given this relationship between permeability and porosity, it is possible to estimate permeability as a function of kerogen conversion by first calculating the porosity created by kerogen conversion. A comparison of measured and calculated porosities is shown in Fig. 5. This simple volume balance calculation uses the correlation between organic content and oil shale grade shown in Fig. 6, an assumed kerogen composition of 81% carbon, kerogen density of 1.05 g/cm3, mineral density of 2.73 g/cm3, and coke density of 1.4 g/cm3, which means that 82% of the kerogen volume is converted to porosity. The created porosity is then added to the

Linear (R0-R3)

40

30

20

y = 2.178x - 1.30 10

0 0

5

10

15

20

25

30

TOC, wt% Fig. 6. Relationship between total organic carbon and Fischer-Assay oil shale grade for the Garden Gulch Member of the Green River Formation in the Piceance Basin.

initial porosity (raw shale) as a function of grade, which was assumed from the CMR log to be

u ¼ 0:04 þ 0:12egpt=3

Fig. 4. Comparison of calculated permeability as a function of porosity with measurements from Thomas [17] and Kibodeaux [25]. The samples of Thomas were retorted under confinement, and the samples of Kibodeaux were recovered from core both inside and outside an in-situ retort at an unspecified depth. Orientation of the measurements is not well documented and could account for some of the scatter.

ð2Þ

where gpt is the oil shale grade in gal/ton and the porosity is in volume fraction. The resulting permeability versus fraction retorted for various oil shale grades calculated from the modified KozenyCarman relation, assuming 63% of the immature kerogen porosity is closed, is given in Fig. 7. Note in Fig. 7 that the initial permeability ranges from 1 to10 lD depending on the initial porosity determined by oil shale grade. The measured range of permeability in Fig. 4 varies more, presumably due to variability in porosity at constant grade and the possibility of some fracture permeability. For rich oil shale, the permeability of retorted shale increases to about 20 mD, which is consistent with the largest permeabilities of Thomas [17] shown in Fig. 4. In contrast, the increase in permeability as a function of

212

A.K. Burnham / Fuel 203 (2017) 208–213

1.E+02

Porosity 0.51 0.42

1.E+01

Permeability, mD

0.28 1.E+00

0.19 0.11

1.E-01 5 10 15 25 35 45 55 65

1.E-02

1.E-03

1.E-04 0.0

0.2

0.4

0.6

0.8

1.0

Fraction retorted Fig. 7. Permeability calculated from porosity using the modified Kozeny-Carman relation shown in Fig. 4. The legend is the oil shale grade in gal/ton. The right-hand scale gives the final porosities for some oil shale grades. The calculation assumes that the pore size increases along with porosity in a way that can be absorbed into the Kozeny-Carman coefficients.

kerogen conversion is relatively small for low oil shale grades, because less porosity is generated. 4. Discussion Extensive measurements of porosity as a function of depth and kerogen content in the Green River Formation using CMR logs (free water content) show that porosity decreases systematically with organic content. This is consistent with the concept that kerogen is more ductile than inorganic grains and presumably deforms into neighboring pore space more easily. The well log trends are qualitatively consistent with sparse, old data from Tisot [1], but log porosities based on free water content indicate more porosity survives at for high organic content. This effect can be simply incorporated into an Athy-law compaction relationship, although the initial porosity and compaction coefficient will depend on lithology. That fossil organic matter softens with temperature and is visco-elastic-plastic is well known [28]. More recently, others have measured the softer nature of kerogen in organic-rich shales using nano-indentation [29,30], although the elastic moduli reported are substantially higher than from macroscopic measurements and are internally variable. Increased compaction with increasing organic content during diagenesis has also been noted [31], although not the quantitative type relationship reported here. Eventually, of course, hardening of the organic matter occurs as it expels oil and gas [10], and the portion of the generated porosity that is preserved causes a positive correlation between porosity and organic content for mature organic-rich shales [32]. Well-established relationships among oil shale grade, oil shale composition, component densities, and pyrolysis stoichiometry can be combined with initial porosity measurements and the measured relationship between porosity and permeability to calculate the permeability of oil shale as a function of grade and fraction retorted (kerogen conversion) at constant rock volume. The con-

stant volume condition is approximately correct for retorting under modest confinement consistent with overburden thickness of 100–700 m depth. The initial permeability is inversely related to oil shale grade due to a greater degree of compaction and smaller porosity for rich oil shale, but the increase in permeability at constant volume is much greater for rich oil shale due to the greater increase in porosity from kerogen conversion. Mbia et al. [33] report that porosity measured by NMR (comparable the CMR used here) were higher than from mercury injection alone and comparable to the combination of helium porosimetry and mercury immersion. They also found that measured permeabilities correlated with porosity and BET surface areas as expected from a Kozeny-Carman relationship. The permeabilities calculated here can be compared to those reported by Aguilera [34] for various unconventional reservoir rocks. In general, the relationship there between porosity and permeability is flatter, presumably because much of the generated porosity is compacted during natural maturation. In this case, the pore throats will remain more nearly constant, whereas for retorting in the absence of compaction, the pore throats will tend to enlarge as a function of kerogen conversion. Nevertheless, the relationship between permeability and pore-throat diameter from Aguilera is consistent with that measured for retorted oil shale. For retorted 22 gal/ton oil shale from the Mahogany zone, mercury porosimetry determined a porosity of 0.34 and a log-mean average pore diameter of about 0.5 lm [35]. The predicted permeability from Aguilera’s curve is 1 mD, which is within a factor of two from that predicted in Fig. 7. Similarly, numerous published scanning electron micrographs of Green River oil shale (e.g., Mehmani et al. [36]) show negligible intergranular porosity as large as 1 lm outside the leached zone. Micro-tomography of a Green River oil shale core with an estimated organic content of 15 wt% and a porosity of 2.5% after drying had no detectable porosity at the lm resolution scale. Similarly, N2 and CO2 gas adsorption isotherms on a Green River oil shale sample with 25 wt% kerogen indicated a porosity of 1.7% between 1 and 300 nm [37] Given that the pore throats determining permeability are only a fraction of the average pore diameter, the relevant diameter for raw shale with significant organic content is likely <0.1 lm. The curves of Aguilera would then predict a permeability of 1 lD for 2% porosity. The variability about that value in Fig. 4 could well be due to variations in organic content or microfractures. Given this general agreement between reported permeabilities for raw and retorted oil shale and correlations of Aquilera, one can have confidence in the permeabilities calculated from porosity generation and the modified Kozeny-Carman relation for conditions where fracture permeability is negligible. Acknowledgments Much of this work was conducted while the author was employed by American Shale Oil, LLC, a joint venture of Genie Energy and Total S.A. The work was subsequently expanded with financial support of Total S.A. in cooperation with the StanfordTotal Enhanced Modeling of Source rock (STEMS) project, a research collaboration between Total S.A. and Stanford University. The author thanks Total S.A. for its support of both ventures. References [1] Tisot PR. Alterations in structure and physical properties of Green River oil shale by thermal treatment. J Chem Eng Data 1967;12:405–11. [2] Braun RL, Burnham AK, Sweeney JJ, Influence of compaction models on petroleum expulsion. In: Petroleum Geochemistry: Industrial Sponsor’s Briefing, Lawrence Livermore National Laboratory Report; 1992.

A.K. Burnham / Fuel 203 (2017) 208–213 [3] Smith JW, Thomas HE, Trudell LG, Geologic factors affecting density logs in oil shale. In: SPWLA 9th Annual Logging Symposium, New Orleans LA, June 1968; Paper P, pp. 1–17. [4] Reeder SL, Kleinberg RJ, Vissapragada B, Macklus M, Herron MM, Burnham A, et al. A multi-measurement core-log integration for advanced formation evaluation of oil shale formations: a Green River Formation case study. Petrophysics 2013;54:258–73. [5] Johnson RC, Mercier TJ, Brownfield ME, Pantea MP, Self JG, An assessment of inplace oil shale resources in the Green River Formation, Piceance Basin, Colorado, U.S. Geological Survey Digital Data Series DDS-69-Y; 2010, Chapter 1, 187 p. [6] Burnham AK, McConaghy JR. Semi-open pyrolysis of oil shale from the Garden Gulch Member of the Green River Formation. Energy Fuels 2014;28:7426–39. [7] Hantschel T, Kauerauf AI. Fundamentals of basin and petroleum systems modeling Chapter 2. Springer-Verlag; 2009. [8] Horsrud P. Estimating mechanical properties of shale from empirical correlations SPE Drilling & Completion, 2001. p. 68–73. [9] Eseme E, Urai JL, Krooss BM, Littke R. Review of mechanical properties of oil shales: implications for exploitation and basin modeling. Oil Shale 2007;24:159–74. [10] Eliyahu M, Emmanuel S, Day-Stirrat RJ, Macaulay CI. Mechanical properties of organic matter in shales mapped at the nanometer scale. Mar Pet Geol 2015;59:294–304. [11] Es-Saheb MHH. The temperature effects on high-density polyethylene (HDPE) pipes. JKAU Eng Sci 1996;8:47–60. [12] Modulus of elasticity or Young’s Modulus and tensile modulus for common materials. http://www.engineeringtoolbox.com/young-modulus-d_417.html [Accessed 12/14/2016]. [13] Compressive strength testing of plastics. http://www.matweb.com/ reference/compressivestrength.aspx. [Accessed 12/14/2016]. [14] Burnham AK, Oil shale rock mechanics—how to deal with huge changes with grade and temperature? In: Proceedings of the 33rd Oil Shale Symposium, Golden CO, October; 2013. [15] White JA, Burnham AK, Camp DW. A thermoplastic model for oil shale. Rock Mech Rock Eng 2017;50:677–88. [16] Grant BF. Retorting oil shale underground—problems and possibilities. Quart Colo Sch Mines 1964;59:39–46. [17] Thomas GW. Some effects of overburden pressure on oil shale during underground retorting. Soc Petrol Eng J 1966:1–8. [18] Sresty GC, Dev H, Snow RH, Bridges JE. Kinetics of low-temperature pyrolysis of oil shale by the IITRI rf process. In: Proceedings of the 15th oil shale proceedings. Golden, CO: Colorado School of Mines Press; 1982. p. 411–23. [19] Sandvik EI, Mercer JN. Primary migration by bulk hydrocarbon flow. Org Geochem 1990;16:83–9.

213

[20] Kobchenko M, Panahi H, Renard F, Dysthe DK, Matthe-Sørrenssen A, Mazzini A, et al. 4D imaging of Fracturing in organic-rich shales during heating. J Geophys Res 2011;116:B12201. [21] Saif T, Lin Q, Singh K, Bijeljic B, Blunt MJ. Dynamic imaging of oil shale pyrolysis using synchrotron x-ray microtomography. Geophys Res Lett 2016;43:6799–807. [22] Tiwari P, Deo M, Lin CL, Miller JD. Characterization of oil shale pore structure before and after pyrolysis by using X-ray micro CT. Fuel 2013;107:547–54. [23] Tisot PR, Sohns HW. Structural response of rich Green River oil shales to heat and stress and its relationship to induced permeability. J Chem Eng Data 1970;15:425–34. [24] Thigpen L, Heard HC. Vertical stress distribution in oil-shale aggregate columns during retorting. Soc Petrol Eng J 1979:97–106. [25] Kibodeaux KR. Evolution of porosity, permeability, and fluid saturations during thermal conversion of oil shale Report SPE-170733-MS. Society of Petroleum Engineers; 2014. [26] Li H, Vink JC, Alpak FO. An efficient multiscale method for the simulation of insitu conversion processes. SPE J 2015:579–93. [27] Shen C. Reservoir simulation of an in-situ conversion pilot of Green-River oil shale Report SPE 123142. Society of Petroleum Engineers; 2009. [28] Van Krevelen DW. Coal Chapter 16. Elsevier; 1993. [29] Zeszotarski, Chromik RR, Vinci RP, Messmer MC, Michels R, Larsen JW. Imaging and mechanical property measurements of kerogen via nanoindentation. Geochim Cosmochim Acta 2004;68:4113–9. [30] Alstadt KN, Katti KS, Katti DR. Nanoscale morphology of kerogen and in situ nanomechanical properties of Green River oil shale. J Nanomech Micromech 2016;6. 040150003-1-10. [31] Milliken KL, Ko LT, Pommer M, Marsaglia KM. SEM Petrography of eastern Mediterranean sapropels: analog data for assessing organic matter in oil and gas shales. J Sediment Res 2014;84:961–74. [32] Milliken KL, Rudnicki M, Awwiller DN, Zhang T. Organic matter-hosted pore system, Marcellus Formation (Devonian), Pennsylvania. AAPG Bull 2013;97:177–200. [33] Mbia EN, Faricius IL, Krogsbøll A, Frykman P, Dalhoff F. Permeability, compressibility and porosity of Jurassic shale from Norwegian-Danish Basin. Pet Geosci 2014;20:257–81. [34] Aguilera R. Flow units: from conventional to tight gas to shale gas to tight oil to shale oil reservoirs Report SPE-165360. Society of Petroleum Engineers; 2013. [35] Burnham AK. Reaction kinetics between CO2 and oil-shale residual carbon. 1. Effect of heating rate on reactivity. Fuel 1979;58:285–92. [36] Mehmani Y, Burnham AK, Tchelepi HA. From optics to upscaled thermal conductivity: Green River oil shale. Fuel 2016;183:489–500. [37] Daigle H, Ezidiegwu S, Turner R. Determining relative permeability in shales by including the effects of pore structure on unsaturated diffusion and advection Report SPE-175019-MS. Society of Petroleum Engineers; 2015.